University of Southern California Dissertations and Theses

Variations in physician practice patterns for eye care under the National Health Insurance of Taiwan

(USC Thesis Other)

Variations in physician practice patterns for eye care under the National Health Insurance of Taiwan

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO U SERS
This manuscript has been reproduced from the microfilm master. UMI films
the text directly from the original or copy submitted. Thus, some thesis and
dissertation copies are in typewriter face, while others may be from any type of
computer printer.
The quality of this reproduction is dependent upon the quality of the
copy submitted. Broken or indistinct print, colored or poor quality illustrations
and photographs, print bleedthrough, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand comer and continuing
from left to right in equal sections with small overlaps.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6" x 9” black and white
photographic prints are available for any photographs or illustrations appearing
in this copy for an additional charge. Contact UMI directly to order.
ProQuest Information and Learning
300 North Zeeb Road, Ann Arbor, Ml 48106-1346 USA
800-521-0600
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
VARIATIONS IN PHYSICIAN PRACTICE PATTERNS FOR EYE CARE
UNDER THE NATIONAL HEALTH INSURANCE OF TAIWAN
Copyright 2000
by
Joanna Chih I Yu Chang
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PUBLIC ADMINISTRATION)
December 2000
Joanna Chih I Yu Chang
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 3041549
Copyright 2000 by
Yu Chang, Joanna Chih I
All rights reserved.
___ ®
UMI
UMI Microform 3041549
Copyright 2002 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UNIVERSITY OF SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES, CALIFORNIA 90007
This dissertation, written by
..ij&aacn.fli ..... .....................................
under the direction of h.*X Dissertation
Committee, and approved by all its members,
has been presented to and accepted by The
Graduate School, in partial fulfillment of re
quirements for the degree of
DOCTOR OF PHILOSOPHY
Dean of Graduate Studies
Qafe December 18, 2000
DISSERTATION COMMITTEE
Chairperson
< S L....
/
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
T a b l e of C o ntents
List of T ables................................................................................................................... vi
List of Graphics................................................................................................................ ix
Abstract...............................................................................................................................x
1. INTRODUCTION..................................................................................1
2. BACKGROUND AND SIGNIFICANCE.........................................5
2.1 N atio n a l Health In su ra n ce of Taiw an (N H I)............................... 5
2.1.1 Information about Beneficiaries, Enrollment and
Contracted Providers............................................................................5
2.1.2 Benefit Coverage o f National Health Insurance o f
Taiwan....................................................................................................6
2.1.3 Economic Background in National Health Insurance
o f Taiwan................................................................................................7
2.1.4 Payment Policy in National Health Insurance.................................... 8
2.1.5 Expenditure o f National Health Insurance in Taiwan,
1997........................................................................................................9
2.1.6 Distribution o f Ophthalmologists in Taiwan......................................10
2.2 C a ta ra ct Su r g e r y ...................................................................................11
2.2.1 Epidemiology o f Cataracts.................................................................. 12
2.2.2 The Reasons fo r Rapid Increase o f Cataract Surgery...................... 13
ii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.2.2.1 Longer Life Expectancy...................................................................................... 13
2.2.2.2 Loosened Criteria for Cataract Surgery.............................................................14
2.2.2.3 Safety and Cost Effectiveness of Outpatient Surgery.....................................15
2.2.2.4 Affordability o f Cataract Surgery.......................................................................15
2.2.2.5 Cost and Utility Ratio..........................................................................................16
2.2.2.6 Backlog o f Cataract Surgery...............................................................................17
2.3 E pisode o f C a r e ...........................................................................................18
2.4 G eographic V a r ia t io n ............................................................................ 20
2.5 V ariation of R eso u r c e U se and P h y sic ia n P ractice
St y l e s ..................................................................................................................... 23
2.5.1 Physician Profiling................................................................................ 24
2.5.2 What is Physician Profiling?................................................................25
2.5.3 The Objective o f Profiling.................................................................... 26
2.5.3.1 Clinic Measurement.............................................................................................27
2.5.3.2 Economic Measurement.....................................................................................27
2.5.3.3 Patient Satisfaction..............................................................................................28
2.5.4 The Criteria o f Establishing Profile.................................................... 29
2.5.5 How to Encourage Physicians to Buy into Profiling.........................30
2.6 Resea rch Qu e s t io n s................................................................................ 31
3. CONCEPTUAL FRAMEWORK.....................................................33
3.1 M odel of Su pplier-Induced D em a n d................................................34
3.2 Factors A ffectin g P hysician D ec isio n-M a k in g ........................35
3.2.1 Cognitive F actors.................................................................................. 35
3.2.2 Behavioral Factors................................................................................ 35
3.2.3 Sociological M odel................................................................................ 36
3.3 V a riables A ffec tin g P hysician D e c isio n-M a k in g.................... 37
3.3.1 Patient Demographic Characteristics................................................ 37
3.3.1.1 Patient A g e ...........................................................................................................37
3.3.1.2 Patient G ender..................................................................................................... 38
iii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.3.2 Physician Characteristics.................................................................... 39
3.3.2.1 Physician Age......................................................................................................40
3.3.2.2 Physician Gender................................................................................................ 41
3.4 Pra ctice S e t t in g ........................................................................................41
3.5 H y p o t h e s e s..................................................................................................43
4. STUDY POPULATION AND DATA SOURCES................ 48
4.1 V a ria ble Specifica tio n........................................................................... 50
4.1.1 Episode o f Care..................................................................................... 50
4.1.1.1 Defining an Episode of Care.............................................................................. 50
4.1.1.2 Grouping Claims into an Episode.....................................................................52
4.1.2 Dependent Variables.............................................................................55
4.1.2.1 Prevalence R ate...................................................................................................55
4.1.2.2 Payment, Charge, and Cost................................................................................ 55
4.1.2.3 Total per Visit Paid by the Bureau to Providers.............................................56
4.1.2.4 Total per Person Paid by the Bureau to Providers......................................... 56
4.1.2.5 Total Paid by the Bureau for Each Diagnosis................................................. 57
4.1.2.6 Percent of Drug Cost in One V isit....................................................................57
4.1.2.7 Percent of Lab Cost in One V isit...................................................................... 57
4.1.2.8 Total Paid by the Bureau per Cataract Surgery Episode............................... 58
4.1.3 Independent Variables.......................................................................... 58
4.1.3.1 Practice Setting....................................................................................................58
4.1.3.2 Patient Demographic Characteristics............................................................... 59
4.1.3.3 Diagnosis as Treatment-Related Information................................................. 59
4.1.3.4 Physician Characteristics....................................................................................60
4.1.3.5 Geographic Region............................................................................................. 60
4.2 A ppro a ch to E mpirical A n a l y s is.......................................................61
4.3 L im it a t io n s..................................................................................................63
5. RESULTS.............................................................................................66
5.1 D escriptiv e Re s u l t s.................................................................................66
5.1.1 Results fo r All Eye Care....................................................................... 66
5.1.2 Results fo r Cataracts.............................................................................81
iv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5.1.2.1 Prevalence of Cataract by Age and Country.................................................... 81
5.1.2.2 Patient Behavior: Pre- and Postoperative Treatment in
Cataract Surgery..................................................................................................85
5.1.2.3 Demographics and Regional Distribution......................................................... 88
5.1.2.4 Subsequent Surgery............................................................................................. 96
5.1.3 Results for Physician Characteristics in All Eye Care.....................97
5.2 An a ly tic Re s u l t s.................................................................................... 102
6. APPLICATION OF THE FINDINGS: PHYSICIAN
PROFILING IN ALL EYE CARE......................................................109
7. DISCUSSION...................................................................................... 114
7.1 Im p l ic a t io n ................................................................................................. 114
7.1.1 All Eye Care............................................................................................714
7.1.2 Cataracts.................................................................................................77 7
7.1.3 Physician Characteristics.....................................................................725
7.1.4 Physician Profiling................................................................................727
7.2 Im plica tio n s and R eco m m en d a tio n s.................................................I 28
7.3 C o n c l u s io n .................................................................................................. 134
BIBLIOGRAPHY..................................................................................... 137
v
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
L ist o f T a b l e s
Table 1. Distribution of Persons, Visits, Total Cost, and Averages
by Seven Eye-Related Diagnoses........................................................... 68
Table 2. Cost, Visits, and Prevalence Rate per 1,000 Categorized
by Gender and Ten Age Groups for All Eye Care............................... 70
Table 3. Distribution and Annual Rate o f Outpatient Visits for All
Eye Care-Related Diagnoses by Patient Age and Sex in
the United States and Taiwan.................................................................. 72
Table 4. Percentage of Frequency for Visits Categorized by
Diagnosis in Eye Care.............................................................................. 73
Table 5. Percentage of Frequency in Visits Categorized by
Institutions in Eye Care............................................................................ 74
Table 6. Average Cost per Visit, per Person, Total Cost,
Percentage of Cost, Percentage of Patient Volume,
Number o f Visits per Person, and Average Cost per
Person Categorized by Four Levels of Institutions for All
Seven Eye-Related Diagnoses.................................................................76
Table 7. Average Cost per Visit, Average Lab and Surgery Cost
per Visit, Average of Drug Cost per Visit in Seven Eye-
Related Diagnoses Categorized by Four Levels of
Institutions................................................................................................. 78
vi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 8. Comparison of Outpatient Charge ($US) for Glaucoma
Among Seven Countries.......................................................................... 81
Table 9. Prevalence Rate o f Cataracts by Twelve Countries..............................82
Table 10. Prevalence of Cataracts in the United States and Taiw an....................84
Table 11. Prevalence of Cataracts in India, China, and Taiwan............................85
Table 12. Information on Costs, Visits Prior to First Surgery, and
Follow-Up Visits to Cataract Surgery, 1997.........................................86
Table 13. The Percentage o f Patients with Pre- or Postoperative
Visits Categorized by Institutions in Cataract Surgery....................... 87
Table 14. Length of Time Between Last Visit Prior to Surgery and
Surgery Categorized by Institutions in Cataract Surgery....................88
Table 15. Costs Attributed to Cataract Surgery...................................................... 89
Table 16. Patient Age Distribution of Cataract Surgery by
Institution................................................................................................... 90
Table 17. Cataract Surgery Rate Between Taiwan and the U.S............................90
Table 18. Patient Gender Distribution of Cataract Surgery
Categorized by Institutions......................................................................91
Table 19. Number of Ophthalmologists by Regions in Taiwan........................... 92
Table 20. Relationship Between Region and Costs for Cataract
Surgery.......................................................................................................93
Table 21. Comparison of Number of Ophthalmologists, Average
Charge o f Cataract, Number o f Cataract Surgeries
Between Region 5 and 6 Regions Categorized by
Institutions................................................................................................. 94
vii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 22. Ratio of Gender in Cataract Surgery Performance in
Regions..................................................................................................... 95
Table 23. Number of Cataract Surgeries Performed Categorized by
Four Institutions........................................................................................95
Table 24. Cost Comparison Between Patients Changing
Ophthalmologists or Without Changing
Ophthalmologists, Counting the Number of Pre- and
Postoperative Visits in the Second Cataract Surgery...........................97
Table 25. Volume of Visits Categorized by Physician Gender and
Institutions................................................................................................98
Table 26. Cost Component per Visit Categorized by Physician
Gender...................................................................................................... 99
Table 27. Distribution of Visits Across Physician Gender and Age
Groups...................................................................................................... 99
Table 28. Distribution of Physician Age Across Gender.....................................100
Table 29. Average Cost per Person and Cost Component per Visit
Categorized by Physician Age.............................................................. 101
Table 30. Number of Physicians in Various Institutions Across Age
Groups.....................................................................................................102
Table 31: Multivariate Statistics (GLM) Testing Cost ($NT) per
Visit Influenced by Various Independent Variables..........................104
Table 32. Scenario 1: Comparison of the Cost per Visit for 65-Year-
Old Patients Between Medical Center and Physician
Office in Glaucoma Provided by 45-Year-Old Male
Physicians3.............................................................................................. 108
Table 33: Scenario 2: Comparison of the Cost per Visit Between
Medical Center and Physician Office in Glaucoma to
Male Patient Age 65 Provided by 45-Year-Old Female
Physician3............................................................................................... 108
viii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 34. Number o f Physicians by Annual Volume o f Visits for
Eye C are.................................................................................................. 109
Table 35. Profile of the 17 Physicians with the Highest Number of
Visits.........................................................................................................111
Table 36. Distribution of Patient Re-Visit Rate to Same Physician....................112
Table 37. Top 13 Physicians with High Charge per Visit....................................113
G r a p h ic
Graphic 1. Cost per Person and Prevalence Rate in Eye Care,
National Health Insurance of Taiwan, 1997..........................................70
:x
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A b s t r a c t
Taiwan implemented its National Health Insurance program (NHI) in 1995.
It covers 20 million people; the enrollment rate is 96%. NHI uses a single-pipe,
fee-for-service payment system. All physicians practice at four levels of
institutions: medical centers, regional hospitals, district hospitals, and physician
offices. NHI has a low co-payment, easy access to all facilities, and no referral
system.
Health expenditures grew by more than 10% in each o f the past four years
while NHI’s premium revenues increased 2.36% yearly between 1995 and 1998.
The average outpatient visit is 15 per beneficiary per year. In 1998, NHI ran a
deficit due to slow reform action and low premium collections.
This dissertation is based on Taiwan’s NHI outpatient claim data for
calendar year 1997. The analyses are limited to services related to eye care.
Differences among physicians in medical care practice patterns have been well
documented. Various practice settings, geographic variations, and physician
characteristics should explain the treatment and charge variations in addition to the
patients’ clinical severity and characteristics.
x
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The study’s conceptual framework is based on the supplier-induced-demand
model, cognitive model, behavioral model, and sociological model. Statistical
techniques utilized are paired t test, chi square, and general linear regression. Four
components (all eye care study, cataract surgery episode study, physician
characteristic study, and physician profiling) comprise the study. The study’s
purposes are to provide detailed baseline information on utilization and expenses.
Additionally, comparative data with other countries have epidemiological,
economic values. Application of the findings acknowledges how well NHI meets
beneficiaries’ needs. In addition, the findings can be applied to policy to control
rising health costs through utilization review and physician profiling.
The major findings are
1. The volume of cataract surgery rate matches the WHO’ standard.
2. The charge per visit is much higher in medical centers than physician
offices after controlling the patients’ ages and genders.
3. Region five has highest density of ophthalmologists explaining the high
volume of cataract surgery compared to other regions.
4. Male physicians charge 2.8% higher than female physicians.
Dissertation Committee
Glenn M elnick, Ph.D., committee chair
Blair LaV onna Lewis, Ph.D.
Denise G lobe, Ph.D.
xi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1. In t r o d u c t io n
Differences among physicians in medical care practice patterns have been
well documented. Variation can result from financial motivation, fear of
malpractice and uncertainty inherent in medical care (Eisenberg 1986). While
proposals to change the physician’s practice patterns have been suggested, they
may be too simple to shed light on why some medical practices need to be changed.
Many factors influence medical decision-making, including the payment and/or
reimbursement policy; patient’s medical condition; patient’s demographic data;
physician’s interest in finance, research or relationships with the patient;
physician’s characteristics; practice setting; and geographic area. Since we always
lack the most important factor of information on a patient’s severity of illness in the
medical resource utilization review, for fully capturing the variations in resource
use, we shift the focus to physicians who are largely responsible for determining
the resource use.
Given that medical utilization is largely determined by the physicians’
decisions for the diagnosis and treatment o f patients, it is important to understand
the factors that contribute physicians’ practice variations. The aggregated analyses
1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
in the literature allude to the relationship between practice styles and level of
uncertainly related to the appropriate use of certain procedures that contribute to
geographic variations among geographic areas (Chassin et al. 1986, 1987). A
second approach, used by health plan and provider groups to profile physician
practice styles, has focused on individual physician variations to identify physicians
whose resource consumption was outside the range of peers or benchmarks for
specific conditions (Welch et al. 1993; Kassirer 1994). While individual physician
profiling provides a useful management tool in an organizational setting,
commonalties that define physician groups, such as specialty training or practice
setting, can also be informative for broader policy applications (Phelps et al. 1994).
As Piland and Lynam (1999) state, the most effective way to improve
quality or value is to reduce unexpected or undesired variations in the delivery of
healthcare services. The study presented in this dissertation is based on Taiwan
National Health Insurance Outpatient Eye Care Claim data for the calendar year
1997. Besides total eye care, the study focuses on cataracts in particular since they
are the most common cause of blindness in the world and consume a big chunk of
medical resources of eye care. This dissertation presents a theoretical and
empirical framework for studying the effects of the fee-for-service payment policy,
lack of referral system, low co-payment, generous reimbursement reality, and easy
access in Taiwan’s medical care environment. Patient demographic characteristic,
practice setting, geographic area and physician characteristics are also examined.
2
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The theoretical model considers how supplier-induced demand in the fee-for-
service insurance environment can affect resource allocation. Since physicians
control most of the medical resource use, the physician’s cognitive, behavioral, and
sociological model can explain how physicians practice differently based on those
factors. Analysis of the models suggests an empirical test of whether or not various
physician practice patterns are related to payment policy, the patient’s demographic
characteristics, practice setting, and geographic area, and whether their own
characteristics in eye care create a demand for excessive services which ultimately
lead to economic waste.
Physicians’ variations in resource use for managing similar conditions
provide valuable information to decision makers on how to manage medical
resources to produce the most cost-efficient care. Additionally, it should be
extremely useful because this study on eye care uses a national health database
instead of a sampling database. It should provide plenty of information for other
countries interested in epidemiological and economic comparisons.
There are 1,223 (20% female) board-certified ophthalmologists in Taiwan
(increasing 50 new ones every year). More than 40% of them work in medical
centers or regional hospitals with the responsibility of resident training. Twenty
percent work in district (community) hospitals or joint clinics, which are more or
less like the group practice model in the United States. The other 40% work in sole
practice with various staff sizes, including pharmacist, nurses, technicians, and
3
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
administrators. The practice styles are very different in these three groups of
ophthalmologists. In general, the ophthalmologists working in the tertiary hospitals
have more responsibility in teaching and research. The sole practitioners put all
their hours (sometimes up to 65 hours per week) into clinical services and
operations.
Since 1998, the National Health Insurance program in Taiwan has run a
deficit. Therefore, all the findings demonstrating where excess services exist or
abnormal services exist should assist the Government in Taiwan in making more
feasible policies that protect scarce medical resources.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2. B a c k g r o u n d a n d S ig n if ic a n c e
2.1 National Health Insurance of Taiwan (NHI)
2.1.1 Information about Beneficiaries, Enrollment and Contracted
Providers
Taiwan implemented National Health Insurance (NHI) in 1995. The NHI
covers 20 million lives. Three years after the NHI program was implemented, the
enrollment rate was 96%. For management purposes, the Bureau of National
Health Insurance divided beneficiaries into six regions based on geographic
allocation. They are Region 1, the Taipei Branch, which includes 7,403,981
beneficiaries; Region 2, the Northern Branch, which includes 2,684,177
beneficiaries; Region 3, the Central Branch, which includes 3,778,125
beneficiaries; Region 4, the Southern Branch, which includes 2,955,508
beneficiaries; Region 5, the Kao-Ping Branch, which includes 3,172,809
beneficiaries; and Region 6, the Eastern Branch, which includes 497,717
beneficiaries. A total of 15,872 (91%) hospitals and physician offices nationwide
have a contract with the Bureau. Among these institutions are 14 medical centers
5
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(usually with more than 600 beds), 56 regional hospitals (usually with more than
200 beds), 534 district hospitals (usually with less than 200 beds) and 15,268
clinics (also referred to as physician office). All inpatient and outpatient treatments
are rendered in these four institutions as medical center, regional hospital, district
hospital, and physician office.
All Taiwan citizens who have established a registered domicile for at least
four months in the Taiwan area should be enrolled in NHI. Active military
personnel and the incarcerated are excluded from NHI, since their medical care is
provided by a different delivery system. Foreigners possessing Alien Residence
Certificates may enroll in NHI through their employer or their spouse. The
premium charged is based on income. The employer contributes 70% for private or
not-for-profit workers and the government contributes 60% for governmental
workers and 100% for low-income households and veterans.
2.1.2 Benefit Coverage o f National Health Insurance o f Taiwan
NHI provides beneficiaries with comprehensive medical services at
contracted medical care institutions in the event of illness, injury, or childbearing.
The insurance benefits encompass most kinds of treatment, including care needed
for illness or injury such as examinations, checkups, tests, consultations, surgery,
medications, medical supplies, therapies, nursing, and hospitalization. To meet the
needs of all kinds of beneficiaries, payments are made for diversified services in a
6
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
wide range of categories: outpatient services, inpatient services, Chinese medical
services, dental services, childbirth, physical therapy, preventive health care, home
care, rehabilitation for chronic mental illness, and so forth. Nearly all provider
payments are made on a fee-for-service basis, except some procedures paid by a
prospective payment system, such as appendectomy, Caesarian operation, and
natural delivery.
2.1.3 Economic Background in National Health Insurance o f Taiwan
In 1997, the average income for all persons covered under this form of
health insurance was $12,423 per capita in the year of 1997 (National Income per
Capita 2000). Only 5.5% of the gross domestic product are spent on health care in
Taiwan compared with the US (13.9%), Canada (9.2%), Germany (10.7%), and the
UK (6.8%) in 1997 (OECD 1998). Thus, the average annual per capita health care
expenditure in Taiwan is only $708 compared to $4,095 in the USA, $1,831 in
Canada, $2,733 in Germany and $1,482 in the UK.
NHI’s total medical expenditures in 1997 were US$8 billion, or an average
of US$400 for each beneficiary (NHI Profile 1998). Outpatient expenses and
inpatient expenses accounted for 67.72% and 32.28%, respectively.
The premium collection is based on payroll. There are 29 payroll scales,
which begin from US$512 to US$1,847. The percentage of NHI premium is 4.25%
of the payroll. The employers or government share from 10% to 100% o f the
7
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
premium depending on which category is insured. Insured have been grouped to six
categories: civil servants, teachers, and self-employed as category one;
occupational union members are category two; farmers and fishermen are category
three; military dependents are category four; low-income households are category
five; and veterans, veteran dependents, and the unemployed as category six. The
amount of payroll scales is adjusted based on the actual salary. The percentage of
payroll contributed to premium will be adjusted also based on the actual NHI
expenditures. However, health expenditures grew by more than 10% in each of the
past four years while NHI premium revenue increased 2.36% yearly between 1995
and 1998 (NHI Profile 1999).
2.1.4 Payment Policy in National Health Insurance
The payment model in Taiwan is a single pipe, fee-for-service model. Fee-
for-service has been criticized because it provides financial incentives for
physicians to do more and earn more (Feldstein 1970). The Bureau of National
Health Insurance has tried different ways to contain the cost, such as increasing co
payment, utilizing case payment, and global budgeting in dental service.
Nevertheless, these methods have not resulted in effective cost control. For
instance, people in Taiwan average 14.49 office visits per person per year
(compared to 4.0 in the US). This big difference in utilization rates could be
caused by the fee-for-service payment policy, lack of referral system, low co-
8
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
payment, easy access, and generous reimbursement policy in tertiary hospitals in
Taiwan.
In the fee-for-service payment system, providers often find it financially
rewarding to do more, such as see more patients, do more tests, and perform more
operations to make more money (Hillman et al. 1989). The payment policy in
Taiwan hasn’t set any limitation on the availability o f medical services for patients.
Patients can go to western medicine service in the morning and to Chinese
medicine service in the afternoon in a same street without any appointment or
referral needed. Since medical services in Taiwan are both very convenient and
very accessible, the fee-for-service payment policy can lead to abuse and fraud both
by patients and physicians.
2.1.5 Expenditure o f National Health Insurance in Taiwan, 1997
Outpatient visits account for 67.7% of the total National Health Insurance
expenditure in Taiwan. Total eye care expenditure in 1997 was US$296.4 million
in Taiwan compared with USS 16 billion in the USA (DHHS 1987). In Taiwan,
US$264 million was spent on outpatient eye care, accounting for 4.43% of total
outpatient expenditure. Inpatient eye care expenditure was US$32.4 million, or
4.26% of all inpatient expenditure. Outpatient visits accounted for most of the
expenditures in eye care (89%) (NHI Profile 1998).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Despite the growth in the use of ophthalmologic care over the past decade,
little is known about the use of eye care services and patterns of physician contact
across population sub-groups. As the population grows older, such information is
crucial in planning strategies for treatment and prevention of eye disorders as well
as for identifying potential problems such as access issues and availability of eye
care services.
2.1.6 Distribution o f Ophthalmologists in Taiwan
In Taiwan, ophthalmologists along with physicians in general and family
practice, pediatrics, internal medicine, and obstetrics and gynecology are among the
five most visited physician specialties (Hsu 1999). This can be partly explained by
the rapid growth of ophthalmologists in Taiwan from only 779 in 1990 to 1,223 in
1997, for an average of 5.8 ophthalmologists per hundred thousand population
compared with the ideal number between 5.0 to 5.5 per hundred thousand (Hsu
1999). In addition, there has been a significant growth in the number of female
ophthalmologists leading to a disproportionate share o f females in ophthalmology
compared to other medical specialties. As of 1990, the female ophthalmologist
comprised 15.9% of all ophthalmologists. By 1998, this figure rose to 21.3%.
These numbers are particularly high compared to the overall average number of
female physicians in 1990 (6.5%) and 9.3% in 1998 (Hsu 1999). One explanation
is that eye care work is relatively lighter than other specialties and attracts more
10
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
physicians, especially female ones. Another major reason is NHI’s fee schedule.
The NHI overall payment in eye care is quite generous compared to other
payments. In effect, the volume of ophthalmologists shows that while the
distribution of ophthalmologists across the region may be an issue, the absolute
number of ophthalmologists is not a real concern in Taiwan. For example, while
there is an average of 13.4 ophthalmologists per 10,000 in the capital of Taipei
City, there is roughly only 0.9 per 100,000 in the rural area of Chia-I County (Hsu
1999).
2.2 Cataract Surgery
Cataracts are the most common cause of blindness in the world, accounting
for nearly half of 40 million of the blind (Sommer 1996). This is especially
startling because cataract extraction is one of the most common surgical procedures
performed in elderly people in developed countries. In the United States,
approximately 1.5 million cataract operations were performed by Medicare
beneficiaries in 1996 (Schein et al. 2000), making it the single largest expenditure
and cost for the Medicare system. Under the Health Care Financing Administration
(HCFA), Medicare spends about $3.4 billion per year for the treatment of patients
with cataracts. In addition, over $150 million is spent for preoperative
ophthalmologic diagnostic tests and $204 million on postoperative ophthalmologic
diagnostic tests per year (Schein et al. 1993 and Schein et al. 2000).
11
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Cataract surgery, which involves the implantation of an intraocular lens, is
extraordinarily safe and highly effective, providing tremendous improvement in
visual acuity and subjective visual function (Applegate et al. 1987; Brenner et al.
1993; Mangione et al. 1994). The technology o f cataract surgery procedure is quite
mature with fast speed, so surgery is almost always done on an outpatient basis
instead of a hospital surgery for cost efficiency in nearly all areas of the world
where it is performed. However, for many people in undeveloped countries who
still cannot afford it, cataracts cause 50% of all blindness.
2.2.1 Epidemiology o f Cataracts
Typically, cataracts are defined as lens opacities associated with some
degree of visual impairment (West and Valmadrid 1995). For epidemiological
studies, cataracts may be classified, according to anatomic location, into nuclear,
cortical, posterior subcapsular (PSC), or mixed types. In an advanced or
hypermature cataract, the whole lens becomes opaque. Differentiation of the
morphologic types is important, as risk factors appear to be different for each type
of cataract (West and Valmadrid 1995).
The prevalence rate is measured at one point in time or over a short period
of time in the study subjects (Kelsey, Thompson, and Evans 1986). For example,
the prevalence rate of cataracts is 2.8% in the worldwide population according to
the Nepal Blindness Survey (Brilliant et al. 1985).
12
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Current evidence suggests that persons over 50 experience a gradual
deterioration in social, psychological, and physical functioning, including various
aspects of vision (Brenner et al. 1993). Conversely, gradual deterioration in vision
is associated with a decline in physical and mental function (Cutler and Grans
1988; Branch et al. 1989). Fortunately, when vision problems arise because of
cataracts, studies have shown that cataract surgery can significantly improve the
quality of life. However, since the volume of cataract surgery has already
increased dramatically around the world over the past 20 years, governments,
regional health authorities, health insurance groups, and ophthalmologists
themselves have recently been questioning whether all cataract procedures
performed are actually appropriate. The discretionary nature of cataract extraction
makes its use susceptible to financial and organizational incentives. For example,
this operation has increased approximately fourfold in just in the United States in
only 10 years (Stark 1989). Over half of all ophthalmic surgeries are cataract
surgeries in many countries, and the cataract procedure has become the most
common elective surgical procedure (Javitt et al. 1993).
2.2.2 The Reasons fo r Rapid Increase o f Cataract Surgery
2.2.2.1 Longer Life Expectancy
First, people are living longer in most developed countries. For example,
the number over 65 years old will more than double in the next 20 years (World
13
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Health Organization 1999). Representative population-based data from the Visual
Impairment Project in Australia show that the prevalence of cataracts increases
dramatically after someone turns 40. At least half of all people who are 70 or older
will have cataracts. Nearly all people who are 90 or older will develop cataracts
(W estetal. 1998).
2.2.2.2 Loosened Criteria fo r Cataract Surgery
In addition to the rapidly increasing elderly population, the recent
popularity of cataract surgery can also be explained by the changing criteria
required for this surgery, making the procedure an option to more people and to
people with less severe conditions. Perfect eye acuity is 20/20 which is noted as
1.0 in some countries. Eye acuity of 6/60 is equal to one-tenth of the vision acuity
of 20/20. The physiological meaning is that the resolution ability o f the eye can
differentiate objects apart down to only one-tenth of normal. That means normal
people can differentiate the chart at a distance of 60 feet, but the patient with 6/60
can only differentiate it at a distance of 6 feet.
While the threshold for cataract surgery with intraocular lens implantation
was set at bilateral blindness in the past, the threshold has loosened from 6/60 to
6/24, then to 6/18 and is now often 6/9 or better (Schein 1995; Klein 1997;
McCarty 1999).
14
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.2.2.3 Safety and Cost Effectiveness o f Outpatient Surgery
The substantial rise in outpatient surgery is also attributed to the safety and
cost-effectiveness of the procedures, which are important because there is no
proven prevention or medical treatment other than surgery for cataracts. Surgical
removal of cataracts remains the only proven therapy that successfully restores
vision in over 94% of persons without other concurrent eye disease (Javitt et al.
1996). Therefore, this technology is widely accepted in developed countries and is
increasingly available in the developing ones. In 1993, the World Bank ranked
cataract surgery in the “most highly cost-effective category.” In fact, it was the
only surgical intervention to achieve that rank (i.e., less than $25 per disability
discounted health year of life added) (Javitt et al. 1993).
2.2.2.4 Affordability o f Cataract Surgery
Cataract surgery is generally only performed on persons who can afford to
pay for the procedure either through insurance, out of pocket, or other payment
methods. In Taiwan and other countries with national health insurance, people are
eligible to get the surgery with little to no out-of-pocket costs. In the USA, most
public and private health plans cover this surgery, including Medicare and
Medicaid. Using the comparison rates for cataract surgery, researchers cited 5,700
cataract procedures per 1 million persons in the USA, 6,300 in Australia, 4,000 in
Sweden, 2,700 in the UK (Sletteberg et al. 1995), and less than 500 per one million
15
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
persons in China (Foster 1999). It is apparent that there is a relationship between
an individual’s socioeconomic status and country’s resources to decide a cataract
operation.
2.2.2.5 Cost and Utility Ratio
Fifth, the cost-utility ratio for cataract surgery is high because cataract
surgery is associated with improved vision function and quality of life rather than
life prolongation. Therefore, its economic return is evaluated in terms of the cost
per quality-adjusted life years (QALYs) saved. Javitt’s (1993) study showed that
regaining functional vision through cataract surgery generated 1500% of the cost of
surgery in increased economic productivity during the first year following surgery.
Cataract surgery was ranked one of the most cost-effective interventions compared
with other common public health interventions like vaccinations for measles, polio,
and tetanus as well as treatment for tuberculosis. It results in an economic return of
US$4,318 per QALY gained (assuming ten-year survival after the surgery)
(Drummond 1992). The QALY value is much lower than other interventions. For
example, hospital haemodialysis is US$33,755 per QALY, breast cancer screening
US$8,384 per QALY, and CABG for mild angina US$30,223 per QALY.
16
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.2.2.6 Backlog o f Cataract Surgery
There are about 8 million cataract operations performed each year
worldwide, yet there are about 10 million people who are awaiting this operation.
Most of the “backlogs” are in areas where the appropriate cataract surgery services
and financial resources are unavailable. The only exception is some
ophthalmologists for controlling their operation volume. They keep partially
eligible cataract patients in their backlog and release them gradually for surgery in
order to keep up their standard and schedule. The World Health Organizations
(WHO) realized the urgent need for such services in some areas and has since
focused on cataract surgery promotion in developing countries. In 1987, the WHO
estimated that the prevalence of cataract-associated blindness (using the definition
that the better eye has visual acuity of <20/400 as legally blind), ranged from a low
of 14 per 100,000 persons in the most highly developed countries to as many as
2,000 per 100,000 persons in parts of the developing world (WHO 1987).
Globally, an estimated 20 million people are blind due to cataracts, creating the
world’s cataract backlog (WHO 1997). India alone accounts for 24% (8.9 million
people) of all cases of blindness worldwide due to cataracts. In addition, the
increasing Iife-expectancy in India means that an additional 3.8 million people
become blind each year due to lack of treatment for cataracts in that country (Singh
et al. 2000). Through its global initiative to eliminate avoidable blindness, the
WHO’s Vision 2020 aims to increase the global volume of cataract surgery to 32
17
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
million per year by the year 2020 to meet these global needs (WHO 1997 and
Pararajasegaram 1999). India was one of the first countries to receive finding from
the World Bank credit to develop the eye camps to pay for cataract surgery. The
target cataract surgery rate set by the WHO for established market economics for
the year 2000 was 3500 new surgeries per year.
With the cataract surgery on the first eye, there is some indication that an
improvement in visual function and in stereoacuity occurs after the second
procedure, suggesting that the best outcomes can be achieved when patients
undergo surgery in both eyes at different times (Castells 2000). However, there is
some debate over the need for surgery in the second eye. With limited cataract
treatment resources available in many areas of the world, perhaps the money
available for this procedure should be designated only to the first eye surgery rather
than to surgery in both eyes (Hanning and Lundstrom 1998).
2.3 Episode of Care
The “episode of care” concept was first defined by Solon et al. (1967) as “a
block of one or more medical services received by an individual during a period of
relatively continuous contact with one or more providers of service, in relation to a
particular medical problem or situation” (p. 402). In essence, an episode of care
refers to a group of health care services provided within a specific time period that
are needed for management of a specific illness or condition.
18
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The specific type o f health care episodes are defined by Hombrook,
Hurtado, and Johnson (1985). The boundaries placed around the events or services
are included within the episodes. An episode can also be described by one o f three
different perspectives: the patient (episode o f illness), the provider (episode of
disease), and the payer (episode of care). For individuals who are ill, the episode
begins when they become sick, not when they seek medical care. However, since
most health services research issues examine the use of health care services, the
patient perspective as described above is not the appropriate perspective to use to
study an episode of care. Advantages of using episodes of care make it easier to
define the beginning and ending of an episode and the ease of examining the use of
health care services (Wingert et al. 1995). The concept of episode of care can be
used as the basis for a medical care price index. As Scitovsky (1985)
demonstrated, an episode of care approach reveals significant changes in resources
used and costs that are not readily apparent through traditional (by each claim)
input measures. Other economic analyses have used episode of care methodology
to evaluate variation in resource utilization between physician and nurse
practitioners (Salkever et al. 1982) and between ambulatory care delivery sites
(Gold 1981). It also has been used to study physician-induced demand (Fahs
1992). In addition, episodes of care methodology have been used for quality
assessment studies (Nutting, Shorr and Burkhalter 1981; Kristensen et al. 1993;
Meyboom-De Jong 1993), cost-effectiveness research (Lasdon and Sigmann 1977),
19
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
reimbursement purposes (Showstack et al., 1987), and to better understand how to
manage of specific conditions (Klinkman, Stevens, and Gorenflo, 1994).
2.4 Geographic Variation
Geographic variation in health care across the United States is a persistent
finding (Welch, Miller and Welch 1993). Contributory causes that have been
explored include differences in practice patterns resulting from physicians’
uncertainty about how and in what setting to treat certain conditions (Wennberg et
al. 1982, 1989), differences in the supply of hospital and nursing home beds
(Knickman 1985), geographic unevenness in the market penetration of health care
organizations (e.g., HMO) whose financial incentives prompt different patterns of
hospital use (Hillman et al. 1989), and differences in access to care, resulting in
variability in the rates of avoidable hospitalizations (Weissman et al. 1992).
The practice style is often driven by financial incentives. Several studies
show that a higher physician-to-population ratio was, in fact, associated with more
visits and that these visits were physician-initiated. Rossiter and Wilensky (1983)
found that 39% of ambulatory visits to physicians and 43% of all visits (inpatient
and outpatient) were initiated by the physician (1.5 visits per person per year).
Root (1983) found that when new general practitioners (some of whom did
surgery) moved into rural areas in Canada, surgical utilization increased by 17%.
20
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
One would expect there to be less geographic variation in hospital use in the
Department of Veterans Affairs (VA) health care system than in the private sector
(Friss et al. 1989). From a sociodemographic standpoint, the VA caseload is
relatively homogeneous. Eligibility for care is set by federal law. In a VA study by
Ashton (1999), a substantial geographic variation existed in hospital use for all
eight cohorts of patients and all the years studied. Variations in the numbers of
hospital days per person-year among the networks were greatest among patients
with chronic obstructive pulmonary disease during a given year and smallest
among patients with angina. Levels of hospital use were highest in the Northeast
and lowest in the West. The variation in the rates of clinic visits for principal
medical care among the networks ranged from a factor of approximately 1.6 to a
factor of 4.0; variations in the rates were greatest among patients with chronic renal
failure and smallest among patients with congestive heart failure.
Large variations in the medical and surgical hospitalization rates of adults
have been described across rural and urban areas (Wennberg et al. 1982). The rates
of hospitalization of children have been examined less carefully. Perrin et al.
selected three urban communities (Boston, Rochester, New Haven) where teaching
hospitals provide most of their hospital-based care. In addition, the rates of
hospitalization varied as much for conditions for which there is little discretion in
deciding on admission (e.g., bacterial meningitis) as for those in which there are
more choices. The results show that Boston children were hospitalized at more
21
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
than twice the rate of Rochester’s for most medical diagnostic categories. The rate
for New Haven was intermediate.
Javitt et al. (1995) conducted a study of geographic variation in utilizing
cataract surgery. In this study, besides regional variation, personal variation also
was examined. This was first analysis in which the explanatory power of a regional
variation model had been compared to that of a person-based model to examine
personal and regional covariates associated with utilization of cataract surgery.
Compared with the geographic variation in providing other surgical procedures, the
variation in cataract surgery across large geographic areas observed in this analysis
was relatively low. A person-based analytic strategy offers considerable
advantages over traditional small-area-variation approaches in measuring the
association between personal characteristics and likelihood of receiving particular
health services. The person-based model demonstrates that increased likelihood of
undergoing cataract surgery was associated with increasing age from 65 to 94
years, the white race, and living in a zip-code area with a mean income greater than
$15,000. In addition, the decreased likelihood of undergoing cataract surgery
among black Medicare beneficiaries was a concern, particularly in light of data
showing a fourfold higher prevalence of cataract blindness among black
Americans.
22
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.5 Variation of Resource Use and Physician Practice
Styles
Physicians do not all practice alike. When one physician might suggest
surgery, another might prefer medical treatments. When one physician suggests a
very expensive lab test, another might prefer observation for a few weeks. There
is no right or wrong because medical knowledge is ambiguous and few services are
absolutely necessary. In other words, physicians have few ironclad rules for
practicing medicine (Eisenberg 1986). Often, there are two situations concerning
physicians’ clinical practice. First, some clinicians have unusual practice habits
compared to the norm. Second, physicians’ practice patterns are the same as their
colleagues’, but who may deviate from an optimal pattern as defined by expert or
scientific evidence.
Researchers have investigated the variation in the rates of service use and
identified the variation in physician practice styles as one of the most logical
starting points for explaining both the high cost o f health care and its
appropriateness. When all usual explanations have been controlled for, variation
still exists; there is no explanation for the difference except the practice styles of
individual providers. According to studies, physician practice patterns can be
influenced by financial incentives, practice setting, and personal characteristics.
23
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.5.1 Physician Profiling
Some unexplained variations in practice styles among physicians suggest
that some medical care may be inappropriate, especially under certain payment
systems as fee-for-service. Indeed, when explicit criteria have been applied to
medical practice, instances of both overutilization and underutilization have been
found. Even when differences in patient preference, case mix, and severity of
disease have been considered, there are potential improvements in both the cost and
quality of care that might result from changes in the practice pattern of some
physicians (Eisenberg 1987). John Wennberg and colleagues have demonstrated a
significant variation in physician practice patterns (Wennberg et al. 1982). This
research on variation is the underpinning o f physician profiling. By identifying
sources of meaningful variations in practice patterns, physicians can be profiled to
distinguish best practices. Insurers can then provide quality reimbursements to the
best practices and implement quality improvement programs targeting the practices
that fell short o f that mark (Cyr and Malloy 1999). Physician profiling is the first
step to understanding practice pattern, upon which the practice guideline can then
be established. Through the practice guideline, physician behavior can be
monitored in a scientific way. The ultimate goal is to utilize the scarce medical
resources more effectively.
24
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Creating physician profiling is a way to monitor physician practice pattern
and is an educational tool to improve outcome, reduce variation, enhance
performance, and lower costs.
2.5.2 What is Physician Profiling?
Profiling is an epidemiologic technique that focuses the assessment of
health care delivery on patterns of care rather than on individual occurrences of
care. Information obtained from large databases is used to identify a provider’s
pattern of practice and compare it with those of similar providers or with an
accepted standard of care. Moreover, profiling enables comparisons between the
current performance of individual physicians against what could be achieved if
physicians followed evidence-based clinical guidelines (Boren 1999) and other
quality improvement initiatives.
Theoretically at least, profiling can be an effective and relatively
nonintrusive way to identify over- and underutilization of services, to uncover
problems with the efficiency and quality of care, and to assess provider
performance. This makes it an attractive and potentially useful tool for health care
professionals, patients, payers, medical educators, and policymakers (Lasker et al.
1992). Physician profiling is a somewhat relative term because physicians, payers,
patients, and purchasers all use the same term while sometimes referring to
different definitions within and between the groups. Terms such as “profiling,”
25
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
“report card,” and “best practices” have been commonly used labels for a new
methods of measuring physician performance in health care (Gevirtz 2000).
Regardless of the exact term used or even the exact definition, physician profiling
refers to the collection o f data to analyze physician performance and document
patterns, utilization, and progression of care over time (McNeil et al. 1992; Sandy,
1999).
2.5.3 The Objective o f Profiling
The ultimate objective o f profiling is to obtain the best outcomes with the
most efficient processes. It can educate physicians and create accountability by
eliminating unexplained clinical variation and promoting uniform nation-wide
standards of care. Profiling can also be shared with physician associations,
insurers, and the public, who are demanding an end to unexplained clinical
variation (Chesanow 1999). Thus, this public benefit comes from understanding
the strengths and limitations of certain physicians and can allow people to select
physicians based on the documented quality of their services.
Actually, physician profiling has different objectives in which the profile
designing should accompany the objective. Currently, three objectives have been
identified.
26
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.5.3.1 Clinic M easurem ent
The clinic measurement is different from resource management. Clinic
measurement is concerned with quality improvement and outcome. In health care,
the problem is indicated by poor outcome. For improving quality, physicians have
to change their performance. Profiling can theoretically play a role in this concern.
First, it can be used to target potential problem areas by identifying conditions or
procedures where there are large variations in outcome. Such profiles would
measure differences in the frequency with which specific outcomes occur when
patients with a particular condition or patients undergoing a particular procedure
are cared for by different providers. Second, profiling can be helpful in determining
how and by whom performance should be changed to improve outcomes (Lasker et
al. 1992). Third, physicians can use profiling as an educational tool. For instance,
physicians ask themselves questions such as “Are patients of a certain type who are
supposedly receiving similar care achieving similar outcome?” (Nadzam et al.
1997). Physicians use this information to change their inefficient and ineffective
practice patterns to improve both the quality and delivery of care they provide.
2.5.3.2 Econom ic M easurement
Profiling has traditionally focused on the quantity of services delivered,
expressed either in dollars or in units of service. Money amount is typically used
when the objective is profiling resource management. Economic efficiency refers to
27
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a population-based, case-mix-adjusted approach to analyze the overall financial
impact of the manner in which physicians practice medicine. There are two classes
of economic measurement. The first measures the ratio of provider efficiency to
productivity. Operational outcomes such as efficiency and productivity measure
quantity in relation to resources used, such as staff minutes/in-clinic Relative Value
Units (RVUs) (Gevirtz 1999). The second one is the gross revenue the physician
generates. The above two manage the appropriateness of physician reimbursement
and scrutinize the efficiency of physician practice.
2.5.3.3 Patient Satisfaction
In addition to clinic measurement and economic measurement, patient
satisfaction— as consumer-driven quality of care measurement— is becoming more
common today. Patient satisfaction profiles measure the quality o f care received
without regard to resources used. Profiles meet requirements for accreditation, like
compiling the Health Plan Employer and Data Information Set (HEDIS), or seek to
reinforce employer group satisfaction and member retention. RAND’s 1960s
Health Insurance Experiment found that patient satisfaction clearly demonstrated
the validity of health care quality. The data demonstrated a strong association
between patient satisfaction and continuity of care. For a one-point decrease on a
general satisfaction scale, there was a 3.4 percentage increase in the probability of
an individual’s changing health providers (Marquis 1983). In the United States,
28
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
health plans and physician groups commonly use satisfaction ratings to profile
physicians and impose financial incentives based on their findings. Moreover,
regulators now recognize patient satisfaction as a legitimate measurement of health
care quality (Ross et al. 1995). Research is beginning to illuminate the connection
between satisfied patients and quality health care (Fleming 1981; Carmel 1985).
2.5.4 The Criteria o f Establishing Profile
To establish profile, the reliability, validity, and relevance of criteria are the
required. Thus, profiling requires three types of information: first, clinically
meaningful indicators of the process or outcome of care; second, information to
identify similar providers or to account for important differences across providers,
such as differences in the patient case mix; and third, data that can be used to set
utilization and outcome rates reflecting appropriate, cost-effective care. The ability
to develop valid and relevant profiles, therefore, depends on the availability of
accurate and reliable sources of these types of information (Brand and Quam 1992).
For example, for making a fair comparison across physicians, the profile must be
risk-adjusted for differences in patient populations. Case-mix adjustment is a
statistical way to allow meaningful comparisons among providers based on their
patient population. The purpose is to correct for differences in the distribution of
one or more confounding factors (health risk factors) in the patient populations
29
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
which are being assessed in order to measure differences in treatment effectiveness
of their respective physicians (Gevirtz 1999).
2.5.5 How to Encourage Physicians to Buy into Profiling
Profiles only effectively improve the delivery and outcomes of health care if
physicians accept them as effective sources of information. Successful profiles
create desired incentives and change behavior in measurable ways. There are three
criteria. First, actionable by physicians, the profile must be quantifiable and address
an area that is under the direct control of a particular physician. Second, relevant to
the physician, the information profiles should be useful for improving the quality of
care. Third, meaningful, the measures should help the provider understand the
clinical and economic significance and differences between the provider’s
performance and benchmarks (AMA 1999).
The special importance is to invite physicians into the development process.
The key to compliance and enhancing physician performance is to gain their
support and acceptance by involving physicians in any profiling planning meetings.
This helps physicians become more comfortable with profiling instead of leaving
them angry, defensive and less motivated to try and improve the quality of health
care (Bindman 1999).
Profiles should not be used as part of contract negotiations (Hofer et al.
1999). The most successful profiles select realistic, achievable, and appropriate
30
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
outcome measures. The most effective profiles promote best clinical practices with
minimal provider burden.
2.6 Research Questions
This study addresses the following questions:
1. How do Taiwan people utilize National Health Insurance in outpatient
eye care?
2. Do Taiwan people have similar eye and cost consumption problems
than other countries?
3. What is the treatment for a cataract surgery episode in Taiwan?
4. Do patient demographic data explain the variation of resource use after
controlling for institution and geographic region?
5. Do physician characteristics explain the variation o f resource use after
controlling for patient, institution, and geographic region?
6. Does physician practice setting explain the variation o f resource use
after controlling for patient characteristics, physician characteristics,
and geographic region?
31
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7. How do various geographic areas make the difference in distribution of
ophthalmologists and rates of cataract surgery?
8. What does ophthalmologist profiling look like in Taiwan?
32
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3. C o n c e p t u a l F r a m e w o r k
In this study, the payment policy, cost consumption, patient characteristics,
physician characteristics, and practice setting issues all cause variation of
resource use. In an ideal world, physicians are technicians par excellence of the
human body whose medical knowledge and scientific methods foster a rational
approach to the management o f disease. Variations in management decisions
under this belief would be attributable solely to the clinical differences among
patients. However, research in the last 30 years has shown that physicians are
sensitive to a multitude of other stimuli in the process of decision-making.
Eisenberg (1979) synthesized early research on physician behavior to
demonstrate that physicians respond to non-clinical factors. He postulated that
decision-making was subject to sociological influences which he characterized
as the sociological characteristics of patients and physicians, the physician’s
interpersonal relationship with the patient, the physician’s interaction with his
profession, and the structure o f the health care system. The following three
theories can explain physician behavior.
33
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.1 Model of Supplier-Induced Demand
Under the traditional fee-for-service payment model, most physicians are
paid retrospectively for services already rendered. Rarely do patients, insurers, or
employer groups question the physicians’ performance or the claim associated with
the services rendered. This payment model rewards physicians for providing more
care, more services, and more tests. Physicians are regarded as multiproduct firms
that maximize profits and whose products include different types of operations,
physician visits, and consultations (Escarce 1993). In the supplier-induced demand
model, the physician, however, is a supplier of medical services. As such, there is
a concern that because the physician has a financial interest in supplying more
services, the information provided to the patient can be biased. When the physician
modifies the diagnosis and treatment recommendation to include a favorable effect
on his own economic well being, then the physician is an “imperfect” agent. In a
fee-for-service payment system, this imperfect agency relationship has been
referred to as “supplier-induced demand” (Feldstein 1970). The greater the
quantity of induced demand, the greater the physician’s income. One variant of
how much demand a physician will induce is based on what “target income” the
physician desires (Feldstein 1970). The target income is often related to the local
income distribution, particularly with respect to the income of other physicians in
their community.
34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.2 Factors Affecting Physician Decision-Making
Fineberg and colleagues (1985) summarized the empirical literature by
pointing to three underlying theories pertaining to the determinants of physician
decision-making: cognitive, behavioral and sociological factors.
3.2.1 Cognitive Factors
In the cognitive model, the physician acts primarily as a processor of
clinical information and patients’ desires. Variation in practice style in this model
would be due to factors such as errors in perception of the patients’ desires,
inaccurate or incomplete information on the disease condition, or in the way that
the physicians process the information.
3.2.2 Behavioral Factors
Physicians are highly motivated individuals due to their professional
requirements. In the medical field, physicians have to deal with ambiguity and
operate independently. Under these circumstances, physicians should believe that
effort will lead to good performance, and good performance will lead to certain
outcomes or rewards. This fits the expectancy theory. Thus, for a physician to be
motivated, the outcome or reward must be valued by the person, she or he must
believe that additional effort will lead to higher performance, and higher
performance will subsequently result in greater rewards or outcomes (Bowditch
35
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and Buono 1994). The supplier-induced demand economic model explains how
physicians motivate themselves to meet personal income target. They make efforts
to generate more patient visits, testing, prescriptions, and surgery and expect to get
great financial reward.
3.2.3 Sociological Model
Fineberg and colleagues (1985) proposed that physicians are regarded as
members of social groups, rather than as individuals in the theory o f operant
conditioning and reinforcement theory that human behavior is determined and
maintained by the person’s environment. Although behavior may be random at
first, as an individual explores the environment and reacts to it, certain behaviors
are reinforced and subsequently repeated (Skinner 1969). Social learning theory
can explain their behavior even better. In this theory, physicians’ behavior is
conceptualized as an integration of the environmental determinism o f operant
conditioning and organizational behavior modification (where behavior is viewed
as a function or external cues and consequences) and the self-deterministic view of
traditional motivation theory such as Maslow, Alderfer, and McClelland (where
behavior is caused by internal needs, satisfaction, and/or expectations). Essentially,
social learning theory is viewed as a function of internal and external cues and
consequences and cognitive functioning (Kreitner and Luthans 1984). This
approach is based on the premise that physicians acquire much of their behavior by
36
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
observing and imitating others within a larger social context. Such modeling,
however, is not a one-way process as physician behavior and environment
influence each other.
3.3 Variables Affecting Physician Decision-Making
According to the theoretical framework proposed by Clark et al. (1991), the
most important variables influencing physician behavior include patient
demographics, physician characteristics, practice setting, and the health care
system.
3.3.1 Patient Demographic Characteristics
From an epidemiological perspective, patient demographics play a role in
clinical decision-making because of the relationship between disease susceptibility
and genetics. However, in many societies, factors other than demographic
differences, such as social characteristics and affluence, play just as important if
nor a more important role in clinical decision-making.
3.3.1.1 Patient Age
Several studies have shown that physicians prefer to treat younger patients.
It has been suggested that this bias exists because older patients tend to be sicker
and are less likely to have a successful treatment outcome compared to a younger
37
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
population. Thus, many physicians have a different attitude about treating older
patients. This bias has been shown in studies demonstrating that (1) older patients
spend less time with their doctors during a clinical encounter than their younger
counterparts (Keeler et al. 1982), (2) physicians choose different prescribing
patterns for the same condition based on age (Linn and Linn 1994); and (3) the
elderly are likely to be referred to another physician (Ifiidu et al. 1999)
3.3.1.2 Patient Gender
Another factor that tends to influence physician decision-making is patient
gender. The impact of gender on clinical decision-making (both assessment and
interpretation) comes from physician and patient. From a clinical standpoint, males
have historically been the only participants in clinical research. Thus, medical
textbooks have provided information on medical illnesses and their treatments.
Females expression o f illness-related behavior is expected to be different
and traditionally has been given more emotional emphasis in clinical evaluation
compared to males where the emotional factors are not given as much
consideration. The social roles and cultural expectations affect the social
interaction in clinical encounters. Evidence of this process is found in studies that
gender differences in the prescription of medical care are in excess of the expected
differences in clinical risk (Thurau 1997). One study found that women receive
coronary artery bypass surgery at an older age and at a more advanced disease stage
38
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
than men (Khan et al. 1990). Others have found lower referral rates for coronary
catheterizations, pulmonary symptoms, and psychiatric conditions despite similar
clinical indications as men (Tobin et al. 1987; Wells and Feinstein 1988; Hohmann
1992; Franks and Clancy 1997). A recent study that explored the possibility that
differences are in patient preferences and not due to physician bias surveyed
eligible patients’ willingness for invasive cardiac procedures by gender. The study
found females expressed greater willingness than males to accept a physician’s
recommendation o f cardiac procedures, refuting the hypothesis that the differences
are due to patient preferences (Saha, Komaromy, et al. 1999).
3.3.2 Physician Characteristics
In addition to patient characteristics, individual physician characteristics
such as age and gender also affect how medical decisions are made. Historically,
individual physician characteristics have not been considered a major factor
influencing physician decision-making because there was little variation in
physician characteristics. In recent years, however, the physician workforce has
become very diversified. Therefore, the relevance of age and gender as it relates to
physician decision-making is now important to consider. Formal physician training
and education also influence clinical decision-making because physicians learn
particular patterns and habits from the collective norms of the particular
educational and training facilities. For example, the culture and philosophy of
39
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
some medical schools influence the physician’s practice style. Moreover, the
length of time a physician has been practicing medicine (including formal and
informal training) as well as the ongoing continuing medical education efforts
(such as updating equipment and attending training seminars to stay up to date on
prevailing issues) are likely to influence the physician’s decision-making style.
3.3.2.1 Physician Age
Physicians who lacked the clinical maturity may acquire it with age and
experience. Usually, the different experience levels create different practice
patterns (Eisenberg 1986). There is an overlap in the underlying sociological
process of clinical experience with physician age. For example, Eisenberg (1986)
found that younger physicians are more likely to prescribe and more likely to
provide services overall. A study on the use of diagnostic tests found that the use
of X-rays and laboratory tests decreased as physician age increased (Eisenberg and
Nicklin 1981). Rosenblatt and Moscovice (1984) reported that older physicians
are also more likely to hospitalize their patients, regardless of the patient’s age.
Other studies have shown that newer graduates are more empathetic (Barnsley et al.
1999), more aware of, and more likely to use, recent innovations (Becker 1970),
and prescribe medications appropriately (Stolley et al. 1972; Hemminki 1974).
40
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.3.2.2 Physician G ender
The effect of physician gender on the social context of the clinical
encounter has received considerable interest recently. For example, male and
female physicians have very different communication styles (Roter and Hall 1998;
Hall et al. 1994). The greater affinity for patient participation in clinical encounters
with female physicians was observed by Bamslsy et al. (1999). Female physicians
spend more time with their patients and are more likely to engage in a collaborative
relationship. Collaborative styles have been known to foster greater satisfaction
and greater compliance among patients, leading to better outcomes (Weisman and
Teitelbaum 1985; Kaplan et al. 1996; Bertakis et al. 1995). In addition, female
physicians tend to promote the use of female-specific preventive services such as
mammograms and pap smears (Lurie et al. 1997) as well as specific therapeutic
interventions such as estrogen replacement therapy (Seto et al. 1996) more than
male physicians. While the evidence suggests differences in practice styles, the
effect of physician gender on resource use decision in non-preventive medical care
is not clear.
3.4 Practice Setting
The physician’s practice environment can also influence physician decision
making. For example, physicians are influenced by both formal and informal
interactions with other physicians, as well as the practice setting and the
41
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
organization of the health care system (Eisenberg 1979). From a sociological
perspective, Friedson (1970) suggested that in addition to the peer interactions
there are two types of medical practice influenced by setting, each providing a
different level of professional and peer level interaction and socialization. First,
there is the client-dependent practice model. In the client-dependent model, the
physician is isolated from professional colleagues and decision-making is primarily
influenced by the physician-patient interaction. Next, in the colleague-dependent
model, interaction and cooperation with colleagues are considered the practice
norm. In this model, individual physician decisions must meet the approval and
acceptance of colleagues so the practitioner feels a sense of belonging to the
“group” practice. Formally organized practices are more likely to face peer
pressure or other internal professional standards that contribute to a “group”
practice style. Thus, physicians in solo practices may behave differently than
physicians who belong to group practices because they do not have any external
influences or pressures.
Empirical evidence on the relationship between group practice styles and
resource use is limited. The Medical Outcomes Study (MOS) provides evidence
that there exists practice variation differences between solo practitioners and group
practice practitioners. The study examined both the use of specific ambulatory
services and rates of hospitalization across practice type and practice settings. Solo
practitioners had the highest rate of hospital admissions and prescribed the most
42
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
drugs per patient compared to all other fee-for-service and prepaid group practices.
While the study found that financial incentives played an important role in
decision-making, the presence of formal peer review or regulation found in
organized group practices are also powerful influences on physician behavior
(Pineault 1976; Hillman et al. 1992).
The study adapts the above framework such as supplier-induced demand,
physician behavior, physician characteristics, and patient characteristics to examine
variations in resource use among physicians treating patients for eye-related
diseases and conditions. By analyzing claims data, this study examines physician
practice patterns to determine how they affect payment and reimbursement policies.
The biomedical and epidemiological nature o f eye diseases are examined by
analyzing physician socialization, various geographic regions, patient demographic
characteristics, individual physician characteristics, and type of physician practice
setting.
3.5 Hypotheses
This study addresses the following questions:
I. Do Taiwan people have similar eye, utilization and cost consumption
problems than other countries?
43
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Hypothesis 1.1: The epidemiology of the ocular problem in Taiwan
people is same as that in other countries.
Hypothesis 1.2: In eye care, older people utilize more medical
resources than younger people.
Hypothesis 1.3: Females more eye problems and seek more
treatments than males.
2. How do Taiwan people utilize National Health Insurance in outpatient
eye care?
Hypothesis 2.1: Since the economic development and national
health insurance implementation in Taiwan, people have
more eye treatments than other developing countries.
Hypothesis 2.2: With the fee-for-service payment system and easy
access to eye care institutions, Taiwan people generate more
outpatient visits than the insured in other countries.
3. What is the treatment fo r a cataract surgery episode in Taiwan?
Hypothesis 3.1: The rate of cataract operation is not lower than the
WHO standard after national health insurance
implementation and for economic reasons.
44
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Hypothesis 3.2: Since the intraocular lens has to be imported, the
cost of intraoperative cataract surgery in Taiwan is not much
lower than in the USA.
Hypothesis 3.3: Under the fee-for-service payment system and lack
of rigid monitoring system, the percentage of total cost in
pre- and postoperative surgery id higher than in America.
Hypothesis 3.4: The distribution of age and gender of cataract
patients is the same as in other countries.
Hypothesis 3.5: The rate for second eye cataract surgery is the same
as for developed countries.
Hypothesis 3.6: Patients use the same physicians for the second eye
cataract surgery.
4. Does patient demographic data explain the variation o f resource use
after controlling fo r practice setting and geographic region?
Hypothesis 4: Physicians make significantly different treatments
based on a patient’s age and gender.
5. Do physician characteristics explain the variation o f resource use after
controlling fo r patient, institution, and geographic region?
45
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Hypothesis 5: Physician characteristics contribute significantly to
variations in eye care services after controlling for patient
characteristics, practice setting, and geographic region.
6. Does physician practice setting explain the variation o f resource use
after controlling fo r patient characteristics, physician characteristics,
and geographic region?
Hypothesis 6.1: Medical centers consume higher medical resources
than other institutions.
Hypothesis 6.2: The components of charge (e.g., drugs, lab,
procedure, diagnosis) contribute more significantly to the
variation in medical centers than in other institutions.
Hypothesis 6.3: The number of cataract surgeries is different in the
four institutions.
7. Is there a difference in distribution o f ophthalmologists and rate o f
cataract surgery by geographic area?
Hypothesis 7: The number of ophthalmologists in various regions is
significantly associated with the variation in cataract surgery.
46
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8. What does ophthalmology profiling look like in Taiwan?
Hypothesis 8.1: Certain physicians have a higher than average
number o f charges.
Hypothesis 8.2: Some physicians generate more visits than other
physicians.
Hypothesis 8.3: Some physicians have a relatively higher patient
volume than other physicians.
47
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4. S t u d y P o p u l a t io n a n d D ata S o u r c e s
The analysis is based on the Taiwan National Health Insurance Outpatient
Claim system for calendar year 1997. The analyses are restricted to patients with
eye-related diagnoses.
Eye care was defined as all services paid by claims reporting (1)
ophthalmic-related procedures and (2) ophthalmic-related diagnoses.
Ophthalmologists were specifically coded as such in the physician specialty field,
allowing us to extract all eye care data from the Bureau’s outpatient claims
database. However, the other specialties were included as well, if they claimed for
non-surgical eye care. The total number of outpatient claims for eye care 1997 was
12,989,686. Total payment for eye care in the 1997 was 296.7 million. As high
technology was applied to ophthalmology popularity, 89% of eye care expenditure
was spent on an outpatient basis. Thus, this study only focuses on outpatient
treatment that still captured 89% of all eye care expenditure. Beneficiary visit
frequency was examined using an inclusive definition for what constitutes a visit:
each claim for eye care was taken to represent a separate provider visit, even if on
the same date another claim was submitted for the same patient by another
48
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ophthalmologist or primary care physician. Moreover, a single claim covering
services on multiple days by one or more providers was considered to represent
multiple visits (even if all services were by the same provider). In this study—
categorized by seven diagnoses according to the payment coding book in the
Bureau o f NHI—the prevalence rate was counted separately with different
diagnosis. These seven diagnoses were Glaucoma (230), Cataract (231), Blindness
and Low Vision (232), Conjunctivitis (233), Lacrimal system disease (234),
Strabismus and Eye Movement Disease (235), and others (239; 236 of Retina has
been merged in others). All diagnoses were done in four different institutions.
Part of this dissertation focuses on cataract-related services and code 231.
The cataract surgical cases were identified by cost: if the total amount reported on
the claimed data was greater than USS600, it was considered a cataract surgery.
The fee schedule of cataract operation was US$300, which was only for cataract
extraction and intraocular lens implantation. Another reimbursement was for the
intraocular lens itself and the surgical materials, which were above USS300. There
were a total of 83,582 cases of cataract surgeries in the Taiwan claim data for 1997.
The cases with complications were excluded from the study: vitrectomy,
trabeculotomy, peripheral iridectomy, and removal o f dislocated lens combined
with vitrectomy, etc., which are identified either by procedure code or charge
amount. Similarly, individuals whose cataract surgery was combined with comeal,
retinal, or glaucoma were excluded from the study. A total of 12,416 cases were
49
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
excluded. After these exclusions, there were 59,823 individuals in our sample who
had strong evidence of having undergone at least one extracapsular or intracapsular
cataract extraction during 1997. O f these 59,823 individuals, 11,299 (19%)
underwent a second or third surgery, resulting in 71,166 total cases of cataract
surgery that formed the basis of our analyses. AH cataract surgeries were
conducted in one o f four levels of institutions. In a case o f cataract surgery, there
are pre-, intra- and post-surgery costs. For studying the total cost in one case from
the beginning to the end, we adopted the episode methodology. By combining all
costs related to the procedure into episodes, it was possible to avoid some
arbitrariness in how follow-up charges, tests, drugs, and supplies are priced,
because aggregating over the episode removes some uncertainties.
4.1 Variable Specification
4.1.1 Episode o f Care
4.1.1.1 D efining an Episode o f Care
Using the “episode of care” approach, one can group spending into episodes
that contain all the line charges on the claim forms from providers associated with a
given diagnosis (Keeler et al. 1982). To calculate the spending related to the given
diagnosis, all claims dating back to the first time the patient presented for the
disease or illness for cure were summed together. There are several advantages to
using the episodic approach instead of calculating annual costs for the study of
50
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
cataract surgery (Keeler et al. 1988). First, health care is usually delivered during
an episode of illness. Health care services are obtained for a certain purpose in a
certain sequence and are interrelated with the goal o f resolving a medical problem.
Grouping a sequence of services together into an episode of care captures this
natural history of medical care. Thus, the episode of care or surgery is the
“natural” unit for analyzing the effects of price either on the total number of
procedure performed or the number of people who can afford the procedure.
Second, data related to a specific episode contain more information than data that
add the costs of all the episodes in a year together. Specially, data provide the
information needed to do comprehensive economic analyses. For example, this
information can be used to evaluate the effect of replacing inpatient cataract
surgery with outpatient cataract surgery. Aggregating costs at the annual level does
not provide detailed information such as the timing and intensity of the episode of
treatment. Third, an episode analysis can provide more precise estimates of
transitory changes in demand that have important policy implications, such as
“catch-up” demand in response to new and permanent insurance coverage. An
episode analysis examines behavior within a period of time, especially the first few
months of health insurance coverage under a new policy. Therefore, it has the
potential to detect transitory changes. Forth, episode of care captures the entire
process of care and allows quality issues to be studied more easily (Donabedian
1968; Gamick et al. 1994). The recent emphasis on patient care outcomes and cost-
51
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
effective analysis by policymakers and providers has served to highlight the need
for an episode of care approach. Fifth, the episode of care methodology can
provide useful information that can be used for reimbursement purposes. Payment
based on an episode discourages dividing service packages into ever smaller
discrete units to maximize revenues. Instead, a bundled payment related to an
episode of care provides an incentive to improve the cost effectiveness of care
through better management of resources (Wingert et al. 1995). The cost of each
episode combined with the number of episodes within a certain period of time can
estimate the total spending for that episode.
Information about the cost distribution and the frequency of episodes is
needed for simulation models that use the episodic approach. In these models,
individual streams of medical expenses can be created by first generating episode
occurrences and then generating the cost of the occurrence pattern. These create a
simple model of how individuals behave. More generally, the overall analysis is
simplified by separating the cost-of-episode analysis from the number-of-episodes
analysis.
4.1.1.2 Grouping Claims into an Episode
The episode processing system transforms different claims into either dated
or typed episodes. In this study, we only focus on cataract surgery episodes. In
working with claims data, each episode included charges for several services, as
52
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
long as that claim was related to cataracts. Each o f these charges is referred to as a
“line charge” (Note that we are linking each charge to a diagnosis and evidence of
cataract surgery in 1997). The episodic procedures use these line charges as the
basic building blocks to create episodes (Keeler et al. 1982).
For cataract surgery episodes, line charges were linked together by
diagnosis code for cataract (231). This diagnosis code had at least one surgery
charge for a total amount above $600. The amount of $600 was considered the
minimum charge for an extracapsular (intracapsular) lens extraction (ECCE or
ICCE) and an intraocular lens implantation. In addition to this amount, charges for
preoperative and postoperative visits were also included under the episode of care
in cataract surgery. Generally, charges related to an episode begin after the patient
makes initial contact (e.g., January 1, 1997) and are ended after a certain period of
time (e.g., December 31, 1997).
Part two of cataract study focuses on cost and time period analyses of
surgery. Because cataract surgery episodes are often a process that begins before
and ends after the operation, this analysis only looked at “surgery” performed
between April and September, 1997 instead of from January to December, 1997.
For finding more preoperative visits, the study included January to March. For
finding more postoperative visits, the study went beyond September and extended
to December. This analysis only included surgeries where the first and second
were done between April and September. If the second surgery was done after
53
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
September, this episode was deleted and only remained a first surgery. As a result,
all visits after the first surgery were considered postoperative visits for the first
surgery. If the second cataract surgery was performed within 5 months following
the patient’s first cataract surgery, then we assumed that there was no preoperative
evaluation performed in connection with the second cataract surgery. Once the
percent of total costs attributed to preoperative and postoperative visits were
realized, we extended our findings and examined the whole year.
Part three o f cataract study focuses on the individual characteristics of
patients who had cataract surgery. The data were broken down by age and gender.
This part also looked at cataract surgery costs on 71,166 cases. The percentage of
pre- and postoperative visit costs obtained from study one was applied to the total
cost related to cataract surgery and to each episode of cataract surgery to better
understand exactly how much was spent in this area in 1997. The costs of drug
therapy were also studied. In addition, the prevalence rate of cataract surgery per
ten thousand individuals categorized by six regions was examined. Moreover, the
number of cataract surgeries performed by each physician in different institutions
was analyzed.
To derive the total cost per episode and catch as many pre- and post-visits
as possible within 1997, the percent of total costs attributed to pre- and post-costs
was calculated for the period between January and December. Next, the average
54
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
intra-operative surgery cost for all surgeries between April and September was
added to the pre- and post-costs derived between January and December.
4.1.2 Dependent Variables
4.1.2.1 Prevalence Rate
The prevalence rate is measured at one point in time or over a short period
of time in the study subjects. This variable counts the total number of people who
had at least one visit eye care during the year. There are seven outpatient eye-
related diagnoses, and the number of persons seeking eye care for each diagnosis
was counted. The data base for this study covered all utilization for 96% the
population. As a result, we calculated national level prevalence rates. Measuring
this variable has epidemiological and biomedical ramifications when compared
with prevalence rates in other countries. In addition, the number of persons with
more than two eye-related problems was also important to study.
4.1.2.2 Payment, Charge, and Cost
The data related to money paid are referred to as payment, charge, or cost.
In claims data, all the dollar amounts are fees charged by providers. These charges
are based on the single-pipe payment schedule provided by the Bureau. The
Bureau bases the payment schedule on two factors, the total premiums collected
from the insured and the total volume of services rendered without any profit.
55
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Since the Bureau of NHI is a government entity, personnel and administrative fees
are allocated within the government’s budget. Therefore, the money paid out to
providers is the cost to the Bureau. Thus, the terms “payment,” “charge,” and
“cost” are interchangeable in this context.
4.1.2.3 Total p e r Visit P aid by the Bureau to Providers
The payment amount in the claim data was the payment made by the
Bureau to providers. A claim can contain multiple services, including diagnosis
fees, tests, drugs, dispensing fee, and surgery fees. This variable measures the total
amount the Bureau spent for each outpatient visit. All reimbursement rates for
outpatient services are based on a fixed fee schedule set by the Bureau. The fee
schedule varies for each type of institution.
4.1.2.4 Total p e r Person Paid by the Bureau to Providers
Each beneficiary may make more than one eye care visit in a year. This
variable measures the total amount the Bureau spends on one person in one year, all
visits included. Additionally, this measures the total spent on one person for each
diagnosis. This information can also be used as a baseline for physician profiling
in order to understand physicians’ practice patterns.
56
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.1.2.5 Total Paid by the Bureau fo r Each Diagnosis
This variable aggregates the amounts taken from the total payment-per-visit
variable in each diagnosis. In the analysis, we calculate the total payment for each
person per year in a single diagnosis. Here we add the individual amount to a
lump sum by diagnosis. This information provides useful information for resource
allocation and can help decision-makers to form new payment policies.
4.1.2.6 Percent o f D rug Cost in One Visit
This variable records drug cost as a percentage of the total cost o f one visit.
Every year, drug fees account for 26% of the Bureau’s total expenditures.
Compared to the 8% spent on drugs in the USA, Taiwan’s drug expenditure is
extremely high. Therefore, the amount spent on drugs for each outpatient visit is
critical information. For example, it may help explain whether physicians in
tertiary hospitals prescribe more expensive drugs to patients.
4.1.2.7 Percent o f Lab Cost in One Visit
This variable records lab cost as a percentage of the total cost o f one visit.
Many studies show that tertiary hospitals own more lab equipment and therefore
utilize the equipment on patients more than is actually necessary. Tatchell (1983)
states that tertiary hospitals perform more ancillary services in general simply
because they usually have access to more advanced technology. This could be one
57
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
reason why the average cost per visit is higher in tertiary hospitals than in other
institutions. This information can narrow down which components make up a large
part o f the cost of per visit.
4.1.2.8 Total Paid by the Bureau p er Cataract Surgery Episode
This variable computes the payment made by the Bureau for each cataract
surgery episode. There are three components in each cataract surgery episode: the
preoperative visits, the intraoperative surgery, and the postoperative visits.
Cataract surgery is the most frequently performed eye care surgery. It accounts for
12% of all Medicare expenditure in the United States (Javitt et al.1995). Because it
is a significant expense, information related to cataract surgery is especially
important for resource planning the government and institutional levels.
4.1.3 Independent Variables
4.1.3.1 Practice Setting
The Bureau recognizes four types of institutions: medical centers (more
than 600 beds), regional hospitals (200-600 beds), district hospitals (less than 200
beds), and the physician offices. The reimbursement rates for outpatient visits vary
for each type of institution. Sixty percent of Taiwan’s physicians are employed in
one of these four practice settings; the other 40% of the physicians are owners of
the institutions such as the physician’s office. For physicians whom Friedson
58
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(1970) describes as “colleague-dependent” or who desire the approval o f their
peers, the influence o f fellow professionals is particularly strong and produces a
sort of group practice style.
4 .1.3.2 P atient Demographic Characteristics
Patient demographic characteristics (age and gender) were derived from the
Bureau’s claims data. The patient’s age is calculated from his/her birth date. A
second set of patient variables includes utilization-associated variables from claim
data, such as patient identification number, date of treatment, and insured
identification number. Patient gender and age were adjusted for case mix purpose.
Date of treatment is the variable of calculating the distribution of cataract surgery
cases by month, how long patients have been waiting for cataract surgery, etc.
4.1.3.3 D iagnosis as Treatment-Related Inform ation
In this category, a diagnosis code is used to identify a patient’s disease.
“Simplified A Code” was the diagnosis book utilized by the Bureau in 1997. This
code book is relatively general, containing only about 400 diagnostic codes. This
limits many advanced research studies. However, it is welcomed by most
physicians because it helps to simplify the large volume of outpatient visits.
The second set of treatment-related variables is major procedural codes,
methods of dispensing drugs, and the number of days per prescription. Many
59
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
physicians do not enter a major procedural code because o f the high volume of
visits. In Taiwan, physicians are allowed to directly dispense drugs. The manner
of dispensation is recorded on the claim and determines how much of the
dispensing fee is reimbursed. Usually, the reimbursement of dispensing fees is
higher for drug stores and lower for physicians in order to discourage physicians
from directly dispensing medications. The Bureau and institutions record the exact
number o f days o f each prescription prescribed. This information can be
crosschecked to determine whether different physicians provided medication to the
same patient at the same time.
4.1.3.4 Physician Characteristics
Variables related to physician characteristics include physician specialty,
gender, and age. Physicians’ gender and age were taken from the physician profile
database, which is separate from the outpatient claims database.
4.1.3.5 G eographic Region
For managerial purposes, the Bureau divided Taiwan into six non
overlapping geographic regions. Indicator variables were used to control for
regional differences in supply and composition of health care services. The
location o f the institution was identified from hospital files and used to designate
whether patients received care in a rural or urban area.
60
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.2 Approach to Empirical Analysis
The empirical analysis contains descriptive and multivariate analyses. The
descriptive analyses include the prevalence rate of eye problems in Taiwan, the
prevalence of cataracts, and the average age of cataract surgery, as well as a cost
comparison of outpatient glaucoma treatment between Taiwan and other countries,
how geographic location and the institutional setting of the practice affect physician
practice styles, and how physician characteristics such as gender and age affect
variation in the use of resources. The multivariates include physician age and
gender, patient age and gender, institutional setting of the practice, geographic
region of the practice, diagnosis, whether or not cataract surgery is performed, and
complications in cataract surgery. A generalized linear model is used determine
how each of these nine independent variables affects use of resources. A
generalized linear model (GLM) can assist in adjusting a patient’s age and gender,
whether or not surgery is performed, and whether or not there are complications.
Through GLM, patient age and gender, institutional settings, regional locations,
physician age and gender, and various diagnoses can be diminished in the use of
resource. A variable representing age squared was added to account for possible
non-linear age effects on utilization.
The general linear regression model in this study is specified as follows:
RU= f (PAT, GR, FI, D, PHY, S, C)+a+£j
61
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Where
RU is the resource use per outpatient visit.
PAT is a vector of patient demographic characteristics (e.g., age, gender).
GR is a vector of geographical factor dummies.
FI is a vector of hospitals (e.g., four levels o f hospitals).
D is a vector of diagnosis (e.g., seven diagnoses in the outpatient eye
care).
PHY is a vector of physician characteristics (e.g., age, gender, practice
setting).
S is an indicator of surgery (e.g., whether patients have cataract
surgery).
C is an indicator of whether complication occurred or comorbidity
(e.g., whether patients have complications during cataract surgery or
have more than two eye problems).
s is the error that is N(~0,1).
a is a constant.
The equation is as follows:
Y = a+PiPATage+P2PATage2+p3PATsex+p4RG+P5FL+P6D+
p7PHYAGE+p8 PHYSEX+p9S+p i0 C+£i
62
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.3 Limitations
1. Not all beneficiaries are candidates for a particular medical treatment, so
taking a percentage o f treated patients from the entire beneficiary pool
does not accurately describe treatment rates. However, to ascertain
what portion of the entire beneficiary pool were not candidates for
treatment requires longitudinal or historic procedural data from the
years preceding 1997, which were not available for this study, and in the
absence of this longitudinal patient data, we reported on one year (1997)
of case incidences using diagnostic groupings. This was a definite
underestimate of actual diagnostic prevalence because not all patients
with a particular diagnosis are examined within any one year and, if
they are examined, it may not necessarily be for the diagnosis of
interest.
2. Health care use will be influenced by changes in reimbursement policies
and other utilization rules. These changes will directly affect some
services and indirectly affect other services as providers seek to shift
care services among different services.
3. Data for this study were limited to one year. Thus, although we
confined pre- and post-study to six months for the cataract surgery
63
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
episode, it is possible that some preoperative and postoperative visits
were not accounted for.
4. While the Bureau of NHI requires two postoperative visits as part of the
total cataract surgery fee, some providers do not follow this rule because
50% of their patients have follow-up visits on the day after surgery.
The Bureau does not strictly observe these practice patterns during their
payment review period, making it difficult to assess the number of
average visits and the cost of postoperative visits.
5. The outpatient diagnosis code system is very simple. There are only
eight codes used for eye care diagnosis, only seven of which show claim
data in this sample. Outpatient volume is overwhelmingly high, and the
bureau did not enforce proper use of diagnosis codes in the claims. The
oversimplicity of the diagnosis codes prevents researchers from
identifying what procedures and tests have been done. The amount
charged for treatment is the only way to distinguish complications or
comorbidities.
6. Since this study only examined services provided to individuals who
actually underwent cataract surgery, individuals who had cataracts but
did not have surgery were not included in the analysis. Thus, the true
64
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
frequency and cost of services provided in connection with actual
cataract surgery episode were underestimated.
65
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5. R e s u l t s
The total number of beneficiaries in the national health insurance database
of Taiwan was 20,492,317 in 1997. O f these beneficiaries, 4,474,750 (22%) had at
least one paid claim for eye care services.
5.1 Descriptive Results
5.1.1 Results fo r All Eye Care
Table 1 shows the total number of patients who received eye care, the total
number of visits and the costs, the percentage o f visits and costs, average visits per
person, average cost per person, and average cost per visit for seven eye-related
diagnoses. Row 1 provides estimates of the number of people who sought eye care
for each diagnosis. Row 2 shows the percentage of people who sought eye care for
each diagnosis. Row 3 shows the total number of visits made relating to each
diagnosis. Among these eye diagnoses, conjunctivitis constitutes the highest
prevalence (46%) and visits (44%), respectively. However, the total cost for
66
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
conjunctivitis comprises only 24% of all eye care-related costs as shown in Row 6.
The total cost for cataract procedures accounts for 33% of all eye-related costs.
Each patient costs $216 ($842 for surgical treatment and $41 for non-surgical
treatment1 ) versus only $19 per visit for blindness and low vision. Moreover, there
were an average o f 3.3 cataract visits per year compared to 1.7 visits per year for
blindness and low vision as shown in Row 7. Glaucoma patients make an average
of 4.1 visits per year, the highest number of visits with a cost of $116 per person
per year.
Cost for surgical and non-surgical cataract treatment differs widely. Total cost for surgical treatment is S73
million (87%); S14 million (13%) for non-surgery. The average cost per person is S842 for cataract surgery
and S41 per person for non-surgical cataract treatment.
67
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 1. Distribution o f Persons, Visits, Total C ost, and Averages by Seven
Eye-Related Diagnoses
Glaucoma Cataracts
Blindness
and Low
Vision
Conjuncti
vitis
Lacrimal
System
Disease
Strabismus
and Eye
Movement
Disease
Others Total
Number
of
Patients
113,955 403,441 143,350 2.764,342 122,303 79.778 2,357,563 5,984,732
% of Total
Patients
1.9% 6.7% 2.4% 46.2% 2% 1.3% 39.4% 100%
Number
of Visits
471,152 1,312,387 241,228 5,671,096 234,398 133,943 4,925,483 12,989,686
% of Total
Visits
3.6% 10.1% 1.9% 43.7% 1.8% 1.0% 38% 100%
Cost of
Treatment
(S)
1,324,846 87,148,490 2,694,667 62.517,281 3,598,949 2,112,927 92,201,069 263,600,000
% of Total
Cost
5% 33% 1% 24% 1.4% 0.8% 35% 100%
No. of
Visits per
Person
4.1 3.3 1.7 2.1 1.9 1.7 2.1 2.2
Average
Cost per
Person
$116 S216 $19 $23 $29 $27 $39 $44
Average
Cost per
Visit
S28 $66 $11 $11 $15 $16 $19 $20
Note: When calculating the total number of patients who received eye care, a patient was counted each time he visited the
physician for a different problem. Thus, the total number of patients who received eye care in this analysis is 5,984,732,
while the actual number of different persons who received eye care is 4,474,750.
Table 2 presents the distribution of cost and visits across gender and age.
Spending is high before age 14 and goes down between age 15 to 64. It increases
again at age 65. The 75 to 84 age group shows the highest spending of the teen age
68
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
groups, comprising 20.2% of total expenditures. The prevalence rate of eye care-
related outpatient visits is 218 visits per 1,000 persons per year. The highest
prevalence rate for males is in the 5 to 14 age group. For females, the highest
prevalence rate is in the 65 to 74 age group. Overall, females have a higher
prevalence rate (252/1000) than males (201/1000). In column 6, it is obvious that
the average cost per visit cost increases as age increases. In column 7, the average
cost per person per year also increases as age increases. Similarly, the number of
visits increases as age increases. Females average 3.1 visits per year, while males
average 2.7 visits per year. However, there is no significant difference in the cost
per visit and cost per person between genders.
Using data from Table 2, Graphic 1 shows that costs per person and number
of visits increase dramatically after age 46.
69
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 2. C ost, Visits, and Prevalence Rate per 1,000 C ategorized by Gender
and Ten A ge G roups for All Eye Care
Age Total Cost for
NHI
% of Total
Cost
% of Total
Visits
Prevalence
Rate/1,000
Population
Avg. Cost per
Visit
Avg. Cost per
Person per
Year
No.
Visit/Person
Gender Male Female Male Female Male Female Male Female
0-4 SO,418,397 3.6% 6.1% 290 261 $12 $12 $23 $24 2.0 2.0
5-14 S24,892,422 9.4% 16.7% 294 310 $12 $11 $27 $27 2.3 2.4
15-24 $18,378,811 7.0% 10.8% 158 237 $14 $13 $28 S27 2.0 2.1
25-34 S18.189.295 6.9% 10.2% 150 211 $15 $13 $30 $27 2.1 2.1
35-44 $19,626,115 7.5% 9.9% 153 190 $16 S14 $36 $31 2.2 2.2
45-54 $21,951,313 8.3% 8.8% 156 220 $21 S18 $55 $51 2.7 2.9
55-64 $45,292,521 17.2% 13.2% 207 329 $28 $25 $108 $108 3.9 4.2
55-74 $71,317,037 27.1% 16.7% 274 388 $33 $33 $167 $187 5.1 5.8
75-84 $31,454,272 11.9% 6.7% 255 338 S38 $35 $207 $217 5.5 6.1
Over 85 $3,001,606 1.1% 0.7% 151 185 $39 S34 $211 S191 5.0 5.3
Total $263,000,000 100% 100% 201 252 $21 S20 $56 S57 2.7 3.1
450
• 400
• 350
- 300 g
■ 250 g
' 2 0 0 |
• 150 £
Q.
• 100
- 50
SO -------------------------------------- - ................. • -------------- --------- --------- ----------------------- 0
0 -4 5-14 15-24 25-34 35-44 45-54 55-64 65-74 75-84 Above 85
Age group
Male cost per person — ♦—Female cost per person - * — Male prevalence per 1000 — • —Female prevalence per 1000
Graphic 1. Cost p er Person and Prevalence Rate in E ye Care, National
Health Insurance o f Taiwan, 1997.
70
S250
S200
S150
$100
S50
JT
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 3 compares eye care outpatient data between the United States and
Taiwan, categorized by eight age groups. The second and third columns are male
and female’s percentage of visits to the total visits for each age group in the US.
The fourth and fifth show the same data for patients in Taiwan. The remaining four
columns are the visits per 1,000 people in the US and Taiwan across age and
gender groups. Americans age 65 and older have a nearly sevenfold higher annual
rate of eye care-related outpatient visits to younger Americans. On the other hand,
the outpatient visits in Taiwan do not differ so much between elderly and younger
age groups in terms of number of visits than the USA. However, older people
indeed consume more visits than younger ones in Taiwan as well. It is obvious that
the number of visits in Taiwan is much higher than America no matter what the age
group.
Table 4 shows the frequency of visits by different diagnoses. This provides
data to establish patient behavior, severity of disease, and the pattern of physician
treatment. Table 4 is divided by frequency of visits and types o f diagnosis. Over
50% of the patients only sought eye care once within the year. Over 70% sought
care twice a year. Glaucoma patients rendered higher visits than other diagnoses.
71
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 3. D istribution and A nnual R ate o f Outpatient V isits for All Eye C are-
Related D iagnoses by Patient A ge and Sex in the U nited States3 and Taiwanb
USA Taiwan USA Taiwan
Age Groups Visits (%) Visits (%) Visits/1,000 Visits/1,000
Male Female Male Female Male Female Male Female
0-4 2.0% 1.8% 3.4% 2.7% 103.2 98.1 510 578
5-14 2.7% 2.5% 8.4% 8.3% 75.0 70.4 679 728
15-24 2.6% 2.1% 4.0% 6.9% 69.7 57.8 321 500
25-44 6.3% 8.9% 8.8% 11.5% 77.7 108.4 325 429
45-64 9.7% 13.5% 8.2% 13.9% 213.8 275.6 580 944
65-74 7.9% 14.2% 7.5% 9.3% 482.5 692.8 1,3755274 2,236
75-84 6.9% 13.3% 2.9% 3.8% 918.3 1,080.5 1,413255 2,076
85 and older 1.7% 3.8% 0.3% 0.4% 972.2 849.3 818 1,049
Total 40%% 60% 43% 57% 177.0 216.0 545 724
Sources:
* Chiang, Y., et al. 'Office visits to ophthalmologists and other physicians for eye care among the US population, 1990,' Public Health
Report, 1995, pp. 147-153.
b “National Health Insurance Profile, 1998, Republic of China, Taiwan’ (Bureau of National Health Insurance, Department of Health,
Taipei, Taiwan).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 4. Percentage of Frequency for Visits Categorized by Diagnosis in Eye
Care
Diagnosis/
No. of
Visits
Glaucoma Cataracts
Blindness
and Low
Vision
Conjunc
tivitis
Lacrimal
System
Disease
Strabismus
and Eye
Movement
D isease
Others
1 Visit 40.1% 43.3% 67.7% 64.5% 64.5% 68.5% 63.3%
2 Visits 15.1% 17.7% 18.1% 17.2% 17.2% 16.8% 18.4%
3 Visits 9% 10.1% 6.6% 7.4% 7.4% 6.6% 7.5%
4 Visits 6.4% 6.9% 3.2% 3.8% 3.8% 3.4% 3.8%
5 Visits 5.1% 5.0% 1.6% 2.2% 2.2% 1.8% 2.2%
6-10 Visits 14.5% 12.3% 1.5% 2.6% 2.6% 1.5% 2.5%
11-20 Visits 8.5% 3.3% 0.3% 0.8% 0.8% 0.3% 0.8%
2 1 + Visits 1.3% 0.4% 0% 0.1% 0.1% 0% 0.1%
Table 5 presents the number of visits in different institutions. Each type of
institution has its own practice norm. The frequency of visits varies significantly
between the different institutions. Patients also go to different institutions for the
same eye problems. About 50% of the patients visited a medical center once a
year. Thus, another 50% of the patients went to the medical center twice or more.
Regional hospitals have the highest rate for single visits, and physician offices have
the highest multiple visit rate among these four institutions. In conclusion, the
highest volume of visits occurred in physician offices. The lowest volume of visits
is rendered in regional hospitals.
73
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 5. Percentage of Frequency in Visits Categorized by Institutions in Eye
Care
Institution/
No. of Visits
Medical
Center
Regional Hospital District Hospital Physician Office
1 Visit 50% 57.4% 55.9% 52.9%
2 Visits 17.5% 17.4% 18.4% 20.2%
3 Visits 9.1% 8.1% 8.5% 9.5%
4 Visits 6.1% 4.7% 4.8% 5.2%
5 Visits 4.1% 3.1% 3.1% 3.2%
6-10 Visits 7.2% 4.9% 4.8% 2.2%
11-20 Visits 2.6% 2.2% 2.0% 4.5%
Above 20 Visits 0.3% 0.3% 0.4% 0.5%
In order to verify whether the number of visits was influenced by type of
institution, the statistical model of Chi-square were applied. Two hypotheses were
considered:
Null hypothesis: Institution types and visit frequency distribution are
independent.
Alternative hypothesis: Institution types and visit frequency distribution are
not independent.
Chi-square 0.005 value is 41.4010 at degree of freedom 21. The grand total
for the data is 79,349.96, which is much bigger than 41.4010. Therefore, we reject
74
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the null hypothesis and accept the alternative hypothesis that visit frequency is
dependent upon type o f institution.
Table 6 presents the average cost per visit and per person among four
institutions in Rows 2 and 3; total cost of each institution within 1997 in Row 4;
percentage of cost across four institutions in Row 5; percentage of patient volume
in each institution in Row 6; average lab and (or) surgery cost for each visit and
percentage of lab and (or) surgery cost for each visit in Rows 7 and 8, respectively;
average drug costs and percentage of drug costs for each visit in Rows 9 and 10,
respectively. Row 1 1 gives the number of visits per person in different institutions.
The last row shows the number of physicians in each of the four institutions.
Sixty-five percent of total costs is charged by physician offices, which is out
of proportion to other institutions in terms of the number of physicians. It shows
that physicians working in physician offices concentrate more on eye care. The
percentage of patient volume is 7.8% in medical centers; however, the percentage
of total cost (13.4%) in medical centers is double the patient volume. The average
cost per visit and cost per person in medical centers are higher than in the physician
offices, 75% and 96%, respectively. This means each person or visit costs almost
twice as much as in medical centers than in other institutions. Additionally, the
frequency of visits to medical centers is slightly higher than the frequency o f visits
to other institutions.
75
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 6. A verage Cost per V isit, per Person, Total C ost, Percentage of Cost,
Percentage o f Patient Volum e, N um ber o f Visits per Person, and Average Cost
per Person Categorized by Four Levels o f Institutions for All Seven Eye-
Related D iagnoses
Medical Center District Hospital Regional
Hospital
Physician
Office
Avg. Cost/Visit $31.4 $27.2 $22.9 $17.9
Cost per Person $88.9 $65.2 $56.5 $45.3
% of all Eye Cost in
Each Institution
13.4% 9.7% 11.7% 65.2%
% of Patient Volume in
Each Institution
7.8% 7.7% 10.6% 74%
Avg. Labs/ Surgery
Cost per Visit
$16.3 $13.1 $10.9 $6.8
% of Avg. Labs/
Surgery Cost per Visit
52% 48% 48% 38%
Avg. Drug Cost per
Visit
$7.2 $6.0 $4.2 $3.8
% of Avg. Drug Cost
per Visit
23% 22% 18% 21%
# of Visits /Person 2.8 2.4 2.5 2.5
No. of Physicians 768 668 683 1,450
(22%) (19%) (19%) (41%)
Regarding cost comparisons, the percent of total costs on labs and (or)
surgery is higher in medical centers than in other institutions, but there is no
significant difference in the percent of total costs accounted for by drug costs
across the four levels of institutions.
76
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Besides, data also show that only eight physicians worked in all four
different types of institutions. One hundred and ten physicians (3%) have worked
in three types of institutions and three hundred seventy eight physicians (11%)
worked in two types of levels of institutions. However, most of the 2,451
physicians (83%) only worked in one type institution.
Table 7 presents the cost components for each diagnosis within the different
types of institutions. In all institutions, labs, surgery, and drugs comprise more
than one-third of all costs for treatment of Glaucoma. Surprisingly, the average
cost for treatment o f cataracts in a physician’s office ($77.80) is much higher than
the average cost for the cataract treatment in medical centers ($50.20). However,
later studies show that 70.7% of cataract surgeries were performed in physician
offices, explaining the higher average cost of treatment. The cost of labs and/or
surgery is relatively high, making up two-thirds of the total costs for cataract
treatment, with the majority of cost spent on surgery. For blindness and low vision,
the majority of cost comes from diagnosing these conditions, as opposed to the
costs associated with labs, surgery or drugs. For conjunctivitis, the average cost of
$14.80 in medical centers is higher than the average cost of $13.40 in regional
hospitals, $11 in district hospitals and $10.80 in physician offices. For diagnoses
of the lacrimal system, medical centers spent more on labs than did other
institutions. For Strabismus and eye movement disease, medical centers and
regional hospitals spent a much higher amount of money on labs and/or surgery and
77
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a relatively low amount of money on drugs. Both district hospitals and physician
offices spent more money on drugs and less money on labs and/or surgery. For
other conditions, the pattern is similar for most diagnoses: spending was higher for
labs and/or surgery in medical centers and regional hospitals. However, drugs were
most commonly prescribed in physician offices compared with other institutions.
T able 7. A verage C ost per V isit, A verage Lab and Surgery C ost per Visit,
A verage o f D rug C ost per Visit in Seven Eye-Related D iagnoses Categorized
by Four Levels o f Institutions
Glaucoma Medical
Center
Regional
Hospital
District
Hospital
Physician
Office
Total
Avg. Costs $32.4 $32.4 $29.3 $24.6 $28.1
Avg. Cost for
Labs/Surgery
$13.2 $11 $11 $8.7 $10.4
Avg. Cost for
Drugs
$11.6 $13.4 $11 $9 $10.4
Cataracts Medical
Center
District Hospital Regional
Hospital
Physician
Office
Total
Avg. Costs $50.2 $50 $53.2 $77.8 $66.4
Avg. Cost for
Labs/Surgery
$33.5 $33.5 $38.5 $61.4 $48
Avg. Cost for
Drugs
$8.6 $8.1 $7 $9.2 $8.6
Table continues next page
78
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 7 (C ontinued). A verage C ost per Visit, A verage Lab and Surgery C ost
per Visit, A verage o f D rug C ost per Visit in Seven Eye-Related D iagnoses
Categorized by Four Levels o f Institutions
Blindness and
Low Vision
Medical
Center
District Hospital Regional
Hospital
Physician
Office
Total
Avg. Cost $11.3 $11.2 $11.7 $11.1 $11.2
Avg. Cost for
Labs/Surgery
$1.8 $1.4 $1.9 $0.5 $1.1
Avg. Cost for
Drugs
$2.1 $2.1 $2.2 $3.2 $2.6
Conjunctivitis Medical
Center
District Hospital Regional
Hospital
Physician
Office
Total
Avg. cost $14.8 $13.4 $11 $11 $11
Avg. Cost for
Labs/Surgery
$2.8 $1.4 $0.7 $0.2 $1.1
Avg. Cost for
Drugs
$3.8 $3.9 $2.5 $3.2 $3.4
Lacrimal
System
Medical
Center
District Hospital Regional
Hospital
Physician
Office
Total
Avg. cost $20.7 $20.1 $15.7 $12.8 $15.4
Avg. Cost for
Labs/Surgery
$5 $3.5 $3.1 $1.6 $2.9
Avg. Cost for
Drugs
$7.5 $8.4 $4.7 $3.9 $5.5
Table continues next page
79
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
T able 7 (Continued). Average Cost per V isit, Average Lab and Surgery Cost
per Visit, Average o f Drug Cost per V isit in Seven Eye-Related Diagnoses
Categorized by Four Levels of Institutions
Strabismus
and Eye
Movement
Disease
Medical
Center
District Hospital Regional
Hospital
Physician
Office
Total
Avg. Cost $31.5 $17.6 $13.5 $12.9 $15.8
Avg. Cost for
Labs/Surgery
$22.9 $7.1 $3.3 $2.3 $8.2
Avg. Cost for
Drugs
$1.5 $2.7 $2.4 $3.1 $2.6
Others Medical
Center
District Hospital Regional
Hospital
Physician
Office
Total
Avg. cost $32.8 $27.2 $18.6 $15.1 $18.7
Avg. Cost for
Labs/Surgery
$17.3 $13.6 $7.3 $4.6 $9.3
Avg. Cost for
Drugs
$7.5 $5.5 $3.6 $3.2 $4.6
After cataracts, glaucoma is the second most costly diagnosis to treat.
Table 1 also shows that the average number of visits (4.1) for glaucoma is the
highest of the seven eye-related diagnoses.
Table 8 presents outpatient charges for glaucoma among eight countries.
This table includes the charge for a visit to an ophthalmologist, a visit to a nurse or
technician, a visit to a general practitioner, and telephone consultation, all depending
80
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
on the payment system for each country. For example, in Taiwan, nurses are not
paid separately, and no payment is charged for telephone consultations.
Table 8. C om parison of O utpatient Charge ($US) for G laucom a A m ong Seven
Countries
USA
(1995)
Can
(1996)
Ger
(1997)
UK
(1996)
NL
(1997)
Swe
(1996)
Taiwan
(1997)
Visits to
Ophthalmologist
59 49 19 11 40 71 28
Visits to Nurse or
Technician
23 17 Ind.* 27 Ind.* 23 Incl.*
Visits to General
Practitioner
31 17 12 27
-
69 28
Telephone
Consultation
39 18 2 5 0 18 0
Note: *lncl. - The exist of the nurse is included in the cost to visit the ophthalmologist.
Sources: Bengt, J. et al., Glaucoma, 1999 (San Francisco: Jossey-Bass Publishers), pp. 123-156, for the first
6 countries.
National Health Insurance o f Taiwan, 1997, for the data in Taiwan.
5.1.2 Results fo r Cataracts
5.1.2.1 Prevalence o f Cataract by A ge and Country
A cataract is an opacity o f the natural, crystalline lens of the eye and
remains the most frequent cause o f blindness in the world today. The World Health
Organization (WHO) estimates that cataracts are the cause of blindness for 50%
(17million) of the world’s blind (Thylefors et al. 1995). Table 9 presents the
81
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
prevalence rate of cataracts among 12 countries, from India (2.01) at the high end
to Germany (0.002) on the low.
Table 9. Prevalence Rate o f C ataracts by Tw elve C ountries
Country Population (millions) Prevalence Rate of Cataracts
Norway 4.1 0.001
Sweden 8.0 0.001
Germany 61.4 0.002
Taiwan 20.0 0.02
United States. 233.7 0.06
Japan 119.0 0.08
Kenya 18.8 0.14
Ethiopia 33.7 0.20
Egypt
44.5 0.47
Indonesia 159.0 1.28
China 1,040.0 1.41
India 732.0 2.01
Sources: World Health Organization, Programme for the Prevention of Blindness, 1987.
National Health Insurance of Taiwan, 1997.
Prevalence of cataracts and lens opacity has been reported in four
population surveys in the United States: the National Health and Nutritional
Examination Survey (NHANES)2, the Framingham Eye Study (FFE)3, the
2 by Hiller et al., 1986.
82
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Maryland Waterman Study (Waterman)4 and the Beaver Dam Eye Study (BDES)5.
The term “Cataracts” required lens opacity that was associated with, or
accompanied by, vision loss. The lens opacity was detected by slit-lamp
photography in these four surveys. The criteria for significant vision loss varied
among surveys, with a vision of 20/30 or worse in NHANES, 20/25 or worse in
FES, and 20/32 or worse in BDES. The published criterion for vision loss in
Taiwan is 20/50, much higher than in other countries. Because most Taiwanese do
not drive, vision requirements are less stringent.
Table 10 presents the prevalence of cataracts (percent of population with
cataract) in the US and Taiwan for residents age 45 to 85. For the 45 to 54 age
group, these five studies are quite similar between the US and Taiwan. For the 55
to 64 and 65 to 74 age groups, the prevalence of cataracts in Taiwan is much higher
than in the US. For the 75 to 85 age group, Waterman’s study shows that 59% of
Americans have cataracts, slightly higher than in Taiwan.
3 by Kahn et al., 1977.
4 by Adamsons et al., 1991.
5 by Klein et al., 1992.
83
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 10. Prevalence o f Cataracts in the United States and T aiw an
Age
USA
Taiwan3
NHANES FES Waterman BDES
45-54 2.6 - 3b - 1.6C 2.7
55-64 10.0 4.5d 10b 5e 7.2 18.3
65-74 28.5 18.0 38b 25e 19.6 49.7
75-85 - 45.9 - 59e 43.1 54.2
a National Health Insurance of Taiwan, 1997.
b FES definition of cataracts and FES age group (52-64) applied.
c Age: 43-54-year-olds.
d Age Group: 52-64.
e NHANES definition of cataract and NHANES age group applied.
Table 11 shows the prevalence of cataracts in different age groups in India,
China, and Taiwan. The calculation of cataract prevalence is based on the
population in each age group. The prevalence of cataracts accompanied by visual
acuity o f 20/60 or worse was 4.3% for all age groups and 15.3% among persons
age 30 or older in Punjab, India. In the suburban county of Beijing, China, the
prevalence of cataracts was defined as lens opacity associated with vision of 20/30
or worse. The prevalence of cataracts was 6.0% in the overall population and
18.6% among those age 30 and above. In Taiwan, diagnosis of cataracts was
specified by visual acuity of 20/50 or worse, and prevalence was 2% for all age
groups and 11.8% among persons age 30 or older.
84
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 11. Prevalence o f C ataracts in India, C hina, and Taiwan
Age
Punjab, India3
%
Beijing, China3
%
Taiwan
30-39 0.2 - 0.3
40-49 2.2 0.4 1.2
50-59 14.7 6.83 8.1
60-69 42.0 25.79 35.6
70-79 55.7 59.95 56.3
80+ 87.8 83.02 41.7
Total for Ages 30+ 15.3 18.58 11.8
Total for All Populations 4.3 6.0 2.0
Sources: a Javitt, J and Wang, Fang, “Blindness due to cataract: Epidemiology and prevention,"
Annual Review Public Health, 1996.
b National Health Insurance of Taiwan, 1997.
5 .1.2.2 Patient Behavior: Pre- and P ostoperative Treatment in Cataract
Surgery
There were 36,414 cataract surgeries performed between April and
September. However, there were no preoperative surgery claims found between
January to December for 45% (16,529) of the patients. Table 12 shows the average
number of pre- and postoperative visits, average and median number of days
between the last preoperative visit and the surgery, average and median number of
days between the surgery and the first postoperative visit, average cost of pre- and
85
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
post-operative visits, and the cost o f pre- and postoperative visits as a percentage of
the total cost for cataract surgery.
The average number of preoperative visits was 1.9 for the persons with at
least one preoperative visit (S.D. 1.8). The average length of time between the last
preoperative visit and surgery was 17 days, and the median was 6 days. However,
75% of patients underwent surgery within 13 days of their last preoperative visit.
Thirty-eight percent (13,888) of the patients did not make postoperative visits. The
average number of postoperative visits for patients who made at least one post-
surgical visit was 5 (S.D. 3.8). There was an average of 8 days between the surgery
and the first follow-up visit (S.D. 26 days). Seventy-five percent of the patients
made postoperative visits within 3 days; the median was 1 day. The average cost
of preoperative treatment was $14 (1.73% of total cost) and $37 (4.47%) for post
operative treatment.
Table 12. Inform ation on Costs, Visits Prior to First Surgery, and Follow-Up
Visits to C ataract Surgery, 1997
Average Number
of Visits
Mean (SD)
Number of Days
Between Visit
and Surgery
Median Number
of Days Between
Visit and Surgery
Average Cost
per Episode
(Percent of Total
Cost)
Preoperative 1.9 visits 17 days(32) 6 days $14(1.73%)
Postoperative 5 visits 8 days(26) 1 days $38 (4.47%)
86
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Between the group of patients who sought pre- and postoperative care and
the group of patients who did not, there was little difference in age or gender. Both
groups were 43% male vs. 57% female, and the average age was 69.45 years old
for both groups. Therefore, neither gender nor age was an important determinant of
the patient’s behavior in seeking pre- or postoperative treatment.
Table 13 presents the percentage of patients who made pre- and
postoperative visits in the four types of institutions. Regional hospitals have the
highest preoperative visit rate (68.2%). Only 52.3% of physicians’ office patients
claimed preoperative visits. Regional hospitals again have the highest
postoperative visit rate (68.62%). Medical centers claimed postoperative visits for
only 57.71% of cataract surgery patients.
Table 13. The Percentage o f Patients w ith Pre- or Postoperative Visits
C ategorized by Institutions in Cataract Surgery
Medical Center Regional
Hospital
District Hospital Physician Office
Preoperative
Visits
64.05% 68.02% 54.31% 52.3%
Postoperative
Visits
57.71% 68.62% 64.23% 61.16%
Table 12 shows that the average number of preoperative visits is 1.9. The
length of time between surgery and the last preoperative visit can serve as a rough
87
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
estimate of the waiting period for patients to receive surgery. Table 14 shows the
percentage of patients who either waited less than 13 days for surgery or more than
13 days among these four institutions. As mentioned before, 75% of the patients
underwent cataract surgery within 13 days after their last preoperative visit. This
shows that patients wait longer for surgery in medical centers than in other
institutions. Patients in regional hospitals have the shortest wait for cataract
surgery.
T able 14. Length o f Tim e Betw een Last Visit Prior to Surgery and Surgery
C ategorized by Institutions in C ataract Surgery
Medical Center Regional Hospital District Hospital Physician Office
< 13 Days 63.03% 79.77% 75.83% 77.01%
> 13 Days 36.97% 20.23% 24.17% 22.99%
5.1.2.3 Demographics and R egional Distribution
Table 15 shows the cost of each component of cataract surgery. Data from
Table 12 show that pre and postoperative surgery account for 1.7% and 4.48%,
respectively, of the total cost of surgical treatment. Adding these costs to the Intra
costs shows that the total cost per episode is $842. The Bureau spent a total of $60
million on cataract surgery in 1997.
88
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 15. Costs Attributed to Cataract Surgery
Pre-Cost Post-Cost Intra-Cost Total Cost per
Episode
Total Cost (Pre-
+lntra+ Post-)
Cataract Surgery
$14(1.7%) $38 (4.48%) $790 (93.5%) $842 $60 million
Table 16 shows the age distribution by institution where cataract surgery
was performed. The first group is 0-44 years, then 5 years per group. Between 60-
75 years, the number of people seeking cataract surgery increases sharply. At 75,
the need falls. The most surgery was performed in the 70-74 age group (22.7%)
significantly higher than in the 65-69 age group (14%), the next highest percentage
group. Physician offices performed 70.6% of the surgeries, by far the highest
percentage of the four types of institutions.
Table 17 shows the rate of cataract surgery by age between Taiwan and the
US. Schein’s research (2000) takes a sample in the US between June 1, 1995, and
June 30, 1995. The results show that cataract surgery in Taiwan is more
concentrated in the 60— 79 age group. Cataract surgeries performed in the US were
evenly distributed in the age groups 60 and over.
89
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 16. Patient Age Distribution of Cataract Surgery by Institution
Age Group Medical
Center
Regional
Hospital
District Hospital Physician Office Total (%)
0-44 197 125 211 1,039 1,572 (2.2%)
45-49 135 115 159 906 1,315(1.8%)
50-54 204 149 304 1,467 2,124 (3.0%)
55-59 425 354 675 3,854 5,308 (7.5%)
60-64 749 643 1,257 7,343 9,992 (14%)
65-69 1,217 1,091 2,383 11,480 16,171 (22.7%)
70-74 1,248 1,178 2,524 11,389 16,339 (23%)
75-79 883 759 1,751 7,691 11,084(15.6%)
80 over 480 511 1,136 5,134 7,261 (10.2%)
Total 5,538 (7.8%) 4,925 (7%) 10,400(14.6%) 50,303 (70.6%) 71,166 (100%)
Table 17. Cataract Surgery Rate Between Taiw an and the U.S.
Age Group Taiwan USA
50-59 10.5% 7.0%
60-69 36.7% 22%
70-79 38.6% 22.9%
Over 80 10.2% 24.3%
Source: “ The Value of Routine Preoperative Medical Testing Before Cataract Surgery," by Schein,
Katz, et al., NEJM, 2000.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 18 displays the distribution o f cataract surgery patients by gender in
the four types of institutions. It shows that female patients are more likely to
choose medical centers and male patients are more likely to select district hospitals
when undergoing this surgery. It also shows that females have a higher rate of
cataract surgery than males, 56% to 44%, respectively.
Table 18. Patient G ender Distribution o f C ataract Surgery C ategorized by
Institutions
Gender Medical
Center
Regional
Hospital
District
Hospital
Physician
Office
Total (%)
Male 2,470 2,432 5,143 21,327 31,372(44%)
Female 2,850 2,385 4,159 22,893 39,794 (56%)
Table 19 demonstrates the distribution of ophthalmologists in Taiwan by
region. In Region 1, the density of ophthalmologists is high since there are many
medical centers located there. For Region 5, the density of ophthalmologists is
even higher than Region 1 with seven ophthalmologists to every ten thousand
residents.
91
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 19. N um ber o f O phthalm ologists by Regions in Taiwan
Region Population No. of Ophthalmologists
Region 1 (Taipei Branch) 7,403,981 490
Region 2 (Northern Branch) 2,684,177 131
Region 3 (Central Branch) 3,778,125 199
Region 4 (Southern Branch) 3,172,809 159
Region 5 (Kao-Ping Branch) 3,172,809 222
Region 6 (Eastern Branch) 497,717 22
Source: Regional population provided by Bureau of National Health Insurance, Department of
Health, Taipei, Taiwan, 2000.
Table 20 presents the number of cataract surgeries, surgery rate per ten
thousand residents, total cost, average cost per episode, cost of drugs, and drug cost
as percentage of total cost in 6 Regions. Region 5 has the second highest cataract
surgery rate among 6 Regions, which corresponds with the high density of
ophthalmologists in that Region. Region 6 comprises the highest density of
cataract surgery since it is a rural area containing an elderly population. Region 1
is the capital city of Taiwan, which contains a high population of youths, giving it
the lowest surgery rate. Since there is a fixed fee schedule, the payments made by
the Bureau do not differ significantly among regions.
92
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 20. Relationship Between Region and Costs for Cataract Surgery
Region No. of
Ophthalmologists
per 10,000
Residents
Number of
Cataract
Surgeries
Surgery Rate
per Ten
Thousand
Population
Total Cost Cost per
Episode
Cost of Drugs Drug Cost a s a
Percentage of
the Total Cost
Region 1 6.6/10,000 20,642 278/ten
thousand
517,214,302 5834 51,566,501 9.1%
Region 2 4.9/10,000 9,517 355/ten
thousand
57,827,930 5822 5774,965 9.9%
Region 3 5.3/10,000 13,179 349/ten
thousand
510,882,956 5826 S1,044,763 9.6%
Region 4 5.0/10,000 11,602 393/ten
thousand
59,726,786 5838 5953,225 9.8%
Region 5 7.0/10,000 13,671 431/ten
thousand
511,420,736 S835 51,096,391 9.6%
Region 6 4.4/10,000 2,555 513/ten
thousand
52.087,957 5817 5160,773 7.7%
Table 21 shows the number of ophthalmologists, the average charge for
cataract surgery, the number of cataract surgeries between Region 5 and the
remaining 6 regions. Region 5 has the highest density of ophthalmologists.
Ophthalmologists in Region 5 performed 25% and 22% of cataract surgeries in
district hospitals and physicians’ offices, respectively. These percentages are out of
proportion with the number of ophthalmologists in the region, 10% in district
hospitals and 7% in physicians’ offices. Ophthalmologists in Region 5 charge
significantly less than average for cataract surgery.
93
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 21. Comparison o f N um ber o f O phthalm ologists, A verage Charge o f
C ataract, Num ber o f C ataract Surgeries Between R egion 5 and 6 Regions
C ategorized by Institutions
Institutions No. of Physicians Average Charge No. of Cataract
Surgeries
Region 5 6 Regions Region 5 6 Regions Region 5 6 Regions
Medical Center 71
(18%),
396 $938 $1,480 528
(9%)
5,815
Regional
Hospital
21
(3% )
692 $913 $1,492 97
(2%)
5,376
District Hospital 105
(10%)
1,010 $1,484 $1,562 2,640
(25%)
10,553
Physician Office 128
(7% )
1,732 $2,236 $2,353 12,766
(22%)
57,928
Table 22 shows the ratio between the number of cataract surgeries
performed by male ophthalmologists and those performed by females. Although
15% of ophthalmologists in Region 5 are female, they account for only 3% of the
surgeries. In the 6 other regions, 19% of ophthalmologists are female, and they
perform 11% of cataract surgeries.
94
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 22. Ratio of Gender in Cataract Surgery Performance in Regions
No. of
Ophthalmologists in
Region 5
No. of
Ophthalmologists in 6
Regions
No. of Cataract
Surgeries Performed
in Region 5
No. of Cataract
Surgeries Performed
in 6 Regions
Male 202 (85%) 1,002(81%) 15,537 (97%) 71,194 (89%)
Female 37(15%) 242 (19%) 494 (3%) 8,478 (11%)
Total 239 (100%) 1,244(100%) 16,031 (100%)
79,672a
Note: a Includes cases with complications.
Table 23 displays the number o f cataract surgeries performed by
ophthalmologists in the four types of institutions. Ophthalmologists in physicians’
offices perform a high number of surgeries compared to physicians practicing in
other institutions. Ophthalmologists in physicians’ offices perform a cataract
surgery on average once every 2.5 days, compared to those in medical centers who
perform a cataract surgery only once every 10 days.
Table 23. Num ber o f Cataract Surgeries Perform ed C ategorized by Four
Institutions
Type of Institution Medical Center District Hospital Regional Hospital Physician Office
Total No. of
Cataract Surgeries
5,538 (7.8%) 4,925 (6.9%) 10,400 (14.6%) 50,303 (70.7%)
No. of
Ophthalmologists
Performing Surgery
163 (21%) 154 (20%) 117(15%) 345 (44.3%)
Average Number of
Operations per
Ophthalmologist
34 32 89 146
95
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5.1.2.4 Subsequent Surgery
The fourth part of the study focuses exclusively on claims related to second
surgeries for cataracts. A total of 11,277 patients (18.8%) had a second cataract
operation and 22 patients underwent third cataract surgeries, most likely because
one of the surgeries was a failure. In the Medicare population, 22% of patients had
a second cataract procedure within two years (Steinberg 1993). This study includes
demographic characteristics of patients who underwent second surgeries and the
length of time between the first and second surgery of the 11,277 patients who
underwent a second surgery within a year. The average interval between the first
and second surgeries is 48 days. The median is 23 days. Among these 11,299
second cataract surgeries, 9,766 (86.4%) were performed by same the
ophthalmologist who performed the first surgery; only 13.6% of patients (1,533)
sought care from more than one ophthalmologist. However, only 3.3% (372) of the
second surgeries was performed in a different type of institution than the first.
Table 24 compares the cost between second surgeries performed by the
same ophthalmologist who performed the first surgery and those performed by a
different ophthalmologist. It also compares the number of pre- and postoperative
visits patients made with the same or different ophthalmologist. There is very little
difference in cost between these two groups of patients; median and mode costs are
the same. Even the highest cost of the two groups differs little. However, the
96
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
average number o f pre- and postoperative visits differs significantly between
patients who did or did not change ophthalmologists. The group of patients who
changed ophthalmologists has a higher pre- and postoperative visit rate than the
group who did not change ophthalmologists for their second cataract surgery.
Table 24. Cost C om parison Between Patients C hanging O phthalm ologists or
W ithout Changing Ophthalm ologists, C ounting the Number o f Pre- and
Postoperative V isits in the Second C ataract Surgery
Mean of Cost
of Surgery
(SD)
Mode of Cost
of Surgery
Median of
Cost of
Surgery
Highest Cost
of Surgery
Number of
Pre- and
Postoperative
Visits
Changed
Ophthalmolo
gists
$839 ($39) $826 $835 $1,549 7.2
Non-
Changed
Ophthalmolo
gists
$837 ($28) $849 $834 $1,688 5.8
5.1.3 Results fo r Physician Characteristics in A ll Eye Care
After omitting physician records with incomplete information (such as
gender or birthday information), there are a total of 39,082 contracted physicians in
the files of the Bureau. O f these, 2,947 physicians submitted claims for eye care
services in 1997. There is only complete profile information for 2,664 of these
2,947 physicians; data for 283 physicians are missing.
97
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Among these 2,664 physicians, 12.5% are female physicians and 87.5% are
male. Some 14.6% of all eye care visits were made to female physicians, and
85.6% were made to male physicians.
Table 25 shows the practice setting by physician gender. It demonstrates
that female physicians are more likely to work in medical centers than physician
offices. Physician office has a higher percentage as the working setting for male
physicians than female physicians.
Table 25. V olum e o f Visits C ategorized by Physician G ender and Institutions
Medical
Center
Regional
Hospital
District
Hospital
Physician
Office
Total
Female 32.36% 24.82% 20.61% 10.66% 14.61%
Male 67.64% 73.18% 79.39% 89.34% 85.39%
Total 8.86% 7.32% 9.94% 73.88% 100.00
Table 26 demonstrates the charge made by male and female physicians for
diagnosis, surgery lab tests, and drugs for each visit. There is no significant
difference between male and female physicians’ fees for surgery, lab, diagnosis,
and drugs.
98
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 26. Cost Component per Visit Categorized by Physician Gender
Average of Surgery and
Lab Fee (S.D.)
Average of Diagnosis
Fee (S.D.)
Average of Drugs' Cost
(S.D.)
Female $8.3 ($57) $6.9 ($0.9) $4.5 ($8.6)
Male $8.6 ($63.8) $6.9 ($1) $4.3 ($9.2)
Table 27 displays the distribution of visits by physician gender and age
groups. It separates physicians into three age groups: younger than 40, 40-60 years
old, and older than. More than half the visits (52.2%) were made to physicians
younger than 40 years old. Some 43.22% of the visits were made to physicians
between the ages of 40 and 60. Only 4.58% of the visits were made to physicians
over 60. No visits were made to female physicians over 60 years old.
T able 27. Distribution o f Visits Across Physician G ender and Age Groups
Female Male Total
<40 9.79% 42.4% 52.2%
18.76% 81.24%
67.03% 49.66%
>= 40 < 60 4.82% 38.41% 43.22%
11.14% 88.86%
32.97% 44.98%
>= 60 0 4.58% 4.58%
100%
5.37%
Total 14.61% 85.39% 100%
99
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 28 shows the distribution of physician age across gender. The
average age of female physicians (38) is relatively lower than the average age of
male physicians (41). The oldest male physician is 78 years old; the oldest female
physician is 57 years old. Both male and female physicians start practicing at a
similar age, 24 and 25, respectively.
T able 28. Distribution of Physician A ge Across G ender
Age Female Physician Male Physician
Mean (S. D.) 38 (5.5) 41 (8.7)
Median 38 40
Mode 39 42
Oldest 57 78
Youngest 24 25
Table 29 demonstrates how physicians’ fees for diagnosis, surgery and labs,
and drugs differ across the three physician age groups. Younger physicians
charged much more than older physicians. Younger physicians also charged a high
proportion (59%) of fees for surgery and lab. In contrast, surgery and lab is a
relatively small proportion (24%) of the fees charged by older physicians. The
major cost in older the physician is diagnosis fee and relatively high in the drug
component.
100
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 29. A verage C ost per Person and Cost C om ponent per V isit Categorized
by Physician A ge
Age
Groups
Diagnosis Surgery & Lab Drugs Total
Mean % of Total Mean % of Total Mean % of Total Mean % of Total
<40 $8.1 23% $20.3 59% $6.1 18% $34.5 100%
>= 40 < 60 $18.4 27% $15.3 54% $5.3 8.7% $28.2 100%
>= 60 $7.9 59% $2.2 16% $3.3 25% $13.4 100%
Table 30 presents the number of physicians in various institutions across the
three age groups. A total of 1,550 physicians are younger than 40 years old, 902
physicians are between 40 and 60, and 213 are older than 60. Among these 2,664
physicians, the highest proportion of younger physicians works in medical centers.
Older physicians are likelier to work in physicians’ offices.
101
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 30. Number of Physicians in Various Institutions Across Age Groups
Age Groups/
Institution
Medical Center Regional
Hospital
District
Hospital
Physician
Office
Total
<40 562 464 370 563 1,959
17.3% 14.29% 11.39% 17.33% 60.31%
28.69% 23.69% 18.89% 28.74%
77.95% 74.84% 59.77% 43.71%
>= 40 < 60 156 149 212 544 1,061
4.80% 4.59% 6.53% 16.75% 32.64%
14.70% 14.04% 19.98% 51.27%
21.64% 24.03% 34.25% 42.24%
>= 60 3 7 37 181 228
0.09% 0.22% 1.14% 5.57% 7.02%
1.32% 3.07% 16.23% 79.39%
0.42% 1.13% 5.98% 14.05%
Total 721 620 619 1,288 3,248s
22.20% 19.09% 19.06% 39.66% 100.00%
5.2 Analytic Results
The above descriptive results show many variations in type of institution,
regions, as well as physician gender and age. How do these independent variables
s Some physicians work more than one institution, so the total number of physicians is higher than
2,664.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
really affect the cost per visit? To examine the results, the generalized linear
regression model was adopted. This model keeps all the independent variables
constant and focuses on one independent variable to determine how it affects the
cost per visit. For example, the region, the patients’ age and gender, and the
physicians’ age and gender can be kept constant in order to conclude the
differences between cost per visit in medical centers and physicians’ offices.
Table 31 displays the utilization cost differences among independent
variables such as patients’ age, gender, region, type of institution, diagnosis,
physicians’ characteristics, cataract surgery, and complications (or comorbidity) by
general linear model. The dependent variable is cost per visit.
All the independent variables significantly influence the cost per visit
(p=0.0001). With all other independent variables kept constant, the cost per visit is
US$11 (35%) higher in medical centers than in physicians’ offices. In addition, the
cost per visit is US$7 higher in regional hospitals than in physicians’ office.
Regarding the variation in cost per visit, Region 1 is slightly higher (0.7%) than
Region 6 after controlling other independent variables. Region 2 is 1.3% lower,
Region 3 is 4.9% lower, Region 4 is 1.2% lower and Region 5 is 4.8% lower than
Region 6, keeping other independent variables constant. Glaucoma, the most
expensive illness to treat, costs an average of $5.60 more per visit than other
diagnoses. Moreover, glaucoma visits cost $14 more than non-surgical cataract
visits. Table 31, transformed by log, shows that male physicians charge 3% more
103
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
per visit than female physicians when other independent variables are constant. In
addition, when other variables are constant, older physicians charged less than
younger physicians at a rate o f 0.13% less per year older. Thus, physicians’ gender
and age did influence the use o f resources.
T able 31: M ultivariate Statistics (GLM ) T esting Cost ($N T ) per Visit
Influenced by Various Independent Variables
Table 3 1 A
Dependent Variable is the Total Cost per Visit
Coefficient t Value p Value S.D.
INTERCEPT 28,230 4,459 0.0001 6.3
PATGENDER Female -10 -20 0.0001 0.5
Male 0
PATAGE 2.12 50.50 0.0001 0.042
PATAGESQ 0.007 14.22 0.0001 0.0005
CON_TYPE Medical Center 333.71 358.16 0.0001 0.9317
Regional Hospital 211.31 214.47 0.0001 0.9853
District Hospital 49.31 58.33 0.0001 0.8453
Physician office 0
REGION Region 1 3.18 1.79 0.073 1.7754
Region 2 29.37 15.94 0.0001 1.8423
Region 3 -14.14 -7.86 0.0001 1.7991
Region 4 26.88 14.87 0.0001 1.8078
Region 5 -18.095 -10 0.0001 1.8096
Region 6 0
DIAGNOSIS A230 168.39 123.55 0.0001 1.3629
A231 -264.45 -272.54 0.0001 0.9703
Table continues next page
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 31 (Continued). Multivariate Statistics (GLM) Testing Cost ($NT) per
Visit Influenced by Various Independent Variables
Dependent Variable is the Total Cost per Visit
Coefficient f Value p Value S.D.
A232 -230.01 -123.35 0.0001 1.8647
A233 -186.44 -334.41 0.0001 0.5575
A234 -161.79 -85.61 0.0001 1.89
A235 4 0 .2 5 -16.07 0.0001 2.5046
A239 0
PHYSEX Female -18.663 -25.95 0.0001 0.7193
Male 0
PHYAGE -0.96282 -32.3 0.0001 0.02981
COMP Without complications 4,425.46 -899.66 0.0001 4.919
With complications 0
CATSUR Without cat. surgery -23,356.66 -7,405.6 0.0001 3.1539
With cat. Surgery! 0
Unit NTS, R2=0.8284.
The following Table 3 IB shows that the dependent variable was
transformed from dollar amount to percentage by log.
T a b le 31 B
Dependent Variable is the Total Cost per Visit
Coefficient t Value p Value S.D.
INTERCEPT 12.62 4552.19 0.0001 0.00277
PATGENDER Female -0.02497 -114.31 0.0001 0.0002
Male 0
PATAGE 0.00445 242.43 0.0001 0.000
PATAGESQ -0.00003 -129.33 0.0001 0.0002
Table continues next page
105
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 31 (Continued). Multivariate Statistics (GLM) Testing Cost ($NT) per
Visit Influenced by Various Independent Variables
Table 31 B
Dependent Variable is the Total Cost per Visit
Coefficient (Value p Value S.D.
C O N JY P E Medical Center 0.3487 854.46 0.0001 0.00041
Regional Hospital 0.1929 447.05 0.0001 0.00043
District Hospital 0.0316 85.55 0.0001 0.00037
Physician office 0 0 0 0
REGION Region 1 0.0068 8.79 0.0001 0.00078
Region 2 -0.0134 -16.66 0.0001 0.0008
Region 3 -0.0487 -61.76 0.0001 0.0008
Region 4 -0.0125 -15.85 0.0001 0.00079
Region 5 -0.0478 -60.25 0.0001 0.0008
Region 6 0 0 0 0
DIAGNOSIS A230 0.3954 662.25 0.0001 0.0006
A231 -0.1678 -394.87 0.0001 0.00043
A232 -0.2775 -339.73 0.0001 0.00082
A233 -0.1969 -806.22 0.0001 0.000244
A234 -0.0358 -43.19 0.0001 0.00083
A235 -0.1186 -108.11 0.0001 0.0011
A239 0 0 0 0
PHYSEX Female -0.0275 -87.27 0.0001 0.0003
Male 0 0 0 0
PHYAGE -0.00134 -102.9 0.0001 0.00001
COMP Without complication -2.50363 -1161.95 0.0001 0.00215
With complication 0 0 0 0
CATSUR Without cat. Surgery -4.1783 -3024.41 0.0001 0.0014
With cat. surgery 0 0 0 0
106
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
To further understand how the cost per visit differs between different types
of institutions and regions, the various scenarios are applied in the GLM equation.
Equation
Y=a+PiPATage+P2PATage2 +p3PATsex+P4RG+P5FL+p6 D+P7PHYAGE+
p8PHYSEX+p9S+p ,0C+8i
For example, to compare the cost of an office visit for a 65-year-old male
diagnosed with glaucoma without complications to a 45-year-old male physician in
a medical center to the same visit made to a physicians’ office, the following
formula was used:
Y=NT$28,230+NTS2.12*65+(NTS0.007)2*65+NT$3.18+NT$168.39+
(NT$-0.96282*45)+(NTS-4,425)+(NTS-23,357)
Y=NT$714.24
NT$714.24-NT$333.71=NT$380.53
NT$714.24/NT$380.53=188%
107
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Result: The cost per visit to a medical center is 88% higher than the cost to
visit a physician’s office for a 65-year-old male patient diagnosed with glaucoma in
Region 1.
T able 32. Scenario 1: C om parison o f the Cost per V isit for 65-Year-O ld
Patients Between M edical C enter and Physician O ffice in G laucom a Provided
by 45-Y ear-O ld Male Physicians3
Region 1 Region 2 Region 3 Region 4 Region 5 Region 6
88% 82% 92% 83% 93% 88%
Note: a Percentage stands for the average difference in cost per visit between medical
centers and physicians’ office.
T able 33: Scenario 2: C om parison o f the C ost per V isit Between M edical
C enter and Physician Office in G laucom a to M ale Patient Age 65 Provided by
45-Y ear-O ld Fem ale Physician3
Region 1 Region 2 Region 3 Region 4 Region 5 Region 6
92% 86% 97% 87% 103% 98%
Note: a Percentage stands for the average difference in cost per visit between medical
centers and physicians’ office.
108
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6. A p p l ic a t io n o f t h e F in d in g s :
P h y s ic ia n P r o f il in g in A l l E y e C a r e
A total of 1,492 physicians claimed less than 120 visits in 1997. Most of
them were not ophthalmologists but other specialists who provided eye care. These
1,492 physicians were excluded from the comparison showing the variation
between practices.
Table 34 shows the volume of visits by number of physicians. Among these
1,172 physicians, the total visits were 12,566,487. On average, each physician
claimed 894 visits per month and 10,722 visits per year. Analysis shows that most
physicians claim between 1,201 and 24,000 visits per year. Seventy-two percent of
the physicians claimed number of visits in this range.
Table 34. Number o f Physicians by Annual Volum e o f Visits for Eye Care
No. of
Visits
1-120 121-
1,200
1,201-
6,000
6,001-
12,000
12,001
24,000
24.000
36.000
36,001
48,000
48,000
and
Above
All
No. of
MDs
1,492 183 309 277 261 107 26 9 2,664
109
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 35 presents the profiles of the 17 physicians who claimed the highest
volume of visits. These 17 physicians claimed more than 3,400 visits per month
and 41,000 visits per year, almost 10 times the number of average visits for the
entire group. The average charge per patient among these physicians is higher than
average. However, the average charge per visit among this group of 17 physicians
is close to the average for the entire group. Fifty percent o f them even charged less
than the average.
Table 36 presents the patient re-visit rate to the same physician and the
charge per patient and charge per visit among those physicians with high re-visit
rates. A total of 4,474,750 beneficiaries sought eye care in 1997. The average
number of visits is 2.9 per person for the year. Furthermore, the average number of
visits to the same physician is 1.94. Table 4 shows that 58% of patients only
sought eye care once a year. Table 36 exhibits the top 12 physicians whose
patients’ re-visit rates are almost double or higher than double compared with the
others. As seen in Table 35, the two physicians with the highest re-visit rates have
a lower than average charge per visit ($20). The 7 physicians with the highest re
visit rates only charged half the average fee per visit.
110
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 35. Profile of the 17 Physicians with the Highest Number of Visits
MD
Obv
No. of
Unduplicated
Patients
No. of
Visits
Avg. Visits
/Month
Total Cost Avg.
Charge
/Person
Avg.
Charge
/Visit
1 21,310 53,177 4,431 $1,144,924 $54 $22
2 16,766 52,510 4,376 $1,233,955 $74 $23
3 16,563 52,102 4,342 $1,166,692 $70 $22
4 19,193 51,104 4,259 $1,083,492 $56 $21
5 16,553 50,216 4,185 $795,253 $48 $16
6 25,959 49,760 4,147 $580,300 $22 $12
7 30,302 49,548 4,129 $425,502 $14 $9
8 25,967 48,580 4,048 $865,387 $33 $18
9 18,618 48,086 4,007 $1,029,355 $55 $21
10 28,345 47,137 3,928 $1,029,921 $36 $22
11 14,580 44,250 3,688 $1,105,255 $76 $25
12 13,253 43,964 3,664 $1,000,724 $76 $23
13 13,162 43,613 3,634 $747,115 $57 $17
14 28,328 42,554 3,546 $511,537 $18 $12
15 20,461 41,580 3,465 $546,579 $37 $13
16 14,948 41,580 3,465 $546,579 $37 $13
17 15,973 41,332 3,444 $633,111 $40 $15
---
Total
MDs =
2665
5,984,732 12,989,686 406 $263.6 M $44 $20
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 36. Distribution of Patient Re-Visit Rate to Same Physician
1
Visit
2
Visits
3
Visits
4-6
Visits
7-9 IQ-
12
13-
18
19-
24
25-
36
37
48
49
96
No. of
Undupli
cated
Patients
No. O f
Visit
Re-
Visit
Cost/
per
Person
Cost
/ per
Visit
1 2828 1432 841 1404 693 383 358 142 102 20 10 8203 37241 4.54 $81 $19
2 1062 378 214 372 180 98 116 41 34 6 2 2503 10673 4.26 $79 $19
3 1761 573 404 943 607 268 150 13 1 1 4721 18893 4.0 $154 $38
4 1539 581 329 614 354 291 130 14 2 3854 14972 3.88 $156 S40
5 1075 527 268 385 171 92 93 39 20 5 2 2677 10234 3.82 $156 $40
6 2625 1692 1042 1346 483 230 141 60 40 17 8 7684 27593 3.59 $32 $10
7 2315 1033 563 832 320 160 139 66 36 8 3 5475 19333 3.59 $36 $10
S 1129 557 366 635 228 91 53 16 1 1 2 3088 10647 3.45 $143 $41
9 3571 1496 761 796 313 175 220 89 67 16 8 7512 25308 3.37 $102 $30
10 816 304 133 190 115 75 50 20 4 1 1708 5738 3.36 $151 $45
1 1 6473 2550 1214 1460 616 308 313 156 118 28 17 13253 43964 3.32 $76 $23
12 6112 2568 1257 1670 662 313 309 137 100 22 12 13162 43613 3.31 $57 $17
Table 37 presents 13 physicians whose charge per visit is three standard
deviations higher than the average ($20). The physician with the highest charge per
visit charges ten times more than average. These physicians did charge much more
per visit and per patient than physicians with a high number o f visits and re-visits.
Three of them claimed more than one million dollars from the bureau in one year.
112
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 37. Top 13 Physicians with High Charge per Visit
MDOBS Total No. of
Unduplicated
Patients
Total No. of
Visits
Total Cost Avg. per Person Avg. per
Visit
1 272 346 $77,423 $285 $224
2 407 453 $61,008 $150 $135
3 4585 7669 $985,865 $215 $129
4 167 175 $18,450 $110 $105
5 4041 10666 $1,074,914 $266 $101
6 7575 19067 $1,608,943 $212 $84
7 183 183 $14,850 $80 $80
8 4700 9479 $741,631 $158 $78
9 5924 11067 $824,018 $139 $74
10 3754 8320 $591,026 $157 $71
11 2209 4046 $270,321 $122 $67
12 6208 18927 $1,123,224 $181 $59
13 1016 1317 $75,513 $74 $57
....
Total 5,984,732 12,989,686 263.6 Million $44 $20
113
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7. D is c u s s io n
7.1 Implication
7.1.1 All Eye Care
The value and uniqueness of this dissertation are based on the National
Health Insurance database. Because NHI covers everyone with little or no cost
sharing, it excludes the bias problem in sample selection. In addition, a detailed
study of eye care in Taiwan has never been done. Eye problems are related to
aging, making this a relevant study for Taiwan, a country whose elderly population
is growing; more than 7% of people are older than 65. The epidemiology of eye
conditions and expenditures in eye care is critical for eye care-related professional
associations and the NHI Bureau as well.
Only 22% of Taiwan’s beneficiaries paid claims for eye care services in
1997 compared with over one-third of Americans who suffer from eye disease or
functional abnormalities in one or both eyes (DHHS 1983; National Center for
Health Statistics 1983). However, only about half o f those needing treatment are
receiving it in the USA (DHHS 1987).
114
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Counting the frequency of visits for different diagnoses and types of
institutions helps us to understand the severity of each disease, patient behavior,
and physician practice patterns. O f these seven diagnoses, glaucoma and cataracts
average the highest number of visits. More than 60% of all patients with eye
problems visit their physicians only once a year for blindness and low vision,
conjunctivitis, lacrimal system disease, strabismus, and eye movement disease.
Table 3 shows that the average age of patients who made eye care visits was much
younger in Taiwan than in the USA. However, after age 65, the visits per person
making eye care visits in the USA increases steeply. This could be caused by
Medicare health insurance availability.
For easier cost comparison with other countries, charges have been
converted to US dollars at a rate of 30NTS to 1USS. The average cost per visit
(S20) in Taiwan is relatively low compared with the cost of visits in the US: $61
for a routine eye exam and $62 for a visual field test (Gauer et al. 1994). For
glaucoma, the average fee for outpatient visits, including the nurses’ fee, is $28 in
Taiwan compared with $59 in the USA, $49 in Canada, and $71 in Switzerland, not
including the nurses’ fee. The charge per visit in Taiwan is much cheaper than in
those developed countries, while the number of visits per year is higher in Taiwan.
There are several possible explanations for the high volume of visits, such as fee-
for-service as payment policy, the lack of a referral system, low co-payments,
generous reimbursements in tertiary institutions, and easy access to eye care. In
115
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
fee-for-service payment policies, physicians set their target income and induce
demand (Einsenberg 1986). In Taiwan, physicians spend less than 5 minutes on
average with each patient since patients have low expectations from their
physicians. In fact, this explains how physicians in Taiwan can easily manage the
number of visits needed to maintain their desired income level.
The lack of a referral system in Taiwan allows patients to seek care at any
type of institution. The only penalty for visiting a higher institutional level is a
small increase in co-payment. For example, the co-payments are S1.60 for a
physician office visit, $3.33 for a visit to a district hospital, S5 for a visit to a
regional hospital, and $6.60 to visit a medical center. Due to the Bureau’s
reimbursement norm, patients believe that they will receive better care and better
drugs in medical centers than in physicians’ offices. To verify this statement, Table
6 shows that medical centers claim $31.40 per visit, 75% higher than in physicians’
offices. Cost per patient in a medical center is 96% higher than in a physician’s
office ($88.90 vs. $45.30). Moreover, drug cost in a medical center is almost
double that in a physician’s office ($7.20 vs. $3.80). The number of visits (2.8)
patients make to medical centers is not much higher than the number of visits made
to other types of institutions (2.4, 2.5, and 2.5, respectively). In Table 31, after
adjusting for age, gender, region, and diagnosis by GLM model, there are no age,
gender, diagnosis, and region differences among these four institutions. We still
see that the cost per visit in medical centers is $11 higher than in the physicians’
116
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
offices, even after controlling for possible confounding variables. For a 65-year-
old male patient visiting a 45-year-old male physician to treat glaucoma, the cost
per visit in a medical center is 88% to 93% higher than in a physician’s office. For
the same patient visiting a 45-year-old female physician, however, the cost per visit
in a medical center is 87% to 103% higher than in a physician’s office. Thus, in a
system without referrals, with low co-payments and generous reimbursements,
tertiary hospitals probably generate medical resource waste.
Regarding accessibility, almost all institutions provide evening clinic hours
for working people. This allows working people to make medical appointments
that do not interrupt their jobs. Since people do not have to take time from work to
seek medical treatment, evening clinic hours may help create unnecessary demand
for beneficiaries to utilize more services.
7.1.2 Cataracts
In the second descriptive analysis for general cataract and cataract surgery,
the data showed that the Bureau spent US$ 108 millions in 1997 for both inpatient
and outpatient cataract services (NHI internal report). A total of USS87 million
went towards cataract outpatient services (US$14 million for non-surgical services,
US$73 million for surgical services), where intra-operative surgery accounted for
US$81.6 million, and the remaining US $5.4 million was spent on pre- and
postoperative procedures.
117
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The prevalence of cataracts in Taiwan is very similar to that in other
countries. For example, as people get older, the prevalence of cataract gets
increases. The Nepal Blindness Survey found a prevalence rate of 2.8% in the
overall population (Brilliant et al. 1985). The NHI data presented here show a 2%
prevalence rate in Taiwan. The prevalence rate is lower in Taiwan than in
developing countries. However, the number of visits in Taiwan is much higher
than the USA in all age groups.
According to the 1997 NHI claims data, there are 21,774 inpatient cataract
surgeries and 83,582 outpatient cataract procedures. Improvements in safety and
cost-effectiveness have expedited the outpatient cataract surgery. Since 1984,
cataract surgery in the United Stated has been performed almost exclusively as an
outpatient procedure, using local anesthesia in conjunction with intravenous
sedation (Schein et al. 2000). The rates of intraoperative morbidity and mortality
associated with cataract surgery are low for outpatient procedure (Breslin 1973;
Quigley 1974). In Scotland, only 1% of cataract extractions were performed as day
cases in 1992. The Study Group on the Management of Ophthalmology Services in
Scotland recommended that this figure should rise to 30% by the end o f 1993 and
to 80% by the end of 1997 (Scottish HSAC 1992). In Taiwan, 81% of the cataract
surgeries are done on an outpatient basis. Performing this surgery on an outpatient
basis is not only cost-effective but also time-effective to cataract patients and
family members.
118
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A change in threshold for cataract surgery can have a dramatic impact in the
number of people who require surgery. The number of cataract operations
increases 2.5 times as the acuity criterion changes from less than 6/60 to less than
6/24 and fivefold if it goes to less than 6/12 in the U.S.A. (Taylor 2000). In one
prospective study, cataract surgery with lens implantation resulted in a visual acuity
of 20/40 or better in 88% of the cases studied (Straatsma et al. 1983). In many
developed countries an acuity of less than 6/9 is regarded as the threshold since the
ability to drive a car which requires visual acuity is often essential for the elderly
(and other potentially disabled groups) to live independently in their community
(Monestam and Wachmeister 1997). In Taiwan, however, driving a car is not
required for independent living among the elderly. Thus, the threshold in Taiwan
can be as low as 6/15 for someone eligible for cataract surgery.
To get information on a complete episode of care for related to cataract
surgery, a six-month time period was used. The goal of this study was to include
all cataract-related visits, including visits before and after surgery, as well as
surgery itself. With the exception of January (6.3%) and February (4.7%), the
Chinese new year time period, the volume of cataract surgery remained relatively
stable during the remaining months (9%) with the exception of August (6.1%).
Thus, to prevent distorting the data and assure a fair distribution of cataract surgery
across several months, we examined only the middle six months of the year.
119
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In Taiwan, access to the medical institutions where cataract surgery
availability is extremely convenient with the exception of some rural area.
Therefore, the time between the preoperative visit and the surgery should not be too
long. This study showed that in fact 75% of the patients had surgery within 13
days of their last visit prior to surgery, with an average of 1.9 preoperative visits.
However, one surprising result was that there were no claims for preoperative visits
for 45% of the patients. There was no significant difference in age or sex between
those with and without a preoperative visit. In fact, the only difference is the
location which physician office has the lowest preoperative visit (52%). There are
several possible explanations for this situation. First, some patients utilized other
medical providers besides ophthalmologists to determine their cataract problem.
Second, ophthalmologists, especially in the physician offices, have absorbed the
cost. It could be understood that if ophthalmologists hadn’t done any
comprehensive tests in preoperative visits, they were willing to offer free service in
order for patients to save co-payments. This could be a gesture to please potential
clients. Third, another possibility is that some patients got the cataract surgery in
the first visits since outpatient cataract surgery only takes 20 minutes. Finally, we
didn’t acquire data on some pre- and postoperative visits due to our one-year data
limitation, but this number is likely to be small. For testing the hypothesis in which
only April to September cataract surgery cases have been assessed in this study,
another hypothesis was adopted in which cataract surgery cases were calculated
120
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
between February and December. The data showed 48% of cataract surgery
patients without any preoperative visits. Comparing these two hypotheses (45% vs.
48%), it is likely that big percentage of patients indeed haven’t been claimed by
providers for preoperative visits.
Most cataract surgery patients are elderly patients with serious co-existing
illnesses. As a result, many physicians believe that a medical examination with
laboratory testing must be performed before a patient can be considered eligible for
surgery. However, a recent study found that routing medical testing before cataract
surgery does not increase the safety of the surgery and, for the most part, is
unnecessary (Schein et al. 2000). Moreover, it has been demonstrated that 60% to
70% of laboratory tests ordered preoperatively are not required according to the
history or the results of the physical examination (Kaplan 1985). It seems that
physicians in Taiwan went by Schein’s finding.
For 38% of the patients without postoperative surgical claims, it could be
possible that these patients received care by ophthalmologists who did not charge
the Bureau for these follow-up visits. The Bureau did promulgate a rule that two
visits right after surgery are part of the surgical cost. According to the data,
medical centers followed the Bureau rule the best. However, we found that if
patients did have postoperative visits, 50% of these visits occurred the day after
surgery, demonstrating that some ophthalmologists didn’t obey the rule and the
Bureau couldn’t screen this situation out.
121
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The annual cost o f cataract surgery is about $34 billion in the United States
for 12% of the Medicare budget (Javitt et al. 1995). The high cost associated with
cataract surgery is of concern for Medicare officials. It is also a concern for
officials in Taiwan where the total spending in cataract surgery was 23%
($60million) of total spending ($264 million) on eye care. After adding the costs of
the preoperative visit, intra-operative surgery and postoperative visit, the total cost
is $842 per episode. The structure of the cost is as follows: 1.7% for the
preoperative visit and 4.48% for the postoperative visit compared with 4% in pre-
and 6% in postoperative visits and the cost of episode $2,500 in the Medicare
patient in the year of 1991 (Steinberg 1993).
In Table 17, the age distribution of cataract surgery in Taiwan is compared
with the United States (the data from June 1, 1995 to June 30, 1997). It is obvious
that the average age for cataract surgery is younger for the Taiwanese patients
receiving cataract surgery than Americans. Over half (56%) of Taiwanese patients
receiving cataract surgeries were female, while in America 61% o f cataract surgery
patients were female (Schein et al. 2000).
For the geographic region variation study, 2.5% of the beneficiaries went
for surgery in a different region versus their first cataract visit. Compared to
Medicare beneficiaries 5% of the cataract surgery was performed in a contiguous
state and 2% of the cataract surgery was performed in a non-contiguous state
relative to the beneficiary’s state of residence. (Javitt et al. 1995). Tables 19 and 20
122
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
show that the number of cataract surgeries performed in Region 5 is consistent with
the number o f ophthalmologists in the region (the highest density of
ophthalmologists in Taiwan). The number o f ophthalmologists in district hospitals
and physician offices performed cataract surgeries is out of proportion to other
regions. The distribution number between male and female ophthalmologists
shows no difference compared to other regions. However, 97% o f the cataract
surgery has been done by male ophthalmologists in Region 5 versus 89% in total. It
generates an interesting question o f why female ophthalmologists haven’t
performed proportionally in this competitive region.
The capital o f Taiwan is in Region 1. The high density of ophthalmologists
in Region 1 can be explained by the pleasant living environment and larger number
of medical centers. These factors provide an attractive working environment for
ophthalmologists. However, in general, people in this region are younger, working
individuals, which suggests why the cataract surgery rate is much lower than in
other regions. Since Region 6 is the most rural and isolated area, fewer
ophthalmologists are willing to work even though that population is much older
than other regions where there are more ophthalmologists. Yet among all six
regions, the cataract surgery rate is the highest in Region 6.
More than 70% of patients are seen in the physician’s office for cataract
surgery, causing this group of ophthalmologists to be very busy compared to
ophthalmologists practicing in other institutions. This situation can be understood
123
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
that physicians are working in the tertiary hospitals spending more time on the
research and teaching duties with less incentive to make a profit. While there are
1,223 ophthalmologists registered in Taiwan, only 779 (64%) of ophthalmologists
performed cataract surgery in 1997. Since there is no sub-specialty medical system
in Taiwan, it is hard to make more specific comments on this situation.
A recent study showed that 51.7% of patients would undergo a second
cataract surgery during the two years observation period (Castells et al. 2000).
Patients under 65 were more likely to undergo a second eye surgery as compared
with persons 65-74 (odds ration 3.9 versus 1.8) (Castells et al. 2000). Therefore,
since the age for cataract surgery is younger in Taiwan than in the USA, the Bureau
should expect there might be a higher number of second surgeries needed than
other places.
As reported made by Javitt (1993), patients that underwent cataract surgery
in both eyes reported greater improvement in subjective visual function than did
those who underwent surgery in only one eye. Some 18.8% of the patients went to
a second surgery within a year, according to the database in Taiwan. Since
physician accessibility is so convenient, physician loyalty has become an
interesting question. The data show that only 13.6% o f the patients went to
different physicians for their second surgery. However, only 3.3% of the surgeries
were done in different levels of institutions, which means the quality of institution
is not a major concern for patients but it is for physicians. Regarding gender in this
124
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
issue, male patients are 8% more likely to change physicians than female patients.
There is no significant age difference between patients with changed physician and
non-changed physician. The average age for second surgery patients is 70 years.
Moreover, there is little difference in the amount charged when a different
physician performed the second surgery on the same patients. However, the
number of pre- and postoperative visits varies by 1.4 visits which is different
between the group that went to the same ophthalmologist for both procedures and
the group that saw two physicians. Actually, this variation can be explained since
patients who go to a different physician for the second surgery may need to have
some lab tests performed by both physicians.
7.1.3 Physician Characteristics
Exactly 12.5% of the female physicians render 14.6% of the eye visits
which is 2% more than their population. However, they charge is 3% less than
male physicians. In addition, the female physician is likely to work in the medical
centers that provide better research and teaching environments. After medical
centers, females prefer to work in well-established regional hospitals. The least
liked working environment for females is solo practice. This indicates that female
physician are less interested in entrepreneurial type jobs than male physicians. Age
is also significantly related to resource use and practice setting. The average cost
per visit to a younger physician is more than twice the average cost per visit to an
125
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
older physician. There are two explanations. One is that younger physicians have
performed more surgical procedures. Second, a large percentage o f the younger
physicians work in medical centers where more resources are consumed. From
Table 31, we found that younger physicians charged 0.13% (per year old) more
than older physicians per visit. These results match the findings by Eisenberg
(1986) who showed that younger physicians are more likely to use more services.
Seventy-nine percent of the physicians older than 60 work in their own
offices, and more than half (51.27%) o f physicians between 40 and 60 years old
work in offices. Only 29% of the physicians younger than 40 work in physician
offices. In medical centers, 78% o f the physicians who see eye-related patients are
younger than 40. This is a trend in Taiwan because younger physicians work in the
tertiary hospitals to get better training. After 40, physicians are ready to run their
own businesses. It also explains another market strategy, showing that younger
physicians expect to gain both professional experience and long lasting patient
relationships in tertiary hospitals. Later on, these physicians can establish their own
businesses since the outpatient and inpatient volume is huge in the tertiary
hospitals. Younger physicians see patients less than older physicians because
younger physicians are hospital employees in the tertiary hospitals, and thus don’t
have strong financial incentives to see more patients. In addition, younger
physicians spend more time in research and teaching.
126
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A relatively low number of physicians (17%) will practice in more than one
institution during their career. In Taiwan, physicians are either solo practitioners or
are hired from hospitals. Group practice is a relatively new development there.
However, this might change later on if the pressure is high from the Bureau o f NHI.
In theory, group practice can fully utilize manpower and better plan for overhead
costs. Specifically, group practices can invest more in the claim process and get
higher reimbursements. Before implementation of the NHI, the income o f the
physician was between USS66,000 and US$100,000 per year, 12 to 18 times
greater than the average income. After implementing NHI, the physician income
increased from US$159,000 (1995), US$175,400 (1996) to US$185,900 (1997),
respectively, (Lin 1999) after adjusting the incomes using the consumer price
index. The increase of income is unbelievably high which helps explain why the
Bureau has run a deficit since 1998.
7.1.4 Physician Profiling
The high volume rendered by physicians does not necessarily equate with
higher charges per visit. In fact, 50% of them actually charged less than the
average of other physicians. However, since many of them see more than 100
patients per day, 30 days a month, the quality o f care is not expected to be too high.
Physicians with a high rate o f re-visits do not necessary charge higher prices either.
Obviously, the major factor generating the income is volume, not the unit charge.
127
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Although Table 37 shows the physicians with high charges, most of these
physicians have a higher volume o f surgical cases. This profile helps Bureau or
managers in hospitals recognize which physicians have high unit charges and total
income from the NHI. Physician profile can be very complex through the use of
case mix adjustment in order to be more accurate to make judgements. Since
multivariate statistical analysis was used to find the results shown in the Table 31,
many findings have been disclosed after the adjustment of independent variables.
Thus, these physician profiles do not incorporate the case mix adjustment method.
In Asian societies, most physicians work more than 40 hours per week
taking care of patients and, thus, achieve a high income. However, the tradeoff is
that physicians remain under constant pressure and work long hours.
7.2 Implications and Recommendations
The evidence presented in this study supports the idea that payment policy,
practice setting, geographic difference, patient characteristics, and physician profile
are important in explaining the variations in resource use in eye care. The findings
in this study, however, raise important issues for payment reform and further
investigation. These are presented below.
1. The baseline information of eye care can be used in three ways. First
of all, the Bureau can use these findings to design a new cost-
128
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
containment payment policy. Issues examined include finding ways to
prevent medical centers from overcharging and controlling the cataract
surgery by episode of care. Through the use of physician profiles, the
Bureau can monitor some outlier physicians closely in order to avoid
over-utilization. Additionally, the Department o f Health can utilize the
epidemiological findings to improve public health efforts. For
example, out of seven eye-related diagnoses, the highest visit is in
conjunctivitis. This can be explained in part by the hot and humid
weather in Taiwan. However, finding a solution and creating a
hygienically safe environment to decrease eye infection can be a major
issue in public health. Thirdly, the ophthalmology association can use
these findings to train their members. Since the resources are scarce,
the association can use physician profiles to control their members
internally by creating internal peer pressure, which is one of the most
efficient methods to manage professionals. The association also can
use this information to bargain with the Bureau when adjusting the fee
schedule.
2. While representing only 7.8% of the total persons treated in medical
centers, the total cost of treatment in medical centers is almost twofold
(13.4%). Obviously, medical centers utilize more resources than other
institutions. Due to the lack of a referral system and low co-payment,
129
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
medical centers not only waste resources, but they also misuse medical
manpower and equipment. The Bureau should set new policies in
terms of a referral system in order to prohibit beneficiaries’ resource
abuses. However, the Bureau should set some mechanism or incentive
for medical centers to encourage them to focus more on inpatient care,
research, teaching duties, and severe disease care.
3. Outpatient visits account for 67.7% of the total National Health
Insurance expenditure in Taiwan (NHI Profile 1998). The average is
14.49 visits per person per year compared to the US’s 4 visits per year.
It is obvious that outpatient services are overutilized. The findings
presented here show, that the unit cost is not significantly different
between institutions, except in medical centers. However, medical
centers only consumed 13.4% of the total cost, and the remaining
86.4% of the resources was spent in the other three institutions. Thus,
the volume of visits is the major concern. There are three
recommendations to contain this situation. First, the Bureau should
design a system similar to Ambulatory Patient Groups (APG),
Ambulatory Visit Groups (AVGs), Ambulatory Care Group, (ACG),
etc., and apply the method to the outpatient payment system. This
outpatient diagnosis-related grouping can do case mix adjustment
which can prospectively place the outpatient treatment resources
130
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
efficiently and effectively. Second, the Bureau should implement a
capitation payment system so that services are prepaid on a per
beneficiary, per month basis. In other words, capitation is paid the
same amount of money every month for a beneficiary regardless of
whether that beneficiary receives services and regardless of how
expensive those services are. This reimbursement method needs to be
carefully designed to determine how to calculate capitation for each
beneficiary based on age, gender, and region. Third, the Bureau should
decide the total budget for NHI in advance and eliminate the fee
schedule. Thus, volume will decide the unit price. This will be easy
for the Bureau to do in the beginning but, if the peer pressure is not
strong enough, the volume can be abused even more and damage the
health of beneficiaries due to poor care quality.
4. The cataract surgery rate is high in Taiwan. The average age for
people receiving cataract surgery is younger in Taiwan than in the US.
The Bureau should check the threshold often and monitor the system
closely to prevent over-utilization. It is necessary to do follow up
studies in order to assess effectiveness of the surgeries.
5. For post-cataract surgery visits, the first two visits are supposed to be
included in the surgery charge already. However, many
ophthalmologists didn’t follow this rule and were still reimbursed.
131
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
This management defect will encourage providers to play around the
rules. The Bureau should set up a computer-monitoring program and
load all the rules into the program in order to review claims strictly.
Doing this will not only reinforce authority, but it will also contain the
cost.
6. The results show that the number of cataract surgeries performed in
Region 5 is consistent with the number of ophthalmologists in the
region (the highest density of ophthalmologists in Taiwan). The
Bureau should work with the Department of Health and the Association
of Ophthalmology to control the distribution of ophthalmologists. In
addition, these 2 entities should do a sophisticated physician profile in
this region in order to know more about the practice pattern there and
find solutions.
7. The results also showed that 58% of the patients sought eye care only
once. More than 18% of the patients sought eye care twice a year.
Only a small percent of patients went to different physicians for eye
care. The reason for such high volume of visits is most likely
concentrated on a small group of patients. It is necessary for the Bureau
to generate patient profiles as well as physician profiles. Through
patient profiling, patients with a high utilization pattern should be put
into a case management scheme.
132
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8. As a large percentage of medical resources is handled by physicians,
physician profiling can be a very powerful tool to control resource use.
However, case-mix adjustment has to be inserted into the profiling
design to correct for variation in the distribution of one or more
confounding factors (health risk factors). Case-mix adjustment is a
statistical way to allow meaningful comparisons among providers
based on their patient population.
133
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7.3 Conclusion
National Health Insurance is an excellent public policy for people of all
countries. All persons, no matter how rich or poor, should be covered by the NHI.
This study on eye care services is the first detailed analysis of its type from an NHI
database. It disclosed baseline information that can be used by the government to
design or refine policies later on. It also can be used by the association of
ophthalmology for training and as a bargaining tool with the Bureau for setting new
fee schedules. The incidence rate rendered in different eye-related diagnoses can
be provided to the public health bureau. It also provides information for
international comparisons. Cataract surgery comprises the highest cost in eye care.
This study provides useful information for the Bureau to prepare the budget for
cataract surgery every year by using an episode of care approach.
For every 1 million people living in Taiwan, there were 4,700 cataract
surgeries performed in 1997. Thus, Taiwan was able to reach the goal set by WHO
because Taiwan exceeded the recommended minimum of 3,500 cataract surgeries
per 1 million before the year 2000. However, further research is needed to measure
the effectiveness of cataract surgery. For example, it would be useful to know how
much improving vision increases the quality of life for patients with cataracts in
this analysis. The analysis shows that the age for people receiving cataract surgery
in Taiwan is younger than for people in the US. Cost efficiency and effectiveness
134
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
for people undergoing early stage cataract surgery is another interesting research
topic.
Since 1998, cataract surgery has been reimbursed on a case-by-case basis
instead of in a retrospective fee-for-service fashion. Therefore, an interesting and
useful follow-up study is needed to determine if there is a difference in total
expenditures related to the new vs. traditional cataract surgery reimbursement
system in Taiwan.
In many respects, the data presented in this paper support the conclusion
that there is some waste in the NHI due to the payment reimbursement policy, non
referral system, low co-payment policy, generous reimbursement policy to tertiary
hospitals and easy access for beneficiaries to go to the institution of their choice.
The fee-for-service payment policy is always embedded in the supplier-induced
demand problem. However, the Bureau changed the cataract surgery
reimbursement policy from fee-for-service payment to case payment.
Nevertheless, for most services, the fee-for-service payment policy still remains in
effect. Moreover, the problems of a referral system and low co-payments haven’t
been solved yet, most likely due to political considerations. As a result, total
expenditures in the NHI are increasing while the average number of outpatient
visits remains the same. Developing an ambulatory diagnosis-related group, such as
APG or ACG, is urgent for controlling excessive outpatient visits. Setting global
budgeting is one method to control costs. However, if the mechanism is not
135
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
designed well, it might create more obstacles for providers and beneficiaries.
Diversifying the insurance coverage and privatizing the NHI are two additional
recommendations which can create more accountability and entrepreneurship in the
system.
In 1998, the Bureau ran a deficit due to slow reform action. Although the
satisfaction rate for the NHI is relatively high (76%), the financial feasibility will
be the most compelling reason for NHI to benefit people eternally.
136
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
B ib l io g r a p h y
Adamsons, I., Munoz, B., Enger, C., Taylor, H.R. 1991. Prevalence of lens
opacities in surgerical and general population. Archives Ophthalmology
109:993-97.
AMA (American Medical Association). 1999. Available from http://www.ama-
assn.orq/med-sci/a.apsite/index.htm: INTERNET.
Applegate, W.B., Miller, S.T., Elam, J.T., Freeman, J.M., Wood, T.O.,
Gettlefinger, T.C. 1987. Impact of cataract surgery with lens implantation on
vision and physical function in elderly patients. The Journal o f the American
Medical Association 257:1064-66.
Barnsley, J., Williams, A.P., Cockerill, R. 1999. Physician characteristics and the
physician-patient relationship. Impact of sex, year of graduation, and specialty.
Canadian Family Physician 45:935-42.
Becker, M.H. 1970. Sociometric location and innovativeness: Reformation and
extension of the diffusion model. American Sociology Review 35:267-82.
Bengt, J. et al. 1999. Glaucoma. Pp 123-156. San Francisco: Jossey-Bass
Publishers.
Bertakis, K.D., Helms, L.J., and Callahan, E.J. 1995. The influence of gender on
physician practice style. Medical Care 33(4):407-16.
Bertakis, K.D., and Robbins, J.A. 1989. Utilization of hospital services: A
comparison of internal medicine and family practice. Journal o f Family
Practice 28( 1 ):91 -6.
Bindman, A.B. 1999. Can physician profiles be trusted? The Journal o f Medical
Association 281)22):2142-43.
137
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Bowditch, J., Buono, A. 1994. A primer on organizational behavior. Pp. 201-25.
New York: John Wiley and Sons, Inc.
Branch, L.G., Horowitz, A., and Carr, C. 1989. The implications for everyday life
of incident self-reported visual decline among people over age 65 living in the
community. Gerontologist 29:359-365.
Brand, D., Quam, L., Sheila, L. 1992. Data needs of profiling systems in physician
payment review commission. Conference on Profiling. Vol. 5. Pp. 92-102.
Washington, D.C.: Physician Payment Review Commission.
Brenner, M.H., Curbow, B., Javitt, J.C., Legro, M.W., and Sommer, A. 1993.
Vision change and quality of life in the elderly. Archives Ophthalmology
111:680-85.
Breslin, P.P. 1973. Mortality in ophthalmic surgery. Internal Ophthalmic Clinic.
13:215-26.
Brilliant, L.B., Pokhrel, R.P., Grasset, N.C., Lepkowski, J.M., Lolstad, A., et al.
1985. Epidemiology o f blindness in Nepal. Bull., WHO 63:375-86.
Castells, X., Alonso, J., et al. 2000. Factors associated with second eye cataract
surgery. British Journal Ophthalmology 84:9-12.
Chassin, M.R., Kosecoff, J., Park, R.E., et al. 1986. Variations in the use of
medical and surgical services by the Medicare population. New England
Journal o f Medicine 314(5):285-90.
Chassin, M.R., Kosecoff, J., Park, R.E., et al. 1987. Does inappropriate use explain
geographic variations in the use o f health care services? A study of the
procedure. The Journal o f the American Medical Association 258( 18):2533-7.
Chesanow, N. 1999. Your report card is about to go public. 1999. Medical
Economics 12(4):220-235.
Chiang, Y.P., Wang, F., Javitt, J. 1990. Office visits to ophthalmologists and other
physicians for eye care among the U.S. population. Public Health Reports
110(2): 147-53.
Clark, J.A., Potter, D.A., McKinlay, J.B. 1991. Bringing social structure back into
clinical decision making. Social Science and Medicine 32(8):853-66.
138
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Cutler, S.J., and Grans, S J. 1988. Correlates o f self-reported everyday memory
problems. Journal o f Gerontology 43:S82-S90.
DHHS (U.S. Department of Health and Human Services). 1983. U.S. Eye
conditions and related need for medical care among persons 1-74 years o f
age: United States, 1971-72. Vital Health Statistics. Series 11:228. DHHS
Pub. No. (PHS) 83-1678. Washington, D.C.: National Center for Health
Statistics.
DHHS (U.S. Department o f Health and Human Services). 1987. Vision research: A
National Plan. NIH Publication. No. 87-2755. Bethesda, Maryland 1987:1.
Donabedia, A. 1968. Promoting quality through evaluating the process of patient
care. Medical Care 6:181-202.
Eisenberg, J.M. 1979. Sociologic influences on decision-making by clinicians.
Annals o f Internal Medicine 90(6):957-64.
---------------- . 1986. Doctors decisions and the cost o f medical care: The reasons
fo r Doctors' practice patterns and the way to change them. Ann Arbor, MI:
Health Administration Press.
----------------- .1995. Commentary: Are differences in outcome due to differences in
doctors or their patients? Health Services Research 30(2):291-94.
Eisenberg, J.M., and Nicklin, D. 1981. Use o f diagnostic services by physicians in
community practice. Medical Care 19:297-309.
Ellwein, A., Leon, B., Friedlin, V., McBean, A.M., Lee, P. 1996. Use of eye care
service among the 1991 Medicare population. Ophthalmology 103.11:1732-43.
Escarce, J. 1993. Effects of the relative fee structure on the use o f surgical
operations. Health Service Research 28(4, Oct.) 479-502.
Fahs, M.C. 1992. Physician response to the United Mine W orkers’ cost sharing
program: The other side of the coin. Health Service Research 27: 25-45.
Feldstein, M.S. 1970. The rising price of physicians’ services. Review o f
Economics and Statistics 52(5); 121 -33.
139
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Feldstein, P J. 1998. Health care economics. 5th ed. Pp. 255-66. Albany, NY:
Delmare Publishers.
Fineberg, H.V., Funkhouser, A.R. and Marks, H. 1985. Variation in medical
practice. A review of the literature. Industry and Health Care 2:143-68.
Foster, A. 1999. A global perspective: Output, outcome and outlay. Eye 13(7):449-
53.
Franks, P., and Clancy, C.M. 1997. Referrals of adult patients from primary care:
demographic disparities and their relationship to HMO insurance. Journal o f
Family Practice 45( 1 ):47-53.
Friedson, E. 1970. Profession o f Medicine: a study o f the sociology o f applied
knowledge. Chicago, IL: University of Chicago Press.
Gamick, D.W., Fowles, J., Lawthers, A.G., Weiner, J.P., et al. 1994. Focus on
quality: Profiling physicians’ practice patterns. Journal o f Ambulatory Care
Management 17 (3): 44-75.
Gauer, B.B., Erickson, T., Ulland, R.L., and Bleything, W.B. 1994. Access,
provision and cost of routing eye care: A comparison of Oregon optometrists
and ophthalmologists. Journal o f the American Optometric Association 4: 240-
47.
Gevirtz, Felicia. 1999. Physician Profile, unpublished.
Gold, M. 1981. Effects of hospital-based primary care setting on internists’
treatment of primary care episodes. Health Service Research 16: 383-405.
Hall, J.A., Irish, J.T., and Roter, D.L., et al. 1994. Gender in medical encounters:
An analysis of physician and patient communication in a primary care setting.
Health Psychology l3(5):384-92.
Hanning, M., Lundstrom, M. 1998. Assessment of the maximum waiting time
guarantee for cataract surgery: The case of a Swedish population. Internal
Journal Technology Assess Health Care 14:180-93.
Hemminki, E. 1974. The effect of a doctor’s personal characteristics and working
circumstances on the prescribing of psychotropic drugs. Medical Care
12(4):351-7.
140
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Hiller, R., Sperduto, R.D., Ederer, F. 1986. Epidemiologic associations with
nucleate, cortical, and posterior subcapsular cataract. American Journal o f
Epidemiology 124:916-25.
Hillman, A.L., Pauly, M.V., Kerstein, JJ. 1989. How do financial incentives affect
physicians’ clinical decisions and the financial performance of health
maintenance organizations? New England Journal o f Medicine 321:86-92.
Hillman, A.L., Welch, W.P., and Pauly, M.V. 1992. Contractual arrangements
between HMOs and primary care physicians: Three-tiered HMOs and risk
pools. Medical Care 30(2): 136-48.
Hofer, T.P., Hayward, R.A., Greenfield, S., Wagner, E.H., Kaplan, S.H., and
Manning, W.G. 1999. The unreliability o f individual physician report cards
for assessing the cost and quality of care o f a chronic disease. The Journal o f
the American Medical Association 277(5):379-382.
Hohmann, A.A. 1992. Gender bias in psychotropic drug prescribing in primary
care. Medical Care 30(2): 136-48.
Hombrook, M.C., Hurtado, A.V., and Johnson, R.E. 1985. Health Care Episode:
definition, measurement and use. Medical Care Review 42: 63-218.
Hsu, W.M. 1999. The demography and transition of ophthalmologists in Taiwan.
Journal o f Taiwan Ophthalmology 9(3): 138-164.
Ifiidu, O., Dawood, M., Iofel, Y., et al. 1999. Delayed referral of black, hispanic,
and older patients with chronic renal failure. American Journal o f Kidney
Diseases 33(4):728-33.
Javitt, J.C. 1993. The cost-effectiveness of restoring sight. Archives
Ophthalmology 111: 1615.
Javitt, J.C., Brenner, M.H., Curbow, B., Legro, M.W., and Stree, D.A. 1993.
Outcomes of cataract surgery. Archives Ophthalmology 111:686-91.
Javitt, J.C., Kendix, M., Tielsch, J.M., et al. 1995. Geographic variation in
utilization of cataract surgery. Medical Care 33:90-105.
Javitt, J.C., Wang, F. 1996. Blindness due to cataract: Epidemiology and
Prevention. Annual Review o f Public Health 17:159-77.
141
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Kahn, H.A., Leibowitz, H.M., Ganley, J.P., et al. 1977. The Framinghan eye study.
1. Outline and major prevalence findings. American Journal o f Epidemiology
106:17-32.
Kaplan, E.B., Sheiner, L.B., Boeckmann, A.J., et al. 1985. The usefulness of
preoperative laboratory screening. The Journal o f the American Medical
Association 253:356-81.
Kaplan, S.H , Greenfield, S , Gandek, B , et al. 1996. Characteristics of physicians
with participatory decision-making styles. Annals o f Internal Medicine
l24(5):497-504.
Kassirer, J.P. 1994. The use and abuse of practice profiles. New England Journal
o f Medicine 330(9):634-35.
Keeler, E.B., Buchanan, J.E., Rolph, J.M., Hanley, D.M. 1988. The demand fo r
episode o f medical treatment in the health insurance experiment. Supported by
the U.S. Department o f Health and Human Services. RAND Publication
series:R-3454-HHS. Santa Monica, California: RAND Corporation.
Keeler, E.B., Rolph, J.E., et al. 1982. The demand fo r episodes o f medical
treatment. RAND Publications Series (12). Santa Monica, California: RAND
Corporation.
Keeler, E.B., Solomon, D.H., et al. 1992. Effect o f patient age on duration of
medical encounters with physicians. Medical Care 20(11):1101-08.
Kelsey, J.L., Thompson, W.D., and Evans, A. 1986. Methods in observational
epidemiology. P. 187. London: Oxford University Press.
Kahn, H.A., Leibowitz, H.M., Ganley, J.P., et al. 1977. The Framinghan eye study.
1. Outline and major prevalence findings. American Journal o f Epidemiology
106:17-32.
Khan, S.S., Nessim, S., Gray, R., Czer, L.S., et al. 1990. Increased mortality of
women in coronary artery bypass surgery: Evidence for referral bias. Annals o f
Internal Medicine 112:561-67.
Klein, B.E.K., Klein, R., Linton, K.L.P. 1992. Prevalence of age-related lens
opacities in a population. Ophthalmology 99:546-52.
142
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Klein, B.E.K., Klein, R., Moss, S.E. 1997. Incident cataract surgery. The Beaver
Dam Eye study. Ophthalmology 104:573-80.
Kravitz, R., Greenfield, S., Rogers, W., Manning, W., et al. 1992. Differences in
the mix of patients among medical specialties and systems of care 267:1617-
30.
Kristensen, F.B., Kelstrup, J., et al. 1993. Computer-based longitudinal recording
of episode of care in general practice using the international classification of
primary care (ICPC). Scandinavian Journal o f Primary Health Care 11
(supplement):53-56.
Lasdon, G.S., Sigmann, P. 1977. Evaluating cost effectiveness using episode of
care. Medical Care 15:260-264.
Lasker, R.D., Shapiro, D.W., Tucker, A.M. 1992. Realizing the potential of
practice pattern profiling. Inquiry 29:287-297.
Lee, P.P., et al. 1995. Estimating eye care workforce supply and requirements.
Ophthalmology 102(12): 1964-71.
Lesk, M.C., Sperduto, R.D. 1983. The epidemiology of senile cataracts: A review.
American Journal o f Epidemiology 118:152-165.
Lichter, P.R. 1991. Statement in support o f the citizens’ budget proposal fo r the
National Eye Institute. Pp. 976-77.Departments of Labor, Health and Human
Services, Education, and Related Agencies Appropriations for 1992, Part 8A:
Testimony of members of congress and other interested individuals and
organizations. Washington: U.S. Government Printing Office.
Lin, H.H. 1999. The impact to the productivity and income of physician after
implementing NHI in Taiwan. Hospital 32 (4): 1-14.
Linn, B.S., Linn, M.W. 1994. Patient symptoms and physician prescribing patterns
in the elderly. Social Science Medicine 16(17): 151-8.
Lurie, N., Margolis, K.L., MacGovem, P.G., et al. 1997. Why do patients of
female physicians have higher rates of breast and breast and cervical cancer
screening? Journal o f General Internal Medicine 12( 1 ):34-43.
143
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Macario, A., Rozien, M.F., Thisted, R.A., K.S., Orkin, F.K. 1992. Reassessment of
preoperative laboratory testing has changed the test-ordering patterns of
physicians. Surgical Gynecology Obstetrics 75:539-47.
Mangione, C.M., Phillips, R.S., Lawrence, M.G., Seddon, J.M., Orav, E.J.,
Goldman, L. 1994. Improved visual function and attenuation of declines in
health-related quality of life after cataract extraction. Archives Ophthalmology
112:1419-25.
Marquis, M.S. 1983. Labortory test ordering by physicians: The effect o f
reimbursement policies. RAND Corportation Publication, R-2091-HFA. Santa
Monica, California: RAND Corporation.
McCarty, C.A., Keeffe, J.E., Taylor, H.R. 1999. The need for cataract surgery:
projections based on lens opacity, visual acuity, and personal concern. British
Journal Ophthalmology 83:62-65.
McNeil, B.J., Pedersen, S.H., Gatsonis, C. 1992. Current issues in profiling quality
of care. Inquiry 29:298-307.
Meyboon-De Jong, B. 1993. Episode-oriented data and quality assessment.
Scandinavian Journal o f Primary Health Care 11 (supplement):47-52.
Monestam, E., and Wachmeister, L. 1997. Impact of cataract surgery on car
driving: A population based study in Sweden. British Journal Ophthalmology
81:16-22.
Nadzam, D.M., Nelson, M. 1997. The benefits of continuous performance
measurement. Nursing Clinics o f North America 32(3):543-559.
National Center for Health Statistics. 1983. Vision Survey, United States, 1983.
Vital and Health Statistics. Series 10. No 162. (DHHS Publication No (PHS)
83-1590). Washington, D.C.:Government Printing Office.
National income per capita: Major indicators. 2000. Taipei, Taiwan: Executive
Accountant office, Executive Branch.
NHI (National Health Insurance). 1998. National Health Insurance Profile, 1998,
Republic o f China, Taiwan. Taipei, Taiwan: Bureau of National Health
Insurance, Department of Health.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Nutting, P.A., Shorr, G.I., and Burkhalter, B.R. 1982. Assessing the performance
of medical care systems: A method and its application. Medical Care 19:281-
296.
Pararajasegaram, R. 1999. Vision 2020—The right to sight: From strategies to
action. American Journal o f Ophthalmology 128:359-60.
Phelps, C, Mooney, C. Mushlin, A.L., et al. 1994. Doctors have styles—And they
matter. Unpublished manuscript.
Piland, N., and Lyman, K. 1999. Physician Profiling. San Francisco: Jossey-Bass
Publishers.
Pineault, R. 1976. The effect of prepaid group practice on physicians’ utilization
behavior. Medical Care 14:121-36.
Quigsly, H.A. 1974. Mortality associated with ophthalmic surgery: A 20-year
experience at the Willmer Institute. American Journal o f Ophthalmology
77:517-24.
Root, L.L., Jr. 1983. Supply, workload and utilization: A population-based
analysis o f surgery in rural Manitoba. American Journal o f Public Health
73:414-21.
Roter, D.L., and Hall, J.A. 1998. Why physician gender matters in shaping the
physician-patient relationship. Journal o f W omen’ s Health 7(9): 1093-7.
Rosenblatt, R.A., Moscovice, I.S. 1984. The physician as gatekeeper:
Determinants of physicians’ hospitalization rates. Medical Care 22(2): 150-59.
Saha, S., Komaromy, M., Koepsell, T.D., Bindman, A.B. 1999. Patient-physician
racial concordance and the perceived quality and use of health care. Archives
o f Internal Medicine 10;159(9):997-1004.
Saha, S, Stetting D, Redberg, RF. 1999. Gender and willingness to undergo
invasive cardiac procedures. Journal o f General Internal Medicine 14(2): 122-
25.
Salkever, D.S., Skinner, E., Steinwachs, D.M., Kata, H. 1982. Episode-based
efficiency comparisons for physicians and nurse practitioners. Medical Care
20:143-153.
145
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Sandy, L.G. 1999. The future of physician profiling. Journal o f Ambulatory Care
Management 22(3): 11-16.
Schein, O.D., Katz, J., Bass, E., et al. 2000 (January). The value of routine
preoperative medical testing before cataract surgery. The New England Journal
o f Medicine 168-175.
Schein, O.D., Steinberg, E.P., Cassard, S.D., et al. 1995. Predictors of outcome in
patients who underwent cataract surgery. Ophthalmology 102:817-23.
Scitovsky, A.A. 1968. Changes in the costs of treatment o f selected illnesses.
American Economic Review 57:1182-1195.
Scott, IV, Schein, O.D., West, S., et al. 1994. Functional status and quality of life
measurement among ophthalmic patients. Archives Ophthalmology 112:329-
35.
Scottish Health Service Advisory Council. 1992. Study group on the management
o f ophthalmology services in Scotland. Report on day case surgery fo r
cataract. Pp. 111-34.
Seto, T.B., Taira, D.A., Davis, R.B., et al. 1996. Effect o f physician gender on the
prescription of estrogen replacement therapy. Journal o f General Internal
Medicine 11(4): 197-203.
Showstack, J.A., Gamick, K.E., Osenfeld, K.E., et al. 1987. Episode-of-care
physician payment: A study of coronary artery bypass graft surgery. Inquiry
24:376-83.
Singh, A.L., Gamer, P., Floyd, K. 2000 (January 15). Cost-effectiveness of public-
funded options for cataract surgery in Mysore, India. Lancet 355(1): 180-84.
Skinner, B.F. 1969. The behavior o f organizations. Pp. 23-56. New York:
Appleton-Century.
Sletteberg, O., Hovding, G., and Bertelsen, T. 1995. Do we operate too many
cataracts? Alta Ophthalmology Scandinavia 73:77-80.
Smith, N.S., and Weiner, J.P. 1994. Applying population-based case mix
adjustment in managed care: The Johns Hopkins ambulatory care group
system. Managed Care 2(3): 21-34.
146
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Solon, J.A., Feene, S.H., Jones, R.D., Sheps, C.G. 1967. Delineating episode of
medical care. American Journal o f Public Health 57:401-408.
Sommer, A. 1995. Toward affordable, sustainable eye care. International
Ophthalmology 18:287-92.
----------------. 1996. Public health and eye disease. Annual Review Public Health
17:115-19.
Steinberg, E.P., Javitt, J., et al. 1993. The content and cost of cataract surgery.
Archives Ophthalmology 111: 1041-1049.
Steinberg, E.P., Tielsch, J.M., Schein, O.D., et al. 1994. National study of cataract
surgery outcomes, variation in four-month postoperative outcomes as reflected
in multiple outcome measures. Ophthalmology 101:1131-1141.
Stolley, P.D., Becker, M.H., and Lasagna, L., et al. 1972. The relationship between
physician characteristics and prescribing appropriateness. Medical Care
10( 1): 17-28.
Straatsma, B.R., Meyer, K.J., Bastek, J.V., et al. 1983. Posterior chamber
intraocular lens implantation by ophthalmology residents. Ophthalmology
90:327-335.
Tatchell, M. 1983. Measuring hospital output: A review of the service mix and
case mix approaches. Social Science Medicine 17:871-83.
Taylor, H. 2000. Cataract: How much surgery do we have to do? British Journal o f
Ophthalmology 84:1-2.
Thurau, R. 1997. Perceived gender bias in the treatment of cardiovascular disease.
Journal o f Vascular Nursing 15(4): 124-7.
Thylefors, B., Negrel, A.P., Pararajasegaram, R., Dadzie, K.Y. 1995. Global data
on blindness. Bulletin. 73:115-21. Geneva: WHO.
Tobin, J.N., Wassertheil-Smoller, S., Wexler, J.P, et al. 1987. Sex Bias in coronary
artery bypass surgery. Annals o f Internal Medicine 107:19-25.
Weisman, C.S., Teitelbaum, M.A. 1985. Physician gender and the physician-
patient relationship: Recent evidence and relevant questions. Social Science
and Medicine 20(11): 1119-27.
147
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Weisman, J.S., Epstein, A.M. 1989. Case mix and resource utilization by
uninsured hospital patients in the Boston Metropolitan Area. Journal o f the
American Medical Association 261(24):3572-6.
Welch, W.P., Miller, M.E., and Welch, H.G., et al. 1993. Geographic variation in
expenditures for physicians’ services in the United States. New England
Journal o f Medicine 330(9):621-27.
Wells, C.K., and Feinstein, A.R. 1988. Detection bias in the diagnostic pursuit of
lung cancer. American Journal o f Epidemiology 128:1016-1026.
Wennberg, J.E., and Zubkoff, M. 1982. Professional uncertainty and the problem
of supplier-induced demand. Social Science Medicine 16:811-24.
West, S.K., Duncan, D.D., Munoz, B., et al. 1998. Sunlight exposure and risk of
tens opacities in a population-based study. The Salisbury eye evaluation
project. The Journal o f the American Medical Association 280:714-18.
West, S.K., Guigley, H.Q. 1991. Cataract blindness: What to do? Archives o f
Ophthalmology 109:1765-66.
Wilensky, G.R., Rossiter, L.F. 1982. The relative importance o f physician-induced
demand for medical care. Milband Find Quarterly 61:252-77.
Wingert, T.D., Kralewski, J.E., Lindquist, T.J. Knutson, D J. 1995/1996 (Winter).
Constructing episodes of care from encounter and claims data: Some
methodological issues. Inquiry> 32:430-43.
WHO (World Health Organization). 1987. Program fo r the prevention o f
blindness. Available data on blindness, Geneva 14:1-23.34. Geneva: WHO.
WHO (World Health Organization). 1997. Global initiative fo r the elimination o f
avoidable blindness:An informal consultation. WHO/PBL/97.61. Geneva:
WHO.
148
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Performance management in health care in Iceland

PDF

The 1996 Health Insurance Portability and Accountability Act (HIPAA) and the 1997 Balanced Budget Act (BBA): Implications for organizational change and delivery of care in cardiology group practices

PDF

Toward a theory of networks and network effectiveness: An analytical framework for the study of multilevel networks in health care service delivery

PDF

The effects of Medi -Cal managed care on access to mammography screening for women residing in Los Angeles County

PDF

The effects of case mix, hospital competition, and managed care penetration on hospital costs and revenues

PDF

Physician profiling and clinical pathways: Combining the tools to change physician resource utilization

PDF

Recruitment of family physicians into rural California: Predictors and possibilities

PDF

The construct validity of the Health Utilities Index in patients with chronic respiratory disease in a managed care population

PDF

An analysis of health risk selection and quality of care under Medicare fee -for -service and Medicare managed care health care systems

PDF

Regionalization of fire protection and emergency medical aid services: A comparative case study analysis of economic and socio-political impacts

PDF

Health care finance reform and market failure at King/Drew Medical Center, Los Angeles: An analysis of pediatric social outcomes in an urban public safety -net hospital

PDF

The extent and effectiveness of nursing home regulation in the 50 states

PDF

Assessment of prognostic comorbidity in hospital outcomes research: Is there a role for outpatient pharmacy data?

PDF

Nurse executive leadership and the relationship to organizational commitment among nurses

PDF

Leading change initiatives: communication and bounded agency in a health care organization

PDF

Business political influence on agency regulatory decision making: The efficacy of corporate political activity in Taiwan's CATV license awarding

PDF

A quantitative and qualitative evaluation of the least burdensome provisions of the Food and Drug Administration Modernization Act of 1997

PDF

Care management for the uninsured: A force field analysis of the business case

PDF

Physician adherence to national hypertension guidelines in an elderly Medicaid population

PDF

Locating nearby sources of air pollution using air quality data and wind direction

Asset Metadata

Creator Yu Chang, Joanna Chih I (author)

Core Title Variations in physician practice patterns for eye care under the National Health Insurance of Taiwan

Degree Doctor of Philosophy

Degree Program Public Administration

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag health sciences, health care management,OAI-PMH Harvest,Political Science, public administration

Language English

Contributor Digitized by ProQuest (provenance)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-148447

Unique identifier UC11329255

Identifier 3041549.pdf (filename),usctheses-c16-148447 (legacy record id)

Legacy Identifier 3041549-0.pdf

Dmrecord 148447

Document Type Dissertation

Rights Yu Chang, Joanna Chih I

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA