University of Southern California Dissertations and Theses

Three essays on emerging issues in hemophilia care

(USC Thesis Other)

Three essays on emerging issues in hemophilia care

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Three Essays on Emerging Issues in Hemophilia Care

By Zheng-Yi Zhou

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PHARMACEUTICAL ECONOMICS AND POLICY)

June 2014

ii

DEDICATION

To my dearest parents, Shu and Zhenghuan,
my month-in-law, Xuchuan,
and my loving husband, Jie
for their love and endless support;

and to my lovely son, Haochang,
you are the sunshine of my life
and give me the courage to make dream come true;

and in memory of
Dr. Kathleen A. Johnson

iii

ACKNOWLEDGEMENTS

There are a number of people without whom this thesis might not have been completed
and to whom I am greatly indebted.
I would like to express the deepest appreciation to my committee chair, Dr. Mike Nichol,
for his support, encouragement and invaluable suggestions throughout the last one and a
half years. Thanks for taking me under your wings and helping me go through all the
difficulties.
I am very grateful to my committee, Dr. Jason Doctor, Dr. Sue Ingles, Dr. Joel Hay, and
Dr. Geoff Joyce, for their insightful comments, time, and expertise, as well as helping me
to finish my dissertation.
I would like to thank the HUGS team without which my dissertation would not be
possible. Special thanks to Mimi, Marion, Randy and Brenda, your support and
friendship over the past 7 years, especially through the bad times, was very much
appreciated.
Most importantly, I would like to convey my heartfelt appreciation to Dr. Kathleen
Johnson. You gave me the chance to work on HUGS project and opened the door to this
exciting research field. From 2006 to 2012, you guided me through all the adventures in
research, encouraged me to think independently, supported all my ideas, and tolerated
any mistakes I made. You were not only my advisor on academics but also advisor of my
life. I cherish every moment we shared together.
Finally, this long journey at USC would have been less enjoyable and fulfilling without
my friends and classmates. Thanks Jaejin, Ning, Monica, Vaidy, Quang, Elaine, my twin
sister – Sara, Jiat, and Joanne!

iv

TABLE OF CONTENTS

DEDICATION .................................................................................................................... 2
ACKNOWLEDGEMENTS .............................................................................................. III
TABLE OF CONTENTS .................................................................................................. IV
LIST OF TABLES ............................................................................................................ VI
LIST OF FIGURES ......................................................................................................... VII
ABSTRACT ........................................................................................................................ 1
CHAPTER 1 : INTRODUCTION ...................................................................................... 3
1.1 Epidemiology and Etiology of Hemophilia .............................................................. 3
1.2 Treatment of Hemophilia .......................................................................................... 5
1.2.1. Clotting Factor Replacement Therapy .......................................................................... 5
1.2.2. Efficacy and Effectiveness of Prophylactic versus Episodic Treatment ........................ 6
1.2.3. Prophylaxis – An Individualized Approach ................................................................... 7
1.2.4. Development of Inhibitors to Clotting Factor ............................................................... 9
1.3 Economic Burden of Hemophilia .............................................................................. 9
1.4 Hemophilia Treatment Access ................................................................................ 12
1.5 Purpose of the Study ............................................................................................... 13
CHAPTER REFERENCES .......................................................................................... 15
CHAPTER 2 : ECONOMIC COMPARISON OF APCC VERSUS RFVIIA FOR MILD-
TO-MODERATE BLEEDING EPISODES IN HEMOPHILIA PATIENTS WITH
INHIBITORS .................................................................................................................... 20
ABSTRACT .................................................................................................................. 20
INTRODUCTION ......................................................................................................... 22
METHODS.................................................................................................................... 25
RESULTS...................................................................................................................... 30
DISCUSSION ............................................................................................................... 34
CHAPTER REFERENCES .......................................................................................... 37
CHAPTER 3 : EXPLORING TEMPORAL PATTERNS OF CLOTTING FACTOR USE
AND ASSOCIATED HEALTHCARE UTILIZATION IN HEMOPHILIA ................... 40
ABSTRACT .................................................................................................................. 40
INTRODUCTION ......................................................................................................... 42
METHODS.................................................................................................................... 47
RESULTS...................................................................................................................... 53
DISCUSSION ............................................................................................................... 64
CHAPTER REFERENCES .......................................................................................... 68
v

CHAPTER 4 : SELF-REPORTED BARRIERS TO HEMOPHILIA CARE IN PERSONS
WITH FACTOR VIII DEFICIENCY ............................................................................... 75
ABSTRACT .................................................................................................................. 75
INTRODUCTION ......................................................................................................... 77
METHODS.................................................................................................................... 80
RESULTS...................................................................................................................... 86
DISCUSSION ............................................................................................................... 92
CONCLUSION ............................................................................................................. 94
CHAPTER REFERENCES .......................................................................................... 96

vi

LIST OF TABLES
Table 2.1 Base Case Assumptions .................................................................................... 28
Table 2.2 Threshold Sensitivity Analysis ......................................................................... 31
Table 3.1 Demographic and Clinical Characteristics by clotting factor use pattern clusters
(6-group model) ................................................................................................................ 54
Table 3.2 BIC values and predicted group percentages for GTBM for 2- to 7-group
trajectory solutions ............................................................................................................ 58
Table 3.3 Healthcare resource utilization by clotting factor use pattern clusters (6-group
model) ............................................................................................................................... 62
Table 4.1 Types of barriers to hemophilia care ................................................................ 82
Table 4.2 Comparisons of participant/parent reported barriers to HTC utilization between
children and adults with hemophilia A ............................................................................. 87
Table 4.3 Distribution of characteristics of study population, overall and by self-reported
barriers to HTC utilization ................................................................................................ 88
Table 4.4 Multivariable logistic regression analysis of participant characteristics
associated with barriers to HTC utilization ...................................................................... 91

vii

LIST OF FIGURES
Figure 2.1 Decision Analytic Model Layout .................................................................... 26
Figure 2.2 Tornado Diagram of One-way Sensitivity Analysis ....................................... 31
Figure 2.3 Two-way Sensitivity Analysis ......................................................................... 32
Figure 3.1 Selection of the study cohort ........................................................................... 53
Figure 3.2 Trajectory models using 4–6 group solutions ................................................. 57
Figure 3.3 Bayesian information criterion versus number of clusters used in GBTM ..... 58
Figure 3.4 Trajectories of clotting factor use patterns using 6-group solution with median
trajectory membership probabilities ................................................................................. 59
Figure 3.5 Healthcare resource utilization in hemophilia patients during each year of the
36-month study period, by clotting factor use pattern clusters ......................................... 61
Figure 3.6 All-cause healthcare costs during each year of the 36-month study period, by
clotting factor use pattern clusters (6-group model) ......................................................... 63

1

ABSTRACT

With the increasing availability of clotting factor and development of comprehensive
hemophilia treatment centers (HTCs) in the US, hemophilia care has improved
dramatically during the past four decades. Nowadays children with hemophilia in
developed countries look forward to a normal life span and high quality of life without
crippling joint diseases. Other the other hand, long-life treatment with the high cost of
care including clotting factor concentrates places a considerable burden on patients,
healthcare systems and society. Patients and providers face a number of challenges,
which may affect patients’ access to clotting factor or hemophilia care at HTCs and lead
to undesired outcomes. Development of antibodies (i.e., inhibitors) to clotting factor
concentrates still represents a significant risk for patients with hemophilia, which requires
even higher healthcare resource utilization and staggering high costs.
This three-paper dissertation aims to address some of the emerging economic and public
health issues in hemophilia care. In Chapter 2, an economic evaluation was conducted
and a cost-minimization model was built to compare two bypassing agents, activated
prothrombin complex concentrates (aPCC) versus recombinant factor VIIa (rFVIIa) as
first-line drugs in the home treatment of mild-to-moderate bleeding episodes in
hemophilia patients with inhibitors from a US third party payer’s perspective. The model
used clinically based assumptions and investigated model parameters with extensive
sensitivity analyses. In Chapter 3, we explored the longitudinal clotting factor use
patterns for patients with hemophilia using 6 state Medicaid databases. The aim of the
study was to classify and characterize clotting factor use patterns using group-based
2

trajectory models (GBTMs) and to assess the economic outcomes associated with clotting
factor use patterns. Finally, in Chapter 4, using data collected from six HTCs in the US,
we identified self-reported barriers that might prevent individuals from using HTC
services and studied patient sociodemographic and clinical characteristics associated with
barriers.

3

CHAPTER 1: INTRODUCTION
1.1 Epidemiology and Etiology of Hemophilia
Hemophilia is a chronic congenital bleeding disorder that affects approximately 400,000
persons of all races worldwide, including 20,000 individuals in the United States
(Akerley, Boucher, Bentz, Arbogast, & Walters, 2009; Soucie, Evatt, & Jackson, 1998;
Stonebraker, BOLTON‐ MAGGS, Michael Soucie, Walker, & Brooker, 2010). Patients
with hemophilia are deficient in clotting factor which results in the impairment in body’s
ability to control bleeds and leads to spontaneous bleeding in joints, muscles, and other
soft tissues in the absence of injury. It may also cause excessive post-traumatic or
postsurgical bleeding. In rare cases, bleeding from hemophilia can be fatal. As an X-
linked recessive disorder, it is predominantly occur in males. The most common types are
hemophilia A (factor VIII deficiency) and hemophilia B (factor IX deficiency).
Hemophilia A occurs among about 1 of every 5,000 male births in the US, while
hemophilia B is far less common, occurring in about 1 in 25,000 male births (CDC,
2012).
Hemophilia is classified as mild, moderate, or severe depending on the circulating
clotting factor levels (Carcao & Blanchette, 2010). The frequency and severity of
bleeding episodes are associated with the amount of factor in blood. In normal people, the
plasma levels of factor VIII (factor IX) range from 50% to 150%. Persons with mild
hemophilia have 5% - 40% of normal clotting factor activity, and bleeding episodes only
occur after serious traumas or surgeries (White & Aledort, 2001). In persons with
moderate hemophilia, the plasma clotting factor activity ranges from 1% to 5% of normal
level. Patients tend to bleed following severe injuries but rarely have spontaneous bleeds.
4

About 60% individuals with hemophilia have the severe form, with clotting factor
activity below 1% (Roth et al., 2001). These patients are more likely to have frequent
spontaneous internal (e.g., joint or muscle) bleeding episodes without obvious causes.
Most individuals with hemophilia are diagnosed at a very early age. Based on the Centers
of Disease Control and Prevention (CDC) Universal Data Collection (UDC) reports, the
median age at diagnosis for people with mild hemophilia in the US is 36 months, 8
months for moderate hemophilia, and 1 month for those with severe hemophilia
(Kulkarni et al., 2009). The bleeding patterns among individuals also vary by age (Carcao
& Blanchette, 2010; Kulkarni et al., 2009). Most commonly seen bleeding episodes in the
neonatal period are circumcision bleeding and intra- or extra-cranial hemorrhages, while
infants aged between 1 and 6 months are more likely to experience excessive bruising or
other soft tissue bleeding. As mobility increases over time, older children and adults
predominantly have joint bleeds, accounting for 70% to 80% of spontaneous bleeding
(Srivastava et al., 2013). Knees, elbows, and ankles are the joints most frequently
affected and joint bleeds may occur up to 20-30 times per year. Repeated bleeding into
the same joint (i.e., target joint) results in progressive damage. Left untreated, patients
may develop hemophilic arthropathy, which leads to chronic pain, loss of joint range of
motion, and development of crippling musculoskeletal deformity and disability during
young adulthood (Arun, 2001; Pipe & Valentino, 2007).

5

1.2 Treatment of Hemophilia
1.2.1. Clotting Factor Replacement Therapy
Whereas hemophilia is not curable at present, lessening the frequency and severity of
bleeding and prevention of arthropathy are the main goals of treatment. Hemophilia is
primarily treated with clotting factor replacement therapy, in which clotting factor
concentrates are infused into peripheral veins to replace the clotting factor that's missing
or low. The availability of commercially-prepared clotting factor concentrates, either
plasma derived or recombinant, has transformed hemophilia from a frequently fatal
condition to a clinically manageable disease.
Episodic (or on-demand) treatment, administered after acute bleeding events, has been
widely used to treat patients since the 1960s (Teitel et al., 2004). It is known to
effectively control bleeding, relieve pain, and restore joint mobility in the short term, but
not to prevent arthropathy in most patients with severe hemophilia (Pipe & Valentino,
2007). There is also a risk that bleeding may have already caused damage in the joint
before the episodic therapy is given. The dosage and duration of episodic treatment
depend on the severity of clotting factor deficiency, the location and extent of bleeding,
and the patient’s clinical condition.
Clotting factors can also be infused on a regular schedule (prophylaxis) to prevent
bleeding and the development of arthropathy. Prophylaxis in persons with severe
hemophilia has been recommended by the World Foundation of Hemophilia (Srivastava
et al., 2013) and the National Hemophilia Foundation (NHF, 2007). The purpose of
prophylaxis is to maintain the blood clotting factor activity above 1% in order to mimic
the clinical phenotype of moderate hemophilia (Fischer et al., 2002). It was based on the
6

observation that persons with moderate hemophilia bled only after trauma and had a
fairly normal life expectancy, while persons with severe hemophilia (with <1% clotting
factor activity) had much severer bleeds, early crippling from hemophilic arthropathy,
and a shorter life expectancy (Nilsson, Blomback, & Ahlberg, 1970).
1.2.2. Efficacy and Effectiveness of Prophylactic versus Episodic Treatment
Many studies have been conducted to examine the efficacy and effectiveness of
prophylactic compared with episodic treatment among patients with severe hemophilia.
Observational studies have shown that prophylaxis significantly improves outcomes by
decreasing the frequency of bleeding, thereby delaying the onset of arthropathy and
improves patients’ quality of life (Aledort, Haschmeyer, & Pettersson, 1994; Fischer et
al., 2002; Royal et al., 2002; Stobart, Iorio, & Wu, 2006; Szucs et al., 1998).
Until recent years, three longitudinal randomized controlled trials have been conducted to
compare the long-term outcomes of two treatment patterns in the prevention of joint
disease among young boys and adults with severe hemophilia. In a six-year randomized
clinical trial, Manco-Johnson et al. found that boys who received primary prophylaxis
(n=32) had significantly greater reduction in the risk of joint damage detected by MRI
and fewer joint bleeds compared to those who received episodic therapy (n=33) (Manco-
Johnson et al., 2007a). Similarly, in a ten-year randomized controlled trial conducted in
Italy, children on prophylaxis (n=22) had fewer joint bleeds than children on episodic
treatment (n=19) (Gringeri et al., 2011). The proportion of patients with radiologic signs
of arthropathy was also fewer in prophylaxis group versus episodic group. Early start of
prophylaxis (≤36 months) was more effective with fewer joint bleeds and no radiologic
signs of arthropathy after tens year of treatment. For adults, as shown in the interim
7

results of a randomized controlled trial – SPINART, routine prophylaxis also significantly
reduced joint bleeding as compared with episodic treatment (Manco-Johnson et al.,
2013).
1.2.3. Prophylaxis – An Individualized Approach
Prophylaxis in persons with severe hemophilia has been advocated for almost half a
century (Ahlberg, 1965) and increasingly used in the care of hemophilia. Data from the
CDC UDC project show that around 68% children with severe hemophilia who were
younger than 6 years of age and 86% children aged between 6 and 13 years received
prophylaxis during 2008 and 2010 (Baker, Riske, Voutsis, Cutter, & Presley, 2011).The
NHF MASAC treatment recommendations suggest that prophylaxis should be initiated in
an early age, prior to the onset of frequent bleeding (Abratt, Shepherd, & Memeena
Salton, 1995; NHF, 2007). This type of prophylaxis is also called primary prophylaxis, in
which patients receive long-term, continuous treatment starting either before the age of
two or after the initial joint bleed. In contrast, secondary prophylaxis can be initiated after
the onset of joint disease, usually in response to repeated bleeding that has not been
managed effectively with episodic therapy (Manco-Johnson, Nuss, Geraghty, Funk, &
Kilcoyne, 1994). It can be either continuous (long-term) or intermittent (short-term).
It has been suggested that prophylactic treatment should be personalized based on
individual’s characteristics and treatment experience (Astermark et al., 1999). Given the
high cost of clotting factor and the difference in pharmacokinetic response to factors and
lifestyle among individuals, personalized dosing regimens may optimize prophylaxis in
patients with hemophilia. The half-life of clotting factor concentrates is short. As a result,
prophylactic therapy is most effective when administered frequently. In the US, the
8

National Hemophilia Foundation (NHF)’s Medical and Scientific Advisory Council
(MASAC) recommend that prophylactic therapy is typically administered in a dose of
25–50IU/kg factor VIII three times per week or every other day, or 40-100IU/kg factor
IX two to three times per week (NHF, 2007). In practice, the natural clinical course of
hemophilia may differ substantially between individuals. Usually a higher dose is
suggested for children and there is a potential modification of prophylaxis in early adult
life when they are less physically active.
Moreover, joint bleeds usually become less frequent in adults and adolescence compared
to childhood. Young adults may also encounter significant barriers to adherence with
prophylaxis when they start to take responsibility of self-treatment. Additionally, because
hemophilia is a rare but costly disease, adults are more likely to face difficulties in
finding insurance with adequate coverage, especially for those who turn into young
adulthood and are no longer covered by their parents’ insurance (Johnson et al., 2012).
All these issues may impact the treatment options available to both patients and providers,
resulting in changes in a patient’s treatment patterns and their adherence to treatment,
which in turn may impact their outcomes.
On the other hand, should prophylaxis be continued for life or can it be safely withdrawn
at certain age remains unclear. Study on 80 patients treated in two Danish and one Dutch
treatment centers indicated that about one-third of young adults with severe hemophilia
on prophylaxis switch to episodic treatment in early adulthood, while still maintaining a
low joint bleeding rate and similar joint status after four years (van Dijk et al., 2005). In
another European survey, around 42% of persons with severe hemophilia discontinued
prophylaxis completely during the age of 16-22 years and 28% of those had to restart
9

some form of routine concentrate infusion because of poor outcomes (Richards et al.,
2007). However, the long-term consequence still needs to be further studied.
1.2.4. Development of Inhibitors to Clotting Factor
Whereas proper treatment can prevent bleeding, slow progress of joint damage, and
improve health-related quality of life, complications associated with treatment may also
be severe. The development of persistent antibodies (or inhibitors) to foreign factor
proteins is an ongoing challenge in hemophilia care. Approximately 25–30% of patients
with severe hemophilia A and 4–6% of patients with severe hemophilia B develop
inhibitors during factor replacement therapy (Franchini & Mannucci, 2011). The
development of inhibitors most often appear during the first year of treatment but it can
appear at any time. Such patients, compared with those without inhibitors, are at higher
risk of severe and incapacitating degrees of arthropathy and are more difficult to treat
with surgery (Morfini et al., 2007; Rodriguez-Merchan, 2008). Treatment of acute
bleeding episodes in patients with hemophilia and high responding inhibitors require
higher doses of clotting factors or bypassing agents, which results in even higher medical
costs (Auerswald et al., 2004; Gringeri, Mantovani, Scalone, Mannucci, & Group, 2003).

1.3 Economic Burden of Hemophilia
Treatment with clotting factor is costly and places substantial economic burden on
patients, their families, and healthcare systems. Based on the data from the randomized
controlled trial comparing the clinical outcomes of episodic versus prophylactic factor
replacement treatment, children at age of six received 6,000 IU of factor VIII per
kilogram in the prophylaxis group, as compared with 2,500 IU per kilogram per year in
10

the episodic group (Manco-Johnson et al., 2007b). Given the average cost of recombinant
factor VIII is $1 per unit, the cost of prophylaxis for a child weighing 50 kg could reach
$300,000 per year.
Real world studies show that 45-94% of the total direct medical costs for hemophilia are
attributable to clotting factor use (Carlsson et al., 2004; Globe, Curtis, Koerper, &
Committee, 2004). The study by Globe et al. showed that the annual medical costs of
hemophilia care in the US in 1995was estimated to be $139,102 (median: $55,330)
(Globe et al., 2004). Higher health care costs were significantly associated with severe
hemophilia, arthropathy, history of inhibitor to FVIII, and infusing FVIII concentrate
through a port (vs. intravenous infusion). In a 2004 study in Norway and Sweden,
clotting factor consumption in prophylaxis was two to three times higher than that in
episodic treatment, accounting for 94% versus 74% of total costs respectively (Carlsson
et al., 2004).
In addition, 25–30% of patients with severe hemophilia A and 4–6% with severe
hemophilia B may develop antibodies (inhibitors) and require higher doses of clotting
factors or more frequent hospitalizations, and surgeries (Franchini & Mannucci, 2011).
The annual medical costs were estimated to be 3-6 times higher compared with those who
without inhibitors, with some cases exceeding $1 million per year when patients receive
bypassing agents for prophylaxis (Auerswald et al., 2004; Gringeri et al., 2003).
Four recent studies have reported annual healthcare costs among persons with hemophilia
from the US payers’ perspective. (Guh, Grosse, McAlister, Kessler, & Soucie, 2012a,
2012b; Tencer, Friedman, Li-McLeod, & Johnson, 2007; Valentino et al., 2012) Three
11

studies were conducted using claims data of individuals covered by employer-sponsored
insurance (ESI) (Guh et al., 2012b; Tencer et al., 2007; Valentino et al., 2012), and one
used a multi-state Medicaid database. (Guh et al., 2012a) The annual average annual
healthcare expenditure in 2008 was estimated to be $142,987 (median: $46,737) for
individuals covered by Medicaid and $155,136 (median: $73,548) among ESI enrollees
(Guh et al., 2012a, 2012b). However, using claims data to conduct a cost-of-illness study
has clear limitations. Clinical details, such as disease severity and treatment patterns, are
not available in the claims. In addition, inhibitor development was assumed only if
bypassing agents were used. Thus, costs associated with disease severity could not be
determined, and indirect costs were not characterized in these studies.
Besides direct healthcare costs, studies have also shown considerable indirect costs
related to hemophilia, including individuals’ and caregivers’ lost productivity, caregivers’
unpaid time costs, and hemophilic individuals' disability (Carlsson et al., 2004). In an 11-
year retrospective panel data analysis in Norway and Sweden, Carlsson et al. found that
costs associated with productivity loss among patients with prophylaxis were less than
half of that in episodic treatment (Steen Carlsson et al., 2003). However, the lower non-
factor related indirect costs still could not offset the higher cost of clotting factor for
prophylaxis versus episodic treatment. In this example, the annual total cost (both direct
and indirect), for a typical 30 year-old patient with severe hemophilia, was €51,832 for
episodic treatment and €146,118 for prophylaxis (values are in 2000 Euros) (Carlsson et
al., 2004).
In the United States, the Hemophilia Utilization Group Study (HUGS) assessed the direct
and indirect costs of hemophilia A and have showed similar trends to those found in
12

European studies (Johnson et al., 2012).The indirect costs for patients with mild
hemophilia were $5,195 and $9,043 for moderate patients. Among patients with severe
hemophilia, the indirect costs were estimated to be $16,952 and $8,867 annually for those
receiving episodic and prophylactic treatment respectively (Johnson et al., 2012).
Furthermore, hemophilia is also associated with school absenteeism among children,
which may affect their academic performance, and with considerable indirect costs,
mainly due to functional disability of adult patients and underemployment of parents of
children with hemophilia.

1.4 Hemophilia Treatment Access
Given the rareness and complexity of hemophilia, primary care is usually insufficient to
meet patients’ specialized needs, and treatment in a general hematology department is
often inadequate or fragmented. In addition to medical needs, individuals with
hemophilia and their families often face physical, emotional, social, and financial
challenges throughout their lives. Therefore, they require care that is disease-specific,
comprehensive and multidisciplinary, and which includes both appropriate medical and
psychosocial services.
A multi-disciplinary, team-based care delivery model – a network of federal-sponsored
hemophilia treatment center (HTC) – has been implemented in the US for several
decades. Because HTCs treat a high volume of patients, they are able to develop
experience and expertise in management of this rare condition. The goal of HTCs is to
prevent hemophilia-related complications and to maximize the physical and
psychological functioning and socioeconomic benefits of this rare disorder population
13

(Baker, Crudder, Riske, Bias, & Forsberg, 2005). Currently, there are more than 130
HTCs across the country (CDC, 2012) and studies have shown that HTC care is
associated with lower mortality (Soucie et al., 2000) and fewer bleed-related
hospitalizations (Soucie et al., 2001) compared with care obtained elsewhere. School
absenteeism and high school graduation rates among HTC patients are comparable to
those found in the general population (Drake et al., 2010). It was reported that around 70%
of persons with hemophilia received any hemophilia care from HTCs and they are more
frequently individuals with mild disease or fewer complications (Soucie et al., 2000).
However, no update to this number has been published in the last decade. The barriers to
hemophilia care have not been studied in a comprehensive manner, but were
hypothesized to include lack of awareness of the potential benefits of HTC care,
socioeconomic factors, cultural barriers, transportation and other logistical problems, and
insurance coverage for HTC services (Soucie et al., 2000).

1.5 Purpose of the Study
With the increasing availability of clotting factor and development of comprehensive
HTCs in the US, hemophilia care has improved significantly during the past four decades.
Nowadays patients with hemophilia in developed country look forward to a normal life
span and high quality of life. Other the other hand, long-life treatment with the high cost
of care and requirement of highly specialized treatment place a considerable burden on
patients, healthcare systems and society. Patients and providers face a number of
challenges, which may impact patients’ access to clotting factor or hemophilia care at
HTCs and lead to undesired outcomes.
14

This three-paper dissertation aims to address some of the emerging economic and public
health issues in hemophilia. In Chapter 2, Paper 1 demonstrates a cost-minimization
model comparing two bypassing agents, activated prothrombin complex concentrates
(aPCC) versus recombinant factor VIIa (rFVIIa) as first-line drugs in the home treatment
of mild-to-moderate bleeding episodes in hemophilia patients with inhibitors from a US
third party payer’s perspective. The model used clinically based assumptions and
investigated model parameters with extensive sensitivity analyses. In Chapter 3, we
explored the longitudinal clotting factor use patterns for patients with hemophilia using 6
state Medicaid database. The aim of the study was to classify and characterize clotting
factor use patterns using group-based trajectory models (GBTMs) and to assess the
economic outcomes associated with clotting factor use trajectory subgroups. Finally in
Chapter 4, using data collected from six HTCs in the US, we identified barriers that may
prevent individuals from using HTC services and studied patient socio-demographic and
clinical characteristics associated with barriers.

15

CHAPTER REFERENCES
Abratt, R. P., Shepherd, L. J., & Memeena Salton, D. G. (1995). Palliative radiation for
stage 3 non-small cell lung cancer - A prospective study of two moderately high dose
regimens. Lung Cancer, 13(2), 137-143.
Ahlberg, A. (1965). Haemophilia in Sweden. VII. Incidence, treatment and prophylaxis
of arthropathy and other musculo-skeletal manifestations of haemophilia A and B. Acta
Orthop Scand Suppl, Suppl 77:73-132.
Akerley, W., Boucher, K. M., Bentz, J. S., Arbogast, K., & Walters, T. (2009). A phase II
study of erlotinib as initial treatment for patients with stage IIIB-IV non-small cell lung
cancer. Journal of Thoracic Oncology, 4(2), 214-219.
Aledort, L. M., Haschmeyer, R. H., & Pettersson, H. (1994). A longitudinal study of
orthopaedic outcomes for severe factor-VIII-deficient haemophiliacs. The Orthopaedic
Outcome Study Group. J Intern Med, 236(4), 391-399.
Arun, B. K., C. (2001). Inherited Bleeding Disorders: Haemostasis and Thrombosis (4
ed.). Philadelphia, PA: Lippincott Williams and Wilkens.
Astermark, J., Petrini, P., Tengborn, L., Schulman, S., Ljung, R., & Berntorp, E. (1999).
Primary prophylaxis in severe haemophilia should be started at an early age but can be
individualized. Br J Haematol, 105(4), 1109-1113.
Auerswald, G., von Depka Prondzinski, M., Ehlken, B., Kreuz, W., Kurnik, K., Lenk,
H., . . . Zimmermann, R. (2004). Treatment patterns and cost-of-illness of severe
haemophilia in patients with inhibitors in Germany. Haemophilia, 10(5), 499-508. doi:
10.1111/j.1365-2516.2004.00950.x
Baker, J. R., Crudder, S. O., Riske, B., Bias, V., & Forsberg, A. (2005). A model for a
regional system of care to promote the health and well-being of people with rare chronic
genetic disorders. Am J Public Health, 95(11), 1910-1916. doi:
10.2105/AJPH.2004.051318
Baker, J. R., Riske, B., Voutsis, M., Cutter, S., & Presley, R. (2011). Insurance, home
therapy, and prophylaxis in U.S. youth with severe hemophilia. Am J Prev Med, 41(6
Suppl 4), S338-345. doi: 10.1016/j.amepre.2011.09.002
Carcao, M. D., & Blanchette, V. S. (2010). Work-up of a Bleeding Child Textbook of
Hemophilia (pp. 118-126): Wiley-Blackwell.
Carlsson, K. S., Hojgard, S., Lindgren, A., Lethagen, S., Schulman, S., Glomstein, A., . . .
Lindgren, B. (2004). Costs of on-demand and prophylactic treatment for severe
haemophilia in Norway and Sweden. Haemophilia, 10(5), 515-526. doi: 10.1111/j.1365-
2516.2004.00952.x
16

CDC. (2012). The Centers for Disease Control and Prevention: Hemophilia Data and
Statistics. http://www.cdc.gov/ncbddd/hemophilia/data.html.
Drake, J. H., Soucie, J. M., Cutter, S. C., Forsberg, A. D., Baker, J. R., & Riske, B.
(2010). High school completion rates among men with hemophilia. Am J Prev Med, 38(4
Suppl), S489-494. doi: 10.1016/j.amepre.2009.12.024
Fischer, K., van der Bom, J. G., Molho, P., Negrier, C., Mauser-Bunschoten, E. P.,
Roosendaal, G., . . . van den Berg, H. M. (2002). Prophylactic versus on-demand
treatment strategies for severe haemophilia: a comparison of costs and long-term
outcome. Haemophilia, 8(6), 745-752.
Franchini, M., & Mannucci, P. M. (2011). Inhibitors of propagation of coagulation
(factors VIII, IX and XI): a review of current therapeutic practice. Br J Clin Pharmacol,
72(4), 553-562. doi: 10.1111/j.1365-2125.2010.03899.x
Globe, D. R., Curtis, R. G., Koerper, M. A., & Committee, H. S. (2004). Utilization of
care in haemophilia: a resource-based method for cost analysis from the Haemophilia
Utilization Group Study (HUGS). Haemophilia, 10 Suppl 1, 63-70.
Gringeri, A., Lundin, B., Von Mackensen, S., Mantovani, L., Mannucci, P. M., & The
Esprit Study, G. (2011). A randomized clinical trial of prophylaxis in children with
hemophilia A (the ESPRIT Study). Journal of Thrombosis and Haemostasis, 9(4), 700-
710. doi: 10.1111/j.1538-7836.2011.04214.x
Gringeri, A., Mantovani, L. G., Scalone, L., Mannucci, P. M., & Group, C. S. (2003).
Cost of care and quality of life for patients with hemophilia complicated by inhibitors: the
COCIS Study Group. Blood, 102(7), 2358-2363. doi: 10.1182/blood-2003-03-0941
Guh, S., Grosse, S. D., McAlister, S., Kessler, C. M., & Soucie, J. M. (2012a). Health
care expenditures for Medicaid-covered males with haemophilia in the United States,
2008. Haemophilia, 18(2), 276-283. doi: 10.1111/j.1365-2516.2011.02713.x
Guh, S., Grosse, S. D., McAlister, S., Kessler, C. M., & Soucie, J. M. (2012b).
Healthcare expenditures for males with haemophilia and employer-sponsored insurance
in the United States, 2008. Haemophilia, 18(2), 268-275. doi: 10.1111/j.1365-
2516.2011.02692.x
Johnson, K. A., Zhou, Z. Y., Koerper, M., Riske, B., Baker, J., Ullman, M., . . . Lou, M.
(2012). Burden of illness: Direct and indirect costs among persons with hemophilia A.
Paper presented at the The 2nd National Conference on Blood Disorders in Public Health,
Atlanta, GA, USA.
Kulkarni, R., Soucie, J. M., Lusher, J., Presley, R., Shapiro, A., Gill, J., . . . Evatt, B.
(2009). Sites of initial bleeding episodes, mode of delivery and age of diagnosis in babies
with haemophilia diagnosed before the age of 2 years: a report from The Centers for
Disease Control and Prevention’s (CDC) Universal Data Collection (UDC) project.
Haemophilia, 15(6), 1281-1290. doi: 10.1111/j.1365-2516.2009.02074.x
17

Manco-Johnson, M. J., Abshire, T. C., Shapiro, A. D., Riske, B., Hacker, M. R., Kilcoyne,
R., . . . Evatt, B. L. (2007a). Prophylaxis versus episodic treatment to prevent joint
disease in boys with severe hemophilia. N Engl J Med, 357(6), 535-544. doi:
10.1056/NEJMoa067659
Manco-Johnson, M. J., Abshire, T. C., Shapiro, A. D., Riske, B., Hacker, M. R., Kilcoyne,
R., . . . Evatt, B. L. (2007b). Prophylaxis versus episodic treatment to prevent joint
disease in boys with severe hemophilia.[see comment]. New England Journal of
Medicine, 357(6), 535-544.
Manco-Johnson, M. J., Kempton, C. L., Reding, M. T., Lissitchkov, T., Goranov, S.,
Gercheva, L., . . . Hong, W. (2013). Randomized, controlled, parallel-group trial of
routine prophylaxis vs. on-demand treatment with sucrose-formulated recombinant
factor VIII in adults with severe hemophilia A (SPINART). Journal of Thrombosis and
Haemostasis, 11(6), 1119-1127. doi: 10.1111/jth.12202
Manco-Johnson, M. J., Nuss, R., Geraghty, S., Funk, S., & Kilcoyne, R. (1994). Results
of secondary prophylaxis in children with severe hemophilia. Am J Hematol, 47(2), 113-
117.
Morfini, M., Haya, S., Tagariello, G., Pollmann, H., Quintana, M., Siegmund, B., . . .
Tusell, J. (2007). European study on orthopaedic status of haemophilia patients with
inhibitors. Haemophilia, 13(5), 606-612.
NHF. (2007). National Hemophilia Foundation. MASAC [Medical and Scientific
Advisory Council] Recommendation 179. MASAC Recommendation Concerning
Prophylaxis (Regular Administration of Clotting Factor Concentrate to Prevent Bleeding).
Available at:
http://www.hemophilia.org/NHFWeb/MainPgs/MainNHF.aspx?menuid=57&contentid=1
007.
Nilsson, I. M., Blomback, M., & Ahlberg, A. (1970). Our experience in Sweden with
prophylaxis on haemophilia. Bibl Haematol, 34, 111-124.
Pipe, S. W., & Valentino, L. (2007). Optimizing outcomes for patients with severe
haemophilia A. Haemophilia, 13(s4), 1-16.
Richards, M., Altisent, C., Batorova, A., Chambost, H., Dolan, G., de Moerloose, P., . . .
Rothschild, C. (2007). Should prophylaxis be used in adolescent and adult patients with
severe haemophilia? An European survey of practice and outcome data. Haemophilia,
13(5), 473-479. doi: HAE1478 [pii]
10.1111/j.1365-2516.2007.01478.x
Rodriguez-Merchan, E. C. (2008). Some recent developments regarding arthropathy and
inhibitors in haemophilia. Haemophilia, 14(2), 242-247.
18

Roth, D. A., Tawa, N. E., Jr., O'Brien, J. M., Treco, D. A., Selden, R. F., & Factor, V. T.
T. S. G. (2001). Nonviral transfer of the gene encoding coagulation factor VIII in patients
with severe hemophilia A. N Engl J Med, 344(23), 1735-1742. doi:
10.1056/NEJM200106073442301
Royal, S., Schramm, W., Berntorp, E., Giangrande, P., Gringeri, A., Ludlam, C., . . .
Szucs, T. (2002). Quality-of-life differences between prophylactic and on-demand factor
replacement therapy in European haemophilia patients. Haemophilia, 8(1), 44-50. doi:
10.1046/j.1365-2516.2002.00581.x
Soucie, J. M., Evatt, B., & Jackson, D. (1998). Occurrence of hemophilia in the United
States. The Hemophilia Surveillance System Project Investigators. Am J Hematol, 59(4),
288-294. doi: 10.1002/(SICI)1096-8652(199812)59:4<288::AID-AJH4>3.0.CO;2-I [pii]
Soucie, J. M., Nuss, R., Evatt, B., Abdelhak, A., Cowan, L., Hill, H., . . . Wilber, N.
(2000). Mortality among males with hemophilia: relations with source of medical care.
The Hemophilia Surveillance System Project Investigators. Blood, 96(2), 437-442.
Soucie, J. M., Symons, J. t., Evatt, B., Brettler, D., Huszti, H., Linden, J., & Hemophilia
Surveillance System Project, I. (2001). Home-based factor infusion therapy and
hospitalization for bleeding complications among males with haemophilia. Haemophilia,
7(2), 198-206.
Srivastava, A., Brewer, A., Mauser‐ Bunschoten, E., Key, N., Kitchen, S., Llinas, A., . . .
Poon, M. (2013). Guidelines for the management of hemophilia. Haemophilia, 19(1), e1-
e47.
Steen Carlsson, K., Hojgard, S., Glomstein, A., Lethagen, S., Schulman, S., Tengborn,
L., . . . Lindgren, B. (2003). On-demand vs. prophylactic treatment for severe
haemophilia in Norway and Sweden: differences in treatment characteristics and outcome.
Haemophilia, 9(5), 555-566.
Stobart, K., Iorio, A., & Wu, J. K. (2006). Clotting factor concentrates given to prevent
bleeding and bleeding-related complications in people with hemophilia A or B. Cochrane
Database Syst Rev(2), CD003429. doi: 10.1002/14651858.CD003429.pub3
Stonebraker, J. S., BOLTON‐ MAGGS, P. H., Michael Soucie, J., Walker, I., & Brooker,
M. (2010). A study of variations in the reported haemophilia A prevalence around the
world. Haemophilia, 16(1), 20-32.
Szucs, T. D., Offner, A., Kroner, B., Giangrande, P., Berntorp, E., Schramm, W., & Grp,
E. S. S. (1998). Resource utilisation in haemophiliacs treated in Europe: Results from the
European Study on Socioeconomic Aspects of Haemophilia Care. Haemophilia, 4(4),
498-501.
Teitel, J. M., Barnard, D., Israels, S., Lillicrap, D., Poon, M. C., & Sek, J. (2004). Home
management of haemophilia. Haemophilia, 10(2), 118-133. doi: 853 [pii]
19

Tencer, T., Friedman, H. S., Li-McLeod, J., & Johnson, K. (2007). Medical costs and
resource utilization for hemophilia patients with and without HIV or HCV infection. J
Manag Care Pharm, 13(9), 790-798. doi: 2007(13)9: 790-798 [pii]
Valentino, L. A., Pipe, S. W., Tarantino, M. D., Ye, X., Xiong, Y., & Luo, M. P. (2012).
Healthcare resource utilization among haemophilia A patients in the United States.
Haemophilia, 18(3), 332-338. doi: 10.1111/j.1365-2516.2011.02677.x
van Dijk, K., Fischer, K., van der Bom, J. G., Scheibel, E., Ingerslev, J., & van den Berg,
H. M. (2005). Can long-term prophylaxis for severe haemophilia be stopped in adulthood?
Results from Denmark and the Netherlands. British Journal of Haematology, 130(1),
107-112. doi: 10.1111/j.1365-2141.2005.05546.x
White, G. C., & Aledort, L. M. (2001). Definitions in hemophilia. THROMBOSIS AND
HAEMOSTASIS-STUTTGART-, 85(3), 560-560.

20

CHAPTER 2: ECONOMIC COMPARISON OF APCC VERSUS
RFVIIA FOR MILD-TO-MODERATE BLEEDING EPISODES IN
HEMOPHILIA PATIENTS WITH INHIBITORS
1

ABSTRACT
OBJECTIVE: To construct a cost-minimization model comparing aPCC vs rFVIIa in
hemophilia patients with inhibitors from a US third party payer perspective.
METHODS: A literature-based decision model was used to model inhibitor treatment
costs and outcomes. Since existing clinical trials fail to demonstrate differences in the
relative efficacy or safety of aPCC vs rFVIIa, we assumed the same efficacy for both
products in the base case. Regimens of aPCC (75IU/kg*2 doses) and rFVIIa (90μg/kg*3
doses) were assumed according to manufacturer recommendations. If the first-line
treatment failed, patients chose to continue the current treatment or switch to another
drug. All costs were adjusted to 2009 US dollars. Sensitivity analyses on the infusion
frequency, efficacy, unit price, switch rate, re-bleed rate, and body weight were
performed to assess model robustness.
RESULTS: In the base case, the total medical cost to treat a bleed with aPCC or rFVIIa
as first-line medication was US$25,969 and US$35,838 respectively. One-way sensitivity
analyses showed that results were insensitive to the efficacy of rFVIIa, unit price of
aPCC or rFVIIa, switch rate, re-bleed rate, or body weight. rFVIIa will reach cost-
neutrality when the efficacy of aPCC is as low as 60%, or rFVIIa is infused only twice
for each line, or aPCC is infused three times for each line. Two-way sensitivity analyses

1
The final, definitive version of this paper has been published in HAEMOPHILIA journal, Vol. 17/Issue 5,
September 2011 by WILEY Publications, Wiley & Sons, Inc. All rights reserved.
21

showed that results were quite sensitive to the assumed infusion frequency for both
products.
CONCLUSIONS: First-line aPCC, compared to rFVIIa, can be a cost-saving alternative
for home treatment of mild-to-moderate bleeds in hemophilia patients with inhibitors.

22

INTRODUCTION
The development of inhibitory antibodies to clotting factor VIII or factor IX remains one
of the most serious complications of factor replacement therapy in patients with
hemophilia. Approximately 20-35% of patients with severe hemophilia A and 4-6% of
patients with severe hemophilia B develop factor inhibitors, which makes hemostasis
difficult to achieve and treatment more expensive (DiMichele, 2000; Ehrenforth et al.,
1992; C. R. Hay et al., 1998). Such patients, compared with those without inhibitors, are
at higher risk of orthopedic complications, more likely to have severe and incapacitating
degrees of arthropathy, and more difficult to treat with surgery (Morfini et al., 2007;
Rodriguez-Merchan, 2008).
Treatment of acute bleeding episodes in patients with hemophilia and high responding
inhibitors most often involves the use of bypassing hemostatic agents, such as activated
prothrombin complex concentrates (aPCC, brand name FEIBA, manufacturer Baxter
Healthcare) or recombinant factor VIIa (rFVIIa, brand name NovoSeven, manufacturer
Novo Nordisk). The clinical efficacy of aPCC and rFVIIa for control of bleeding
episodes has been demonstrated in many single-arm studies (Hilgartner, Aledort, Andes,
& Gill, 1990; Kavakli et al., 2006; Key et al., 1998; Negrier et al., 1997). However,
recently two head-to-head clinical trials have been conducted directly comparing the
efficacy of the two alternatives (Astermark et al., 2007; Young, Shafer, Rojas, &
Seremetis, 2008). Notwithstanding a recent Novo Nordisk-sponsored “Bayesian meta-
analysis” of comparative efficacy from mostly single-arm uncontrolled clinical studies
(Treur et al., 2009), an objective summary of the randomized controlled trial comparative
evidence for aPCC versus rFVIIa by the Cochrane Collaboration indicates that there is no
23

conclusive evidence that one product is superior to the other in terms of efficacy or safety
(Iorio, Matino, D'Amico, & Makris, 2010). In fact, it is quite plausible that one of the
agents may work better for some patients and/or for some bleed sites and/or for different
episodes of care than the other product (Teitel et al., 2007). Thus, cost-minimization
analysis comparing the cost of two alternatives with equivalent efficacy and safety is
appropriately used in the economic evaluations of these two bypassing agents.

A number of articles and abstracts have been published comparing the cost-effectiveness
of aPCC and rFVIIa to treat a mild-to-moderate bleed (Dundar et al., 2005; Joshi,
Stephens, Munro, Mathew, & Botteman, 2006; Odeyemi & Guest, 2002a, 2002b; Ozelo
et al., 2007; Putnam, Bohn, Ewenstein, Winkelmayer, & Avorn, 2005; Steen Carlsson,
Astermark, Donfield, & Berntorp, 2008; You, Lee, & Park, 2009). Most of these studies
are sponsored by one of the two competing manufacturers of bypassing agents: Baxter
and Novo Nordisk. These studies made crucial assumptions on the values of treatment
efficacy and cost which drive the findings to favor their products. Not surprisingly, with
one exception (Putnam et al., 2005), most of the studies sponsored by Baxter imply cost
or cost-effectiveness advantages for aPCC, while all studies sponsored by Novo Nordisk
tend to favor rFVIIa (J. W. Hay & Zhou, 2010). However, in contrast to some of these
studies that include manufacturer employees as co-authors, our study is sponsored by
Baxter through a research grant to the University of Southern California which
specifically precludes the sponsor from altering, modifying or censoring our findings. To
explore the validity of prior cost effectiveness findings, we used clinically-based model
assumptions and investigated model parameters with extensive sensitivity analyses in a
24

cost-minimization model comparing the costs of aPCC versus rFVIIa as first-line drugs in
the home treatment of mild-to-moderate bleeding episodes in hemophilia patients with
inhibitors from a US third party payer’s perspective.

25

METHODS
Decision Analytic Model
A literature-based decision analytic model was used to model the costs and outcomes of
treating mild-to-moderate bleeding episodes among patients with severe hemophilia A
and inhibitors. The model structure (Figure 2.1) was adapted from previously published
models and adjusted to reflect the practice patterns in the US (Odeyemi & Guest, 2002a,
2002b). Model structure and parameter assumptions were discussed and verified by a US
hematologist with experience of managing bleeds in hemophilia patients with inhibitors.
The model follows the initial and subsequent treatments until the bleed is resolved, with
consideration of switching from one product to the other product, increasing infusion
frequency, and re-bleed. We assumed that patients are treated with either aPCC or rFVIIa
as first-line agent when bleeding occurs. If first-line therapy fails, the following options
will be chosen based on patient and clinician’s judgment: (1) continue with current
treatment; (2) switch bypassing agent; (3) Continue with same agent but increase infusion
frequency (augmentation). We assumed patients will try the first-line and second-line
treatments at home and be admitted to the hospital if the second-line treatment fails. If the
bleed reoccurs after initial control (re-bleed), the previous treatment is resumed.

26

Figure 2.1 Decision Analytic Model Layout

Model Parameter Assumptions
The base-case parameter assumptions are based on literature review and expert opinion
(see Table 2.1). According to the general recommendation, the dose of aPCC is 50-
100IU/kg body weight and up to 200IU/kg per day; and the dose of rFVIIa is 90μg/kg
given every two-three hours until hemostasis is achieved. Specifically, in the first-line
treatment, the dosage for aPCC was assumed to be two infusions at 75IU/kg, and three
infusions at 90μg/kg for rFVIIa. In the subsequent treatment, the dosage remains the
same in the continuation treatment or is doubled in the augmentation treatment. We
assumed the average patient weight was 75kg in the base-case analysis.
No rebleed
Bleed stops
Bleed stops
Rebleed aPCC (back to 1st line)
Bleed continues
Rebleed aPCC (back to 2nd line)
aPCC Bleed stops
No rebleed
aPCC
Bleed stops
Bleed continues rFVIIa
Bleed continues rFVIIa (augmentation) Bleed stops
Bleed Continues
Rebleed rFVIIa (back to 2nd line)
Bleed stops
No rebleed
rFVIIa
Bleed continues rFVIIa (augmentation) Bleed stops
rFVIIa
Mild-to-moderate
bleed
27

As aforementioned, existing clinical trials fail to demonstrate significant differences in
the relative efficacy of aPCC compared with rFVIIa (Iorio et al., 2010). As no study has
reported the differential efficacy of aPCC or rFVIIa by line of therapy, we assumed the
same efficacy for both agents in each line of treatment in the base case. The probabilities
of switching to another product and developing a re-bleed were adopted from previous
studies (Knight, Paisley, Wight, & Jones, 2003; You et al., 2009) and expert opinion.
Cost
The analysis is from a US third party payer perspective. We consider direct medical costs
only, which include drug costs, hospital costs and physician fee during hospitalization.
Drug costs were based on the average sales price from Medicare Part B payment limits.
The cost of hospitalization was obtained from 2006 National Statistics by Healthcare
Cost and Utilization Project for DRG 397 (coagulation disorders) and multiplied by the
cost-to-charge ratio calculated using U.S. Centers for Medicare and Medicaid Services
(CMS) Medicare prospective payment system reimbursement data (MEDPAR). The
physician fee for hospital care is obtained from the Medicare & Medicaid services
physician fee schedule, using Healthcare Common Procedure Coding System (HCPCS)
code 99231-3. All the costs were adjusted to 2009 US dollars.

28

Table 2.1 Base Case Assumptions
aPCC rFVIIa Reference
Weight (75kg)
Dose 75IU/kg 90μg/kg (Young et al., 2008)
Infusion frequency 2 3 Package Insert
Efficacy 85% 85%
(Astermark et al., 2007;
Young et al., 2008)
Switch rate 50% 50%
(You et al., 2009) and
expert opinion
Re-bleed rate 15% 15%
(Knight et al., 2003) and
expert opinion
Cost (in 2009 US $)
Drug Unit Cost $1.555 $1.308 CMS
a

Hospitalization

$9,692 AHRQ
b

Physician Fee

$267 CMS
c

Notes:
a. Payment allowance limit for Medicare Part B in October 2009
b. HCUPnet, Healthcare Cost and Utilization Project. Year 2006. Agency for Healthcare
Research and Quality, Rockville, MD. http://hcupnet.ahrq.gov/. The value was adjusted
to 2009 US dollar.
c. Medicare Physician Fee schedule (HCPCS 99231-3 )

Sensitivity Analysis
Sensitivity analyses were performed to assess the robustness of outcomes by varying the
model parameters. Specifically, one-way sensitivity analyses, in which one parameter is
varied while all other parameters are held constant, were performed on the efficacy rates,
29

factor unit price, infusion frequency, body weight, re-bleed rate, and switch rate.
Threshold analyses were also conducted to hypothetically determine cost neutrality
between the treatment strategies when holding other variables constant. Two-way
sensitivity analyses were performed by varying the infusion frequency, efficacy and unit
price for both products simultaneously. This maps the cost-effectiveness of treatment
options when different parameter value changes were incorporated simultaneously.
The base-case model and sensitivity analyses were conducted in Microsoft Excel 2007
(Microsoft Corporation, Redmond, Washington, USA).

30

RESULTS
In the base case, the total medical cost to treat a bleed with aPCC or rFVIIa as first-line
medication was US$25,969 and US$35,838 respectively. Compared with rFVIIa, aPCC
as first-line therapy saves US$9,869 per mild-to-moderate bleed.
Figure 2.2 summarizes the results of one-way sensitivity analysis using a tornado
diagram. Holding all other parameters constant, one-way sensitivity analyses showed that
results were insensitive to the efficacy of rFVIIa, unit price of aPCC or rFVIIa, switch
rate, re-bleed rate, or body weight. The model was relatively sensitive to the dose of
aPCC and rFVIIa and the efficacy of aPCC. The threshold analysis (Table 2.2) indicated
that the rFVIIa will reach cost-neutrality when the efficacy of aPCC is as low as 60%, or
rFVIIa is infused only twice for each line, or aPCC is infused three times for each line. If
we increase the unit price of aPCC by 50% (from $1.555 to $2.354), or reduce the rFVIIa
unit price by one-third (from $1.308 to $0.864), rFVIIa will also be a dominant strategy.
In two-way sensitivity analyses, we varied multiple parameters for both products
simultaneously. Efficacy, infusion frequency and unit price were assessed separately. As
shown in Figure 2.3, the grid area indicates the combination of parameter values
favoring aPCC, while the dotted diamond area indicates values favoring rFVIIa. The line
of demarcation indicates cost neutrality. Again, results were quite sensitive to the
assumed infusion frequency for both products. Results were found to be robust in the
two-way analysis of efficacy and unit price.

31

Figure 2.2 Tornado Diagram of One-way Sensitivity Analysis

Notes:
The incremental cost of rFVIIa versus aPCC in dollar is shown on the x-axis. For each
parameter examined, the upper and lower limits of the sensitivity analysis (labels appear
at either end of each band) are based on either ±20% of the base case, or clinically
reasonable range.
Table 2.2 Threshold Sensitivity Analysis
Parameter Baseline value Breakeven value
Infusion frequency: aPCC 2 3
rFVIIa 3 2
Efficacy: aPCC 85% 60%
rFVIIa 85% NA
Factor Unit price: aPCC $1.555 $2.354
rFVIIa $1.308 $0.864
Switch rate

50% NA
Re-bleed rate 15% NA
NA: not available

32

Figure 2.3 Two-way Sensitivity Analysis

aPCC is cost-saving
rFVIIa is cost-saving

aPCC is cost-saving
rFVIIa is cost-saving
33

Notes:
Impact of simultaneously varying the infusion frequency (1-5 times for aPCC and 1-6
times for rFVIIa), efficacy (60-80% for aPCC and 80-100% rFVIIa) and unit price ($0.6-
$1.5 for aPCC and $0.9-$2.4 for rFVIIa). The line of demarcation indicates cost
neutrality. Grid area indicates that using aPCC as first-line is the cheaper option and
dotted diamond area indicate that using rFVIIa is cost-saving.

aPCC is cost-saving
rFVIIa is cost-saving
34

DISCUSSION
Our results indicate first-line aPCC, compared to rFVIIa, can often be a cost-saving
alternative for home treatment of mild-to-moderate bleeds in patients with hemophilia
and inhibitors. Most of the costs are incurred during the first-line therapy. Subsequent
treatments, such as switching, increasing doses, hospitalization, and re-bleed, do not
affect the cost results substantially. Based on the similar model structure, our results
appear to be in contrast to some of the previous studies which found rFVIIa was less
costly (Dundar et al., 2005; Joshi et al., 2006; Odeyemi & Guest, 2002a, 2002b; Ozelo et
al., 2007; You et al., 2009). Parts of the disparities between our findings and other studies
are explained by the different efficacy assumption used in the analysis. Currently, only
two head-to-head clinical trials have been conducted to directly compare the efficacy of
aPCC versus rFVIIa (Astermark et al., 2007; Young et al., 2008). The FENOC study
(Astermark et al., 2007) compared one dose of aPCC (75-100IU/kg) with two doses of
rFVIIa (90-120μg/kg) and showed no statistically significant difference within 48 hours
follow up. In the study by Young et al. (Young et al., 2008), no difference was found
between patients using one dose of 75IU/kg aPCC and patients using three doses of
90μg/kg rFVIIa, which is the FDA approved dosage. Thus, it is inconclusive whether one
agent has superior efficacy to the other. However, most of the previous cost-effectiveness
studies assumed the base-case efficacy strongly favoring rFVIIa using evidence from two
separate single-arm observational studies (79% aPCC efficacy reported in 1990
(Hilgartner et al., 1990) versus 92% rFVIIa efficacy reported in 1998 (Key et al., 1998))
(Joshi et al., 2006; Odeyemi & Guest, 2002a, 2002b; Putnam et al., 2005). Other studies
used efficacy number from regional uncontrolled retrospective studies (Dundar et al.,
35

2005; Ozelo et al., 2007; You et al., 2009). These studies made no attempt to adjust for
observable or unobservable confounders that would lead to differential efficacy results
between these non-comparable aPCC and rFVIIa efficacy values, nor did they adjust to
account for 95% confidence intervals or other measures of statistical imprecision in
efficacy values. In our case, we started from assuming the same efficacy for aPCC and
rFVIIa and conducted extensive two-way sensitivity analysis comparing different
combinations of efficacy for both agents. The threshold analysis indicated that rFVIIa
will be cost-saving if the efficacy of aPCC falls to 60% and the results were insensitive to
the efficacy of rFVIIa.
Other sources of disparities can be explained by the infusion frequency used in each line
of therapy. Previous work often cited the dose of rFVIIa reported in Key et al. trial (Key
et al., 1998), in which an average of 2.3 doses of 90μg/kg rFVIIa was used to achieve
hemostasis (Dundar et al., 2005; Joshi et al., 2006; Odeyemi & Guest, 2002a, 2002b).
However, none of the analyses included an additional maintenance dose of rFVIIa
mandated in Key et al’s study. In fact, three doses of 90μg/kg or a single dose of
270μg/kg rFVIIa are often used in clinical trials (Kavakli et al., 2006; Santagostino et al.,
2006; Young et al., 2008). Using fewer doses of rFVIIa, these studies generated favorable
results for rFVIIa. However, these results were very sensitive to the base case values. As
shown by Odeyemi et al. (Odeyemi & Guest, 2002b), the treatment costs would be
equivalent if the infusion frequency for aPCC falls to 2.5 from 3 doses, or increase to 2.8
from 2.3 doses for rFVIIa. Our model indicated similar results. When holding other
variable constant, two doses of rFVIIa will make it a cost-saving strategy, while three
doses of rFVIIa lead to opposite results. Similar results were found for aPCC. So the
36

clinical optimal number of doses used to resolve a bleed is the key driver of total medical
cost.
Our study is not without limitations. First, there is a paucity of published data directly
comparing the efficacy of aPCC and rFVIIa and optimal dosage to reach such efficacy.
Although the FENOC and Young et al. studies provide some information on the relative
efficacy of aPCC and rFVIIa, they used specific regimens or short follow up evaluation
periods, which may not reflect the actual clinical experience with these agents. Due to the
lack of clinical evidence, we made parameter assumptions based on literature and experts’
opinions and assumed both agents achieve the same efficacy. In order to correct the
potential bias, we conducted extensive sensitivity analysis to assess the robustness and
provide a way to make systematic comparisons among various strategies. Moreover, we
did not include indirect costs in the model. Patient receiving rFVIIa might achieve
hemostasis faster, given its quicker infusion interval compared with aPCC (3 hours vs. 12
hours). Our findings may therefore overestimate the cost-saving of aPCC versus rFVIIa.
However, the degree to which items such as caregiver support time or lost productivity is
likely to be negligible compared with the total costs of medication.
Further head-to-head comparative studies examining the effectiveness and costs of
treatment strategies should be conducted, with considerations of key clinical variables
such as efficacy in each line of treatment, dosage, and re-bleed among the agents. The
full treatment course should be observed to ensure the optimal and most cost-effective
care for patients.
37

CHAPTER REFERENCES
Astermark, J., Donfield, S. M., DiMichele, D. M., Gringeri, A., Gilbert, S. A., Waters,
J., . . . Group, F. S. (2007). A randomized comparison of bypassing agents in hemophilia
complicated by an inhibitor: the FEIBA NovoSeven Comparative (FENOC) Study. Blood,
109(2), 546-551.
DiMichele, D. M. (2000). Inhibitors in haemophilia: a primer. Haemophilia, 6 Suppl 1,
38-40.
Dundar, S., Zulfikar, B., Kavakli, K., Gonen, C., Zulfikar, H., Yilmaz, D., . . . Karamalis,
M. (2005). A cost evaluation of treatment alternatives for mild-to-moderate bleeding
episodes in haemophilia patients with inhibitors in Turkey. Journal of Medical
Economics 8, 45-54.
Ehrenforth, S., Kreuz, W., Scharrer, I., Linde, R., Funk, M., Gungor, T., . . . Kornhuber,
B. (1992). Incidence of development of factor VIII and factor IX inhibitors in
haemophiliacs.[see comment]. Lancet, 339(8793), 594-598.
Hay, C. R., Ludlam, C. A., Colvin, B. T., Hill, F. G., Preston, F. E., Wasseem, N., . . .
Santagostino, E. (1998). Factor VIII inhibitors in mild and moderate-severity
haemophilia A. UK Haemophilia Centre Directors Organisation.[see comment].
Thrombosis & Haemostasis, 79(4), 762-766.
Hay, J. W., & Zhou, Z. Y. (2010). Commentary on Knight et al.: A systematic review of
the cost-effectiveness of rFVIIa and APCC in the treatment of minor/moderate bleeding
episodes for haemophilia patients with inhibitors. Haemophilia, 16(2), 366-368;
discussion 369-371. doi: HAE2117 [pii]
10.1111/j.1365-2516.2009.02117.x
Hilgartner, M., Aledort, L., Andes, A., & Gill, J. (1990). Efficacy and safety of vapor-
heated anti-inhibitor coagulant complex in hemophilia patients. FEIBA Study Group.
Transfusion, 30(7), 626-630.
Iorio, A., Matino, D., D'Amico, R., & Makris, M. (2010). Recombinant Factor VIIa
concentrate versus plasma derived concentrates for the treatment of acute bleeding
episodes in people with haemophilia and inhibitors. Cochrane Database of Systematic
Reviews, (8). Retrieved from
http://www.mrw.interscience.wiley.com/cochrane/clsysrev/articles/CD004449/frame.htm
l doi:10.1002/14651858.CD004449.pub3
Joshi, A. V., Stephens, J. M., Munro, V., Mathew, P., & Botteman, M. F. (2006).
Pharmacoeconomic analysis of recombinant factor VIIa versus APCC in the treatment of
minor-to-moderate bleeds in hemophilia patients with inhibitors. Current Medical
Research & Opinion, 22(1), 23-31.
38

Kavakli, K., Makris, M., Zulfikar, B., Erhardtsen, E., Abrams, Z. S., Kenet, G., &
NovoSeven trial, i. (2006). Home treatment of haemarthroses using a single dose regimen
of recombinant activated factor VII in patients with haemophilia and inhibitors. A multi-
centre, randomised, double-blind, cross-over trial. Thrombosis & Haemostasis, 95(4),
600-605.
Key, N. S., Aledort, L. M., Beardsley, D., Cooper, H. A., Davignon, G., Ewenstein, B.
M., . . . Taft, E. (1998). Home treatment of mild to moderate bleeding episodes using
recombinant factor VIIa (Novoseven) in haemophiliacs with inhibitors. Thrombosis &
Haemostasis, 80(6), 912-918.
Knight, C., Paisley, S., Wight, J., & Jones, M. L. (2003). Economic modelling of
different treatment strategies for haemophilia A with high-responding inhibitors.
Haemophilia, 9(4), 521-540.
Morfini, M., Haya, S., Tagariello, G., Pollmann, H., Quintana, M., Siegmund, B., . . .
Tusell, J. (2007). European study on orthopaedic status of haemophilia patients with
inhibitors. Haemophilia, 13(5), 606-612.
Negrier, C., Goudemand, J., Sultan, Y., Bertrand, M., Rothschild, C., & Lauroua, P.
(1997). Multicenter retrospective study on the utilization of FEIBA in France in patients
with factor VIII and factor IX inhibitors. French FEIBA Study Group. Factor Eight
Bypassing Activity. Thrombosis & Haemostasis, 77(6), 1113-1119.
Odeyemi, I., & Guest, J. (2002a). Modelling the economic impact of recombinant
activated factor VII and activated prothrombin-complex concentrate in the treatment of a
mild to moderate bleed in adults with inhibitors to clotting factors VIII and IX at a
comprehensive care centre in the UK. Journal of Medical Economics 5, 61-64.
Odeyemi, I., & Guest, J. (2002b). Modelling the economic impact of recombinant
activated factor VII compared to activated prothrombin-complex concentrate in the home
treatment of a mild to moderate bleed in adults with inhibitors to clotting Factors VIII
and IX in the UK. Journal of Medical Economics 5, 119-133.
Ozelo, M. C., Villaca, P. R., De Almeida, J. O., Bueno, T. M., De Miranda, P. A., Hart,
W. M., & Karamalis, M. (2007). A cost evaluation of treatment alternatives for mild-to-
moderate bleeding episodes in patients with haemophilia and inhibitors in Brazil.
Haemophilia, 13(5), 462-469.
Putnam, K. G., Bohn, R. L., Ewenstein, B. M., Winkelmayer, W. C., & Avorn, J. (2005).
A cost minimization model for the treatment of minor bleeding episodes in patients with
haemophilia A and high-titre inhibitors.[see comment]. Haemophilia, 11(3), 261-269.
Rodriguez-Merchan, E. C. (2008). Some recent developments regarding arthropathy and
inhibitors in haemophilia. Haemophilia, 14(2), 242-247.
Santagostino, E., Mancuso, M. E., Rocino, A., Mancuso, G., Scaraggi, F., & Mannucci, P.
M. (2006). A prospective randomized trial of high and standard dosages of recombinant
39

factor VIIa for treatment of hemarthroses in hemophiliacs with inhibitors. Journal of
Thrombosis & Haemostasis, 4(2), 367-371.
Steen Carlsson, K., Astermark, J., Donfield, S., & Berntorp, E. (2008). Cost and outcome:
comparisons of two alternative bypassing agents for persons with haemophilia A
complicated by an inhibitor. Thrombosis & Haemostasis, 99(6), 1060-1067.
Teitel, J., Berntorp, E., Collins, P., D'Oiron, R., Ewenstein, B., Gomperts, E., . . . Young,
G. (2007). A systematic approach to controlling problem bleeds in patients with severe
congenital haemophilia A and high-titre inhibitors. Haemophilia, 13(3), 256-263.
Treur, M. J., McCracken, F., Heeg, B., Joshi, A. V., Botteman, M. F., De Charro, F., &
Van Hout, B. (2009). Efficacy of recombinant activated factor VII vs. activated
prothrombin complex concentrate for patients suffering from haemophilia complicated
with inhibitors: a Bayesian meta-regression. Haemophilia, 15(2), 420-436. doi: HAE1956
[pii]
10.1111/j.1365-2516.2008.01956.x
You, C. W., Lee, S. Y., & Park, S. K. (2009). Cost and effectiveness of treatments for
mild-to-moderate bleeding episodes in haemophilia patients with inhibitors in Korea.
Haemophilia, 15(1), 217-226.
Young, G., Shafer, F. E., Rojas, P., & Seremetis, S. (2008). Single 270 microg kg(-1)-
dose rFVIIa vs. standard 90 microg kg(-1)-dose rFVIIa and APCC for home treatment of
joint bleeds in haemophilia patients with inhibitors: a randomized comparison.
Haemophilia, 14(2), 287-294.

40

CHAPTER 3: EXPLORING TEMPORAL PATTERNS OF CLOTTING
FACTOR USE AND ASSOCIATED HEALTHCARE UTILIZATION IN
HEMOPHILIA

ABSTRACT
Bleeding episodes in patients with hemophilia can be managed with clotting factor replacement
therapy either episodically or prophylactically. The purpose of this study was to classify and
characterize clotting factor use patterns using group-based trajectory models (GBTMs) and to
assess the economic outcomes associated with clotting factor use trajectory subgroups.
A retrospective analysis was conducted using Medicaid claims data from California, Florida,
Iowa, Kansas, Missouri, and New Jersey from January 1998 through June 2012. Medicaid
enrollees aged between 2-64 years with ≥1 diagnosis of hemophilia A or B and had used clotting
factor at least once were selected. Patients were followed up for 36 months, and based on
individual’s clotting factor use pattern over time, his probability of belonging to one trajectory of
clot factor use pattern was modeled using GBTMs with up to 10-group solutions. Patient’s
characteristics and healthcare utilization and costs were examined and compared among
trajectory subgroups. Multivariate regressions were performed to assess the impact of
prophylaxis on healthcare utilization by comparing the outcomes versus other subgroups.
A total of 1,035 patients who met all inclusion and exclusion criteria were included in the
analysis. A six-group model with linear trajectory for each subgroup best explained the data,
representing patients: 1) low utilization of clotting factor (25.8% of the study population); 2)
used clotting factor only occasionally across the 36 months (20.6%); 3) switched from
prophylaxis to episodic treatment (7.5%); 4) used clotting factor occasionally at the beginning
41

and had suboptimal prophylaxis later on (14.1%); 5) had high frequent episodic treatment and
switched to prophylaxis later on (17.2%); 6) consistently used prophylaxis (n=152, 14.5%). After
adjusting for baseline characteristics, patients always on prophylaxis had significantly fewer
bleeding-related ER visits or hospitalization compared with those who received episodic
treatment or suboptimal prophylaxis (adjusted incidence rate ratio ranges from 1.9 to 3.3 from
year 1 to year 3, all p-values < 0.05). Annual costs for the study population without inhibitors
averaged $107,420 (SD: 145,915; median: $51,564), of which 93% was attributed by clotting
factor costs.
Patients with hemophilia have distinct temporal patterns of clotting factor use. By identifying
such subgroups, clinician and payers can design more cost effective personalized treatment
regimens to this population.

42

INTRODUCTION
Hemophilia is a rare chronic congenital bleeding disorder that affects approximately 20,000
individuals in the United States (Akerley, Boucher, Bentz, Arbogast, & Walters, 2009; Soucie,
Evatt, & Jackson, 1998). Patients with hemophilia are deficient in one of clotting factor proteins
in blood, leading to the impairment in body’s ability to control bleeding and incidence of
spontaneous bleeds in joints, muscles, and other soft tissues. Left untreated, patients may
develop hemophilic arthropathy, which involves chronic pains, losses of range of motion in the
joint, and crippling musculoskeletal deformity and disability by young adulthood (Arun, 2001;
Pipe & Valentino, 2007).
While still not curable at present, bleeding episodes can be managed with clotting factor
replacement therapy using plasma-derived or recombinant clotting factor concentrates. Since
commercially prepared clotting factor became available, home infusion administered by patients
themselves or caregivers has been the mainstays of hemophilia care given the benefits of rapid
bleed control and reduction in school or work absenteeism (Gargallo, 2000). Nowadays, patients
with hemophilia are expected to have a near normal life expectancy, even among those who with
severe disease (Darby et al., 2007; Tagliaferri et al., 2010).
Factor replacement therapy is often administered after a bleeding episode (i.e., episodic or on-
demand treatment) or on a regular basis (i.e., prophylaxis). Episodic treatment is known to
effectively control bleeding, relieve pain, and restore joint mobility; however, it cannot prevent
arthropathy in the long term (Pipe & Valentino, 2007). Whereas episodic factor replacement
therapy is suitable for most patients with mild or moderate hemophilia, there's still a risk that
joint damage has already taken place before the episodic therapy is given, especially in patients
43

with severe hemophilia. In contrast, clotting factors can be infused on a regular schedule in order
to maintain blood clotting factor activity above 1%, therefore to prevent bleeding and the
development of arthropathy.
To date, many studies have examined the efficacy and effectiveness of prophylaxis compared
with episodic treatment in patients with severe hemophilia. Observational studies have shown
that prophylaxis significantly improves outcomes by decreasing the frequency of bleeding,
thereby delaying the onset of arthropathy and improves patients’ health-related quality of life
(Aledort, Haschmeyer, & Pettersson, 1994; Fischer et al., 2002; Royal et al., 2002; Stobart, Iorio,
& Wu, 2006; Szucs et al., 1998). Until recent years, three longitudinal randomized controlled
trials were conducted to compare the two treatment regimens in the prevention of joint disease in
young boys and adults with severe hemophilia. In a six-year randomized clinical trial, Manco-
Johnson et al. found that boys who were on primary prophylaxis (n=32) had significantly greater
reduction in the risk of joint damage detected by magnetic resonance imaging (MRI) and fewer
joint bleeds compared to those who received episodic therapy (n=33) (Manco-Johnson et al.,
2007). Similarly, in a ten-year randomized controlled trial conducted in Italy (ESPRIT), children
on prophylaxis (n=22) had fewer joint bleeds than children on episodic treatment (n=19)
(Gringeri et al., 2011). The proportion of patients with radiologic signs of arthropathy was also
fewer in prophylaxis group versus episodic group. Early start of prophylaxis (≤36 months) was
more effective with fewer joint bleeds and no radiologic signs of arthropathy after tens year of
treatment. For adults, as shown in the interim results of a randomized controlled trial –
SPINART, routine prophylaxis also significantly reduced joint bleeds as compared with episodic
treatment (Manco-Johnson et al., 2013).
44

Prophylactic treatment has been recommended by the World Foundation of Hemophilia and the
National Hemophilia Foundation for patients with severe hemophilia (NHF, 2007). Preferably,
prophylaxis should be initiated in an early age – prior to the onset of frequent bleeding and the
use of prophylaxis should be life-long (Makris, 2012). On the other hand, treatment in
hemophilia is a highly individualized approach. Different types of prophylaxis have also been
used in clinical practice. Primary prophylaxis is used predominantly in young children, in which
long-term, continuous treatment is initiated before the age of two or after the initial joint bleed.
In contrast, secondary prophylaxis (either continuous or intermittent) can be initiated after the
onset of joint disease, usually in response to repeated bleeds that cannot be managed effectively
with episodic therapy (Manco-Johnson, Nuss, Geraghty, Funk, & Kilcoyne, 1994). Data from the
CDC UDC project indicated that around 68% children with severe hemophilia who were
younger than 6 years of age and 86% children aged between 6 and 13 years received prophylaxis
between 2008 and 2010 (Baker, Riske, Voutsis, Cutter, & Presley, 2011).
Moreover, patients have different pharmacokinetic response to factors and their lifestyle may
impact the adherence of clotting replacement treatment (Astermark et al., 1999). Transition to
adulthood may be a significant barrier to adherence with prophylaxis when patients start to take
responsibility of self-treatment. In addition, joint bleeds become less frequent in adults and
adolescence as compared to childhood and the natural clinical course of hemophilia may differ
substantially between individuals. Should prophylaxis be continued for life or can it be safely
discontinued at certain age remains uncertain. A study on 80 patients treated in two Danish and
one Dutch treatment centers indicated that about one-third of young adults with severe
hemophilia switched to episodic treatment in early adulthood after years on prophylaxis, and still
maintained low joint bleeding rates and comparable joint status for four years (van Dijk et al.,
45

2005). In another European survey, around 42% of persons with severe hemophilia completely
discontinued prophylaxis during the age of 16-22 years and 28% of those had to restart routine
clotting factor replacement therapy because of poor outcomes (Richards et al., 2007). The long-
term consequence of discontinuing prophylaxis is still unknown.
From an economic perspective, treatment with clotting factor can be extraordinarily expensive. A
recent study of Medicare spending on Part B drugs found that clotting factor is among the 55
highest expenditure Part B drugs in 2010. Among 660 Medicare beneficiaries with hemophilia A,
the average expenditure on recombinant clotting factor VIII was $216,800 per patients, making it
the most expensive drug (GAO, 2012). Clotting factor consumption in prophylaxis was estimated
to be two to three times higher than that in episodic treatment, accounting for 94% versus 74% of
total costs respectively (Carlsson et al., 2004). Longitudinal treatment with clotting factor places
substantial economic burdens on patients, their families, and healthcare systems and the high
cost of clotting factor may also affect adherence to prophylaxis.
Whereas many clinicians have discussed the optimal treatment in hemophilia, the longitudinal
clotting factor use patterns in real world and its impact on healthcare utilization and costs have
not been fully studied. Most previous studies compared outcomes among patients receiving
either prophylactic or episodic treatment; however, it is also important to capture the changing
dynamics of treatment based on patient’s experience (e.g., non-adherence to prophylaxis, initiate
prophylaxis due to poor outcomes, switch to episodic treatment with a stable bleeding control).
The present study employed a group-based trajectory model (GBTM) method to account for the
dynamic nature of hemophilia treatment over time. GBTMs have been used in sociological and
medical research to identify developmental trajectories and were recently introduced to assess
the adherence of prescription medications overall time (Franklin et al., 2013; Modi et al., 2010;
46

Modi, Rausch, & Glauser, 2011). The purpose of this study was to classify and characterize
clotting factor use trajectories for patients with hemophilia using GBTMs and to assess the
economic outcomes associated with clotting factor use trajectory subgroups.
47

METHODS
Data Source
Medicaid data from six states - California, Florida, Iowa, Kansas, Missouri, and New Jersey
were used for this analysis. Medicaid is a state-administered government health insurance
program designed to serve low-income citizens, the blind, and those with chronic disabling
conditions. The database contained complete medical and pharmacy dispensing claims for
enrollees during covered years, including Medicare/Medicaid crossovers and can be linked on an
enrollee-specific basis over time. Eligibility information, including enrollment start and end
dates and program type, was available on a monthly basis. Other available variables included
demographics, claims by types of services (e.g., hospitalization, outpatient, emergency room
[ER]), dates of services, units of services (e.g., length of hospital stay), diagnoses and procedures
codes, and outpatient prescription drug dispensing claims. Diagnosis-specific claims were
identified by International Classification of Diseases, 9th Revision (ICD-9) codes. Clotting factor
infusion procedures and physician visits were identified by Current Procedural Terminology
Version 4 (CPT-4) codes or Healthcare Common Procedure Coding System (HCPCS).
Prescription clotting factor or drug claims were identified by National Drug Codes (NDCs).

Study design and sample
This study employed 100% of the Medicaid claims database in six states. Data spanned from
January 2005 through December 2009 for California, and January 1998 through June 2012 for
the rest of the five states (i.e., Florida, Iowa, Kansas, Missouri, and New Jersey). Patients with
hemophilia A or B were selected based on the following inclusion criteria:
48

1) Male;
2) One or more medical visits with a diagnosis of hemophilia A (ICD-9 code: 286.0x) or
hemophilia B (ICD-9 code: 286.1x);
3) At least one claim of clotting factor identified using NDC in the pharmacy claims or
clotting factor injection procedure code in the medical claims (see Appendix 1 for
HCPCS codes and brand name of clotting factors);
4) Were continuously eligible in their health plans for at least 36 consecutive months. The
first eligible episode was selected among all continuous eligibility episodes that met the
above criteria. The start date of eligibility was defined as the index date and the 36
months following index date were defined as the study period;
5) Age 2-64 as of the index date
Patients were excluded from the study if they were female or had at least two medical visits with
diagnosis of Von Willebrand disease (ICD-9 code: 286.4x). Patients with a history of inhibitors
were identified if they had at least one claim of bypassing agent during the entire eligibility
period. These patients were further excluded from the sample as they had very different clotting
factor use patterns and healthcare resource utilization compared with patients who never
developed inhibitors.

Measuring clotting factor use patterns
Clotting factor is infused based on body weight and can either be dispensed via pharmacy or
received during outpatient visits; therefore, traditional approaches using pharmacy claims to
define medication procession ratio (MPR) or proportion days covered by medication (PDC)
49

cannot be calculated because the total units and days of supply are not available if patients
receive clotting factor in the outpatient setting. In this study, we used the observed clotting factor
dispensing each month as a proxy for clotting factor use and the study period included 36 months
after the index date. Because most insurance plans only allow monthly supply for prophylactic
treatment, we expected to observe continuously monthly refills if patients were on prophylaxis,
while episodic treatment would follow a sporadic pattern. For each patient, we created 36 binary
indicators that represent whether each month was covered with clotting factor by counting the
distinct prescription fill dates. If a patient received clotting factor during the last 7 days of the
month following another dispensing during the same month, the later dispensing was extracted
and taken into account in the following month. These 36 binary indicators defined the observed
clotting factor use pattern during the study period. Also, based on these indicators, we calculated
the proportion of months covered (PMC), defined as the number of months with clotting factor
dispensed divided by the total number of follow-up months (i.e., 36 months).

Patient characteristics and outcomes
Patient characteristics were measured as of the index date, including demographics (age and
race/ethnicity), calendar year of index date (i.e., index year), insurance type, and region.
Hemophilia type, inhibitor history, use of desmopressin (medication only indicated for mild or
moderate hemophilia), and hemophilia-related comorbidities (i.e., hepatitis C virus [HCV] and
human immunodeficiency virus [HIV] infection) were determined by claims with diagnosis or
related treatments using the entire available data. Non-hemophilia related comorbidities were
measured during the 12-month following the index date and consisted of Charlson Comorbidity
50

Index (CCI) (Romano, Roos, & Jollis, 1993) and selected Elixhauser comorbidities (Quan et al.,
2005).
Studied outcomes included hemophilia-related healthcare utilization and costs measured during
each year of the 36-month study period. All-cause healthcare utilization (including inpatient, ER,
and outpatient visits) were measured by visit, number of visits, and length of inpatient stay.
Bleeding-related ER or inpatient visits and joint-related outpatient procedures or inpatient visits
were further measured if the claims were associated with a diagnosis of hemorrhage in joint,
muscle, subcutaneous area, gastrointestinal tract or brain. When calculating healthcare costs,
patients who were covered by managed care health insurance plans (e.g., health maintenance
organizations, preferred provider organizations, point of service) were excluded due to the
incomplete information on healthcare costs in these data. All-cause healthcare costs captured the
reimbursement amounts from Medicaid to healthcare providers and the costs were adjusted to
2012 US dollars using the medical care component of Consumer Price Index. Total costs,
clotting factor costs, other medication costs, and medical costs including inpatient, ER,
outpatient, and other costs were analyzed in the study.

Statistical analysis
A semi-parametric, group-based approach, group-based trajectory model (GBTM), was used to
classify patients by their observed clotting factor use patterns (Daniel S. Nagin, 1999). GBTM is
one type of finite mixture models that is designed to identify clusters of individuals (or trajectory
subgroups) following a similar developmental trajectory on an outcome of interest over age or
time (D. S. Nagin & Odgers, 2010). It is particular useful when the number of subgroups and
51

shapes of trajectory curves are unknown (Modi et al., 2010). Trajectory subgroups are used as a
statistical device for approximating the unobserved developmental trajectories across populations
and can be treated as latent longitudinal strata assuming that population variations are captured
by differences across groups in the shape and level of their trajectories. The model assumes that
repeated observations for the same individual are independently conditional on trajectory
subgroup, implying that the within-person correlation structure is explained completely by the
estimated trajectory curve for each group or cluster.
In this study, based on individual’s clotting factor use pattern over time, his probability of
belonging to one cluster of potential clot factor use pattern was modeled using a multinomial
logit regression with no predictors. The monthly indicators of clotting factor use were modeled
as outcomes and 4 to 6-group solutions were estimated and compared. In each model, clot factor
use was modeled as a smooth function of time using up to a fourth order polynomial within each
trajectory group. On the basis of these models, the posterior probability of membership in each
group for each individual was estimated and patients were assigned to one distinct trajectory
subgroup in which he was most likely to be. To further determine the optimal number of
trajectory subgroup that differentiate clotting factor use, the 2 to 10-group solutions were
compared based on model Bayesian Information Criterion (BIC) values, where closest to zero
(largest BIC) indicates the best fit. The final model selection were also based on reasonable
judgment and predicted trajectory group proportions that are sufficiently large (> 5%) (D. S.
Nagin & Odgers, 2010).
Within each cluster of clotting factor use pattern, patient characteristics and outcomes were
examined descriptively using means and standard deviations (SDs) for continuous variables, and
frequency counts and percentages for categorical variables. Unadjusted values were compared
52

among trajectory subgroups using Wilcoxon rank-sum tests for continuous variables and Chi-
square tests for categorical variables. Multivariate regressions were performed to assess the
impact of prophylaxis on healthcare utilization by comparing the inpatient and ER visits for
patients receiving prophylaxis versus other subgroups, adjusting for demographics, index year,
insurance type as of the index date, hemophilia-related comorbidities, and modified CCI. Count
variables (including the number of medical service visits and length of hospital stay) were
analyzed with negative binomial regression models. Incidence rate ratios (IRRs) and 95%
confidence interval (CIs) were estimated.
SAS version 9.3 (SAS Institute Inc, Cary, North Carolina) was used to perform the data analyses.
The GBTM was estimated using PROC TRAJ macro (http://www.andrew.cmu.edu/user/bjones),
a closed-source module developed specifically for use with SAS software.

53

RESULTS
A total of 9,173 patients with at least one diagnosis of hemophilia A or B between January 1998
and June 2012 were identified from the combined state Medicaid databases (Figure 3.1).
Figure 3.1 Selection of the study cohort

Among them, 1,035 patients who met all inclusion and exclusion criteria were included in the
analysis. Additionally, 155 patients were identified with a history of inhibitors to clotting factor
based on their use of bypassing agents and were excluded from the analysis. As shown in Table
3.1, the mean age for the overall population was 15.6 (SD: 14.0) years and more than 70% were
under 20 years of age. Around 80% (n=841) of the patients had hemophilia A and the rest were
with hemophilia B (n=194). More than two-thirds patients received fee-for-service (FFS) health
insurance plans. The distribution of patients among states is consistent with the population in
each state. Thirty-six percent of the selected patients were from California and only 3% were
from Kansas. Around 13% patients had used desmopressin, which indicated that they had a mild
or moderate hemophilia. More than 14% patients were infected with HIV and 27% with HCV.
54

Table 3.1 Demographic and Clinical Characteristics by clotting factor use pattern clusters (6-group model)
a

Overall
(n = 1035)
Clotting Factor Use Pattern Cluster
b

P-value
c

One
(n = 262)
Two
(n = 220)
Three
(n =76)
Four
(n =142)
Five
(n = 183)
Six
(n = 152)

n (%) n (%) n (%) n (%) n (%) n (%) n (%)
Clotting factor use pattern
PMC, mean [SD] 0.4 (0.3) 0.0 (0.0) 0.2 (0.1) 0.5 (0.1) 0.4 (0.1) 0.7 (0.1) 0.9 (0.1) <0.0001 *
Hemophilia type

0.0043 *
Hemophilia A 841 (81.26%) 215 (82.06%) 159 (72.27%) 64 (84.21%) 115 (80.99%) 157 (85.79%) 131 (86.18%)
Hemophilia B 194 (18.74%) 47 (17.94%) 61 (27.73%) 12 (15.79%) 27 (19.01%) 26 (14.21%) 21 (13.82%)
Insurance type

FFS 706 (68.21%) 170 (64.89%) 151 (68.64%) 54 (71.05%) 103 (72.54%) 133 (72.68%) 95 (62.50%) 0.2334
Age at index year

Age in years,
mean [SD] 15.6 (14.0) 15.9 (14.8) 17.1 (15.5) 18.2 (14.7) 17.3 (15.0) 14.3 (11.2) 11.5 (10.5) 0.0144 *
Age category

2 - 5 years 301 (29.08%) 83 (31.68%) 66 (30.00%) 17 (22.37%) 43 (30.28%) 49 (26.78%) 43 (28.29%) 0.6653
6 – 11 years 215 (20.77%) 49 (18.70%) 42 (19.09%) 13 (17.11%) 18 (12.68%) 39 (21.31%) 54 (35.53%) 0.0000 *
12 - 17 years 188 (18.16%) 48 (18.32%) 31 (14.09%) 12 (15.79%) 27 (19.01%) 38 (20.77%) 32 (21.05%) 0.4795
18 - 24 years 106 (10.24%) 22 (8.40%) 19 (8.64%) 16 (21.05%) 16 (11.27%) 27 (14.75%) 6 (3.95%) 0.0005 *
25 - 44 years 170 (16.43%) 43 (16.41%) 45 (20.45%) 12 (15.79%) 28 (19.72%) 27 (14.75%) 15 (9.87%) 0.1140
45 - 65 years 55 (5.31%) 17 (6.49%) 17 (7.73%) 6 (7.89%) 10 (7.04%) 3 (1.64%) 2 (1.32%) 0.0110 *
Race/Ethnicity

White 440 (42.51%) 121 (46.18%) 103 (46.82%) 33 (43.42%) 63 (44.37%) 72 (39.34%) 48 (31.58%) 0.0419 *
Black 140 (13.53%) 40 (15.27%) 33 (15.00%) 14 (18.42%) 25 (17.61%) 16 (8.74%) 12 (7.89%) 0.0301 *
Hispanic 246 (23.77%) 67 (25.57%) 50 (22.73%) 15 (19.74%) 29 (20.42%) 37 (20.22%) 48 (31.58%) 0.1283
Other 209 (20.19%) 34 (12.98%) 34 (15.45%) 14 (18.42%) 25 (17.61%) 58 (31.69%) 44 (28.95%) <0.0001 *
Hemophilia-related
comorbidities

HIV 147 (14.20%) 29 (11.07%) 29 (13.18%) 18 (23.68%) 29 (20.42%) 32 (17.49%) 10 (6.58%) 0.0007 *
HCV 275 (26.57%) 63 (24.05%) 53 (24.09%) 28 (36.84%) 43 (30.28%) 58 (31.69%) 30 (19.74%) 0.0257 *

Severity
Mild condition
d
131 (12.66%) 83 (31.68%) 32 (14.55%) 2 (2.63%) 5 (3.52%) 4 (2.19%) 5 (3.29%) <0.0001 *
Other comorbidities

CCI score
e
, mean[SD] 0.8 (2.0) 0.5 (1.4) 0.8 (2.0) 1.4 (2.5) 1.1 (2.2) 1.1 (2.4) 0.5 (1.4) 0.0008 *
Selected comorbidity
f

55

Abbreviations: FFS=Fee for service; SD = standard deviation; HCV= hepatitis C virus; HIV= human immunodeficiency virus;
PMC=proportion of months covered; CCI = Charlson Comorbidity Index; uncomp=uncomplicated; comp=complicated
Notes:
a.
The demographic and clinical characteristics were measured during the 365-day period (first year) following the index date. HIV and
HCV infections were identified during the whole eligibility period.
b.
Clotting factor use pattern clusters were identified using group-based trajectory analysis which recursively grouped together patients
with similar temporal adherence patterns.
c.
Statistical comparisons were conducted using chi-square tests for categorical variables and Kruskal-Wallis tests for continuous
variables.
d.
Mild condition was identified if patients have used desmopressin, which is indicated only for mild or moderate hemophilia.
e.
Based on Romano et al. (1993).
f.
Based on Quan et al. (2005).

Hypertension 29 (2.80%) 9 (3.44%) 7 (3.18%) 1 (1.32%) 7 (4.93%) 3 (1.64%) 2 (1.32%) 0.3450
Paralysis 22 (2.13%) 7 (2.67%) 4 (1.82%) 2 (2.63%) 1 (0.70%) 4 (2.19%) 4 (2.63%) 0.8297
Other neurological
disorders 21 (2.03%) 8 (3.05%) 5 (2.27%) 1 (1.32%) 1 (0.70%) 2 (1.09%) 4 (2.63%) 0.5521
Chronic pulmonary
disease 93 (8.99%) 20 (7.63%) 16 (7.27%) 9 (11.84%) 12 (8.45%) 19 (10.38%) 17 (11.18%) 0.6207
Diabetes, uncomp 24 (2.32%) 4 (1.53%) 9 (4.09%) 1 (1.32%) 8 (5.63%) 1 (0.55%) 1 (0.66%) 0.0088 *
Diabetes, comp 3 (0.29%) 2 (0.76%) 1 (0.45%) 0 (0.00%) 0 (0.00%) 0 (0.00%) 0 (0.00%) 0.5717
Liver disease 71 (6.86%) 12 (4.58%) 19 (8.64%) 10 (13.16%) 12 (8.45%) 12 (6.56%) 6 (3.95%) 0.0612
Depression 47 (4.54%) 6 (2.29%) 8 (3.64%) 7 (9.21%) 6 (4.23%) 11 (6.01%) 9 (5.92%) 0.1126

56

GBTM analyses were conducted on 2 to 10-group models, with time as the only covariate
representing 36 observation time points. As illustrated in Figure 3.2, distinct patterns of clotting
factor use were identified based on the pre-specified number of groups (examples are shown for
4 to 6-group solutions). Using BIC values as the criterion for statistical significance, results
indicated that a six-group model with linear trajectory for each subgroup best explained the data
compared with solutions using fewer groups (Table 3.2 and Figure 3.3).
Group membership for each trajectory in the six-group model is shown in Figure 3.2 C:
1) low utilization of clotting factor (n=262, 25.8% of the study population);
2) used clotting factor only occasionally (episodic treatment) across the 36 months (n=220,
20.6%);
3) switched from prophylaxis to episodic treatment (n=76, 7.5%);
4) used clotting factor occasionally at the beginning anad had slowly increasing use of
clotting factor later on (suboptimal prophylaxis) (n=142, 14.1%);
5) had high frequent episodic treatment and switched to prophylaxis later on (n=183, 17.2%);
6) nearly always used prophylaxis (n=152, 14.5%).
In addition, the median and interquartile range of the predicted probability of membership among
patients assigned to that subgroup is shown in Figure 3.4. For example, the values next to
subgroup 1 describe the predicted probability of being in that group among those who were
assigned to subgroup 1. All trajectory group assignments were predicted with high probabilities
of membership in their respective trajectory group. The fact that the 95% confidence intervals
overlap for only relatively short periods indicates that the identified trajectory groups are distinct.
57

Figure 3.2 Trajectory models using 4–6 group solutions

Note: In each trajectory plot using 4-6 group solutions (from top to bottom), the observed
clotting factor use patterns are plotted with solid lines and the predicted trajectories for clotting
factor use pattern are plotted with dashed lines. The percent of each group is displayed at the
right of each curve.

Cluster 2: 24.3%
Cluster 3: 24.5%
Cluster 1: 33.4%
Cluster 4: 17.8%

Cluster 5: 18.1%
Cluster 4: 19.2%
Cluster 2: 7.9%
Cluster 3: 22.0%
Cluster 1: 32.7%

Cluster 6: 14.8%
Cluster 5: 17.2%
Cluster 2: 20.6%
Cluster 3: 7.5%
Cluster 1: 25.8%
Cluster 4: 14.1%
A
B
C
58

Table 3.2 BIC values and predicted group percentages for GTBM for 2- to 7-group
trajectory solutions
Number of
groups
BIC
Predicted Group Proportions
1 2 3 4 5 6 7
2 -18395 53.8% 46.2%

3 -17321 41.2% 30.6% 28.2%

4 -17039 33.4% 24.3% 24.5% 17.8%

5 -16855 32.7% 22.0% 7.9% 19.3% 18.1%

6 -16760 25.8% 20.6% 14.1% 17.2% 7.5% 14.8%

7 -16683 20.4% 25.5% 4.2% 15.3% 5.8% 15.4% 13.4%
Bold represents the final model that was selected
Figure 3.3 Bayesian information criterion versus number of clusters used in GBTM

-18500
-18100
-17700
-17300
-16900
-16500
2 3 4 5 6 7 8 9 10
Bayesian Information Criterion
Number of clusters
59

Figure 3.4 Trajectories of clotting factor use patterns using 6-group solution with median
trajectory membership probabilities

Note: The graph displays the predicted trajectories for clotting factor use pattern (solid lines)
with 95% confidence limits (dashed lines). The median and interquartile range of trajectory
membership probabilities by group assignment are displayed at the right.

The average PMC and patients’ characteristics in each trajectory subgroup from the 6-group
model are shown in Table 3.1. There were no significant difference in insurance type and index
year (data not shown in the table) among trajectory subgroups. Patients nearly always on
prophylaxis (subgroup 6) were on average younger, more likely to be growing children, and
living in California. Patients who used clotting factor episodically (subgroup 2) were generally
older, more likely to have hemophilia B. In addition, patients who switched between episodic
and prophylactic treatment (subgroups 3 and 5) were more likely to be young adults and be
infected with HCV or HIV. Most patients with mild hemophilia either barely used any clot factor
(subgroup 1) or only used it occasionally (subgroup 2).
60

Annual healthcare resource utilization in each of the year during the 36-month study period was
further assessed and compared among trajectory subgroups. In addition, multivariate analyses
were conducted to assess the impact of prophylaxis on healthcare utilization by comparing the
outcomes of subgroup 6 (adherence to prophylaxis) with other subgroups (2-5). As shown in
Figure 3.5 and Table 3.3, patients used clotting factor episodically (subgroup 2) and suboptimal
prophylaxis (subgroup 4) had more all-cause ER visits (adjusted incidence rate ratio [aIRR]
ranges from 1.6 to 2.1 from year 1 to year 3, all p-values < 0.05) than those who were always on
prophylaxis. Patients in subgroup 5 (had occasionally factor use at beginning and switched
prophylaxis) had significantly more ER visits during the first two years (aIRR=1.7 [95% CI: 1.1,
2.5] in year 1, aIRR=1.9 [95% CI: 1.3, 2.8] in year 2) and the rate ratio deceased in the third year
(aIRR= 1.5 [1.0, 2.3]) after patients switched to prophylaxis. Similarly trends were found for
bleeding-related ER or inpatient visits (Table 4 and Figure 4). However, there was no significant
difference in inpatient and outpatient visits for patients used prophylaxis versus other treatment
patterns. The healthcare utilization in subgroup 3 remained stable after decreasing the frequency
of clotting factor use.

61

Figure 3.5 Healthcare resource utilization in hemophilia patients during each year of the
36-month study period, by clotting factor use pattern clusters
A. Annual all-cause ER visits

B. Annual all-cause hospitalizations

C. Annual bleeding-related ER visits or hospitalizations

0.0
0.5
1.0
1.5
2.0
2.5
One Two Three Four Five Six
Mean number of ER visits
Clotting factor use pattern cluster
Year 1
Year 2
Year 3
0
0.2
0.4
0.6
0.8
1
One Two Three Four Five Six
Mean number of hospitalizations
Clotting factor use pattern cluster
Year 1
Year 2
Year 3
0.0
0.2
0.4
0.6
0.8
1.0
1.2
One Two Three Four Five Six
Bleeding-related ER visits
or hospitalizations
Clotting factor use pattern cluster
Year 1
Year 2
Year 3
62

Table 3.3 Healthcare resource utilization by clotting factor use pattern clusters (6-group
model)
a

Incidence Rate Ratio [95%CI]
b

[One]/[Six] [Two]/[Six] [Three]/[Six] [Four]/[Six] [Five]/[Six]
All-cause ER
visits
Year 1 0.6 [0.4 ,0.8]** 1.6 [1.0 ,2.4]* 1.6 [1.0 ,2.6]* 1.6 [1.1 ,2.5]* 1.7 [1.1 ,2.5]*
Year 2 0.7 [0.5 ,1.1] 1.8 [1.2 ,2.5]** 1.1 [0.7 ,1.7] 2.1 [1.5 ,3.1]*** 1.9 [1.3 ,2.8]**
Year 3 0.9 [0.6 ,1.3] 2.2 [1.5 ,3.3]*** 1.6 [0.9 ,2.5] 2.1 [1.4 ,3.1]** 1.5 [1.0 ,2.3]
All-cause IP
visits

Year 1 0.6 [0.4 ,0.8]** 0.5 [0.3 ,0.9]* 1.2 [0.6 ,2.3] 0.6 [0.3 ,1.1] 1.0 [0.6 ,1.8]
Year 2 0.5 [0.3 ,0.9]* 0.8 [0.5 ,1.4] 1.4 [0.8 ,2.5] 0.8 [0.4 ,1.4] 1.7 [1.1 ,2.7]*
Year 3 0.5 [0.3 ,0.8]** 0.8 [0.5 ,1.3] 1.4 [0.8 ,2.5] 1.2 [0.7 ,1.9] 1.2 [0.7 ,1.9]
IP length of stay
(LOS), days

Year 1 0.2 [0.1 ,0.6]** 0.3 [0.1 ,0.6]** 1.1 [0.4 ,3.1] 0.3 [0.1 ,0.8]* 0.9 [0.4 ,2.0]
Year 2 0.4 [0.2 ,0.9]* 0.5 [0.3 ,1.1] 1.1 [0.4 ,3.0] 0.6 [0.3 ,1.6] 1.4 [0.7 ,2.9]
Year 3 0.3 [0.1 ,0.8]* 0.6 [0.3 ,1.3] 2.3 [0.9 ,6.4] 1.2 [0.6 ,2.7] 1.0 [0.5 ,2.1]
All-cause OP
visits

Year 1 0.4 [0.3 ,0.5]*** 0.5 [0.4 ,0.6]*** 0.9 [0.8 ,1.1] 0.7 [0.6 ,0.8]** 0.8 [0.7 ,0.9]**
Year 2
0.5 [0.4 ,0.6]*** 0.7 [0.6 ,0.9]** 0.9 [0.7 ,1.1] 0.9 [0.8 ,1.1] 1.0 [0.8 ,1.1]
Year 3 0.6 [0.5 ,0.7]*** 0.9 [0.7 ,1.1] 1.0 [0.7 ,1.3] 1.2 [1.0 ,1.5] 1.2 [1.0 ,1.4]
Bleeding related
ER/IP visits

Year 1 0.6 [0.4 ,1.1] 1.9 [1.0 ,3.5]* 2.2 [1.1 ,4.4]* 2.2 [1.2 ,4.0]** 1.8 [1.1 ,3.0]*
Year 2 0.9 [0.6 ,1.5] 2.6 [1.6 ,4.3]*** 1.2 [0.7 ,2.3] 3.3 [2.0 ,5.5]*** 1.8 [1.1 ,3.0]*
Year 3 0.7 [0.4 ,1.3] 2.3 [1.4 ,3.9]** 1.6 [0.8 ,3.2] 2.1 [1.3 ,3.6]** 1.2 [0.7 ,2.0]
Abbreviations: ER=emergency room; IP = inpatient; OP = outpatient; CI= confidence interval
Notes:
Significance at 0.05 level was indicated with *, 0.01 level was indicated with **, 0.001 was
indicated with ***
a
Healthcare resource utilization was measured during each year of the 36-month study period.
b
Incidence rate ratios (IRRs) and p-values were calculated for number of visits or total length of
stay from negative binomial model with generalized estimating equation. All models controlled
for age, hemophilia type, geographic region, index year, insurance type, CCI, and comorbidities
(HIV or HCV infections). IRRs>1 indicate an increased risk or incidence rates for patients in
subgroup 1 -5 compared with those in subgroup 6.

63

Costs analysis was conducted only among patients received FFS health plans during the entire
36-month study period (n=706). Annual costs for the study population without inhibitors
averaged $107,420 (SD: 145,915; median: $51,564), of which 93% were attributed by clotting
factor costs. Figure 3.6 shows the unadjusted costs among the trajectory subgroups during each
of the year during the 36-month study period and the results were consistent with the clotting
factor use in each subgroup.
Figure 3.6 All-cause healthcare costs during each year of the 36-month study period, by
clotting factor use pattern clusters (6-group model)

0
50,000
100,000
150,000
200,000
250,000
300,000
Y1 Y2 Y3 Y1 Y2 Y3 Y1 Y2 Y3 Y1 Y2 Y3 Y1 Y2 Y3 Y1 Y2 Y3 .
One Two Three Four Five Six
Cost per year (2012 US$)
Clotting factor use pattern cluster (by year)
Clotting factor Other drugs Inpatient Outpatient ER visits
64

DISCUSSION
This retrospective analysis combined administrative claims data from six state Medicaid
databases to identify temporal patterns of clotting factor use using a group-based latent trajectory
model. To our knowledge, this is the first study to characterize longitudinal clotting factor use
patterns among patients with hemophilia using claims database. Pharmacy administrative data
are commonly used to study treatment patterns and adherence and PDC or MPR has been the
standard metric. However, the conventional approaches are not feasible in hemophilia care
because patients can receive clotting factor in both pharmacy and outpatient settings. Trajectory
model provides a flexible alternative and previous study has shown adherence subgroups
identified by the trajectory models are generally more homogenous with respect to observed
adherence than groups based on PDC (Franklin et al., 2013). Moreover, it is important to
consider the dynamic nature of hemophilia treatment over time when characterizing treatment
patterns in hemophilia and a simplified matric such as PDC or MPR cannot fully capture the
temporal patterns of the treatments.
By using GBTMs, six distinctive patterns of clotting factor use were identified. While subgroup
1 characterized patients who hardly used any clotting factor and more likely to be patients with
mild condition, the rest subgroups represented patients received prophylaxis (subgroup 6),
episodic treatment (subgroup 2), suboptimal prophylaxis (subgroup 4), and those who
transitioned between prophylactic and episodic regimens (subgroups 3 and 5). Patients who were
adherent to prophylaxis were more likely to be young children. Besides primary prophylaxis,
which is often initiated before age of two years, some patients may choose to initiate secondary
prophylaxis in response to repeated bleeding that cannot be managed effectively with episodic
regimens in their later age. In addition, we also observed more than 14% patients were non-
65

adherent to their prophylactic treatment and they were more likely to be older. Moreover, joint
bleeds are usually less frequent in adults and adolescence. We observed a proportion of patients
who transitioned to adulthood and switched to episodic treatment when they need to take
responsibility of self-treatment.
Economic outcomes associated with the temporal clotting factor use patterns were further
evaluated among the subgroups identified based on the longitudinal observations. After adjusting
for baseline characteristics, patients who were always on prophylaxis had comparable number of
bleeding-related ER visits or hospitalizations compared with those who had low utilization of
clotting factor. In addition, patients always on prophylaxis had significantly fewer bleeding-
related ER visits or hospitalizations compared with patients received episodic treatment or
suboptimal prophylaxis. With more patients with severe hemophilia, those who were adherent to
prophylaxis had comparable risks of inpatient hospitalizations compared with patients received
episodic regimens.
While the annual costs for clotting factor can reach more than $250,000 for patients nearly
always on prophylaxis, the high cost can be partly offset by the medical cost savings from
preventing bleeding-related ER visits and hospitalizations, and long-term arthropathy. In addition,
non-monetary benefits including improvement in health-related quality of life and preventing
school absenteeism, productivity loss and disability should also be considered when assessing
the benefits of prophylaxis (Feldman et al., 2007; Manco-Johnson et al., 2007).
This study is subject to several limitations. First, due to the lack of prescribing notes in the
claims databases, clotting factor use patterns (e.g., prophylactic or episodic treatments) were
determined by the frequency of clotting factor dispensing observed in the data. The actual
66

patterns may be misclassified because patients can fill prescriptions without actually taking them.
This is one of the common limitations of using claims to study adherence and becomes more
prominent in short-term adherence after treatment initiation. This issue may be less of concern in
our study because prior authorization is required for the use of clotting factor and the observation
period is three years. Second, the study sample is not representative of the hemophilia population
as only people who were covered by Medicaid plans in six states were included in the study.
Generalization of findings to populations beyond Medicaid enrollees should be made with
caution. On the other hand, Medicaid databases provide a more stable population compared with
commercial insurance plans to study treatment patterns and outcomes. In general, state Medicaid
programs serve low-income population and often offer generous benefits through pharmaceutical
assistance programs, both of which may reduce the likelihood that income or ability to afford
treatments—two major unobserved characteristics in claims data—are confounding factors in our
analysis. Third, due to the rareness of hemophilia, even combining data from six states, we still
face a small sample size, especially after temporal patterns were identified in each subgroup.
Additionally, given the relatively short observation period compared with the disease course,
outcomes were only assessed and compared in the each of the year during the 36-month study
period. Longitudinal study to evaluate outcomes over time and the long-term outcomes
associated with temporal patterns of factor use, such as development of arthropathy, were not
studied. Lastly, the economic analysis was from a payer’s perspective and did not include paid
amount from other sources or indirect costs. Future studies are needed to quantify indirect costs
associated with productivity loss and work absenteeism.
In summary, group-based trajectory models were used to characterize the temporal patterns of
clotting factor use among patients with hemophilia and shed a light on the impact of adherence to
67

prophylaxis or changes between treatment patterns on the economic outcomes. The subgroups of
patients with distinct patterns of clotting factor use may represent unique strata of health-seeking
behavior as a result of health consciousness or previous experience of treatments. By identifying
such subgroups, clinician and payers can target different interventions to patients with different
adherence experiences and design more cost effective personalized regimens to this population.

68

CHAPTER REFERENCES
Akerley, W., Boucher, K. M., Bentz, J. S., Arbogast, K., & Walters, T. (2009). A phase II study
of erlotinib as initial treatment for patients with stage IIIB-IV non-small cell lung cancer.
Journal of Thoracic Oncology, 4(2), 214-219.
Aledort, L. M., Haschmeyer, R. H., & Pettersson, H. (1994). A longitudinal study of orthopaedic
outcomes for severe factor-VIII-deficient haemophiliacs. The Orthopaedic Outcome Study
Group. J Intern Med, 236(4), 391-399.
Arun, B. K., C. (2001). Inherited Bleeding Disorders: Haemostasis and Thrombosis (4 ed.).
Philadelphia, PA: Lippincott Williams and Wilkens.
Astermark, J., Petrini, P., Tengborn, L., Schulman, S., Ljung, R., & Berntorp, E. (1999). Primary
prophylaxis in severe haemophilia should be started at an early age but can be individualized. Br
J Haematol, 105(4), 1109-1113.
Baker, J. R., Riske, B., Voutsis, M., Cutter, S., & Presley, R. (2011). Insurance, home therapy,
and prophylaxis in U.S. youth with severe hemophilia. Am J Prev Med, 41(6 Suppl 4), S338-345.
doi: 10.1016/j.amepre.2011.09.002
Carlsson, K. S., Hojgard, S., Lindgren, A., Lethagen, S., Schulman, S., Glomstein, A., . . .
Lindgren, B. (2004). Costs of on-demand and prophylactic treatment for severe haemophilia in
Norway and Sweden. Haemophilia, 10(5), 515-526. doi: 10.1111/j.1365-2516.2004.00952.x
Darby, S. C., Kan, S. W., Spooner, R. J., Giangrande, P. L. F., Hill, F. G. H., Hay, C. R. M., . . .
Williams, M. (2007). Mortality rates, life expectancy, and causes of death in people with
hemophilia A or B in the United Kingdom who were not infected with HIV. Blood, 110(3), 815-
825. doi: 10.1182/blood-2006-10-050435
Feldman, B. M., Aledort, L., Bullinger, M., Delaney, F. M., Doria, A. S., Funk, S., . . . Second
International Prophylaxis Study, G. (2007). The economics of haemophilia prophylaxis:
governmental and insurer perspectives. Proceedings of the Second International Prophylaxis
Study Group (IPSG) symposium. Haemophilia, 13(6), 745-749.
Fischer, K., van der Bom, J. G., Molho, P., Negrier, C., Mauser-Bunschoten, E. P., Roosendaal,
G., . . . van den Berg, H. M. (2002). Prophylactic versus on-demand treatment strategies for
severe haemophilia: a comparison of costs and long-term outcome. Haemophilia, 8(6), 745-752.
Franklin, J. M., Shrank, W. H., Pakes, J., Sanfelix-Gimeno, G., Matlin, O. S., Brennan, T. A., &
Choudhry, N. K. (2013). Group-based trajectory models: a new approach to classifying and
predicting long-term medication adherence. Med Care, 51(9), 789-796. doi:
10.1097/MLR.0b013e3182984c1f
GAO. (2012). Government Accountability Office. Medicare: High-Expenditure Part B Drugs.
October 12, 2012. Available at: http://www.gao.gov/assets/650/649459.pdf
69

Gargallo, B. (2000). Hemophilia treatment. Immune tolerance: the parent's perspective.
Haematologica, 85(10 Suppl), 81-82.
Gringeri, A., Lundin, B., Von Mackensen, S., Mantovani, L., Mannucci, P. M., & The Esprit
Study, G. (2011). A randomized clinical trial of prophylaxis in children with hemophilia A (the
ESPRIT Study). Journal of Thrombosis and Haemostasis, 9(4), 700-710. doi: 10.1111/j.1538-
7836.2011.04214.x
Makris, M. (2012). Prophylaxis in haemophilia should be life-long. Blood Transfus, 10(2), 165-
168. doi: 10.2450/2012.0147-11
Manco-Johnson, M. J., Abshire, T. C., Shapiro, A. D., Riske, B., Hacker, M. R., Kilcoyne, R., . . .
Evatt, B. L. (2007). Prophylaxis versus episodic treatment to prevent joint disease in boys with
severe hemophilia. N Engl J Med, 357(6), 535-544. doi: 10.1056/NEJMoa067659
Manco-Johnson, M. J., Kempton, C. L., Reding, M. T., Lissitchkov, T., Goranov, S., Gercheva,
L., . . . Hong, W. (2013). Randomized, controlled, parallel-group trial of routine prophylaxis vs.
on-demand treatment with sucrose-formulated recombinant factor VIII in adults with severe
hemophilia A (SPINART). Journal of Thrombosis and Haemostasis, 11(6), 1119-1127. doi:
10.1111/jth.12202
Manco-Johnson, M. J., Nuss, R., Geraghty, S., Funk, S., & Kilcoyne, R. (1994). Results of
secondary prophylaxis in children with severe hemophilia. Am J Hematol, 47(2), 113-117.
Modi, A. C., Cassedy, A. E., Quittner, A. L., Accurso, F., Sontag, M., Koenig, J. M., & Ittenbach,
R. F. (2010). Trajectories of adherence to airway clearance therapy for patients with cystic
fibrosis. J Pediatr Psychol, 35(9), 1028-1037. doi: 10.1093/jpepsy/jsq015
Modi, A. C., Rausch, J. R., & Glauser, T. A. (2011). Patterns of nonadherence to antiepileptic
drug therapy in children with newly diagnosed epilepsy. JAMA, 305(16), 1669-1676. doi:
10.1001/jama.2011.506
Nagin, D. S. (1999). Analyzing developmental trajectories: A semiparametric, group-based
approach. [doi:10.1037/1082-989X.4.2.139]. Psychological Methods, 4(2), 139-157. doi:
10.1037/1082-989X.4.2.139
Nagin, D. S., & Odgers, C. L. (2010). Group-based trajectory modeling in clinical research.
Annu Rev Clin Psychol, 6(1), 109-138. doi: 10.1146/annurev.clinpsy.121208.131413
NHF. (2007). National Hemophilia Foundation. MASAC [Medical and Scientific Advisory
Council] Recommendation 179. MASAC Recommendation Concerning Prophylaxis (Regular
Administration of Clotting Factor Concentrate to Prevent Bleeding). Available at:
http://www.hemophilia.org/NHFWeb/MainPgs/MainNHF.aspx?menuid=57&contentid=1007.
Pipe, S. W., & Valentino, L. (2007). Optimizing outcomes for patients with severe haemophilia
A. Haemophilia, 13(s4), 1-16.
70

Quan, H., Sundararajan, V., Halfon, P., Fong, A., Burnand, B., Luthi, J. C., . . . Ghali, W. A.
(2005). Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative
data. Med Care, 43(11), 1130-1139.
Richards, M., Altisent, C., Batorova, A., Chambost, H., Dolan, G., de Moerloose, P., . . .
Rothschild, C. (2007). Should prophylaxis be used in adolescent and adult patients with severe
haemophilia? An European survey of practice and outcome data. Haemophilia, 13(5), 473-479.
doi: HAE1478 [pii]
10.1111/j.1365-2516.2007.01478.x
Romano, P. S., Roos, L. L., & Jollis, J. G. (1993). Adapting a clinical comorbidity index for use
with ICD-9-CM administrative data: differing perspectives. J Clin Epidemiol, 46(10), 1075-1079;
discussion 1081-1090.
Royal, S., Schramm, W., Berntorp, E., Giangrande, P., Gringeri, A., Ludlam, C., . . . Szucs, T.
(2002). Quality-of-life differences between prophylactic and on-demand factor replacement
therapy in European haemophilia patients. Haemophilia, 8(1), 44-50. doi: 10.1046/j.1365-
2516.2002.00581.x
Soucie, J. M., Evatt, B., & Jackson, D. (1998). Occurrence of hemophilia in the United States.
The Hemophilia Surveillance System Project Investigators. Am J Hematol, 59(4), 288-294. doi:
10.1002/(SICI)1096-8652(199812)59:4<288::AID-AJH4>3.0.CO;2-I [pii]
Stobart, K., Iorio, A., & Wu, J. K. (2006). Clotting factor concentrates given to prevent bleeding
and bleeding-related complications in people with hemophilia A or B. Cochrane Database Syst
Rev(2), CD003429. doi: 10.1002/14651858.CD003429.pub3
Szucs, T. D., Offner, A., Kroner, B., Giangrande, P., Berntorp, E., Schramm, W., & Grp, E. S. S.
(1998). Resource utilisation in haemophiliacs treated in Europe: Results from the European
Study on Socioeconomic Aspects of Haemophilia Care. Haemophilia, 4(4), 498-501.
Tagliaferri, A., Rivolta, G. F., Iorio, A., Oliovecchio, E., Mancuso, M. E., Morfini, M., . . .
Castaman, G. (2010). Mortality and causes of death in Italian persons with haemophilia, 1990-
2007. Haemophilia, 16(3), 437-446. doi: 10.1111/j.1365-2516.2009.02188.x
van Dijk, K., Fischer, K., van der Bom, J. G., Scheibel, E., Ingerslev, J., & van den Berg, H. M.
(2005). Can long-term prophylaxis for severe haemophilia be stopped in adulthood? Results
from Denmark and the Netherlands. British Journal of Haematology, 130(1), 107-112. doi:
10.1111/j.1365-2141.2005.05546.x

71

Appendix A. Identification ICD-9, HCPCS codes and brand name of clotting factors for
hemophilia A and B

ICD-9-CM
code
HCPCS code Brand Name
Hemophilia A 286.0 J7180; J7185;
J7189; J7190;
J7191; J7192;
Q2022;
Advate, Alphanate, Bioclate, Factor VIII, Factor
VIII:C,Factorate, Genarc, H.T. Factorate,
Helixate, Helixate FS, Hemofil M, Hemofil T,
Humate-P, Hyate: C, Koate, Koate-DVI, Koate-
HP, Koate-HS, Koate-HT, Kogenate, Kogenate
FS, Kogenate FS with bio-set, Kogenate FS
without bio-set, Melate, Monarc-M, Monoclate-
P, Profilate, Profilate HP,Xyntha

Hemophilia B 286.1 J7193; J7194;
J7195
Alphanine, Alphanine SD, Bebulin VH,
Benefix, Konyne 80, Konyne-HT factor IX
complex, Mononine, Profilnine, Profilnine
SD, Proflex SX-T, Proflex T, Prothar
Bypassing
agents
J7189;J7198;
Q0187; Z5230
Autoplex; Factor VIIa, FEIBA VH, NovoSeven

72

Appendix B. Final model for 6-group trajectory model
Group Parameter Estimate (SE) P
1 Intercept
Linear
-3.705 (0.167)
0.003 (0.008)
<0.0001
0.7154
2

Intercept
Linear
-1.926 (0.098)
0.016 (0.004)
<0.0001
<0.0001
3 Intercept
Linear
-1.260 (0.108)
0.049 (0.005)
<0.0001
<0.0001
4 Intercept
Linear
0.161 (0.102)
0.040 (0.005)
0.1152
<0.0001
5 Intercept
Linear
1.359 (0.154)
-0.073 (0.007)
<0.0001
<0.0001
6 Intercept
Linear
2.423 (0.136)
-0.002 (0.006)
<0.0001
0.707

73

Appendix C. Annual healthcare utilization by clotting factor use pattern (6-group model)
Overall
(n = 1035)
Clotting Factor Use Cluster
b

P-value
c

One
(n = 262)
Two
(n = 220)
Three
(n = 76)
Four
(n = 142)
Five
(n = 183)
Six
(n = 152)

n (%) n (%) n (%) n (%) n (%) n (%) n (%)
All cause ER visit

Year 1, mean [SD] 1.3 (3.7) 0.6 (1.5) 1.5 (5.0) 1.4 (2.6) 1.5 (3.0) 1.8 (5.5) 1.1 (1.8) <0.0001 *
Year 2, mean [SD] 1.3 (4.0) 0.7 (1.4) 1.5 (3.7) 1.0 (1.5) 1.8 (3.7) 2.0 (7.6) 0.9 (1.3) 0.0002 *
Year 3, mean [SD] 1.3 (5.2) 0.8 (1.4) 1.8 (5.1) 1.2 (3.3) 1.5 (3.1) 1.7 (10.1) 0.8 (1.2) 0.0100 *
All cause IP visit

Year 1, mean [SD] 0.4 (1.0) 0.2 (0.6) 0.3 (0.8) 0.5 (0.9) 0.3 (0.8) 0.5 (1.3) 0.5 (1.1) 0.0003 *
Year 2, mean [SD] 0.4 (1.0) 0.2 (0.7) 0.4 (1.1) 0.5 (0.9) 0.3 (1.0) 0.7 (1.3) 0.4 (0.7) <0.0001 *
Year 3, mean [SD] 0.4 (0.9) 0.2 (0.7) 0.3 (0.8) 0.5 (1.1) 0.4 (1.0) 0.6 (1.2) 0.4 (0.8) 0.0073 *
IP length of stay (LOS) (days)

Year 1, mean [SD] 1.5 (5.5) 1.2 (6.5) 0.9 (2.3) 2.6 (6.2) 1.0 (4.0) 2.2 (6.8) 2.3 (6.1) 0.0002 *
Year 2, mean [SD] 1.6 (4.6) 1.2 (4.8) 1.2 (3.2) 1.6 (3.9) 1.1 (3.3) 2.9 (6.3) 1.6 (4.7) <0.0001 *
Year 3, mean [SD] 1.9 (7.4) 2.0 (11.8) 1.3 (4.0) 2.9 (8.1) 1.7 (4.4) 2.3 (5.8) 1.6 (4.5) 0.0051 *
All cause OP visit

Year 1, mean [SD] 17.1 (20.0) 10.1 (10.1) 14.0 (15.6) 25.1 (38.1) 17.6 (27.7) 20.4 (13.9) 25.0 (18.2) <0.0001 *
Year 2, mean [SD] 16.8 (21.1) 11.0 (11.7) 15.4 (22.2) 19.7 (33.6) 19.3 (28.2) 20.7 (17.2) 20.6 (18.1) <0.0001 *
Year 3, mean [SD] 16.4 (23.0) 10.6 (11.5) 16.0 (27.0) 16.7 (22.4) 20.7 (31.3) 20.9 (24.2) 17.5 (18.8) <0.0001 *
Bleeding related ER/IP visit

Year 1, mean [SD] 0.6 (1.9) 0.2 (0.7) 0.7 (3.1) 0.7 (1.7) 0.8 (1.9) 0.7 (1.8) 0.4 (1.2) 0.0002 *
Year 2, mean [SD] 0.6 (1.6) 0.3 (0.8) 0.8 (2.3) 0.4 (1.0) 1.0 (2.2) 0.6 (1.4) 0.3 (0.7) 0.0002 *
Year 3, mean [SD] 0.5 (1.8) 0.3 (0.6) 0.8 (3.5) 0.5 (1.1) 0.6 (1.3) 0.4 (0.9) 0.3 (0.9) 0.0060 *
Joint related treatment

Year 1, mean [SD] 0.8 (2.7) 0.3 (0.9) 0.8 (3.4) 1.0 (3.1) 0.8 (2.0) 1.1 (2.6) 1.1 (3.9) <0.0001 *
Year 2, mean [SD] 0.8 (2.2) 0.4 (1.7) 0.9 (2.6) 0.9 (2.3) 1.1 (2.9) 0.9 (1.8) 0.7 (1.8) <0.0001 *
Year 3, mean [SD] 0.8 (2.6) 0.3 (1.2) 0.7 (3.7) 0.7 (1.8) 1.2 (2.7) 1.1 (2.9) 1.0 (2.0) <0.0001 *
Univariate comparison was made using Wilcoxon signed rank tests for number of visits or total length of stay.
74

Appendix D. Annual healthcare costs by clotting factor use pattern (6-group model)

Overall
(n = 706)
Clotting Factor Use Pattern Cluster
P-value
a

One
(n = 170)
Two
(n = 151)
Three
(n = 54)
Four
(n = 103)
Five
(n = 133)
Six
(n = 95)
n (%) n (%) n (%) n (%) n (%) n (%) n (%)
Medical costs
All cause ER visit
(excluding clotting factor)
Year 1 133 (469) 56 (146) 192 (856) 119 (183) 144 (356) 175 (374) 113 (272) 0.0004 *
Year 2 150 (642) 69 (260) 151 (454) 90 (225) 206 (472) 262 (1272) 112 (254) 0.0001 *
Year 3 135 (487) 69 (190) 210 (770) 81 (245) 132 (347) 176 (623) 113 (194) 0.0300 *
All cause OP visit
(excluding clotting factor)

Year 1 2620 (7059) 2856 (10790) 2360 (7335) 2694 (3428) 2203 (3788) 2838 (4856) 2712 (4772) <0.0001 *
Year 2 2435 (7103) 2961 (11528) 1898 (4745) 1679 (2196) 1762 (2364) 3354 (7633) 2222 (3144) <0.0001 *
Year 3 3377 (14742) 2974 (10682) 4144 (19537) 8871 (35596) 2379 (5744) 2470 (4898) 2111 (3059) <0.0001 *
All cause IP visit

Year 1 3232 (12412) 1460 (5792) 1812 (5201) 3844 (9929) 2729 (10361) 5460 (20969) 5736 (15482) 0.0527
Year 2 3250 (12079) 1944 (9007) 2488 (10084) 2980 (5991) 3514 (16692) 5039 (12064) 4160 (15803) 0.0083 *
Year 3 3513 (13287) 2935 (13421) 3299 (15773) 3349 (10729) 2963 (10371) 4358 (14179) 4391 (11682) 0.0657
Pharmacy costs
Clotting factors

Year 1 93770 (141591) 2886 (14235) 25223 (39082) 221068 (214688) 62887 (68686) 182546 (159738) 202196 (157313) <0.0001 *
Year 2 97222 (134791) 2907 (10710) 24459 (36596) 159315 (144083) 84181 (67160) 203524 (167375) 211670 (147522) <0.0001 *
Year 3 107774 (150721) 1926 (7791) 31389 (57426) 96192 (126422) 115983 (105390) 227029 (173080) 249322 (178092) <0.0001 *
Other medications

Year 1 1537 (3152) 1283 (2498) 1563 (3626) 2405 (3899) 1139 (1880) 2005 (4170) 1229 (2093) 0.0011 *
Year 2 1549 (3097) 1336 (2844) 1354 (3107) 2464 (4087) 1590 (2982) 1879 (3577) 1215 (2021) 0.0002 *
Year 3 1564 (3381) 1431 (2993) 1457 (3661) 2579 (4254) 1482 (3868) 1741 (3162) 1237 (2645) 0.0119 *
Total Medicaid Paid
Amount

Year 1 101291 (145912) 8542 (20633) 31149 (43758) 230130 (219754) 69102 (69099) 193024 (168370) 211986 (158800) <0.0001 *
Year 2 104606 (138148) 9218 (24268) 30350 (42100) 166528 (147051) 91253 (68939) 214056 (169696) 219379 (153258) <0.0001 *
Year 3 116363 (153079) 9336 (22121) 40499 (65041) 111072 (137149) 122938 (105650) 235774 (174887) 257174 (179481) <0.0001 *

Univariate comparison was made using Wilcoxon signed rank test
75

CHAPTER 4: SELF-REPORTED BARRIERS TO HEMOPHILIA
CARE IN PERSONS WITH FACTOR VIII DEFICIENCY
2

ABSTRACT
BACKGROUND: In 1975, a national network of Hemophilia Treatment Centers (HTCs)
was created to increase access to healthcare services for individuals with hemophilia.
Studies demonstrate that care in HTCs improves outcomes and reduces costs. The
objective of the study was to assess the association of demographic, insurance, and
clinical characteristics with self-reported barriers to HTC utilization.
METHODS: Data were collected from six HTCs from 2005 through 2007. Adult
participants and parents of children <18 were interviewed. Barriers were assessed by
asking whether it was difficult to obtain care in the past 12 months. Chi-square test and
logistic regression were used to assess factors associated with self-reported barriers to
care.
RESULTS: Data for 327 participants (50% adult, 64% severe hemophilia) were analyzed
in 2010-2011. Most participants/parents did not report barriers to HTC utilization.
However, 46 participants/parents (14%) reported one to six barriers, and 23 reported one
barrier. Most frequently reported barriers were “distance to the clinic” for children (44%)
and “insurance coverage” for adults (40%). Factors significantly associated with self-
reported barriers were: lower income (<$20,000) (OR=3.11, 95%CI=1.14-8.45),
difficulty finding insurance or obtaining full-year coverage (OR=5.71, 95%CI=2.63-

2
The final, definitive version of this paper has been published in the American Journal of Preventive
Medicine, Vol. 41/Issue 6 Suppl 4, December 2011 by Elsevier, Inc. All rights reserved.
76

12.41), and decreased state Medicaid coverage for low-income, nonelderly individuals
(OR=0.93, 95%CI=0.89-0.98).
CONCLUSIONS: This study indicates that while few persons with hemophilia have
barriers to care at HTCs, those with lower income, difficulty finding or maintaining
adequate insurance coverage, or living in states with lower Medicaid generosity are more
likely to report barriers. Identifying and resolving such barriers may improve care access
and patient-reported outcomes.

77

INTRODUCTION
Inherited chronic disorders that require lifelong clinical management are a public health
concern. Historically, the US healthcare delivery system was designed to provide acute,
episodic and curative care or to treat injuries and has not until recently focused on
providing long-term management for persons with chronic conditions (Grosse et al., 2009;
Wagner et al., 2001). For people with rare chronic genetic diseases, such as hemophilia
and cystic fibrosis, primary care is usually insufficient to meet their specialized needs,
and access to specialty health services is often limited or fragmented. In addition, these
individuals and their families often face physical, emotional, social and financial
challenges throughout their lives. Thus, individuals with inherited disorders require care
that is disease-specific, comprehensive and multidisciplinary, and which includes both
appropriate medical and psychosocial services. The national hemophilia comprehensive
care program in the US has been recognized as one of the most successful comprehensive
care approaches for the care of persons with inherited diseases (Hoots, 2003).
Hemophilia is a rare genetic bleeding disorder that occurs among one in 5,000 male
births and affects approximately 20,000 persons in the US, based on the expected births
and deaths since 1994 (CDC, 2012; Soucie, Evatt, & Jackson, 1998). Persons with
hemophilia either are deficient in or are missing clotting factor VIII or IX, which places
them at high risk of internal, muscular and joint bleeding as well as prolonged bleeding
following trauma or surgery. Repeated hemorrhages, especially in persons with severe
hemophilia (factor activity less than 1% of the normal level), can lead to the development
of chronic arthropathy. Overtime, this condition can cause joint pain, reduction in joint
78

range of motion, crippling musculoskeletal deformity, and disability. Treatment consists
of injecting intravenously the missing clotting factor.
The complexity of treatment and the psychosocial aspects of hemophilia make care in a
general hematology department or practice less desirable. In the US, a network of
federally funded Hemophilia Treatment Centers (HTCs) was initiated in 1975 to provide
comprehensive care for persons with congenital bleeding disorders (Hoots, 2003). Since
the 1980s, the Centers for Disease Control and Prevention (CDC) has provided additional
funding to implement prevention programs, with an emphasis on risk-reduction practices.
Over the years, the original network of 22 HTCs has expanded to more than 130 HTCs,
which are organized into 12 regional networks. Treatment centers are currently funded by
the Health Resources and Services Administration (HRSA) and by the CDC to provide
comprehensive care and preventive services to persons with hemophilia and other
bleeding disorders (Baker, Crudder, Riske, Bias, & Forsberg, 2005). A core HTC team
usually consists of a pediatric or adult hematologist who serves as medical director, a
nurse coordinator, a physical therapist, and a psychosocial professional. Additional
members may include an adult or pediatric hematologist, dentists, orthopedists, genetic
counselors, pharmacists, infectious disease specialists, social workers, and research
coordinators. HTCs also provide extensive infusion education and treatment plan
development. Many of the larger HTCs increase individuals’ access to hemophilia care
by operating satellite clinics in rural areas and by offering telephone counseling to
patients in remote or underserved areas. Many HTCs are covered entities in the federal
340B drug pricing program, allowing them to purchase clotting factor concentrates at
79

discounted prices, which generates program income that is used to maintain and expand
HTC services.
Older studies report 70 percent of persons with hemophilia in the United States receive at
least some of their medical care from one of these HTCs (Evatt, 2006; Soucie et al.,
2000). The benefits of comprehensive HTC care to reduce hospitalization and
unemployment rates, and as a consequence, lower the overall cost of hemophilia are
known (Levine, McVerry, Segelman, Cranford, & Zimbler, 1976; Smith, Keyes, &
Forman, 1982; Smith & Levine, 1984). Based on surveillance data gathered by the CDC,
persons who receive care at HTCs have a 40% deceased risk of death and a 40%
decreased hospitalization rate for bleeding complications compared with persons who
receive care from non-HTC providers, despite the fact that HTCs provide health care to a
disproportionately larger share of individuals with severe complications (Soucie et al.,
2000; Soucie et al., 2001). About 30% of all persons with hemophilia in the US do not
receive care from HTCs. However, this population is difficult to study because each non-
HTC care site has few individuals with hemophilia.
While at least 70 percent of the population receives care at HTCs, the barriers to HTC
utilization have not been formally studied. Identifying barriers to HTC utilization and
implementing strategies to increase access to these services are critical to the
improvement of outcomes and the reduction of long-term disabilities. The objective is to
identify (1) barriers that may prevent individuals from using HTC services and (2) patient
socio-demographic and clinical characteristics associated with barriers to this care.

80

METHODS
Participants and Procedures
This analysis used data from the Hemophilia Utilization Group Study part Va (HUGS
Va), a multicenter, observational study designed to examine the cost and burden of illness
in persons with hemophilia in the US. The details of study design, methodology and
baseline data have been described elsewhere (Zhou et al., 2011). Briefly, data were
collected prospectively among persons with hemophilia A who received care at one of six
HTCs located in California (2 centers), Colorado, Indiana, Massachusetts and Texas.
These sites were selected because they are geographically diverse and representative of
HTCs in the US, located in populous states thus serving relatively large hemophilia
populations allowing for adequate subject enrollment, and had research personnel willing
to participate and conduct IRB review. Participant selection was stratified by level of
factor VIII deficiency based on the CDC surveillance report to obtain a closely
representative sample of individuals with mild, moderate and severe hemophilia in each
state where the HTCs are located. Eligibility criteria included: (a) age between 2 and 64
years; (b) factor VIII level ≤30%, with or without a history of inhibitor; (c) receiving at
least 90% of hemophilia care at the participating HTC; (d) obtaining care at the HTC
within two years prior to enrollment in the study; and (e) English speaking. Individuals
were excluded from participation in the study if they were determined to be cognitively
impaired or had an additional bleeding disorder.
From 2005 through 2007, 329 individuals with hemophilia A were enrolled in the HUGS
Va study. After signing informed consent, adult participants or parents of children <18
years of age completed an initial interview. Individuals completed the survey by self-
81

administration or were interviewed by research staff, so there would be no potential bias
in response to the barriers to care questions in the survey. The interview gathered socio-
demographic information, health insurance status, perceived barriers to hemophilia care,
self-rated joint pain and motion limitation, infusion method, clotting factor utilization,
and co-morbidities. Clinical data abstracted through chart review included factor VIII
activity level, body weight and height, current and historic inhibitor levels, history of
immune tolerance therapy, hepatitis virus serology, infusion method, and treatment
regimen.

Measurement of Barriers to Care
The perception of barriers to HTC utilization among adult participants and parents seen at
HTCs was assessed by examining responses to a single question: “In the past 12 months,
has there ever been a time that you needed hemophilia care but it was difficult to get?”
Ten specific barriers and one open-ended question were assessed for study participants
who reported difficulty receiving care (Table 4.1).

82

Table 4.1 Types of barriers to hemophilia care
Types of Barriers to HTC Utilization (options listed in study questionnaire)
1. Distance to the center
2. Transportation to center
3. Insurance does not pay for comprehensive care at HTC
4. Difficulty getting off of work
5. The clinic hours were not convenient
6. You needed someone to take care of your children
7. You would lose pay from work
8. You had a conflict with the staff at HTC
9. HTC staff is not responsive or receptive to your needs
10. Language barrier
11. Other barrier, specify

Covariates
Several factors hypothesized to be associated with barriers to HTC utilization were
included in the analysis. For some variables, such as marital status, education level, and
employment status, parents’ status for participants <18 years of age were combined with
adult participants’ data. Socio-demographic characteristics included: age; education (high
school or less versus beyond high school); marital status (married or with partner versus
‘other’ status); employment (part-time or full-time employed versus unemployed);
83

income (household income <$20,000 per year versus ≥$20,000 per year); and race (white
versus non-white). To adjust for the effect of co-morbidities and health status, clinical
characteristics (such as hemophilic severity, history of inhibitors, HIV/AIDS infection,
and liver disease/hepatitis) were included in the analysis. Participants in rural areas often
face issues of geographic distance and availability of transportation when seeking health
care. Therefore, “distance to HTC”, which is the distance from the participants’ home zip
code to their regular HTC clinic and HTC outreach clinic, was also included as a variable
in the analysis.
Insurance status was assessed using both individual- and state-level data. Two individual-
level insurance variables were collected: length of insurance coverage (none or less than
12 months coverage versus full-year coverage) and difficulty finding insurance. Because
these two variables measure similar aspects of insurance problems, they were combined
into a new variable, health insurance issues in the multivariable analysis. Studies indicate
that the relative generosity of state Medicaid eligibility can influence access to health
care, not only for those eligible for Medicaid, but also for all low-income, nonelderly
adults who are affected by Medicaid coverage decisions (Weissman, Zaslavsky, Wolf, &
Ayanian, 2008). To account for state variation in Medicaid generosity, state-level
Medicaid coverage rates were examined in the states where centers are located. Using the
methodology developed by Weissman et al., the Medicaid coverage rate for each state
was calculated as the actual number of low-income nonelderly individuals (<200% of
poverty) covered by Medicaid divided by the number of all low-income nonelderly
individuals without private insurance. State data provided by the Kaiser Commission on
84

Medicaid and the Uninsured (KCMU), 2005-2006, were used in the calculation (KFF,
2012).

Statistical Analysis
The proportion of participants who reported any barrier to HTC utilization was calculated
and specific barriers were also identified. Specific barriers reported by adults and those
reported by the parents of children <18 years of age were compared. Due to the small
number of participants who reported each specific barrier, the subsequent analyses
focused on the factors associated with any barrier to HTC utilization. A series of bivariate
analyses (χ
2
or Fisher Exact test and Student t-test) was run to evaluate the association
between overall barriers and each covariate. To identify characteristics associated with
any barrier to care, a series of logistic regression models was developed. Univariate
logistic analysis was conducted to examine each independent variable separately, and
then a multivariable logistic regression containing all independent variables was
conducted to adjust for variation in characteristics among the participants. Because co-
morbidities occur predominantly in adults, the association of patient characteristics with
barriers in children and adult participants was examined separately. Co-morbidities,
including AIDS/HIV infection and liver disease, were considered in the model for adults.
Participants’ history of inhibitor development was not included in the multivariable
model because of the high association with hemophilic severity. For categorical variables,
the group with fewer barriers to HTC utilization was designated as the reference group.
The model’s overall significance was tested by likelihood ratio test (χ
2 Model
). Variables in
85

the model were checked for multicollinearity by using correlations, tolerances, and
variance inflation factors (VIF). Finally, the Hosmer-Lemeshow χ
2
test (χ
2HL
) was used to
test goodness of fit of the model to the data.
All analyses were performed in 2010-2011 using SAS statistical software version 9.2
(SAS Institute, Cary, NC).

86

RESULTS
Of 329 HUGS Va participants, 327 (99.4%) adult participants or parents of children >18
years who completed the barrier to HTC utilization questionnaire were included in the
analysis. Fifty percent (50.2%) of participants were adults, and nearly two thirds (64.2%)
had severe hemophilia. Mean age was 9.7 years for children and 33.7 years for adults. A
total of 32 (9.7%) participants received their usual care at local outreach clinics that are
affiliated with two HTCs. For these participants, the average distance to the outreach
clinic was 47.8 miles, compared to 271.2 miles if they attended the primary clinic
(P<0.0001).
Most of the adult participants or parents of children (86%) reported no barriers to HTC
utilization. However, 46 participants/parents (14%) reported at least one barrier to HTC
utilization (Table 4.2). Adults were more likely to report having barriers to care than
were the parents of children with hemophilia (P=0.03). Among those with perceived
barriers, 23 participants/parents (50%) reported only one barrier, while one participant
(2%) reported six barriers. The most frequently reported barrier to care for parents was
“distance to the clinic”, which was cited by 44% of parents who perceived barriers.
Among adult participants, “insurance coverage” was cited as a barrier to care by 40% of
adults reporting barriers. Adult participants were more likely than parents to report
“insurance coverage” as a barrier to HTC care.

87

Table 4.2 Comparisons of participant/parent reported barriers to HTC utilization
between children and adults with hemophilia A
Barrier, n (%) Total Sample Children Adults

(n=327) ( n=163) (n=164)
Overall Barriers
a
46 (14.1) 16 (9.8) 30 (18.3)*

Specific Barrier
b

(n=46) (n=16) (n=30)
Distance to the center
16 (34.8) 7 (43.8) 9 (30.0)
Clinic hours were not convenient
14 (30.4) 6 (37.5) 8 (26.7)
Insurance does not pay for comprehensive care at
HTC
13 (28.3) 1 (6.3) 12 (40.0)*
Transportation to center
11 (23.9) 5 (31.3) 6 (20.0)
You would lose pay from work
11 (23.9) 3 (18.8) 8 (26.7)
Difficulty getting off of work
8 (17.4) 2 (12.5) 6 (20.0)
You needed someone to take care of your
children
4 (8.7) 2 (12.5) 2 (6.7)
You had a conflict with the staff at HTC
4 (8.7) 2 (12.5) 2 (6.7)
HTC staff is not responsive or receptive to your
needs
4 (8.7) 2 (12.5) 2 (6.7)
Language barrier
c

- - -
Other barriers
4 (8.7) 2 (12.5) 2 (6.7)
Note: Data are presented as frequency (column percentage). Parents of children age
younger than 18 years and adults differ significantly: * P<0.05
a
Response to question “In the past 12 months, has there ever been a time that you needed
hemophilia care but it was difficult to get it?”.
b
The percentages for specific barrier are based on those who reported barrier to HTC
utilization (n=46).
c
No participants reported a language barrier.
HTC, hemophilia treatment center.

The demographic and clinical characteristics of the study population by overall barriers to
HTC utilization are summarized in Table 3.3. Compared with participants who reported
no barriers, those who reported barriers to HTC utilization were more often unemployed
(45.7% versus 29.9% among participants with no barriers, P=0.03), to be from a family
88

with income <$20,000 (35.7% versus 14.7%, P=0.0009), to have no or less than 12
months’ health insurance coverage (31.1% versus 6.2%, P<0.0001), and to have had
difficulty finding insurance (68.9% versus 21.1%, P<0.0001). We also found geographic
variations among individuals who reported barriers. Participants from some states were
more likely to report barriers to care than those from other states (P=0.005) (Table 4.3).
Distance to HTC was not significantly different among participants who reported barriers
to care compared with those with no barriers (P=0.15). Hemophilic severity also was not
associated with participant/parent report of barriers to care. Individuals with liver disease
or hepatitis more frequently reported barriers than those without these co-morbidities
(47.8% versus 31.3%, P=0.03).
Table 4.3 Distribution of characteristics of study population, overall and by self-
reported barriers to HTC utilization
Characteristics, n (%)
Total Sample
(n=327)
Barriers to Care
a

P-value
Yes (n=46) No (n=281)
Age, mean years (SD) 21.6 (15.2) 23.5 (14.1) 21.3 (15.3) 0.3582
Age group 0.0275
Child (2-17 years old) 163 (49.8) 16 (34.8) 147 (52.3)
Adult (18+ years old) 164 (50.2) 30 (65.2) 134 (47.7)
Education
b
0.9424
High school or less 99 (30.7) 14 (31.1) 85 (30.5)
Beyond high school 224 (69.3) 31 (68.9) 193 (69.4)
Marital status
b
0.7444
Married/With partner 192 (58.7) 26 (56.5) 166 (59.1)
Not married 135 (41.3) 20 (43.5) 115 (40.9)
Employment
b
0.0338
Unemployed 105 (32.1) 21 (45.7) 84 (29.9)
Full-time or part-time
employed
222 (67.9) 25 (54.3) 197 (70.1)
Income
b, c
0.0009
<$20,000 54 (17.5) 15 (35.7) 39 (14.7)
≥$20,000 254 (82.5) 27 (64.3) 227 (85.3)
Race

0.9525
White 212 (64.8) 30 (65.2) 182 (64.8)
Nonwhite 115 (35.2) 16 (34.8) 99 (35.2)
89

Note: Data are presented as frequency (column percentage).
a
Response to question “In the past 12 months, has there ever been a time that you needed
hemophilia care but it was difficult to get it?”.
b
For participants or parents of child age <18 years.
c
Data do not add up to n=327 due to non-response.
d
Distance to regular clinic or outreach if outreach clinic is available in the area.
SD, standard deviation; HTC, hemophilia treatment center; HIV/AIDS, acquired
immunodeficiency syndrome/human immunodeficiency virus.

An analysis of barriers by state was also conducted. Univariate analysis indicates that
overall rates of barriers to care by state ranged from 4% to 23% (P=0.005). Ten specific
barriers were compared by state between those who reported barriers and those who did
not report barriers. “Distance to the clinic” was reported as a barrier in participants from
4 of the 5 states analyzed (5% of all participants), with the percentage by state ranging
from 0% to 13% (P=0.01). Regarding “transportation to clinic”, three states had
participants who reported this as an issue (overall rate 3%, range from 0% to 9% by state,
P=0.001). No other specific barriers were found to vary significantly by state.
Insurance coverage
c
<0.0001
None or less than 12 months 31 (9.7) 14 (31.1) 17 (6.2)
Full year 288 (90.3) 31 (68.9) 257 (93.8)
Difficulty finding insurance
c
88 (27.9) 31 (68.9) 57 (21.1) <0.0001
Location of HTCs

0.0045
California 99 (30.3) 8 (17.4) 91 (32.4)
Colorado 61 (18.7) 14 (30.4) 47 (16.7)
Indiana 56 (17.1) 9 (19.6) 47 (16.7)
Massachusetts 53 (16.2) 2 (4.4) 51 (18.1)
Texas 58 (17.7) 13 (28.3) 45 (16.0)
Distance to HTC, mean miles
(SD)
d

45.2 (63.6) 62.8 (94.0) 42.2 (56.8) 0.1530
Hemophilia severity 0.6284
Mild/Moderate 107 (35.8) 15 (32.6) 102 (36.3)
Severe 210 (64.2) 31 (67.4) 179 (63.7)
History of inhibitors 51 (15.6) 7 (15.2) 44 (15.7) 0.9391
HIV/AIDS 44 (13.5) 9 (19.6) 35 (12.5)

0.19
02
Liver disease/Hepatitis 110 (33.6) 22 (47.8) 88 (31.3) 0.0280
90

Table 4.4 illustrates the multivariable logistic regression analyses of the association
between participant socio-demographic and clinical characteristics and having any barrier
to HTC utilization. Only individuals with complete data for all variables in the model
were included in the analyses (n=302). Among the entire sample, individuals with
household income <$20,000 were more likely to report a barrier compared with those
with higher household income (odds ratio [OR] =3.11, 95% confidence interval [CI]
=1.14-8.45). Compared to those without health insurance issues (no difficulties obtaining
insurance and insurance coverage for an entire year), those individuals with insurance
issues (less than full year coverage and/or difficulty finding insurance) had a 470%
higher risk of reporting a barrier to care (OR=5.71, 95%CI=2.63-12.41). Participants
attending HTCs in states with lower Medicaid coverage rates for the low-income
nonelderly were more also likely to report a barrier to HTC care. Each percentage-point
increase in the Medicaid coverage rate for the low income non-elderly resulted in a 7%
decrease in the risk of reporting any barrier to care (OR=0.93, 95%CI=0.89-0.98).
Additional analyses were conducted among groups of children and adult participants
separately. Although sample size was reduced in these sub-analyses, the results indicate
that the generosity of state Medicaid programs was significantly associated with barriers
to HTC utilization for children (OR=0.88, 95%CI=0.80-0.97). Adults with health
insurance issues (less than full year coverage and/or difficulty finding insurance) were
even more likely to report barriers to HTC utilization (OR=10.38, 95%CI=3.36-32.05)
compared with the overall sample (OR=5.71).

91

Table 4.4 Multivariable logistic regression analysis of participant characteristics
associated with barriers to HTC utilization

Model 1
Overall (n=302)
Model 2
Children
(n=149)
Model 3
Adults (n=153)
Odds Ratio
(95% CI)
Odds Ratio
(95% CI)
Odds Ratio
(95% CI)
Age group
Adult (18+ years old)

1.59 (0.64-3.95)

-

-
Marital status
Married/with partner

1.41 (0.59-3.38)

0.82 (0.14-4.69)

1.93 (0.66-5.65)
Employment
Unemployed

1.15 (0.48-2.76)

1.66 (0.41-6.78)

0.85 (0.25-2.96)
Income
<$20,000

3.11 (1.14-8.45)

7.55 (0.99-57.77)

2.82 (0.76-10.51)
Race
White

0.92 (0.47-2.76)

2.10 (0.46-9.55)

0.68 (0.23-2.01)
Insurance issues
a

Yes
5.71 (2.63-12.41)

2.19 (0.59-8.08) 10.38 (3.36-32.05)
State Medicaid coverage, in
percentage
b

0.93 (0.89-0.98) 0.88 (0.80-0.97) 0.95 (0.89-1.00)
Distance to HTC, in miles
c
1.00 (1.00-1.01) 0.99 (0.98-1.01) 1.00 (1.00-1.01)
Hemophilia severity
Severe

1.27 (0.57-2.88)

0.82 (0.21-3.17)

1.80 (0.55-5.89)
Liver disease/Hepatitis
Yes

-

-

1.54 (0.47-4.97)
HIV/AIDS
Yes

-

-

0.62 (0.17-2.24)
Note: For categorical variables, the group theorized to have fewer barriers to HTC
utilization was used as the reference group.
Model 1, χ
2
model

=53.70, P<0.0001, χ
2
HL
=7.95, P=0.44.
Model 2, χ
2
model

=17.41, P<0.05, χ
2
HL
=3.56, P=0.89.
Model 3, χ
2
model

=41.27, P<0.0001, χ
2
HL
=5.25, P=0.73.
a
Coverage less than 12 months and/or difficulty finding insurance.
b
‘State Medicaid Coverage’ refers to the percentage of low-income nonelderly covered
by Medicaid as a percent of all low-income non-elderly without private insurance in
2005-2006. California=47.2%; Colorado=33.6%; Indiana =50.4%; Massachusetts=59.5%;
Texas=34.6%.
c
Distance to regular clinic or outreach if outreach clinic is available in the area.
CI, confidence interval; HTC, hemophilia treatment center; HIV/AIDS, acquired
immunodeficiency syndrome/human immunodeficiency virus.

92

DISCUSSION
To our knowledge, this analysis is the first prospective cohort study to examine in a
comprehensive manner the barriers to HTC utilization among persons with hemophilia A
and individual characteristics associated with those barriers. It is important to note that
although hemophilia is a chronic disease associated with significant physical, social and
financial consequences, only one of seven HTC participants or parents who participated
in this study reported barriers to hemophilia care. Some of the barriers reported by
individuals included those possibly addressed by the individual centers (e.g. clinic hours,
child care) and this feedback was provided. However, other participant characteristics
such as income, insurance issues (lack of full year insurance coverage and/or difficulty
finding insurance), and the generosity of state Medicaid eligibility requirements found to
be significantly associated with barriers to HTC utilization and are policy issues that need
to be addressed more globally.
In the preliminary analysis, it was found that participants from some states were more
likely to report barriers to care than those from other states. We further explored this
geographic variation in terms of generosity of state Medicaid eligibility criteria, distance
to HTC, and provision of emergency care. Previous research has demonstrated that the
generosity of a state’s Medicaid eligibility criteria has an impact on access to health care
for all low-income nonelderly adults who are affected by Medicaid coverage decisions
(Weissman et al., 2008). Medicaid coverage rates for the low-income nonelderly were
used as an indicator of Medicaid generosity and it was found that a 1% increase in the
Medicaid coverage rate results in a 7% decrease in the likelihood of a participant
reporting a barrier to care. Distance to clinic, particularly for those who live in rural areas,
93

can be a barrier to obtaining care, and is not easily overcome due to time and
transportation difficulties involved. However, in the multivariable analysis, measured
distance to HTC (regular clinic or outreach) was not significantly associated with barriers
to HTC utilization after adjusting for other variables. One explanation is that two HTCs
in the study sample provide outreach clinics to individuals who reside in rural areas. The
availability of outreach clinics can substantially reduce the travel distance. This issue
needs further evaluation. Another potential source of differing barriers to care may be
access to emergency care that could differ from center to center. It was found that all
HTC sites in this study shared a common procedure for providing emergency care or
recommending a same-day appointment and all had a 24-hour call number. Thus it did
not appear that the emergency care provided by the various HTCs differed in a substantial
way. Further research using a larger sample size is needed to evaluate geographic
variations among individuals with hemophilia.
The data have limitations, and the study results must be evaluated with these limitations
in mind. Because all participants in the study received care at HTCs, the study results are
only applicable to individuals receiving care from HTCs, and cannot be generalized to all
individuals with hemophilia. Older studies have reported that around 30% of persons
with hemophilia do not receive any hemophilia care from HTCs and they are more
frequently individuals with mild disease or fewer complications (Soucie et al., 2000).
However, no update to this number has been published in the last 12 years. Still, the
overall rate of reported barriers to care in this study may underestimate barriers to care
for the entire hemophilia population due to the exclusion of non-HTC treated individuals.
Although including the non-HTC treated group would be informative, this population is
94

also difficult to identify as the number of individuals seen at non-HTC healthcare sites
are small. Future studies should address the barriers to care faced by the non-HTC
population. Similarly, because only English-speaking participants were included in this
study, these results cannot be generalized to the non-English-speaking population.
Additionally, the study’s small sample size (a common limitation when studying rare
conditions) may result in biased statistical inference. This also makes it difficult to
explore the statistical association of specific barriers with socio-demographic, clinical
and state-specific characteristics. Lastly, the study included 20 households with more
than one child participant. Although these siblings are individuals with unique ages and
clinical characteristics, they shared the same socio-demographics as a household and
most of them (18 households) reported the same level of barriers. After excluding 20
children with the same characteristics as their included siblings, the results of
multivariable analyses remained the same. Therefore, the results from the entire study
population are reported in this manuscript.

CONCLUSION
In conclusion, this study indicates that lower household income, insurance difficulties,
and residence in states with lower Medicaid program generosity are associated with self-
reported barriers to HTC utilization. The availability of outreach clinics that serve rural
communities may reduce the reported barriers related to distance to HTC care, but more
study is needed. Identification of barriers to hemophilia care is the first step in identifying
policies that will increase access to care, potentially improving patient outcomes.
Interventions that promote adequate insurance opportunities for individuals with
95

hemophilia should be a high priority. The Affordable Care Act legislation may resolve
some insurance barriers (eliminating lifetime caps, expanding insurance coverage,
reducing annual spending limits, eliminating preexisting exclusions, and expanding
eligibility for Medicaid) (Johnson & Zhou, 2011). It also increases payments to providers
in rural areas and provides coverage for preventive services (Johnson & Zhou, 2011).
Because of the wide range of financial and professional resources potentially available to
HTCs, these providers are in a unique position to assist persons with hemophilia who
experience barriers to accessing adequate healthcare.

96

CHAPTER REFERENCES
Baker, J. R., Crudder, S. O., Riske, B., Bias, V., & Forsberg, A. (2005). A model for a
regional system of care to promote the health and well-being of people with rare chronic
genetic disorders. Am J Public Health, 95(11), 1910-1916. doi:
10.2105/AJPH.2004.051318
CDC. (2012). The Centers for Disease Control and Prevention: Hemophilia Data and
Statistics. http://www.cdc.gov/ncbddd/hemophilia/data.html.
Evatt, B. L. (2006). The natural evolution of haemophilia care: developing and sustaining
comprehensive care globally. Haemophilia, 12 Suppl 3, 13-21. doi: 10.1111/j.1365-
2516.2006.01256.x
Grosse, S. D., Schechter, M. S., Kulkarni, R., Lloyd-Puryear, M. A., Strickland, B., &
Trevathan, E. (2009). Models of comprehensive multidisciplinary care for individuals in
the United States with genetic disorders. Pediatrics, 123(1), 407-412. doi: 123/1/407 [pii]
10.1542/peds.2007-2875
Hoots, W. K. (2003). Comprehensive care for hemophilia and related inherited bleeding
disorders: why it matters. Curr Hematol Rep, 2(5), 395-401.
Johnson, K. A., & Zhou, Z. Y. (2011). Costs of care in hemophilia and possible
implications of health care reform. Hematology Am Soc Hematol Educ Program, 2011,
413-418. doi: 10.1182/asheducation-2011.1.413
KFF. (2012). Kaiser Commission on Medicaid and the Uninsured. Health insurance
coverage in America: 2006 data update. Table 23 Health Insurance Coverage of Low-
Income Nonelderly Adults by State, 2005–2006. Available at:
www.kff.org/uninsured/upload/2006_DATA%20_UPDATE.pdf.
Levine, P. H., McVerry, B. A., Segelman, A. E., Cranford, C. M., & Zimbler, S. (1976).
Comprehensive health care clinic for hemophiliacs. Arch Intern Med, 136(7), 792-794.
Smith, P. S., Keyes, N. C., & Forman, E. N. (1982). Socioeconomic evaluation of a state-
funded comprehensive hemophilia-care program. N Engl J Med, 306(10), 575-579.
Smith, P. S., & Levine, P. H. (1984). The benefits of comprehensive care of hemophilia:
a five-year study of outcomes. Am J Public Health, 74(6), 616-617.
Soucie, J. M., Evatt, B., & Jackson, D. (1998). Occurrence of hemophilia in the United
States. The Hemophilia Surveillance System Project Investigators. Am J Hematol, 59(4),
288-294. doi: 10.1002/(SICI)1096-8652(199812)59:4<288::AID-AJH4>3.0.CO;2-I [pii]
Soucie, J. M., Nuss, R., Evatt, B., Abdelhak, A., Cowan, L., Hill, H., . . . Wilber, N.
(2000). Mortality among males with hemophilia: relations with source of medical care.
The Hemophilia Surveillance System Project Investigators. Blood, 96(2), 437-442.
97

Soucie, J. M., Symons, J. t., Evatt, B., Brettler, D., Huszti, H., Linden, J., & Hemophilia
Surveillance System Project, I. (2001). Home-based factor infusion therapy and
hospitalization for bleeding complications among males with haemophilia. Haemophilia,
7(2), 198-206.
Wagner, E. H., Austin, B. T., Davis, C., Hindmarsh, M., Schaefer, J., & Bonomi, A.
(2001). Improving chronic illness care: translating evidence into action. Health Aff
(Millwood), 20(6), 64-78.
Weissman, J. S., Zaslavsky, A. M., Wolf, R. E., & Ayanian, J. Z. (2008). State Medicaid
coverage and access to care for low-income adults. J Health Care Poor Underserved,
19(1), 307-319. doi: S1548686908103072 [pii]10.1353/hpu.2008.0021
Zhou, Z. Y., Wu, J., Baker, J., Curtis, R., Forsberg, A., Huszti, H., . . . Johnson, K. (2011).
Haemophilia utilization group study - Part Va (HUGS Va): design, methods and baseline
data. Haemophilia, 17(5), 729-736. doi: 10.1111/j.1365-2516.2011.02595.x

Abstract (if available)

Abstract With the increasing availability of clotting factor and development of comprehensive hemophilia treatment centers (HTCs) in the US, hemophilia care has improved dramatically during the past four decades. Nowadays children with hemophilia in developed countries look forward to a normal life span and high quality of life without crippling joint diseases. Other the other hand, long‐life treatment with the high cost of care including clotting factor concentrates places a considerable burden on patients, healthcare systems and society. Patients and providers face a number of challenges, which may affect patients' access to clotting factor or hemophilia care at HTCs and lead to undesired outcomes. Development of antibodies (i.e., inhibitors) to clotting factor concentrates still represents a significant risk for patients with hemophilia, which requires even higher healthcare resource utilization and staggering high costs. ❧ This three‐paper dissertation aims to address some of the emerging economic and public health issues in hemophilia care. In Chapter 2, an economic evaluation was conducted and a cost‐minimization model was built to compare two bypassing agents, activated prothrombin complex concentrates (aPCC) versus recombinant factor VIIa (rFVIIa) as first‐line drugs in the home treatment of mild‐to‐moderate bleeding episodes in hemophilia patients with inhibitors from a US third party payer's perspective. The model used clinically based assumptions and investigated model parameters with extensive sensitivity analyses. In Chapter 3, we explored the longitudinal clotting factor use patterns for patients with hemophilia using 6 state Medicaid databases. The aim of the study was to classify and characterize clotting factor use patterns using group‐based trajectory models (GBTMs) and to assess the economic outcomes associated with clotting factor use patterns. Finally, in Chapter 4, using data collected from six HTCs in the US, we identified self‐reported barriers that might prevent individuals from using HTC services and studied patient sociodemographic and clinical characteristics associated with barriers.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The impact of treatment decisions and adherence on outcomes in small hereditary disease populations

PDF

Burden of illness in hemophilia A: taking the patient’s perspective

PDF

Selected papers on the evaluation of healthcare costs of prematurity and necrotizing enterocolitis using large retrospective databases

PDF

Impact of pharmacy-based transitional care on healthcare utilization and costs

PDF

Delivering better care for children with special health care needs: analyses of patient-centered medical home and types of insurance

PDF

Effects of a formulary expansion on the use of atypical antipsychotics and health care services by patients with schizophrenia in the California Medicaid Program

PDF

Advances and applications for economic evaluation methods in health technology assessment (HTA)

PDF

Assessing value defects in limb preservation care

PDF

US biosimilar market entry and uptake: US commercial payer preferences in access decision making

PDF

Discriminating changes in health using patient-reported outcomes

PDF

The value of novel antihyperlipidemic treatments in the U.S. healthcare system: Reducing the burden of cardiovascular diseases and filling the gap of low adherence in statins

PDF

Understanding primary nonadherence to medications and its associated healthcare outcomes: a retrospective analysis of electronic medical records in an integrated healthcare setting

PDF

Economic aspects of obesity

PDF

Essays in health economics

PDF

The causal-effect of childhood obesity on asthma in young and adolescent children

PDF

Physician behavior and low-value care

PDF

An evaluation of risk predictors for disease progression in Veteran Affairs patients with chronic hepatitis C

PDF

Essays on the economics of infectious diseases

PDF

Three essays on comparative economic and clinical outcomes of antipsychotics using retrospective claims data

PDF

Impact of medication persistence on clinical and health care cost outcomes in patients with major depressive disorder using retrospective claims data

Asset Metadata

Creator Zhou, Zheng-Yi (author)

Core Title Three essays on emerging issues in hemophilia care

School School of Pharmacy

Degree Doctor of Philosophy

Degree Program Pharmaceutical Economics and Policy

Publication Date 07/15/2014

Defense Date 06/13/2014

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

Tag access to care,burden of illness,economic evaluation,hemophilia,inhibitor,OAI-PMH Harvest,treatment pattern

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Nichol, Michael B. (committee chair), Doctor, Jason N. (committee member), Ingles, Sue Ann (committee member)

Creator Email zhouzhyi@gmail.com,zhouzhyi@hotmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-441318

Unique identifier UC11287700

Identifier etd-ZhouZhengY-2691.pdf (filename),usctheses-c3-441318 (legacy record id)

Legacy Identifier etd-ZhouZhengY-2691.pdf

Dmrecord 441318

Document Type Dissertation

Format application/pdf (imt)

Rights Zhou, Zheng-Yi

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 a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA