Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Assessing value defects in limb preservation care
(USC Thesis Other)
Assessing value defects in limb preservation care
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
ASSESSING VALUE DEFECTS IN LIMB PRESERVATION CARE
by
Hanke Zheng, MS
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
(HEALTH ECONOMICS)
August 2024
Copyright 2024 Hanke Zheng
ii
Acknowledgements
This dissertation is dedicated for my lovely family, particularly my mother and father,
who have been giving me the unconditional love and support on anything that I feel interested in
since the day I was born. Without your unrelenting support, I could not have had the confidence
to start a life in a new country to pursue my passion and become who I am today. Thank you,
mom and dad.
Thank you to my partner, Ally, for your company and everything you have done to make
me feel supported and comforted during the past four years of stress in my PhD journey. Thank
you for always being there for me.
I extend my heartfelt gratitude to my PhD advisor and committee chair, Dr. William
Padula, who has been so supportive and attentive not just in the research front but also in my
professional and personal growth. Your mentorship is invaluable whenever I needed it.
I am thankful for my committee members and professors in the Department of Health and
Pharmaceutical Economics at USC. Special thanks to Dr. Seth Seabury, Dr. John Romley, Dr. J,
Dr. Tze-Woei Tan, and Dr. David Armstrong for providing the constructive advice in my
learning trajectory and research works.
Lastly, I appreciate the funding support that I received from the USC Mann School of
Pharmacy and Pharmaceutical Sciences, and Bristol Myers Squibb Fellowship throughout my
doctoral training.
iii
Abstract
Value defects in health care are the suboptimal treatment behaviors that needlessly reduce
quality of care (clinical effectiveness, patient experience, quality of life, etc.) or lead to
unnecessary financial waste to the healthcare system or society. Due to systematic inefficiency in
quality care delivery and resources allocation, these defects in value prevail across various
disease areas, especially in the field of chronic lower extremity ulcers. Diabetic foot ulcer (DFU)
and venous leg ulcer (VLU) are two leading chronic lower extremity ulcerations in the U.S. that
affect millions of people and result in significant financial toll through direct medical costs and
impaired productivity. The healthcare system has spent extra amount of money on downstream
costs for no-value or low-value care, such as lower extremity amputation (LEA) for one with
chronic lower extremity ulcers, which could have been averted by appropriate and timely
management. However, challenges exit for the key stakeholders to invest on these high value
care as they may cost up front. Therefore, to eliminate the defects in value, it is of great
important to inform healthcare decisions based on value and call for key stakeholders and policy
makers to move towards the care based on value. Leveraging an innovative Value Defect
Framework on a cost-effectiveness scale, this dissertation attempts to comprehensively identify
the unmet need and assess the value of limb preservation strategies in the field of chronic lower
extremity ulcers, focusing on DFU and VLU.
The dissertation work consists of three research studies. The first study revealed the
underutilization of guideline-consistent limb salvage care before above-ankle LEA of DFU
patients, and the variation in the utilization by patients’ social determinants of health and their
clinical characteristics. It highlighted the unmet need and underscored the importance for
decision makers to continue making efforts to facilitate the use of guideline-consistent vascular
iv
care, particularly for the disadvantaged population. The second study compared the clinical
effectiveness of two mainstay revascularization strategies in DFU patients at high risk of aboveankle LEA, using propensity score approach to control for the observable confounding. The
findings supplemented the two recent randomized trials with opposite conclusions by
incorporating the key endpoints from both, thus helping inform the decision making of
clinicians, payers, and policy makers from the real-world evidence perspective. The third study
was an economic evaluation that shed light on the cost-effectiveness of early intervention of
endovenous therapy for Medicare beneficiaries with VLU. It demonstrated the value of timely
intervention of appropriate vascular intervention, such as endovenous therapy, to prevent
reduced quality of life and the downstream healthcare resources utilization due to disease
recurrence and escalation from the Medicare perspective.
v
Table of Contents
Acknowledgements-------------------------------------------------------------------------------------------ii
Abstract -------------------------------------------------------------------------------------------------------iii
List of Tables-------------------------------------------------------------------------------------------------- 1
List of Figures------------------------------------------------------------------------------------------------- 2
Chapter 1: Introduction -------------------------------------------------------------------------------------- 3
1.1 Overview of Chronic Leg Ulcers-------------------------------------------------------------------- 4
1.2 Defects of Value in Limb Preservation Care ------------------------------------------------------ 6
1.3 Difficulty in Supporting Evidence-Based Share Decision Making----------------------------- 7
1.4 Supply-Sensitive Care and Overuse by Providers ------------------------------------------------ 9
1.5 Research Aims--------------------------------------------------------------------------------------- 11
Chapter 2: Variability in Utilization of Guideline-Consistent Limb Salvage Care Before Lower
Extremity Amputation in Patients with Diabetic Foot Ulcers ---------------------------------------- 14
2.1 Introduction ------------------------------------------------------------------------------------------ 15
2.2 Methods----------------------------------------------------------------------------------------------- 17
2.2.1 Data Source and Patient Selection ----------------------------------------------------------- 17
2.2.2 Measures and Covariates---------------------------------------------------------------------- 18
2.2.3 Outcomes---------------------------------------------------------------------------------------- 18
2.2.3 Statistical Analysis----------------------------------------------------------------------------- 18
2.3 Results ------------------------------------------------------------------------------------------------ 20
2.3.1 Patient Characteristics------------------------------------------------------------------------- 20
2.3.2 Utilization of Pre-Amputation Vascular Assessment and Revascularization ---------- 20
2.3.3 Factors Predicting Pre-Amputation Utilization of Vascular testing--------------------- 21
2.3.4 Factors Predicting Pre-Amputation Utilization of Revascularization------------------- 22
2.4 Discussion-------------------------------------------------------------------------------------------- 22
Chapter 3: Assessing the Value Defects in Initial Revascularization Strategy for Patients with
Diabetic Foot Ulcer in the United States: A Propensity Score Approach --------------------------- 37
3.1 Introduction ------------------------------------------------------------------------------------------ 38
3.2 Methods----------------------------------------------------------------------------------------------- 41
3.2.1 Study Design------------------------------------------------------------------------------------ 41
3.2.2 Patient Selection-------------------------------------------------------------------------------- 41
3.2.3 Predictors---------------------------------------------------------------------------------------- 42
3.2.4 Main Outcome Measures --------------------------------------------------------------------- 42
3.2.5 Statistical Methods----------------------------------------------------------------------------- 43
3.3 Results ------------------------------------------------------------------------------------------------ 48
3.3.1 Patient Sample and Characteristics ---------------------------------------------------------- 48
3.3.3 Major LEA and major LEA-Free Survival ------------------------------------------------- 49
3.3.4 Reintervention and MALE-------------------------------------------------------------------- 50
vi
3.3.5 Subgroup analysis------------------------------------------------------------------------------ 51
3.4 Discussion-------------------------------------------------------------------------------------------- 52
Chapter 4: Cost-Effectiveness of Early Endovenous Ablation with Compression Therapy in
Venous Leg Ulcerations for A Medicare Population -------------------------------------------------- 70
4.1 Introduction ------------------------------------------------------------------------------------------ 71
4.2 Methods----------------------------------------------------------------------------------------------- 72
4.2.1 Model Overview ------------------------------------------------------------------------------- 72
4.2.2 Probabilities------------------------------------------------------------------------------------- 73
4.2.4 Health Utilities --------------------------------------------------------------------------------- 76
4.2.5 Sensitivity Analyses --------------------------------------------------------------------------- 76
4.2.6 Budget Impact Analysis----------------------------------------------------------------------- 77
4.2.7 Statistical Analysis----------------------------------------------------------------------------- 77
4.3 Results ------------------------------------------------------------------------------------------------ 78
4.3.1 Sensitivity Analyses --------------------------------------------------------------------------- 78
4.3.2 Budget Impact Analysis----------------------------------------------------------------------- 79
4.4 Discussion-------------------------------------------------------------------------------------------- 79
4.5 Limitations ------------------------------------------------------------------------------------------- 81
4.6 Conclusions ------------------------------------------------------------------------------------------ 82
Chapter 5: Conclusions ------------------------------------------------------------------------------------ 89
5.1 Aims and Hypothesis Revisited ------------------------------------------------------------------- 90
5.2 Final Remarks --------------------------------------------------------------------------------------- 92
References --------------------------------------------------------------------------------------------------- 96
Chapter 1 references------------------------------------------------------------------------------------- 96
Chapter 2 references------------------------------------------------------------------------------------- 98
Chapter 3 references------------------------------------------------------------------------------------ 100
Chapter 4 references------------------------------------------------------------------------------------ 103
Chapter 5 references------------------------------------------------------------------------------------ 106
1
List of Tables
Table 2.1 Patient Baseline Characteristics (N=13,711)................................................................ 31
Table 2.2 Multivariate Logistic Regression of Utilization of Vascular Testing........................... 33
Table 2.3 Multivariate Logistic Regression of Utilization of Revascularization ......................... 35
Table 3.1 Baseline Characteristics by Initial Revascularization Strategy .................................... 63
Table 3.2 Multivariate Analyses of All-Cause Mortality, Major LEA, and Major LEA-Free
survival.......................................................................................................................................... 65
Table 3.3 Multivariate Analyses of Reintervention, and MALE.................................................. 67
Table 3.4 Subgroup Analysis........................................................................................................ 69
Table 4.1 Transition Probability Base Case Inputs and Range for Sensitivity Analyses ............. 83
Table 4.2 Direct Costs of Venous Leg Ulceration Treatment ...................................................... 84
Table 4.3 Base Case Results......................................................................................................... 86
2
List of Figures
Figure 1.1 Domains of Defects in Value in Chronic Lower Extremity Ulcers............................. 13
Figure 2.1 Patient Selection .......................................................................................................... 28
Figure 2.2 Vascular Testing Use in the Year before Major LEA ................................................. 29
Figure 2.3 Revascularization Use in the Year before Major LEA................................................ 30
Figure 3.1 Patient selection........................................................................................................... 59
Figure 3.2 Covariate Balance After Inverse Probability Treatment Weighting ........................... 60
Figure 3.3 Kaplan-Meier curves showing time to primary endpoints after EDVT or open bypass
surgery........................................................................................................................................... 61
Figure 3.4 Kaplan-Meier curves showing time to secondary endpoints after EDVT or open
bypass surgery............................................................................................................................... 62
Figure 4.1 Markov Model Simulating Disease Progression of Venous Leg Ulcers..................... 87
Figure 4.2 Cost-Effectiveness Acceptability Curve of Compression Therapy with Early vs.
Deferred Ablation ......................................................................................................................... 88
3
Chapter 1: Introduction
Hanke Zheng
4
1.1 Overview of Chronic Leg Ulcers
Chronic leg ulcer (CLU) is a form of chronic wound that does not heal in a timely manner,
leading to reduced quality of life, function loss, and severe morbidity. CLU is attributable to
infective disease, venous insufficiency, arterial disease, neuropathy, and other sorts of causes.
CLU commonly occurs in adult patients with diabetes or vascular disease, with higher prevalent
in the elderly.1 The estimated prevalence is over 15% among older adults in the U.S., with 2-3
million people newly diagnosed with CLU annually.2 The dissertation specifically focuses on
DFU and VLU, the two most prevalent forms of CLU in the U.S.3
DFU is a severe skin complication of diabetes mellitus, typically resulting from improper
glycemic control, poor circulation, underlying neuropathy, and improper foot care. It is known as
the leading risk factor preceding 80% of non-traumatic lower extremality amputations (LEA),
which is associated with greater morbidity and mortality.4 The 5-year mortality of patients with
DFU is estimated to be over 25%, and it can reach to more than 50% for patients who undergo
above-ankle (major) LEA.5 To prevent DFU patients from severe infection and LEA, the
International Working Group on Diabetic Foot (IWGDF) 2019 guidelines have recommended a
set of limb preservation strategies to consider in clinical practice.6 First, it is recommended to
perform vascular testing on DFU patients with diagnosis or sign of peripheral arterial disease
(PAD), who are considered high risk for major LEA. Common vascular imaging procedures
include duplex ultrasound, ankle systolic pressure and systolic ankle brachial index (ABI),
ultrasonography, and others. Vascular testing enables providers to examine one’s clinical
presentation in details and determine subsequent treatments suitable for the patient.
Revascularization is deemed as a critical component in the treatment landscape of DFU and
relevant vascular conditions such as PAD and chronic limb-threatening ischemia (CLTI), and
5
thus should be considered for high-risk DFU patients according to the guidelines.6
Revascularization is a type of medical procedure that aims to restore the blood flow of the
affected area of DFU, which may not only improve symptoms but also reduce the likelihood of
adverse limb outcomes such as major LEA.7 Open bypass surgery and endovascular therapy
(EDVT) are the two mainstay revascularization techniques, but it has not been established
whether one technique is superior to the other. Multiple individual factors should be taken into
consideration in defining the appropriate revascularization strategy, including but not limited to
patient risk, symptoms, disease severity, patient preference, and others.6,8
VLU is the most common cause of lower extremity ulceration and are characterized by slow
healing trajectory and frequent recurrence. VLU leads to significant disability, reduced quality of
life, and tremendous economic burden.9,10 The U.S. prevalence of VLU ranges from 0.15% to
0.3%, equating to approximately 600,000 cases per year, and is higher among women and the
elderly.11 Although VLU does not directly influence the mortality, it can lead to severe infections
such as cellulitis that significantly reduce one’s quality of life. Compression therapy and wound
management is the standard of care of VLU. There are more invasive techniques intending to
remove incompetent veins for advanced cases where the wounds are persistent and non-healing,
including open vascular surgery, radiofrequency, and endovenous procedures. A randomized
trial has showed VLU patients who received timely intervention of these more advanced
procedures presented a higher healing rate and lower recurrent rate.12
CLUs pose major challenges to the healthcare system and society by consuming a great deal
of healthcare resources, resulting in significant downstream medical costs and productivity loss.
The estimated Medicare spending on different type of CLU ranged from $28.1 to $96.8 billion
per year.13 DFU alone represents a significant economic burden on society, leading to up to $79
6
billion direct medical costs in the U.S. per year.5 The estimated annual payer burden of VLU
treatment exceeds $14.9 billion.14
1.2 Defects of Value in Limb Preservation Care
Defects in value are not uncommon across various disease areas, which may be the root
causes of the eye-popping healthcare expenditures in the U.S. According to valid calculations,
one third of the healthcare costs is attributable to these defects.15 Due to systematic inefficiency
in delivery of quality care and resources allocation, the healthcare system spends excessive
amount of money on the suboptimal treatment behaviors that needlessly reduce the quality of
care or result in unnecessary financial waste to the healthcare system or society.15
Building on the existing Value Defects Framework in surgical care developed by Pronovost
and colleagues (2021), we expanded the framework into the context of CLU care on a costeffectiveness scale, relative to no-value, low-value, or high-value care (Figure 1.1). On top of the
existing value defects, we introduced ‘preventable major LEA and ‘preventable reintervention’
representing no-value and low-value care in DFU management. Major LEA is a form of no-value
procedure that does not only generate greater downstream costs but also undermine one’s quality
of life and even increase mortality risk. However, such suboptimal treatment procedure is
preventable by early diagnosis and appropriate wound management in a timely manner. ‘Missed
vascular testing’ and ‘open bypass vs. EDVT as the initial revascularization’ are two value
defects set to be quantified in the dissertation. Although it costs more upfront, running vascular
testing helps physicians to carefully gauge DFU patients’ clinical situation and prescribe an
optimal treatment and management strategy to prevent adverse outcomes, not to mention the
subsequent care and healthcare resources needed to treat these events. Open bypass and EDVT
7
are two mainstay revascularization procedures to treat DFU patients at risk of amputation. The
cost of open bypass surgery is knowingly greater than the EDVT, but it remains unknown as to
whether it would generate extra clinical benefits than EDVT as the initial revascularization and
whether these benefits are worth the additional costs.16
As we have discussed in the previous section, the financial tolls resulting from the value
defects in managing DFU and VLU are tremendous. These defects represent huge opportunity
costs in the field of CLU care that could have been invested to deliver clinical benefit to other
patients.17 Therefore, it is imperative to eliminate the defects in value and move towards highvalue preventive care that ultimately benefits patients and healthcare system in the long run.
However, the complexity of the healthcare delivery mechanism and payment system has created
challenges for key stakeholders to prioritize appropriate treatment decisions for individual
patients with DFU or VLU.
1.3 Difficulty in Supporting Evidence-Based Share Decision Making
Recent discussions in the field have revealed the hurdles in evidence-based shared
decision making that facilitates the delivery of high quality and high value care.18 This issue
persists in many disease areas and is highly relevant to our topic in limb preservation care for
DFU and VLU as well.
First, there is a lack of transparency in scientific research and low-quality evidence
synthesis. Solid evidence has yet been established to inform the choice of initial
revascularization strategy, as to which procedure benefits which patient more. Although two
major clinical trials compared the outcomes of CLTI patients receiving EDVT or open bypass
surgery, the conclusions varied between the two trials due to selection of primary endpoints,
8
difference in underlying population, and study design.19,20 Both did not report the detailed limb
status and anatomical patterns of the enrolled patients, making it difficult to interpret the
differences between trials. The majority of existing real-world evidence is of low quality, limited
by small sample size, outdated data sources, insufficient control of confounding, and limited
generalizability, etc.21-23 Plus, these studies mainly examined the broader population with PAD
or CLTI and did not provide the data on effectiveness of revascularization procedures in the
important DFU sub-population. Taken together, clinicians and patients are challenged to make a
fully informed decision due to lack of high-quality evidence.
Second, despite the availability of evidence, some remain inaccessible to physicians and
patients due to inefficient dissemination, hindering the involvement of all parties in the decisionmaking process to sustain patient-centered care.18 For example, the education resources may be
limited for physicians practicing in rural areas to stay up to date about the recent evidence and
guideline recommendations. Some DFU patients, especially those in disadvantaged socioeconomic status, have limited access to the information about importance of preventive care and
treatment options they have. Many patients suffering from CLU have trouble in navigating what
to expect about the disease progression, and who to turn to for help.24 The inaccessibility of
critical information and evidence is a recognized contributor to defects in value, leading to delay
in diagnosis and missed opportunity for timely wound management.18
Third, many have criticized the current performance measurement of wound care services
for failing to capture patient preferences, which are central to patient-centered care.18 Patients
with chronic wounds often have comorbid systemic conditions, making wound care an
interdisciplinary effort that requires a wide range of considerations throughout the treatment
journey. However, under a fee-for-service (FFS) system, healthcare professionals are less
9
incentivized to spend adequate time communicating with patients and thoroughly understanding
their preferences and values. In the following section, we further expand the discussion on how
the current FFS payment system and supply-sensitive care place barriers in achieving efficient
delivery of high-quality and effective wound care.
1.4 Supply-Sensitive Care and Overuse by Providers
Overuse of supply-sensitive care persists particularly in the management of chronic
diseases like DFU and VLU.25 Supply-sensitive care refers to the phenomenon where the types
and amounts of care delivered depend on the availability of healthcare resources rather than the
quality measures. The notion centers around the belief that more care contributes to superior
health outcomes. One-third of the U.S. healthcare expenditure is spent on healthcare services that
produce no meaningful benefits to patients’ health and quality of life, such as duplicative CT
scans, unexplained variation in intensity of end-of-life care, and medical procedures that are
questionable appropriate, etc.26,27 In the management of CLUs, the overuse of unnecessary
advanced wound care therapies and antibiotics without clinical justification has imposed a
significant economic burden to the healthcare system.28-30
When facilities are equipped with
advanced diagnostic tools, advanced vascular testing procedures may be conducted more
frequently than necessary, even to patients at minimal risk. The estimated Medicare spending on
different type of CLU ranged from $28.1 to $96.8 billion per year.13 While these unnecessary
treatments do not offer clinically meaningful benefits, more than 50% of Medicare expenditure is
spent on supply-sensitive care, with remarkable variation across regions and treating
physicians.26
10
There are numerous factors contributing to the phenomenon of overutilization and
overtreatment. Facilities with abundant healthcare resources tend to use these resources more
frequently, regardless of clinical necessity or appropriateness. Under the FFS payment model,
such as in the traditional Medicare, where physicians are compensated based on quantity of
healthcare services provided. Surgeons’ decisions may be influenced by the financial incentives
that prioritize volume over quality of care and patient needs.31,32 When making a treatment
decision about revascularization for patients with DFU, surgeons may prefer EDVT over open
bypass surgery because it requires less time, allowing more procedures to be performed within a
day. This can be problematic when open bypass surgery is the more suitable procedure in certain
cases, or revascularization is not even necessary at certain point. In such situation, pursuing
inappropriate revascularization procedure offers little or no benefits to patients at increased
healthcare costs. Besides, providers may be pressured to order unnecessary care by patients
whose satisfaction is achieved by more care under the belief of “more care is better care.”33,34 It
is likely that some physicians tend to choose the easy way and do as the patient requests rather
than spending valuable time in explaining. Multiple other factors may also influence physicians’
treatment behaviors leading to overutilization in wound care, such fear of malpractice and
difficulty accessing medical records.33
Overuse of supply-sensitive care by providers plays a concerning economic role in setting
the reimbursement mechanism of wound care procedures. To reduce the healthcare spending due
to overtreatment, restrictions may be placed on reimbursement of wound procedures. However,
simply restricting the reimbursement of advanced procedures solely based on costs would also
deviate from evidence-based treatments and cause defects in value. Payers and providers may
prioritize compression therapy and delay advanced therapy like endovenous ablation to treat
11
VLU patients, as compression therapy costs much less. Although endovenous ablation costs
more up front, early ablation is proved to improve the healing trajectory and reduce recurrence
probability for VLU patients that are deemed clinically eligible.12 By prioritizing endovenous
ablation to patients who need it, we avoid paying for the downstream treatment costs associated
with nonhealing and recurrent wounds.
Therefore, decision makers should understand the value using a comprehensive matrix
and develop tailored strategies to provide patient-centered care. Ultimately, the goal is to ensure
the accessibility of optimal treatment that maximizes long-term benefits for both patients and
healthcare system.
1.5 Research Aims
To drive the efforts of eliminating defects in value and promoting high value care, this
dissertation attempts to comprehensively identify the unmet need and assess the value of limb
preservation strategies in the field of CLU, focusing on DFU and VLU.
AIM 1: To evaluate the variability in utilization of guideline-consistent vascular limb
salvage care before major LEA in DFU.
Aim 1.1: We understand the utilization of two selected guideline-consistent limb salvage care,
including vascular testing and revascularization, in the year before major LEA among patients
with DFU.
Aim 1.2: We characterized the variability by patient baseline socio-demographic and clinical
features, and identify factors predicting the utilization of the two selected procedures in the year
before major LEA.
12
AIM 2: To assess the value of revascularization strategies for patients with DFU using
propensity score methods.
Aim 2.1: We compare the long-term clinical outcomes of DFU patients undergoing alternative
revascularization strategies (EDVT-first or open bypass surgery-first approach).
Aim 2.2: We compare patient baseline characteristics of patients initiating with EDVT vs. those
treated with open bypass surgery first.
Aim 2.3: We explore the variation in the association between choice of initial revascularization
procedure and the outcomes by disease severity and patient demographics.
AIM 3: To evaluate the economic value of early endovenous ablation with compression
therapy for VLU patients from the Medicare perspective.
Aim 3.1: We compare the cost-effectiveness of early intervention with endovenous ablation in
combination with compression therapy, compared to deferred ablation, in VLU patients from the
Medicare perspective at 3 years and 5 years, respectively.
Aim 3.2: We estimate the budget impact of early intervention of endovenous ablation as opposed
to deferred ablation in a hypothetical insured population of 1 million members.
13
Figure 1.1 Domains of Defects in Value in Chronic Lower Extremity Ulcers
Notes: EDVT = endovascular therapy. QALY = quality adjusted life year.
14
Chapter 2: Variability in Utilization of Guideline-Consistent Limb
Salvage Care Before Lower Extremity Amputation in Patients with
Diabetic Foot Ulcers
Hanke Zheng, Tze-Woei Tan, Seth Seabury, John Romley, David Armstrong,
William Padula
15
2.1 Introduction
Diabetic foot ulcers (DFUs) are severe complications stemming from diabetes mellitus,
preceding 80% of non-traumatic lower extremity amputations (LEA) that can lead to increased
morbidity and mortality.1 To minimize the risk of theses adverse outcomes, the 2019 clinical
practice guidelines published by the International Working Group on the Diabetic Foot (IWGDF)
recommend that all patients at risk of LEA undergo diagnostic vascular testing procedures,
consisting of a set of vascular imaging technologies used to confirm the diagnosis of peripheral
artery disease (PAD) and evaluate blood flow of the affected area for risk stratification.2 This
enables providers to evaluate their risk and prescribe a patient-centered management strategy,
such as an expedited specialty referral for revascularization, to avoid adverse outcomes.3
Revascularization such as open bypass surgery and endovascular therapy is a form of limb
salvage treatment strongly advised for high-risk DFU patients, as these procedures can restore
the blood flow of affected areas and thus prevent further escalation of DFUs leading to LEA.3
According to Hollman and colleagues, despite guideline recommendations, up to 45% of
patients with an LEA did not receive any form of vascular assessment in the year leading up to
limb-loss; only 23.6%-31.6% of Medicare beneficiaries who underwent major (above-ankle)
LEA received revascularization in 2 years before LEA.4
These figures represent DFU patients as
a critical population that could have avoided LEA and associated complications had they been
screened and well managed in the year prior. Brennan et al. (2023) elucidated the barriers to care
that perpetuates racial disparity in major LEA among DFU patients using the National Institute
on Minority Health and Health Disparities framework.5 They identified common contributing
factors, consisting of under-utilization of appropriate foot care, living in disadvantaged
16
neighborhood, and transportation difficulties, among many. These components interact with each
other, resulting to adverse limb events.
Substantial variability exists in the utilization of limb salvage care before LEA in the
real-world setting. Several research evaluated the use of limb preservation procedures, especially
for vascular testing and revascularization, before LEA among patients with DFU-relevant
conditions such as PAD and chronic limb threatening ischemia (CLTI) using various data
sources.4,6-8 A recent study assessed the association between usage of care and vascular
assessment before major LEA among U.S. veterans diagnosed with PAD, concluding that prior
healthcare utilization and distance to primary care were strong predictors of vascular assessment.
African American veterans were found to be more likely to receive a vascular testing before
amputation, likely due to a lack of data of disease severity and facility-level information.7
Secemsky et al. (2024) demonstrated that individuals with low-income status or ones treated in
safety-net hospitals were less likely to receive vascular testing or revascularization before LEA
among the Medicare beneficiaries with CLTI, and those patients also exhibited a poorer longterm survival.8
Nonetheless, there remains a gap in the current literature regarding research specifically
focused on the DFU population, which is known to have an elevated risk of major complications
than the general PAD population without diabetes.9 Our objective was to evaluate patient
utilization of guideline-consistent limb salvage care, defined as (1) diagnostic vascular testing
and (2) revascularization, and identify factors associated with the utilization of these guidelineconsistent limb salvage management strategy among the DFU population in the year preceding
major LEA.
17
2.2 Methods
2.2.1 Data Source and Patient Selection
We used the Optum’s De-Identified Clinformatics Data Mart (CDM) from January 1,
2010, to December 31, 2021 for this retrospective cohort study. Optum’s CDM is a Health
Insurance Portability and Accountability Act (HIPAA)-compliant closed administrative claims
database consisting of data from millions of lives covered by commercial plans or Medicare
Advantage plans from all 50 U.S. states. It is composed of detailed patient-level information
including member enrollment, medical claims, pharmacy claims, inpatient confinement claims,
emergency visit, office visit, lab results, and other healthcare utilization that is adjudicated
throughout the enrollment. Some level of provider information and hospital data are also
included.
Our sample included DFU patients who underwent major LEA during the study period.
We focused on 1 year before amputation to explore the utilization of pre-LEA diagnostic
vascular testing and revascularization, as it is a critical period when patients could have benefited
from the better limb salvage care to prevent major LEA. Based on a validated algorithm, we
identified DFU patients by a confirmed diagnosis of diabetes in combination with foot ulcers,
osteomyelitis, or gangrene using the International Classification of Disease (ICD)-9 or ICD-10
codes.10,11 Patients were included if they underwent major LEA after the DFU diagnosis. They
were required to have at least 24-month continuous enrollment pre-LEA; we explored the
utilization of limb salvage care during 12 months before major LEA and used the data prior to
that to track patients baseline characteristics and comorbid conditions. Patients who were
younger than 18 or had any forms of LEA before DFU diagnosis were excluded from the
analysis.
18
2.2.2 Measures and Covariates
Patients’ socio-demographics and clinical characteristics were tested as covariates in the
analysis, including: age, male, race/ethnicity (non-Hispanic White, non-Hispanic Asian, nonHispanic Black or African American, or Hispanic), region (Southeast, Midwest, Northeast, or
Southwest, West), Medicare Advantage enrollment, insurance plan type [preferred provider
organization (PPO), managed care (health maintain organization, point of service), or other],
household income level (<$40,000, $40,000-$75,000, or > $75,000), education level (high
school or below, or above high school), comorbidities [hypertension, hyperlipidemia, stroke,
chronic kidney disease (CKD), end-stage renal disease (ESRD), congestive heart failure (CHF),
chronic obstructive pulmonary disease (COPD), acute myocardial infraction (AMI), and cancer],
and diagnosis of PAD or gangrene, as a measure of ischemia severity.
2.2.3 Outcomes
The primary outcomes were (1) the receipt of any vascular testing procedures and (2) the
use of revascularization during the year before the major LEA. The specific vascular assessment
procedures included ankle-brachial index measurement, duplex ultrasound, computed
tomographic angiography, and magnetic resonance angiography. The date of first use of each
procedure was flagged. For revascularization, we recorded the first use of open bypass surgery or
endovascular therapy during the year before major LEA. Those procedures were identified using
the Current Procedural Terminology (CPT) codes.
2.2.3 Statistical Analysis
Descriptive statistics were performed to present patient baseline socio-demographic and
clinical characteristics. The one-way analysis of variance (ANOVA) test was conducted to
evaluate differences in numeric measures, and chi-square test was used for the comparison of
19
categorical variables by use of vascular testing and by use of revascularization, respectively. We
presented the variation in pre-LEA vascular testing and revascularization by selected socioeconomic factors, including age, sex, race/ethnicity, payer type, household income level, and
education.
Following a stepwise approach, we applied multivariate logistic regressions to assess the
factors associated with use of vascular testing and revascularization, respectively. The primary
model is the following logistic regression:
𝑌i= αi+𝛽𝑋i +𝜀i
where Yi is a binary variable denoting whether patient i utilized the pre-amputation limb salvage
intervention of interest; the probability was estimated using the log link function, logit. Xi
represents a set of baseline covariates of patient i. In the primary model, we included categorical
variables for age (<50, 50-64, 65-74, >75), male, race/ethnicity, household income level <
$40,000, education level of high school or below, region, enrollment of Medicare Advantage,
insurance plan type, DFU severity, comorbidities. Age, household income level, and education
were recoded based on the sample distribution. The residual term was denoted by 𝜀i. For
exploratory purposes, the model was further expanded to explore combined effect of the racial
minority and socio-economic disadvantaged population, given the established effect of the
convergence of racial and income disparity.12 We added the interactions between race and
inferior socio-economic factors, measured by household income <$40,000 and education level of
high school or below. Although model with interaction terms can be difficult to interpret, we did
this to captures the potential difference in the effect of inferior socio-economic status on the logodds of receiving vascular testing between non-Hispanic White and other races. For both
endpoints, a final restrictive model was trimmed to eliminate non-socio-demographics covariates
20
with negligible effect in previous expanded models. As many covariates can be strongly related,
removing the ones with non-significant effect allows us to gain independent impact that the
remaining socio-demographic covariates have on the probability of utilizing pre-LEA limb
salvage care.
2.3 Results
2.3.1 Patient Characteristics
The final sample comprised 13,711 eligible DFU patients who underwent major LEA and
had sufficient previous enrollment (Figure 2.1). Presented in Table 2.1, older patients tended to
get the vascular testing more consistently, as well as revascularization in the year preceding
major LEA. Hispanic patients were more likely to receive vascular testing or revascularization.
We also observed greater share Medicare Advantage enrollees in patients who received vascular
testing or revascularization. Comorbidities were more prevalent patients who received vascular
testing or revascularization. Furthermore, more receivers of vascular testing or revascularization
tended to have PAD and gangrene, the most severe presentation of DFU.
2.3.2 Utilization of Pre-Amputation Vascular Assessment and Revascularization
During the year before major LEA, 72.9% of patients received the vascular assessment at
least once. The duplex ultrasound was prescribed the most (50.8%), followed by ankle-brachial
index measurement (49.4%) and other types of vascular imaging procedures. The mean (SD) and
median time (range) between the first vascular screening and the amputation date were 95.6 (99)
and 57 (0-365) days. 63.4% did not receive any revascularization. The use rate increased by age
– more than 75% of patients over 65 got vascular testing pre-LEA (Figure 2.2). In addition,
21
Hispanic was the population with the highest utilization of vascular testing (80%). Although the
variability was small across sex, household income, and education, we observed more Medicare
beneficiaries receiving vascular testing than the commercially insured patients. Among the 5,018
(36.6%) revascularization users, 84.5% used endovascular therapy, 31.6% underwent open
bypass surgery, and 5.9% got hybrid during the year pre-LEA. The mean (SD) and median
(range) time from first observed revascularization and amputation were 107.5 (95.6) and 79 (0-
365) days. Similar trends in the variability of revascularization by key socio-demographics were
observed (Figure 2.3). Only 35.4% of patients utilized both vascular testing and
revascularization prior to major LEA.
2.3.3 Factors Predicting Pre-Amputation Utilization of Vascular testing
In the primary logistic regression (Table 2.2), age was positively associated with
receiving a pre-amputation vascular screening (p<0.01). Hispanic patients had 46% greater
probability of getting vascular testing pre-LEA [Odds ratio (OR)=1.46, p<0.01]. Ones with PPO
were less likely to get any assessment before LEA as opposed to patients with managed care plan
(OR=0.86, p<0.05). Patients who were previously diagnosed with PAD had 25% greater
likelihood of getting assessed before major LEA. Severe DFU (gangrene) significantly doubled
the likelihood of getting vascular testing (OR=2.67, p<0.01). Baseline comorbidities were
consistently associated with greater chance of vascular testing. In the expanded model, while
there was no overall effect detected for being Black or education of high school or below
individually, Black patients were less likely get vascular testing, conditional on having an
education level of high school or below (OR=0.8, p<0.05). The findings of the final restrictive
model were consistent with that of the primary model.
22
2.3.4 Factors Predicting Pre-Amputation Utilization of Revascularization
The primary model suggested age (p<0.01), being Hispanic (OR=1.44, p<0.01), and
living in the Southwest (OR=1.23, p<0.001) were positively associated with pre-LEA utilization
of revascularization (Table 2.3). However, non-Hispanic Asian patients had a reduced possibility
of receiving revascularization (OR=0.71, p<0.05) than the non-Hispanic White. Patients with
higher household income were found to have 10% additional chance of getting revascularization
(OR=1.1 for household income of $40,000-$75,000 vs. < $40,000, p<0.05). As for the ischemia
severity, PAD (OR=1.32, p<0.01) and gangrene (OR=2.67, p<0.01) were both strong predictors
of one receiving revascularization. Like vascular testing, our data showed patients’ baseline
comorbidities positively associated with pre-LEA revascularization, especially hyperlipidemia
(p<0.01). In the expanded model, none of the interaction terms suggested a significant effect.
Final restrictive model showed consistent results as the primary regression model.
2.4 Discussion
Timely intervention with appropriate vascular care procedures has been proven to reduce
the risk of LEA and mortality of severe DFU.13 Although the scientific advance in medicine has
improved the clinical outcomes of DFU over the past decade, racial and ethnic minority patients
and the undeserved ones are consistently subject to greater risk of major LEA.14 This is the first
study that sheds light on the utilization and barriers in DFU patients’ access to guidelineconsistent limb salvage care, including diagnostic vascular testing or revascularization, before
major LEA in the commercially or Medicare Advantage insured population. Our analysis
highlighted the gap of guideline-consistent limb salvage care. During the year before major LEA,
23
there were 28.1% patients who did not get any form of vascular testing and 63.4% of them with
no attempt of revascularization, respectively, representing a missed opportunity for limb salvage
care improvement.
We further revealed variation in use of vascular testing and revascularization across
socio-demographic and clinical factors. Notably, Hispanics were more likely receive vascular
testing or revascularization procedures in the year prior to major LEA, which may be attractable
to the advanced presentation in the course of the disease in the race/ethnicity minority group.15 In
our sample, PAD and gangrene were both more prevalent in the Hispanics and Blacks than the
Whites. Furthermore, older patients and those with comorbidities tended to have greater chance
of receiving vascular testing as well as revascularization in the year before major LEA. As a
reflection of the real-world clinical practice, our data showed that providers tended to be more
attentive and prescribe care to patients in more critical state. It could also result from patients’
reluctance of visiting clinics for an early diagnosis and appropriate vascular management until
their condition has progressed to a severe state. Another important finding was that higher
income level would increase the possibility of getting revascularization whereas there was no
difference in vascular testing by income, after controlling for severity of ischemia and baseline
comorbidities. Low-income individuals with DFU are likely to have limited access to primary or
high-quality preventive care in a timely manner, causing the escalation of DFU disease stage.
Additionally, the procedural cost alone per patient of endovascular therapy is reported to be
$22,200 whereas it can cost up to $42,900 for open bypass surgery.16 While this is an insured
population, the out-of-pocket costs of revascularization can still be devastating to some,
especially disadvantaged patients with low income.
24
Several recent research described the pre-LEA vascular care utilization in various
populations.7,8,17,18 Alabi and colleagues concluded that previous healthcare utilization, distance
to primary care, and region were significant predictors of receipt of vascular imaging procedures
or revascularization among PAD veterans in the year before major LEA.7 They also reported a
greater likelihood of receiving vascular imaging procedures as well as revascularization attempts
in Black veterans than White veterans.17 Although the patient population was distinct, their
finding in racial difference resonated to some extent with our study, showing greater utilization
of pre-LEA vascular procedures in the racial minority group of Hispanic. Similarly, they did not
adjust for PAD severity or whether the LEAs were performed in an emergent setting due to
severe infections so that Black veterans may be more severe at the baseline. Another study
examined the individual, hospital, and regional factors associated with vascular care intensity in
the year before major LEA among Medicare beneficiaries with CLTI. Their analysis suggested
low-income adults were substantially less likely to receive both vascular testing and
revascularization before LEA, which was consistent with our findings for the DFU population.
According to our findings, we call for Decision-Makers to make efforts to facilitate the
use of guideline-consistent vascular care for DFU patients who were at high risk of major LEA.
Younger patients and those with fewer comorbidities often tend to not get the limb preservation
care as they should have received, despite facing significant risk of adverse limb events. Plus,
low-income patients may have limited access to receiving timely and high-quality vascular care
to manage their DFU appropriately. Although we showed that Hispanics were more likely to
receive vascular procedures, it was possible that these racially disadvantaged cohorts were prone
to delayed diagnosis and more advanced disease presentation at diagnosis because of inequitable
investment in healthcare resources on a structural level.19
25
As such, education and training programs should be put in place to strengthen the
awareness of importance of timely vascular testing and appropriate management for both
healthcare professionals and patients. We also encourage the hospitals and vascular care centers
to adopt a proactive approach for timely vascular testing to identify high-risk patients and
intervene with patient-centered care. Healthcare providers should enhance a shared decision
making to ensure patients understand their options and importance of timely vascular testing and
management in preventing major adverse limb events.
Furthermore, payers should revisit the coverage policy to remove access barriers for
high-risk DFU patients to necessary limb preservation care. It is known that health inequity in
access to quality wound management and preventive care, resulting from one’s underlying socioeconomic status, is associated with poor health outcomes in DFU. Patients who are Black or
Hispanic and those residing in rural areas are found to be significantly more likely to experience
major LEA.15,20
Targeted policy efforts should be made, such as reducing co-pays of these
guideline-recommended procedures, particularly for disadvantaged minority patients who are in
need but unable to access the treatment due to economic barriers or lack of information. While
the goal is to ensure adequate patient access, reimbursement approval for treatment should be
strictly based on clinical examination and results that are well justified to avoid the risk of
overtreatment and use of unnecessary procedures. In combination with clinical eligibility, patient
socio-economic status such as their federal poverty level and other appropriate criteria should
also weigh in to determine their eligibility for copay reduction and other financial aid programs
from the insurance plan.
Value-based models such as pay-for-performance (P4P) program for DFU management
may be another solution that pushes to achieve better clinical outcomes and lower the healthcare
26
costs. As opposed to the fee-for-service (FFS) model, the concept of P4P is to financially reward
healthcare providers for achieving pre-defined quality indicators.21 Under a P4P model, vascular
surgeons or general practitioners receive higher payment if they help DFU patient avoid adverse
outcomes such as major LEA, hospitalization due to severe wound infections, etc. Therefore, it
aligns physician’s incentives with quality of care rather than quantify. Healthcare providers are
incentivized to adhere clinical guidelines for limb preservation, which include but not limited to
testing high-risk patients in a timely manner, ensuring consistent follow-up, proactively
communicating with patients and caregivers, and considering appropriate vascular treatment
procedures for those who are clinically eligible. On one hand, the shift in financial incentives
reduces overuse of unnecessary care by providers. On the other hand, it promotes the delivery of
high-quality care, freeing up the monetary and healthcare resources by preventing the need for
extensive infection management and treatment associated with DFU complications.
Continuous efforts should be made by major medical associations such as Society of
Vascular Surgery and American Podiatric Medical Association, including promoting the
adoption and adherence of the best clinical practice guidelines for DFU management among
clinicians managing DFU and relevant vascular conditions. It is essential to consider financial or
technical supports for under-resourced healthcare providers serving low-income and minority
community, which will reduce the disparity and increase utilization of guideline-consistent care
to improve outcomes of DFU.
This study is not without limitations. First, the granular clinical measure of the extent of
infection and facility-level information were not available in the administrative claims data,
which might be critical in the treatment decision making for limb care utilization. Informal
bedside assessments that were not coded in the claims. Second, while vascular testing and
27
revascularization are crucial for limb salvage, other limb salvage management aspects such as
infection treatment, wound care, offloading, and other preventive measures were not explored in
this analysis. Furthermore, due to the nature of observational study, we were unable to capture
the potential association between utilization of limb salvage care and the unobservable factors,
such as patient preference and physicians’ behaviors, and unobserved facility-level variation. In
addition, given the data we used for analysis, this study has limited generalizability to the
population that is uninsured or covered by other health plans.
Thus, future research should leverage data sources with rich clinical information, such as
electronic health records or registry data of DFU patients, to provide detailed insights into the
severity of ischemia and infections, enabling a holistic adjustment to uncover the disparity in
care. It is also of great importance to explore the underlying disparity in DFU population that are
not covered through OPTUM, such as commercial insures other than OPTUM, traditional FFS
Medicare, or Medicaid. Beyond these, it is worth gaining a systematic understanding of patients’
and physicians’ perspectives of challenges that hamper the utilization of guideline-consistent
limb preservation care.
In conclusion, our analysis emphasized the under-utilization of guideline-consistent
vascular care in the DFU population and underscored the variation by socio-demographic and
clinical factors. Healthcare providers, payers, and policy makers must collaborate to mitigate the
disparity and ensure an equitable access to high-quality vascular care for high-risk DFU patients
in the U.S.
28
Figure 2.1 Patient Selection
Adult patients with confirmed
diagnosis of diabetes in
combination with 1 diagnosis
for foot ulceration, gangrene,
or osteomyelitis between Jan
1, 2010 and Dec 31, 2020.
N= 1,066,550
N=24,557
Select patients who received
major lower-extremity
amputation after the confirmed
diagnosis.
(97.7%)
N=24,103
Include patients with
continuous enrollment of 24
months before major lowerextremity amputation
(19.9%)
N=13,711
Exclude patients who received
any forms of amputation before
diagnosis.
(1.8%)
29
Figure 2.2 Vascular Testing Use in the Year before Major LEA
30
Figure 2.3 Revascularization Use in the Year before Major LEA
31
Table 2.1 Patient Baseline Characteristics (N=13,711)
Measure No vascular
testing
With
vascular
testing
P
value No revascularization With
Revascularization
P
value
Unique Patients 37.20% 72.90% 63.40% 36.60%
Age, year
Mean 65.9 69.8 ** 67.5 70.7 **
SD 11.8 10.6 11.6 9.6
Range 24-89 22-89 24-89 22-89
Male 67.70% 66.80% 0.31 67.60% 66.10% 0.07
Race/Ethnicity
Non-Hispanic White 61.10% 55.30%
**
58.60% 53.90%
**
Non-Hispanic Black or
African American 22.00% 22.70% 22.80% 22.10%
Non-Hispanic Asian 1.50% 1.30% 1.50% 1.10%
Hispanic 10.80% 16.10% 12.60% 18.20%
Missing/Unknown 4.60% 4.60% 4.60% 4.60%
Region
Midwest 19.50% 18.10%
*
18.90% 17.80%
*
Northeast 10.30% 9.90% 9.90% 10.20%
Southeast 35.40% 34.20% 35.40% 33.00%
Southwest 16.10% 19.20% 16.70% 21.30%
West 18.60% 18.40% 19.00% 17.70%
Missing/Unknown 0.10% 0.10% 0.10% 0.00%
Medicare advantage
Yes 78.00% 86.00% ** 81.60% 87.60% **
Insurance type
Managed care 28.80% 31.30%
**
31.30% 29.50%
PPO 21.90% 14.70% 18.10% 14.10% **
Other 49.20% 54.00% 50.60% 56.40%
Education
High School or below 41.80% 43.80%
0.03
42.70% 44.10%
Above high school 55.00% 52.80% 53.90% 52.60% 0.12
Missing/Unknown 3.30% 3.40% 3.40% 3.30%
Household Income
Range
<$40,000 38.60% 41.30%
*
40.40% 40.70%
*
$40,000-$75,000 27.90% 28.00% 26.80% 30.00%
>$75,000 20.30% 17.70% 18.70% 17.80%
Unknown 13.20% 13.00% 14.00% 11.50%
32
Comorbidities
Hypertension 80.00% 85.60% ** 82.40% 87.00% **
hyperlipidemia 74.80% 82.00% ** 77.40% 85.40% **
Stroke 21.20% 33.70% ** 27.20% 35.70% **
CKD 48.20% 57.30% ** 52.80% 58.30% **
ESRD 11.80% 19.70% ** 14.90% 22.10% **
CHF 35.90% 47.10% ** 41.80% 48.10% **
OPRC 18.00% 23.70% ** 20.50% 25.10% **
AMI 11.00% 16.60% ** 12.90% 18.80% **
Cancer 17.00% 21.10% ** 18.10% 23.20% **
Severity of ischemia
PAD 69.80% 77.90% ** 73.10% 80.20% **
Gangrene 53.70% 77.20% * 63.80% 83.00% *
Notes: * p < 0.001; ** p < 0.0001. CKD = chronic kidney disease; ESRD = end-stage renal disease; CHF =
congestive heart failure; COPD = chronic obstructive pulmonary disease; AMI = acute myocardial infarction; PAD
= peripheral artery disease.
33
Table 2.2 Multivariate Logistic Regression of Utilization of Vascular Testing
Primary Model Expanded Model Final Restrictive Model
Variable OR P value OR P value OR P value
Intercept 0.35 ** 0.32 ** 0.34 **
Age
< 50 Reference
50-64 1.28 * 1.27 * 1.28 *
65-74 1.78 ** 1.77 ** 1.78 **
>75 2.03 ** 2.03 ** 2.03 **
Race/Ethnicity
Non-Hispanic White Reference
Non-Hispanic Asian 0.80 0.19 0.94 0.81 0.77 0.13
Non-Hispanic Black or
African American 0.97 0.62 1.02 0.86 0.98 0.70
Hispanic 1.46 ** 1.73 ** 1.46 **
Male 1.02 0.63 1.02 0.58
Medicare Advantage 1.01 0.91 1.01 0.88
Insurance plan type
Managed care Reference
PPO 0.86 * 0.86 * 0.87 *
Other 0.99 0.85 0.99 0.87 1.00 0.94
Region
Southeast Reference
Midwest 1.00 1.00 1.00 0.94
Northeast 0.88 0.10 0.90 0.14
Southwest 1.08 0.21 1.10 0.14
West 0.91 0.17 0.92 0.23
Education
High School or Below 0.95 0.31 1.03 0.60 0.97 0.49
Household Income
< $40,000 Reference
$40,000-$75,000 0.93 0.18 0.95 0.43 0.94 0.22
>$75,000 0.92 0.16 0.94 0.36 0.92 0.19
Race * Education of High
School or Below
Non-Hispanic White Reference
Non-Hispanic Asian 1.02 0.95
Non-Hispanic Black or
African American 0.80 *
Hispanic 0.88 0.38
34
Race * Household
Income < $40,000
Non-Hispanic White Reference
Non-Hispanic Asian 0.97 0.94
Non-Hispanic Black or
African American 1.12 0.32
Hispanic 0.88 0.38
Severity of ischemia
PAD 1.25 ** 1.25 ** 1.24 **
Gangrene 2.67 ** 2.68 ** 2.66 **
Comorbidity
Hypertension 1.00 1.00 1.00 0.98
Hyperlipidemia 1.52 ** 1.53 ** 1.52 **
Stroke 1.53 ** 1.53 ** 1.53 **
CKD 0.96 0.36 0.96 0.38
ESRD 1.38 ** 1.38 ** 1.36 **
CHF 1.27 ** 1.27 ** 1.26 **
COPD 1.26 ** 1.26 ** 1.28 **
AMI 1.22 ** 1.22 ** 1.21 **
Cancer 1.09 0.07 1.09 0.07
Notes: * p < 0.05; ** p < 0.01. OR = odds ratio; CKD = chronic kidney disease; ESRD = end-stage renal disease;
CHF = congestive heart failure; COPD = chronic obstructive pulmonary disease; AMI = acute myocardial
infarction; PAD = peripheral artery disease.
35
Table 2.3 Multivariate Logistic Regression of Utilization of Revascularization
Primary Model Expanded Model Final Restrictive Model
Variable OR P value OR P value OR P value
Intercept 0.04 ** 0.04 ** 0.04 **
Age
< 50 Reference
50-64 1.80 ** 1.80 ** 1.79 **
65-74 2.86 ** 2.86 ** 2.84 **
>75 2.91 ** 2.91 ** 2.88 **
Race/Ethnicity
Non-Hispanic White Reference
Non-Hispanic Asian 0.71 * 0.74 0.20 0.71 *
Non-Hispanic Black or
African American 0.97 0.55 0.95 0.54 0.97 0.54
Hispanic 1.44 ** 1.32 ** 1.43 **
Male 0.93 0.08 0.93 0.08
Medicare Advantage 0.98 0.75 0.98 0.75
Insurance plan type
Managed care Reference
PPO 1.11 0.13 1.11 0.13
Other 1.16 ** 1.15 **
Region
Southeast Reference
Midwest 1.04 0.50 1.04 0.53 1.03 0.60
Northeast 0.96 0.57 0.96 0.55 0.96 0.59
Southwest 1.23 ** 1.22 ** 1.22 **
West 0.94 0.31 0.94 0.32 0.90 0.06
Education
High School or Below 1.21 0.30 0.96 0.51 0.98 0.58
Household Income
< $40,000 Reference
$40,000-$75,000 1.10 * 1.11 0.08 1.09 *
>$75,000 1.05 0.45 1.06 0.40 1.04 0.49
Race * Education of High
School or Below
Non-Hispanic White Reference
Non-Hispanic Asian 0.85 0.67
Non-Hispanic Black or
African American 0.95 0.64
Hispanic 1.23 0.07
Race * Household
Income < $40,000
Non-Hispanic White Reference
36
Non-Hispanic Asian 1.02 0.96
Non-Hispanic Black or
African American 1.10 0.36
Hispanic 0.90 0.40
Severity of ischemia
PAD 1.32 ** 1.32 0.00 1.31 **
Gangrene 2.67 ** 2.67 0.00 2.64 **
Comorbidity
Hypertension 0.99 0.94 0.99 0.94
Hyperlipidemia 2.13 ** 2.14 0.00 2.16 **
Stroke 1.22 ** 1.22 0.00 1.21 **
CKD 0.83 ** 0.83 0.00 0.83 **
ESRD 1.38 ** 1.39 0.00 1.38 **
CHF 0.99 0.84 0.99 0.90
COPD 1.25 ** 1.25 0.00 1.26 **
AMI 1.23 ** 1.23 0.00 1.23 **
Cancer 1.12 ** 1.12 0.01 1.12 *
Notes: * p < 0.05; ** p < 0.01. OR = odds ratio; CKD = chronic kidney disease; ESRD = end-stage renal disease;
CHF = congestive heart failure; COPD = chronic obstructive pulmonary disease; AMI = acute myocardial
infarction; PAD = peripheral artery disease.
37
Chapter 3: Assessing the Value Defects in Initial Revascularization
Strategy for Patients with Diabetic Foot Ulcer in the United States:
A Propensity Score Approach
Hanke Zheng, Tze-Woei Tan, Seth Seabury, John Romley, David Armstrong,
William Padula
38
3.1 Introduction
Diabetic foot ulcer (DFU) is a severe form of complication of diabetes mellitus, typically
resulting from improper glycemic control, poor circulation, underlying neuropathy, and improper
foot care. It is known as the leading risk factor preceding 80% of non-traumatic lower
extremality amputations (LEA), which is associated with greater morbidity and mortality.1 DFUs
also represent a significant economic burden on society, leading to up to $79 billion direct
medical costs in the U.S. per year.2 Peripheral artery disease (PAD) is commonly comorbid with
DFU in approximately 50% of patients. DFU patients with PAD or the even more severe chronic
limb-threatening ischemia (CLTI) are known to have an elevated risk of non-healing of their
ulcers, leading to LEA if not well managed.3 Revascularization is critical treatment component
that aims to restore blood flow to the affected area to improve the symptom and prevent adverse
limb events in patients with ischemic DFUs and relevant conditions such as PAD and CLTI.3,4
The International Working Group on Diabetic Foot (IWGDF) 2019 guidelines recommend
considering revascularization for DFU patients at high risk of LEA.5 Two common
revascularization procedures are open bypass surgery and endovascular therapy (EDVT).
Historically, open bypass surgery offers durable healing by creating a detour around the blocked
artery using a graft to improve the blood flow. Although it is subject to greater surgical infection
risks, it is recognized to be more suitable for patients with advanced disease.6 Technical
advances have led to the availability of minimally invasive EDVT. EDVT is a catheter-based
technology, including balloon angioplasty and stenting, which is used to widen blocked arteries
and clear the blockage through small incisions in the patient’s groin or upper thigh. It is preferred
by some since it typically does not require hospitalization and has shorter timeframe for
recovery. The selection of the initial treatment is contingent on many considerations, including
39
the extent and location of the disease, patient anatomy, surgical risk, prognosis, patient
preference, physicians’ expertise, and insurance coverage.
However, the optimal initial revascularization strategy has not been well established
between open bypass surgery and EDVT. Discordant findings were concluded in two recent
major randomized controlled trials comparing the predominant EDVT and open bypass surgery
for lower extremity PAD and CLTI.7,8 The BEST-CLI trial recruited patients from over 133 sites
in the U.S, 12 in Canada, and 5 in other high-income countries. Two parallel randomizations
were conducted for individuals who had a single segment of great saphenous vein needed for
surgery (cohort 1) and those who needed an alternative bypass conduit (cohort 2). While no
outcome difference was detected in cohort 2, they reported a significantly lower incidence of
major adverse limb events (MALE), defined as above-ankle (major) LEA or any additional
revascularization of the index limb, or death in the surgical group in cohort 1. In contrast, the
BASEL-2 trial, which primarily involved the U.K. population, reached the opposite conclusion –
patients initiated with EDVT achieved better clinical outcomes, particularly in amputation-free
survival, driven by fewer deaths.
A broad range of factors can drive the divergence of the results. For instance, the
definition of primary endpoints of the two trials differed.9 While amputation-free survival is a
critical to measure patient outcomes, it does not reflect the burden of revascularizations in
solving unhealed or recurrent symptoms, which is captured by MALE that provides the
understanding of quality-of-care matrix. In addition, the underlying population enrolled in the
BEST-CLI trial differed from those in the BASIL-2 trial in the baseline demographic and clinical
characteristics, resulting from difference in recruitment criteria. For instance, the location of the
lesions where revascularization was performed different - the BEST-CLI trial included patients
40
undergoing infrainguinal revascularization while patients underwent infrapopliteal
revascularization in BASIL-2 trial.10 Besides, the detailed anatomical complexity of the disease
remained unclear in both trials.
Real-world evidence is equally important as these two randomized trials do not
necessarily reflect the real-life practices and outcomes. A systematic literature review evaluated
the effectiveness of revascularization among diabetic patients with PAD using data collected
from clinical trials and observational studies conducted between 1980 to 2022.11 The primary
outcomes reported in those studies include wound closure, minor (below-ankle) LEA, major
LEA, all-cause mortality, post-operative complications, and need for another revascularization
due to non-healing condition or wound recurrence. Many of the results from observational
research varied and were inconclusive as to which revascularization procedure, open bypass or
EDVT, was more effective.6,12-14 Low quality of these real-world evidence studies was due to
lack of control of confounding factors in treatment selection, short length of follow-up, and
limited generalizability due to use of local institutional data. In addition, some data dated back
over 20 years. Thus, there is still a lack of high-quality and up-to-date comparative real-world
evidence for the specific population with ischemic DFU.
As described in Chapter 1, we implemented the Value Defect Framework in DFU care as
the conceptual framework for this research. The current aim of the dissertation was to identify
the value defects associated with the choice of initial revascularizat6,12-14ion by comparing the
value matrixes of the two alternative revascularizations based on the Value Defect Framework.
The study compared patients’ clinical outcomes, including major LEA, overall survival,
amputation-free survival, and reinterventions.
41
3.2 Methods
3.2.1 Study Design
This is a retrospective cohort study comparing the effectiveness of alternative initial
revascularization strategy, open bypass surgery vs. EDVT, among DFU patients. In this analysis,
we obtained the administrative medical and pharmacy claims data from the Optum’s DeIdentified Clinformatics® Data Mart (CDM) Database from years 2009 to 2021. Optum’s CDM
is a national-level closed claims database that covers millions of lives insured by commercial or
Medicare Advantage plans in all 50 states in the U.S. It is compliant to the Health Insurance
Portability and Accountability Act (HIPAA). In this analysis, we utilized the information from
the member enrollment file, medical claims, pharmacy claims, office visits, and other healthcare
utilization that is adjudicated throughout patients’ enrollment period. The Date of Death (DOD)
file was used to obtain the mortality data. This observational study was exempt from institutional
review board review as it involved no human research participants.
This study was exempt from institutional review board review because it did not involve
participation of human subjects.
3.2.2 Patient Selection
According to the Chronic Conditions Data Warehouse category and validated selection
algorithms, we identified adult DFU patients by requiring them to have at least one
inpatient/skilled nursing facility/home health diagnosis or two outpatient diagnoses on separate
dates for diabetes in combination with at least one diagnosis for foot ulceration, osteomyelitis, or
gangrene after diabetes diagnosis between Jan 1, 2010, and Dec 31, 2021.15-17 We selected
patients who underwent either EDVT or open bypass surgery after the confirmed DFU; the date
42
of initial revascularization procedure was anchored as the index date. Patients must have at least
12 months pre-index continuous enrollment to capture their baseline characteristics. Finally, we
excluded ones who underwent any form of revascularization or LEA during the 12-month preindex enrollment.
3.2.3 Predictors
The key explanatory variables that influence patients’ outcomes as well as predict the
type of initial revascularization procedure - EDVT-first or open bypass-first – depended on a
number of factors. Patients baseline demographic and clinical characteristics were measured
during the 12 months pre-index, including age, sex, race/ethnicity, region, Medicare Advantage
enrollment, specific insurance type, pre-operative medication use to reflect whether the patient
was actively managing their disease through medicine, comorbidities, and the ischemia
presentation, proxied by the diagnosis of PAD and/or gangrene. Note that we excluded the
diagnosis of PAD as a covariate in the multivariate analysis due to lack of variability, as it was
present in majority of the sample.
3.2.4 Main Outcome Measures
The primary outcomes consisted of all-cause mortality, major LEA, and major LEA-free
survival. The secondary outcomes included reinterventions, meaning the patient received another
revascularization after the index procedure, and MALE, defined as major LEA or reintervention
during the entire follow-up frame.
43
3.2.5 Statistical Methods
Patients were classified to two groups based on their initial revascularization – EDVTfirst or open bypass-first cohorts. Their baseline characteristics were presented and compared
using t-test for continuous variables or chi-square test for categorical variables at the 95%
confidence level.
Inverse probability of treatment weighting (IPTW)
We applied the IPTW to account for the treatment selection bias based on measured
confounders. IPTW is a common propensity score-based approach to control for observable
confounders by balancing the distribution of baseline characteristics, mimicking the balancing
quality of randomized trials.18,19 Compared to traditional multivariable regression analysis, IPTW
explicitly aims to balance covariates between treatment groups, which can lead to better control
of confounding and allows for direct comparison of outcomes. In addition, traditional
multivariable regression requires specifying a model for the outcome variable, which can be
complex and prone to misspecification. IPTW, on the other hand, focuses on estimating the
probability of treatment assignment based on covariates, which is simpler and less prone to
model misspecification.
As the first step, we ran a multivariate logistic regression of the actual revascularization
procedure patients received on their baseline characteristics. As advised by the Good Practice of
Propensity Score Methods, we included all baseline characteristics into the logistic model
predicting the receipt of open bypass surgery first, regardless of the p value and statistical
significance.20 By doing so, we captured the influence of any observable factors that were
relevant or meaningful to determine the treatment assignment holistically. Second, we used the
model estimates to generate the propensity score for each individual as the probability of being
44
treated with open bypass surgery first, 𝑒𝑖=Pr(T𝑖=1|𝑋𝑖), conditional on baseline measured
covariates, where T𝑖 is the revascularization choice for individual I and 𝑋𝑖 refers to the
observable baseline confounders for individual I that were associated with treatment selection
and/or patient outcomes.
Using the propensity score (𝑒𝑖), we calculated the IPTW as 1/𝑒𝑖 for patients receiving
open bypass surgery first, and 1/ (1- 𝑒𝑖) for those receiving endovascular therapy first, attaching
greater weights to open bypass-first patients with a lower probability of receiving open bypass,
as well as EDVT-first patients who had a greater probability of receiving open bypass. A critical
methodology issue with IPTW is that weights can be extremely high or low for patients with a
low or high probability of receiving open bypass first, consequently inflating the variance of the
estimate.21 To overcome this and improve the precision, we stabilized the weights by replacing
the numerator with the crude probability of exposure, which is the proportion of patients in the
open bypass-first cohort (Popen) in our case. As such, the stabilized weights became Popen/𝑒𝑖 for
patients receiving open bypass first, and (1- Popen) / (1- 𝑒𝑖) for those receiving endovascular
therapy first. Finally, the standardized mean difference (SMD) for each covariate in the
propensity score model was calculated for balance check before and after weighting:
𝑆𝑀𝐷 =
𝑋1− 𝑋2
√(𝜎1
2+𝜎2
2)/2
(Eq. 1)
Where 𝑋1 and 𝑋2 are sample means of the open bypass-first and the EDVT therapy-first
cohorts, and 𝜎1
2
and 𝜎2
2
represent the variance of the variable in the two cohorts, respectively.
Regardless of the unit of measurement, it enables the comparison of the mean in the unit of
standard deviation (SD) of variables on different scales and can be calculated for both continuous
and categorical variables. As a rule of thumb, an absolute SMD < 0.1 was indicative to achieve a
sufficiently balanced sample.21,22
45
Time-to-Event Analysis
Kaplan Meier (KM) curves were produced to estimate survival distributions of time to
overall survival, major LEA, major LEA-free survival, time to reintervention, and time to
MALE; Mantel-Cox log-rank test was used to compare the survival curves between open bypassfirst and EDVT-first cohorts.
Applying the stabilized IPTW, we first conducted the multivariate Cox Proportional
Hazard (PH) regression with fixed effect to assess the impact of choice of initial
revascularization procedure on the risk of all-cause mortality throughout the follow-up period.
While Cox regression model is semi-parametric with no assumption about baseline hazard
(distribution of survival times), it fundamentally assumes a PH, which means that the covariate
shifts the hazard rate proportionally in any period. The hazard might increase or decrease over
time, but the proportion of each two subjects remain time-invariant. To ensure an unbiased
estimate, we tested and confirmed the PH assumption for each Cox model as the transformed
survival curves are parallel by treatment type.
The model was specified in the Eq. 2 below:
H(t)I = H0(t)i exp[β1* InitialRevascularizationi + β2 * IschemicSeveirityi + + β4*
Cormobiditiesi + β5*Agei + β6*SexIi + β7*Regioni+ … +εi](Eq. 2)
where
• Patients were censored if they were lost to follow-up
• t represents the survival time until the event of death or censoring
• H(t) represents the expected hazard of death at time t
• H0(t) represents the baseline hazard
• the coefficients (β1, β2,...) measure the impact (i.e., the effect size) of covariates.
46
The hazard ratio (HR) for each covariate, equal to exp(β1), and its 95% confidence
interval (CI) was estimated. For interpretation, a HR of 1 suggests lack of association, a HR < 1
suggests a reduction in hazard of major amputation, and a HR > 1 suggests an increase in hazard
of major amputation.
For the analysis of major LEA and reintervention, however, the traditional Cox PH
regression is likely subject to a caveat in the presence of competing risk of all-cause mortality, as
death completely precludes the occurrence of these two events if it occurs. Treating mortality as
censoring assumes that the risk of major LEA/reintervention is equal for patients censored
because of competing event (death) and others, which does not hold true and thus violates the
independent censoring assumption for the Cox PH model to generate valid estimates. Patients
who died at the ‘censoring’ point would never have the possibility of experiencing major LEA or
reintervention. Therefore, we implemented the competing risk regression for these two endpoints
using the Cumulative Incidence Function (CIF) method, developed by Fine and Grey.23 For
example, the incidence of experiencing the event of interest, major LEA, can be expressed as:
𝐼𝐿𝐸𝐴(𝑡𝑓) = 𝑆(𝑡𝑓−1) × ℎ𝐿𝐸𝐴(𝑡𝑓) (Eq.3)
where ℎ𝐿𝐸𝐴(𝑡𝑓) denotes the cause-specific (sub-distributional) hazard of major LEA at time 𝑡𝑓,
and 𝑆(𝑡𝑓−1) denotes the probability of overall survival before the occurrence of event of interest.
As reflected by Eq.3, the subject must have survived to time 𝑡𝑓−1
to have the possibility of major
LEA. The CIF can be simply estimated by cumulating the incidence function of major LEA.
Note that the CIF is always smaller than (1-KM) in the presence of competing risk. Thus, instead
of simply censoring the competing risk event, the CIF estimates the marginal probability of the
competing event, all-cause mortality, to inform the probability of major LEA. The incidence
from the CIF means the probability of developing the event of interest, conditional on not
47
experiencing either event of interest or the competing event. The Fine and Grey modeling
mimics the Cox PH model to some extent by sharing the same general hazard ratio function,
except that it derives the hazard function from the CIF rather the survivor function to incorporate
the competing risk factor. As such, it can be interpreted in the similar but slightly different way
as the probability of undergoing major LEA, conditional on not experiencing major LEA and
death until the time. The IPTW was also implemented in the competing risk analysis.
Subgroup analysis
Additionally, we conduced multivariate subgroup analyses of all-cause mortality, major
LEA, and reintervention, to explore the variation in treatment effect in subgroups of interest. The
subgroups were selected as they were known to have an influence on both treatment selection
and patient outcomes, including age, and diagnosis of gangrene. For age, we categorized patients
to two groups by age of 65.
48
3.3 Results
3.3.1 Patient Sample and Characteristics
A total of 28,608 patients met the inclusion criteria for this analysis. During the study
period, 24,314 (85%) patients underwent EDVT as their first revascularization treatment while
4,294 (15%) received open bypass surgery first after the diagnosis of DFU (Figure 1).
As presented in Table 1, patients treated with EDVT-first were slightly older than those
treated with bypass surgery first (mean age=73.2 years in EDVT-first cohort vs. 72.2 in open
bypass surgery-first cohort, p<0.001). Both cohorts were predominantly male (58.5% in EDVTfirst cohort vs. 60.3% in open bypass-first cohort, p<0.05). There were significantly more nonHispanic whites in the open bypass-first cohort in contrast to the EDVT-first cohort (59.9% vs.
64.7%, p<0.001). As for the pre-operative medication use, Statin was the most used in both
cohorts, but there was a significantly greater P2Y12 use in the EDVT-first cohort (46% vs
34.5%, p<0.001). The prevalence of baseline comorbidities was mixed between the two
treatment groups. Hypertension and hyperlipidemia are two most commonly prevalent
comorbidities in both groups of patients. Greater than 98% of patients had the diagnosis of PAD
before revascularization. More with severe ischemia presentation, proxied by diagnosis of
gangrene, was observed in the open bypass-first cohort than the EDVT-first cohort (34.7% vs.
23.6%, p<0.001). Adjusted by the IPTW, we achieved the baseline covariate balance between
the two treatment groups, as indicated by the absolute SMD < 0.1 (Figure 2).
3.3.2 All-Cause Mortality
The trend of all-cause mortality was similar by the choice of initial revascularization. The
median survival was 3.46 years for EDVT-first cohort, and 3.55 years for open bypass-first
49
cohort (Figure 3). Adjusted for patients baseline covariates using IPTW, there was no
statistically significant difference in the mortality risk between the two treatment alternatives
(HR=0.96, 95% CI: 0.94-1.03). Gangrene, known to be the most severe form of ischemia DFU,
significantly increased one’s mortality by nearly 40% (HR=1.39, 95% CI: 1.32-1.47). The
utilization of pre-operative medicine reduced the mortality risk by 10% (HR=0.89, 95% CI: 0.83-
0.95). No racial disparity was detected in the mortality risk. Other factors including aging and
comorbidities especially congestive heart failure (CHF) and chronic kidney disease (CKD) were
found to be significant risk factors significantly associated with increased mortality regardless of
which procedure patients received (p<0.01, Table 2).
3.3.3 Major LEA and major LEA-Free Survival
Among the subgroup of patients who were followed up for at least 12 months, 12.7%
experienced major LEA in the EDVT-first group as opposed to 17.6% in the open surgery-first
cohort at 12 months (p<0.001). The unadjusted risk of major LEA amongst EDVT-first patients
was significantly smaller than open bypass-first patients (Figure 3). According to the adjusted
competing risk analysis, in which we considered mortality as a competing event of major LEA,
the sub-distributional hazard ratio (SHR) of open bypass-first approach was 1.23, indicating 23%
greater risk of undergoing major amputation than ones treated by EDVT (SHR=1.23, 95% CI:
1.12-1.35, Table 3). Presence of gangrene was another risk factor that doubled DFU patients’
probability of major LEA even after the intervention of revascularization (SHR=2.33, 95% CI:
2.11-2.57). Older patients tend to not get their lower extremity amputated (p<0.01). Additionally,
the major LEA risk was 30% higher among male patients (SHR=1.32, 95% CI: 1.19-1.46).
Compared with White, we noted a significantly higher chance of major LEA for Black
50
(SHR=1.41, 95% CI: 1.24-1.6) and Hispanic patients (SHR=1.2, 95% CI: 1.02-1.4) whereas
Asian American tended to have a more favorable outcome, presenting with 44% lower risk of
major LEA (SHR=0.56, 95% CI: 0.36-0.85). The effect of comorbidities varied. While
hyperlipidemia, chronic obstructive pulmonary disease, and cancer were negatively associated
with major LEA (p<0.05), those with CKD or CHF were more likely to be amputated above the
ankle (p<0.01).
If evaluating mortality and major LEA in tandem, the major LEA-free survival turned out
to be close between the two treatment groups (HR=1.02, 95% CI: 0.98-1.07), likely driven by the
similar all-cause mortality that was much higher than the major LEA rate. The median time of
being free from major LEA and death was 2.78 and 2.9 years for patients treated with EDVT or
open bypass first.
3.3.4 Reintervention and MALE
We observed a greater share of patients in the EDVT-first cohort undergoing at least one
reintervention throughout the follow-up period (Figure 4). The 12-month reintervention rate was
significantly higher in EDVT-first patients than the open bypass-first cohort (59.3% vs. 40.5%,
p<0.001), among the subgroup of patients who were still enrolled. Controlling for the baseline
covariates, one has a significantly 30% lower risk of reintervention if treated by open bypass
other then EDVT first (SHR=0.7, 95% CI: 0.66-0.74, Table 3). DFU severity, proxied by
gangrene, significantly contributed to reintervention regardless of the revascularization
procedure one received (SHR=1.56, 95% CI: 1.49-1.64). Older patients tended to not receive
another round of revascularization (p<0.001).
51
As for the MALE, in which we focused on both major LEA and reintervention, open
bypass-first approach was associated with a lower likelihood of a MALE among the DFU
patients (SHR=0.76, 95% CI: 0.72-0.81). Although adapting the EDVT-first approach
diminished the risk of major LEA, the increase in reintervention probability was stronger,
resulting in an escalated probability of MALE overall.
3.3.5 Subgroup analysis
We examined the variation in treatment effect of alternative revascularizations across
selected subgroups. No statistically significant difference in mortality was found between the
choice of initial revascularization in both younger (aged < 65) and older cohorts (aged > 65).
However, after adjusting for baseline covariates, initiating with open bypass surgery was
associated with greater risk of major LEA to the elderly compared to younger patients
(SHR=1.25 for patients aged > 65, p<0.001 vs. SHR=1.18, p=0.07; Table 4). The reduction in
reintervention risk associated with open bypass was similar to both age groups (p<0.01).
Interestingly, open bypass surgery significantly decreased the mortality risk for patients
with gangrene (HR=0.85, p<0.001), but no survival benefit was observed for patients without
gangrene. While EDVT-first substantially reduced the major LEA risk by 42% for patients
without gangrene (SHR=1.42, p<0.001), this benefit was not observed in patients with gangrene
(SHR=1.04, p=0.58). The lower reintervention risk for those treated with open bypass first was
consistently observed regardless the gangrene status (p<0.001).
52
3.4 Discussion
This retrospective large cohort analysis illuminates the long-term comparative
effectiveness of alternative revascularizations for high-risk DFU patients enrolled in commercialor Medicare Advantage insurance. Although treating with EDVT first did not generate the
benefit in extending life expectancy, it was significantly associated with a reduced likelihood of
major LEA, compared with the open bypass-first approach. The trend in amputation-free survival
was similar between the two treatment groups, which was largely driven by the higher mortality
rate that was observed in both cohorts. Despite fewer than half of the patients experiencing major
LEA during follow-up, the median survival time was 3.46 years for EDVT-first cohort, and 3.55
years for open bypass-first, respectively. The observed high mortality may be attributed to
fragility of the population (mean age > 70 years), where the all-cause mortality is naturally
elevated by age and various risk factors. However, we still highlight the quality-of-life benefit
enhanced by EDVT-first approach through preserving the limbs among the DFU patients. In
other words, patients treated with EDVT first were more likely to survive with an intact limb.
Another notable observation was the diminishing effectiveness of revascularization in
limb salvage for patients diagnosed with severe DFU (gangrene) at baseline, underscoring the
importance of early diagnosis and timely intervention. Our data also revealed the racial disparity
in limb preservation outcomes in the DFU patients – Black and Hispanic patients were
significantly more likely to experience a major LEA even after a revascularization procedure was
performed. For a holistic understanding of treatment quality, we further dissected the probability
of reintervention and MALE after the initial revascularization to measure the burden of
recurrence and unhealed DFU. We found that the likelihood of requiring at least one additional
53
revascularization (reintervention) for those initially treated with EDVT significantly exceeded
that of the open bypass-first group, resulting in a higher overall incidence of MALE.
Our findings contribute robust real-world evidence to supplements the findings of the two
major randomized clinical trials, by incorporating the pivotal clinical and quality-of-life
outcomes from both trials into our analysis. While the two trials focused on the broader CLTI
population requiring revascularization, it is worth comparing with our analyses with trial
findings, given that DFU represents a challenging and complex subgroup of CLTI with increased
risk of adverse outcomes.24 Similar to BASIL-2 trial, our findings advocate for the greater
clinical benefits achieved by the EDVT-first approach. However, differences existed as we
examined the outcomes in detail. The BASIL-2 trial revealed a more favorable major LEA-free
survival associated with the EDVT-first approach, resulting from fewer deaths during the followup. This is different from our results showing similar major LEA-free survival, but a higher risk
of major LEA associated with open bypass. Our results also resonate with the BEST-CLI trial to
some extent, which focused on MALE or death as the primary outcome. While the difference in
all-cause mortality and major LEA was not statistically significant between the two treatment
groups, they reported a greater incidence of MALE or death, as well as reintervention, in the
EDVT-first cohort. The higher reintervention in the EDVT-first cohort was also reported in the
BASIL-2 trial. Thus, our data further confirmed the conclusion regarding the reduced burden of
needing another revascularization to treat the wound if the patient received open bypass surgery
first. Another observation worth noting was that our sample presented a greater mortality than
that in the two clinical trials, which was expected since only 68% and 71.8% patients had the
diagnosis of diabetes in the BASIL2 cohort and the BEST-CLI cohort 1, respectively.
54
Interpreting the variance in results requires consideration of the distinct underlying
patient populations and demographics across trials and our analysis. Subgroup analyses of the
trial data are needed to render comprehensive understanding of impact of EDVT- vs. open
bypass-first strategy on clinical outcomes for the DFU population. Several studies compared the
effectiveness of alternative revascularization procedures in the DFU population or those with
diabetes and PAD in the real-world setting, yielding mixed results11,25-27. However, these studies
are often limited by small sample size, outdated data sources, or insufficient control for treatment
selection bias.
In addition to the trials and existing real-world evidence, our findings are of great value
to support the decision making of key healthcare stakeholders, aiming to enhance the limb
preservation trajectory by informing the choice of revascularization deemed appropriate for
individual DFU patients. In the decision-making process, providers must weigh the reduced
major LEA risk associated with EDVT against the increased probability of reintervention it
presents, compared to an open bypass surgery-first strategy. To clarify which procedure benefits
which patients and aid providers in decision-making, our subgroup analyses evaluated the
variation in treatment effects by age and gangrene diagnosis. For older and fragile patients with
DFU, losing a limb can cause devastating, even fatal, consequences. Compared with those aged
< 65, we found that older patients had a significantly higher risk of major LEA, if starting with
open bypass surgery. In combination with the surgical risks of post-operative infections
associated with bypass surgery, providers should consider prioritizing EDVT for older patients
who tend to be more fragile. The ischemia presentation of the wound has always been critical in
determining the type of initial revascularization procedure. Our analysis demonstrated the
superior effect of open bypass surgery in treating severe DFU patients who were presented with
55
gangrene, which significantly decreased the mortality by 15%. Therefore, providers may
consider using open bypass to treat patients with more extensive and severe disease presentation
when it is feasible.
There is no one-size-fits-all solution as not all patients are the same - a patient-first
strategy should be prioritized. We encourage payers to adopt a pay-for-performance (P4P)
payment model for revascularization to treat DFU. Under this model, payers reward or pay
providers more if they reach the pre-defined quality indicators, such as positive health outcomes
and/or economic measures, after initial revascularization. Key value drivers that are commonly
recognized by the clinical community include preventing minor LEA, major LEA,
hospitalizations due to severe infections, or death within 90 days and at 1 year.28
There are also
P4P programs with penalty design, in which hospitals or providers get penalized for
complications that are considered as preventable, which could be even more effective in
improving quality in surgical care.29,30 However, one of the caveats is patient selection. Providers
may intentionally avoid patients with severe DFU who are more likely to negatively influence
the performance measures. Payers should be cautious in implementing such penalty-based
payment models. P4P model encourages high-value care by incentivizing physicians to adopt
value-based limb preservation strategy, such as choosing the most suitable revascularization
procedure for the individual patient, effective wound management, and consistent monitoring of
patients’ status, leading to more coordinated and patient-centered management. By promoting
the high-quality wound care in DFU, the P4P model would save the downstream healthcare
resources and costs by preventing the occurrence of major LEA, hospitalizations, and other
adverse outcomes in the long run.
56
While our study generated valuable insights that help make an informed revascularization
decision, we acknowledge that the decision-making process is a complex one that involves a
wide set of components, such as patient preference. Most patients prefer the EDVT because it is
less invasive, takes shorter period of times, and typically does not require hospitalization after
the procedure. Although our data demonstrated the benefit of EDVT in reducing the major LEA
risk for DFU patients, open bypass surgery can improve the chance of limb preservation and
survival for patients with severe DFU if they are deemed to be eligible for surgery. Therefore,
physicians and healthcare providers should facilitate informed shared decision-making by
maintaining transparent communication and ensuring patients fully understand their treatment
options, including the associated benefits and risks.
By promoting a value-based decision making in DFU care, it will not only benefit
patients in achieving better clinical outcomes and quality of life, but also leads to efficient
resources allocation with cost savings for the payers and healthcare system by eliminating the
value defects in the DFU management process.
Our study has several limitations. We first recognize that our conclusions do not imply
causality due to retrospective nature of our evaluation. In addition, our data were drawn from the
administrative claims database, which precludes the inclusion of unmeasured patient-level
variables that may have influenced both the treatment decision and patients’ long-term outcomes.
For instance, despite that we have used diagnosis of gangrene to proxy severe DFU, we were
unable to capture the clinically relevant features regarding the vein or systematic disease severity
measure such as the Site, Ischemia, Neuropathy, Bacterial Infection, and Depth (SINBAD)
system, as guided by the IWGDF.
5 Other unobservable factors may include patient preference,
facility volume, and surgeon experience, so we were unable to know the reasons why one
57
treatment was chosen by the provider. Additionally, the administrative claims database does not
allow us to discern whether the second revascularization (reintervention) was performed on the
index limb, potentially inflating the reintervention rate. As this is a relatively old population,
some may have transitioned to the traditional Medicare so that we had a limited follow-up to
understand the long-term survival comprehensively. Furthermore, we did not evaluate other
important outcomes to measure the quality of care, such as morbidity, adverse events, and
healthcare utilization and costs. Lastly, our findings have limited generalizability to the
commercially insured or Medicare Advantage enrollees through the OPTUM system.
Future research is necessary to validate our findings and explore other important value
components, such as morbidity, major adverse events (e.g., infections), and the utilization and
costs of healthcare resources borne by payers and physicians. Beyond the clinical endpoints, we
encourage researchers to conduct qualitative studies to gain a systematic understanding of patient
preference and factors that are viewed important by providers in the decision-making process.
Future research should explore the value of alternative revascularization strategies using other
data bases to expand the generalizability of our findings, such as traditional Medicare claims
data, Medicaid claims data, claims data of other major national private insurers, and electronic
health record or registry data with rich clinical information.
In conclusion, our study added valuable real-world evidence on the long-term
comparative effectiveness of EDVT vs. open bypass surgery in patients with DFU. While no life
extension benefit was observed, EDVT-first approach is associated with higher likelihood of
limb preservation, albeit with an increased likelihood of re-intervention. Healthcare providers
and payers should consider the benefits of the EDVT-first approach when making treatment
decision or designing coverage policies. However, the treatment effect may vary by patients’
58
age, DFU severity, and other factors. Thus, the benefits and risks of each treatment should be
carefully evaluated for individual patients based on their specific situation. Payers should
implement a pay-for-performance model to enhance value-based DFU care.
59
Figure 3.1 Patient selection
Notes: EDVT = endovascular therapy.
Adult patients with confirmed
diagnosis of diabetes in
combination with 1 diagnosis
for foot ulceration, gangrene,
or osteomyelitis between Jan
1, 2010 and Dec 31, 2020.
N= 1,066,663
N=47,424
Exclude patients who did not
receive open bypass surgery or
endovascular therapy post
diagnosis of diabetic foot ulcer.
(95.6%)
N=35,735
Exclude patients who did not
have continuous enrollment for
12 months before the index
date.
(24.6%)
Exclude patients had history of
lower limb amputation before
index date
(19.9%)
N=28,608
EDVT-first
N=24,314
(85%)
Open
surgery-first
N=4,294
(15%)
60
Figure 3.2 Covariate Balance After Inverse Probability Treatment Weighting
Notes: The forest plot visualizes the maximum value of absolute standardized mean difference across
covariate pairs. AMI = acute myocardial infarction, CHF = congestive heart failure; CKD = chronic kidney
disease; COPD = chronic obstructive pulmonary disease; DFU = diabetic foot ulcer; PAD = peripheral arterial
disease.
61
Figure 3.3 Kaplan-Meier curves showing time to primary endpoints after EDVT or open
bypass surgery
Notes: The figures in the first row visualize the unadjusted time to events of interest; the figures in the
second row visualize inverse probability treatment weighted time to events of interest. CI = confidence
interval; IPTW = inverse probability treatment weighted; LEA = lower extremity amputation.
62
Figure 3.4 Kaplan-Meier curves showing time to secondary endpoints after EDVT or open
bypass surgery
Notes: The figures in the first row visualize the unadjusted time to events of interest; the figures in the
second row visualize inverse probability treatment weighted time to events of interest. CI = confidence
interval; IPTW = inverse probability treatment weighted; MALE = major adverse limb events.
63
Table 3.1 Baseline Characteristics by Initial Revascularization Strategy
Measure
EDVT-first Open Surgery-first
P value
N % N %
Unique Patients 24314 85% 4294 15%
Age, year
Mean 73.2 72.2
SD 9.9 9.6 <0.001
Range 21-90 25-90
Age by group, year
<65 4488 18.5% 913 21.3%
65-74 8196 33.7% 1496 34.8% <0.001
75+ 11630 47.8% 1885 43.9%
Male 14228 58.5% 2588 60.3% 0.03
Race/Ethnicity
Non-Hispanic White 14559 59.9% 2778 64.7%
<0.001
Non-Hispanic Black or African
American 4201 17.3% 759 17.7%
Non-Hispanic Asian 439 1.8% 67 1.6%
Hispanic 3961 16.3% 496 11.6%
Unknown/Missing 1154 4.7% 194 4.5%
Region of care
East North Central 3054 12.6% 530 12.3%
<0.001
East South Central 1111 4.6% 242 5.6%
Middle Atlantic 2778 11.4% 609 14.2%
Mountain 1759 7.2% 297 6.9%
New England 880 3.6% 200 4.7%
Pacific 2543 10.5% 612 14.3%
South Atlantic 5561 22.9% 846 19.7%
West North Central 1572 6.5% 370 8.6%
West South Central 5045 20.7% 586 13.6%
Unknown/Missing 11 0.0% 2 0.0%
Medicare advantage 21259 87.4% 3630 84.5% <0.001
Insurance type
Managed Care 6297 25.9% 1398 32.6%
Preferred Provider Organization 3576 14.7% 722 16.8% <0.001
Other 14441 59.4% 2174 50.6%
Pre-Medicine use
64
Statin 18391 75.6% 3176 74.0% 0.02
Warfarin 3773 15.5% 701 16.3% 0.18
ACE inhibitors 14230 58.5% 2519 58.7% 0.86
P2Y12 11195 46.0% 1483 34.5% <0.001
Xarelto 1069 4.4% 123 2.9% <0.001
Comorbidities
AMI 4832 19.9% 910 21.2% 0.05
CHF 12546 51.6% 2154 50.2% 0.08
CKD 14572 59.9% 2454 57.1% <0.001
COPD 10989 45.2% 2206 51.4% <0.001
Cancer 6560 27.0% 1186 27.6% 0.38
Stroke 9077 37.3% 1821 42.4% <0.001
Hyperlipidemia 21897 90.1% 3842 89.5% 0.24
Hypertension 21583 88.8% 3920 91.3% <0.001
Ischemia presentation
PAD 24005 98.7% 4217 98.2% <0.001
Gangrene 5733 23.6% 1488 34.7% <0.001
Notes: This table presents patients’ baseline characteristics before applying propensity score weighting
approach. ACR = angiotensin-converting enzyme; AMI = acute myocardial infarction; CHF = congestive
heart failure; CKD = chronic kidney disease; COPD = chronic obstructive pulmonary disease; DFU = diabetic
foot ulcer; EDVT = endovascular therapy; PAD = peripheral arterial disease; SD = standard deviation.
65
Table 3.2 Multivariate Analyses of All-Cause Mortality, Major LEA, and Major LEA-Free
survival
Covariate
All-cause mortality Major LEA Major LEA or Death
HR P
value 95% CI SHR P
value 95% CI HR P
value 95% CI
Open bypass vs
EDVT 0.98 0.42 0.93 1.03 1.23 <0.01 1.12 1.35 1.03 0.29 0.98 1.07
Age
65-75 vs. < 65 1.1 <0.05 1.02 1.19 0.75 <0.01 0.65 0.85 1.01 0.84 0.94 1.09
> 75 vs < 65 1.61 <0.01 1.49 1.74 0.72 <0.01 0.62 0.83 1.37 <0.01 1.27 1.47
Male 1.01 0.72 0.96 1.06 1.31 <0.01 1.18 1.45 1.06 <0.01 1.01 1.11
Medicare
enrollment 1.24 <0.01 1.13 1.36 1.55 <0.01 1.29 1.87 1.3 <0.01 1.19 1.41
Insurance type
Other 1.07 <0.05 1.01 1.14 0.97 0.62 0.87 1.09 1.06 0.07 1 1.12
PPO 1.08 0.07 0.99 1.18 1.14 0.15 0.95 1.37 1.1 <0.05 1.01 1.19
Race/ethnicity
Non-Hispanic
Asian vs.
White
1.05 0.64 0.86 1.28 0.56 <0.05 0.36 0.85 0.95 0.61 0.79 1.15
Non-Hispanic
Black vs.
White
0.91 <0.05 0.85 0.98 1.42 <0.01 1.25 1.61 1.04 0.19 0.98 1.11
Hispanic vs.
White 0.89 <0.05 0.82 0.97 1.2 <0.05 1.03 1.4 0.95 0.26 0.88 1.04
Comorbidities
Hypertension 0.98 0.75 0.88 1.1 0.84 0.07 0.7 1.01 0.89 <0.05 0.8 0.99
Hyperlipidemia 0.89 <0.01 0.82 0.96 0.84 <0.05 0.73 0.98 0.87 <0.01 0.81 0.94
Stroke 1.08 <0.01 1.03 1.14 1 0.95 0.91 1.11 1.08 <0.01 1.03 1.13
CKD 1.5 <0.01 1.42 1.58 1.16 <0.05 1.04 1.29 1.4 <0.01 1.33 1.47
CHF 1.71 <0.01 1.62 1.8 1.24 <0.01 1.11 1.38 1.57 <0.01 1.49 1.65
COPD 1.12 <0.01 1.07 1.18 0.85 <0.01 0.77 0.94 1.07 <0.05 1.02 1.12
AMI 1.28 <0.01 1.2 1.36 1.12 0.08 0.99 1.27 1.22 <0.01 1.15 1.29
Cancer 1.09 <0.01 1.03 1.15 0.89 0.05 0.79 1 1.06 <0.05 1.01 1.12
Ischemia
severity
Gangrene 1.39 <0.01 1.32 1.47 2.33 <0.01 2.12 2.56 1.57 <0.01 1.49 1.65
Region
66
East North
Central vs.
West South
Central
1.04 0.41 0.95 1.15 0.96 0.64 0.79 1.16 1.03 0.59 0.94 1.12
East South
Central vs.
West South
Central
1.09 0.17 0.96 1.23 1.31 <0.05 1.05 1.64 1.13 <0.05 1 1.27
Middle
Atlantic vs.
West South
Central
1.02 0.65 0.93 1.13 0.86 0.13 0.72 1.04 0.99 0.8 0.9 1.08
Mountain vs.
West South
Central
1.01 0.89 0.89 1.14 1.13 0.29 0.9 1.42 1.03 0.66 0.91 1.15
New England
vs. West South
Central
1.06 0.37 0.93 1.21 0.87 0.31 0.66 1.14 1.02 0.81 0.9 1.15
Pacific vs.
West South
Central
0.94 0.23 0.86 1.04 0.99 0.91 0.83 1.18 0.97 0.49 0.89 1.06
South Atlantic
vs. West South
Central
1.06 0.21 0.97 1.16 1.06 0.49 0.9 1.26 1.06 0.21 0.97 1.15
West North
Central vs.
West South
Central
1.09 0.14 0.97 1.21 0.89 0.31 0.72 1.11 1.07 0.24 0.96 1.18
Preoperative
medication use 0.89 <0.01 0.83 0.95 0.9 0.12 0.78 1.03 0.9 <0.01 0.84 0.96
Notes: AMI = acute myocardial infarction; CHF = congestive heart failure; CKD = chronic kidney disease;
COPD = chronic obstructive pulmonary disease; EDVT = endovascular therapy; HR = hazard ratio; LEA = lower
extremity amputation; PPO = preferred provider organizations; SHR = sub-distributional hazard ratio.
67
Table 3.3 Multivariate Analyses of Reintervention, and MALE
Covariate
Reintervention MALE
SHR P
value 95% CI SHR P
value 95% CI
Open bypass vs EDVT 0.70 <0.01 0.66 0.74 0.77 <0.01 0.73 0.81
Age
65-75 vs. < 65 1.03 0.43 0.96 1.11 0.95 0.19 0.89 1.02
> 75 vs < 65 0.87 <0.01 0.81 0.95 0.85 <0.01 0.79 0.92
Male 1.05 0.09 0.99 1.11 1.10 <0.01 1.05 1.16
Medicare enrollment 0.92 0.06 0.84 1.01 0.99 0.79 0.91 1.08
Insurance type
Other 0.99 0.66 0.92 1.05 0.98 0.52 0.92 1.04
PPO 1.04 0.46 0.95 1.13 1.05 0.27 0.96 1.15
Race/ethnicity
Non-Hispanic Asian vs.
White 0.97 0.80 0.78 1.21 0.90 0.31 0.73 1.11
Non-Hispanic Black vs.
White 1.01 0.83 0.94 1.09 1.10 <0.05 1.03 1.18
Hispanic vs. White 1.01 0.77 0.93 1.11 1.06 0.21 0.97 1.15
Comorbidities
Hypertension 1.06 0.27 0.96 1.17 0.97 0.58 0.89 1.07
Hyperlipidemia 1.04 0.43 0.95 1.13 1.01 0.78 0.93 1.10
Stroke 1.06 <0.05 1.01 1.12 1.05 0.10 0.99 1.10
CKD 0.91 <0.01 0.86 0.97 0.95 0.09 0.90 1.01
CHF 1.03 0.41 0.97 1.09 1.06 0.06 1.00 1.12
COPD 0.97 0.27 0.92 1.02 0.96 0.09 0.91 1.01
AMI 1.00 0.97 0.93 1.07 1.00 0.92 0.94 1.07
Cancer 0.99 0.82 0.93 1.06 0.96 0.19 0.91 1.02
Ischemia severity
Gangrene 1.14 <0.01 1.08 1.21 1.37 <0.01 1.30 1.44
Region
East North Central vs.
West South Central 0.92 0.12 0.83 1.02 0.93 0.15 0.84 1.03
East South Central vs.
West South Central 1.01 0.86 0.89 1.15 1.08 0.24 0.95 1.22
Middle Atlantic vs. West
South Central 0.94 0.19 0.85 1.03 0.93 0.15 0.85 1.03
Mountain vs. West South
Central 1.01 0.89 0.89 1.14 1.05 0.40 0.94 1.18
New England vs. West
South Central 0.99 0.92 0.86 1.15 0.96 0.54 0.84 1.10
Pacific vs. West South
Central 0.94 0.26 0.85 1.04 0.96 0.40 0.87 1.06
68
South Atlantic vs. West
South Central 0.91 <0.05 0.83 0.99 0.97 0.45 0.89 1.06
West North Central vs.
West South Central 0.97 0.58 0.86 1.09 0.96 0.43 0.86 1.07
Pre-operative medication
use
0.99 0.73 0.91 1.07 0.97 0.40 0.90 1.04
Notes: AMI = acute myocardial infarction; CHF = congestive heart failure; CKD = chronic kidney disease;
COPD = chronic obstructive pulmonary disease; EDVT = endovascular therapy; MALE = major adverse limb
event; PPO = preferred provider organizations; SHR = sub-distributional hazard ratio.
69
Table 3.4 Subgroup Analysis
Subgroup
All-cause mortality Major LEA Reintervention
HR 95% CI SHR 95% CI SHR 95% CI
Aged < 65 1.11 0.98 1.26 1.18 0.98 1.43 0.73** 0.65 0.82
Aged > 65 0.96 0.92 1.02 1.25** 1.13 1.39 0.69** 0.65 0.74
With
gangrene
0.85** 0.89 0.92 1.04 0.91 1.18 0.68** 0.62 0.75
No gangrene 1.04 0.98 1.11 1.42** 1.25 1.61 0.71** 0.66 0.76
Notes: ** p value < 0.001; The presented estimates represent the impact of an open bypass surgery-first approach on
the selected endpoints, adjusted for patients’ baseline covariates. CI = confidence interval; HR = hazard ratio; SHR
= sub-distributional hazard ratio; LEA = lower extremity amputation.
70
Chapter 4: Cost-Effectiveness of Early Endovenous Ablation with
Compression Therapy in Venous Leg Ulcerations for A Medicare
Population
Hanke Zheng, Gregory Magee, Tze-Woei Tan, David Armstrong, William Padula
Placed in verbatim the original version published in JAMA Network Open.
2022;5(12):e2248152.
71
4.1 Introduction
Venous leg ulcers (VLU) are the most common cause of lower extremity ulceration and
are characterized by slow healing trajectory and frequent recurrence. VLU leads to significant
disability, reduced quality of life, and tremendous economic burden.1,2 The U.S. prevalence of
VLU ranges from 0.15% to 0.3%, equating to approximately 600,000 cases per year, and is
higher among women and the elderly.3,4 The estimated annual cost of VLU treatment exceeds
$3.5 billion.5
The traditional standard of care for VLU is compression therapy, which has been
demonstrated as clinically beneficial.6 However, adherence to compression therapy is poor due to
uncomfortableness.7,8 Alternative treatments including surgical interventions have been
proposed. In addition to superficial venous surgery, the minimally invasive endothermal
treatments including endovenous laser ablation, radiofrequency ablation, and mechanochemical
ablation have shown valid effectiveness in healing VLU.9
In a recent study, cyanoacrylate
adhesive ablation was found to be cost-effective compared to surgical stripping for treating
varicose veins from the societal perspective in Spain, considering the opportunity costs of
medical leave.10
The Early Venous Reflux Ablation (EVRA) trial conducted in the U.K. suggests that
early ablation with compression therapy substantially increases the healing rate and reduces the
chance of recurrence of VLU.11,12 Early ablation has also shown to be cost-effective in a long
term from the U.K. healthcare sector perspective.12 Given the uniqueness and complexity of the
U.S. healthcare system, particularly in the Medicare and dual-eligible (i.e. Medicare+Medicaid)
populations, we aim to assess the cost-effectiveness of early endovenous ablation with
compression therapy among the elderly with VLU from the U.S. Medicare perspective. These
72
economic data can be used by payers that participate in Medicare programs to cover and rank
early ablation for VLU with respect to other alternative forms of treatment.
4.2 Methods
4.2.1 Model Overview
Our study analyzed the cost-effectiveness of compression therapy with early vs. deferred
endovenous ablation among VLU patients aged 65 and older from the U.S. Medicare
perspective, following methods prescribed by the U.S. Panel on Cost-effectiveness in Health and
Medicine and the Consolidated Health Economic Evaluation Reporting Standards (CHEERS)
reporting guidelines.13,14 The treatment assignment and the clinical features of the patients
included in the model were based on the U.K. EVRA trial.11,12 Specifically, the early intervention
was defined to receive compression therapy and undergo early endovenous ablation performed
within two weeks after becoming clinically significant. Whereas patients receiving deferred
intervention would receive compression therapy alone, and deferred the ablation until the ulcer
has healed, or after 6 months if the ulcer has not healed.
Patients entered the model with an open VLU for a period of between 6 weeks and 6
months, an ankle–brachial index of 0.8 or higher, and primary or recurrent superficial venous
reflux that is deemed by the treating clinician to be clinically significant. We developed a
Markov model with three mutually exclusive health states, unhealed VLU, post-VLU (healed),
and death, to simulate the disease progression of VLU (Figure 4.1). Patients began in the
unhealed VLU state and could stay unhealed or transition to post-VLU (healed) or death states
based on their assigned transition probabilities.
73
Monthly cycles were used to assess the costs and outcomes associated with the three
health states. The time horizon used for the base case is 3 years. Both costs and health outcomes
were discounted at an annual rate of 3%. All monetary terms were converted to 2021 USD using
the Medical Component of the Consumer Price Index.15 The primary outcomes of the model
included the costs associated with VLU treatment and management, and quality-adjusted life
years (QALYs) gained per patient. These data were used to derive the incremental net monetary
benefits (NMB) at a cost-effectiveness threshold of $100,000 per QALY. This study was
exempted from the institutional review board approval and informed consent because no human
subjects were involved.
4.2.2 Probabilities
Transition probabilities between unhealed VLU and post-VLU (healed) were calculated
based on the healing rate and recurrence rate from the EVRA trial (Table 4.1).11,12,16 The EVRA
trial reported the healing rate at 6- months and 12-months, and the recurrence rate from 1 year up
to 3 years after the treatment initiation. Most of the patients got healed within 6 months [85.6%
in the early ablation group vs. 76.3% in the deferred ablation group]; more got healed within 12
months [93.7% in the early ablation group vs. 85.8% in the deferred ablation group]. The
recurrence rates at 3 years were 24.5% for the early ablation group and 29.9% for the deferred
group. Based on the EVRA trail data, we calculated the monthly between-state transition
probabilities applying the DEALE methods.17 The probability of healing after 12 months was
assumed to be half of that between 6 to 12 months to reflect the reality that some patients might
have smaller chance to heal. Given the insufficient evidence for an increased risk of mortality
74
associated with VLU, we used all-cause mortality for the general population age 65 and older in
the U.S.16
4.2.3 Costs
Direct medical costs associated with VLU treatment were considered in the model (Table
4.2). Specific cost components included costs of endovenous ablation, compression therapy, pain
medication, additional home health, and hospitalization due to infections and complications of
VLU. The Medicare national average reimbursement rates in accordance with the Current
Procedural Terminology (CPT) codes and diagnosis-related groups (DRGs) identified for the
VLU-related medical procedures and services were sourced from the Center for Medicare &
Medicaid Services (CMS) datasets and published literature.18-21 We derived the total costs as the
product of quantity used and the relevant unit costs. The CPT codes were identified to target
costs for the endovenous ablation [36473 & 36474 for mechanochemical ablation, 36475 &
36476 for endovenous radiofrequency, and 36478 & 36479 for endovenous laser].22 In the
EVRA trial, the endovenous treatment decision was left to the discretion of the clinicians, and no
granular data about the specific procedure utilization was reported. Thus, according to the expert
opinion of Magee (2021), we assumed 40% of patients to be treated with mechanochemical
ablation, 40% to receive endovenous radiofrequency, and 20% to receive endovenous laser.
Among them, 10% were assumed to have more than one vein treated on a single extremity. Since
the Medicare reimbursement rate for these three ablation procedures were nearly equivalent
(Table 4.2), the assumption about specific procedure utilization was not expected to result in
great variations in the calculated costs.
75
Costs of compression stocking, outpatient visits associated with debridement and
compression therapy were also considered. For the outpatient visits, we assumed one visit per
week to reflect a typical frequency of visits until the wound healed.20 According to the CMS
reimbursement policy, no compression billing is allowed if debridement occurs.20 Taking the
established data from the existing cost-effectiveness analyses, the estimated likelihood of
debridement per week was 12.5%, and it was assumed to only occur within the first 3 months,
representing a reasonable debridement frequency of the VLU treatment.20,23 After that, only CPT
codes for an established clinic visit for compression were used to generate costs for outpatient
visits. Additionally, 25% of the patients would have home health care one visit per week to
change dressing.20 We used the code C2F2S1 to obtain the cost of home health within a 60-day
episode for Medicare.
We incorporated the costs of hospitalization due to VLU and common infections such as
cellulitis into the model.24-26 For patients with unhealed VLU, the annual rate of hospitalization
for VLU visits and infections were referenced from a cost-effectiveness analysis assessing the
management of chronic VLU, which translated to a monthly probability of 0.83% for
hospitalization due to skin debridement with complications and/or due to skin ulcer with
complications, 0.08% for cellulitis with major complications, and 0.33% for cellulitis without
complications. The DRGs codes for VLU were extracted to derive the hospitalization costs,
including DRGs 571 (skin debridement with complications and comorbidities) and 593 (skin
ulcer with complications and comorbidities).20,21 DRGs 602 and 603 (cellulitis with and without
major complications) were used to determine the hospitalization costs for infections.20,21 The
national average Medicare hospitalization payment was referenced to calculate the costs to the
76
payer.19 We omitted the costs associated with antibiotics for infection because it was low and
negligible.20
Given the fact that patients with VLU commonly exhibit pain caused by the disease, our
study considered the costs of pain management directly associated with VLU, including
Amitriptyline (40%), Gabapentin (10%) and Hydrocodone (50%).20,27 For patients healed from
VLU, we included the costs of compression stockings as it is considered as a standard care in the
management post-VLU.28
4.2.4 Health Utilities
The utilities measuring patients’ quality of life (QOL) were assigned by health state,
based on data reported from the Euro-QOL 5-Domain (EQ5D) index of U.S. nationally
representative QALYs reported by Sullivan and Ghushchyan (2006).29 We took the utility score
of people impacted by chronic ulcer of skin, based on ICD-9 codes 707.0 for the unhealed VLU
state (0.69) (Supplement). The utility score for the healed VLU state (0.75) (Supplement) was
derived from another study that assessed the impact of VLU on QOL in the U.K. population.30
To account for the impact of aging on people’s preference of QOL, we adjusted these utility
scores to estimate the perception of individuals aged between 65 to 74 using the U.S. population
disutility for aging.31
4.2.5 Sensitivity Analyses
To understand the impact of changes in value of a specific parameter on the model
results, univariate sensitivity analysis was performed by varying each parameter within its ± 20%
range. Probabilistic sensitivity analyses (PSA) were carried out to test concurrent uncertainty of
77
the base case results based on model structure, parameter sourcing and sampling simultaneously.
We generated 10,000 Monte Carlo simulations for the PSA by varying all model inputs
according to their given distributions. We then created a cost-effectiveness acceptability curve
using the simulated cases, which allows for the visualization of the likelihood that the early
ablation is cost-effective at varying willingness-to-pay thresholds. The expected value of perfect
information (EVPI) per person was calculated to inform the amount to invest for future research
to eliminate the uncertainty for the recommended optimal strategy.
4.2.6 Budget Impact Analysis
To estimate the monetary impact of implementing early ablation for patients with VLU
from the payer’s perspective, we conducted a budget impact analysis in a hypothetical population
of 1 million members, assuming 1,000 VLU cases among the members having clinical
characteristics as presented in the model. We reported the total budget impact of early
endovenous ablation and the per-member-per-month (PMPM) amount at 1 year, 3 years. and 5
years post the intervention.
4.2.7 Statistical Analysis
Descriptive statistics of costs and QALYs over 1 year and 3 years were presented.
Microsoft Excel, 2016 was used for all statistical calculations, simulations, and figures
production. Data were analyzed from September 2021 to June 2022.
78
4.3 Results
Compression therapy with early endovenous ablation was the dominant option by
yielding more QALYs at a lower cost. The total cost of early intervention was $12,527 and the
total QALYs gained were 2.011 per person at a 3-year time horizon from the U.S. Medicare
perspective (Table 4.3). In contrast, compression therapy with deferred ablation yielded a total
cost of $15,208 and 1.985 QALYs per person. Thus, at the cost-effectiveness threshold of
$100,000/QALY, the incremental NMB of early ablation was $5,226 per person at 3 years. The
base case results of early ablation were dominant at both 1 year and 3 years, with an increasing
incremental NMB with longer time horizon.
4.3.1 Sensitivity Analyses
In the univariate sensitivity analysis, the parameter showing the greatest impact on the
incremental NMB was the probability of healing, followed by the probability of recurrence. If
varying the probability of healing for patients receiving early ablation in the first 6 months
within its + 20% range, the incremental NMB would range from -$17,192 to $27,691 per patient.
Based on the 100,000 cases simulated in the PSA, the average incremental NMB
generated by early intervention is $5,286 per person. The cost-effectiveness acceptability curve
derived from the PSA illustrates a greater likelihood of early ablation being cost-effective
regardless of the cost-effectiveness threshold (Figure 2). At $100,000/QALY thresholds, the
compression with early endovenous ablation is cost-effective in 59.2% of the 100,000 simulated
cases, and this probability drop a bit to 57.4% if applying a $150,000/QALY threshold.
79
At $100,000/QALY, the EVPI was $6,341 per person from the Medicare perspective,
meaning that investing $6,341 in research on each person with VLU would increase our
confidence in the recommendation of optimal strategy.
4.3.2 Budget Impact Analysis
Assuming 1,000 VLU patients in a hypothetical 1-million-member health plan,
compression therapy with early endovenous ablation generated a total cost saving of $636,238 at
1 year and $2,680,246 at 3 years, which was equivalent to a PMPM difference of $0.053 at 1
year and $0.075 at 3 years. Therefore, early endovenous ablation for patients with VLU was
cost-saving from a payer’s perspective,
4.4 Discussion
Our study addressed the unmet need for an economic evaluation of early intervention of
endovenous ablation for patients with VLU. According to our analysis, compression therapy with
early endovenous ablation dominates deferred ablation since it provides care for VLU at a lower
cost and improved patient quality of life over a 3-year period from the Medicare perspective.
This finding is robust as supported by PSA results. These cost-effectiveness results apply
specific to a simulated cohort of individuals aged 65 and older, where the highest concentration
of VLUs occur in the U.S.3, 4 Payers wishing to deliver greater health benefits for their patients,
while saving costs, should place early endovenous ablation at the top of their formulary for VLU
patients. While the calculated budget saving is insignificant, it increases the likelihood for payers
to adopt early ablation without displacing resources for other patients in need.
80
Medicare programs stand to benefit from the findings of the EVRA trial combined with
this economic evaluation. Medicare spends over $1.0 billion per year on the chronic management
of VLUs.32 These findings imply that this spending could be conserved if more interventional
procedures such as early endovenous ablation are undertaken decisively for VLU patients who
are clinically significant to mitigate downstream costs of chronic would care, not to mention
gains in clinical benefits for patients through intact skin.
Health systems that implement these types of early interventions need to be properly
incentivized by Medicare payments and performance measures to be sustainable. Currently,
health systems reap approximately $100 billion per year in chronic wound management because
outcomes such as VLU, pressure injury, diabetic foot ulcer, etc., require chronic wound care in
outpatient and long-term care settings.32,33 Health systems that transition to measures such as
early endovenous ablation may lose money by providing procedures up-front for a fraction of the
cost to avert downstream chronic wound care costs. As a result, Medicare should work with
health systems to increase the reimbursement rate for early endovenous ablation while maintain
its dominance as cost-effective to increase incentives for its use and compensate losses from less
chronic wound care in the long run. Medicare could also develop a pay-for-performance
incentive that would reward health systems that improve population health by effectively
managing the prevention of escalated VLU cases in the chronic phase to leverage value of care,
as has been piloted with other types of health outcomes including cardiovascular disease,
diabetes and cancer.34-37
81
4.5 Limitations
This study has several limitations. First, data on the clinical efficacy of early endovenous
ablation are based on a trial from the U.K. and do not necessarily represent outcomes that pertain
to a U.S. patient population. For example, in the EVRA trial, the average age of the participants
randomized with early intervention is 67 years, and it is 68.9 years among those assigned to
deferred intervention. The U.S. Medicare population is likely to be older than the trial population
as the majority of Medicare beneficiaries are aged 65 and older.38 Second, the EVRA trial
examines clinical benefits of early-ablation compared to delayed ablation with compression
therapy in a controlled setting. Since we know of compression therapy to have low adherence,
the results of this economic evaluation represent a lower-bound since patients in real-world
settings gain fewer clinical benefits from delayed ablation with compression therapy when they
lack adherence. Third, in the EVRA trial, the type of ablation technology to apply was left to the
discretion of the clinical team. Due to the lack of data, assumption was made about the share of
different ablation procedures to account for the ablation-related costs under the U.S. context
based on expert opinion, but this assumption was not expected to result in great variation of the
results because the costs of these ablation procedures borne by Medicare are close (Table 2).
Furthermore, the healing rate after 12 months was assumed to be half of that between month 6-
12 because it was not reported in the EVRA data. We captured the uncertainty of the assumption
in our sensitivity analyses. In addition, the economic model does not control for variability in the
VLU patient population, of which there are many sociodemographic causes. Patients in rural
areas, or patients who are predisposed to health disparities that perpetuate challenges to
accessing specialty care for VLU, may have less predictable outcomes.33,39 However, one could
make that argument that a swift and effective treatment such as early ablation would be more
82
efficient, particularly for individuals facing health disparities, than treatments that require
consistent follow-up in the long run for chronic wound management.
4.6 Conclusions
In this economic evaluation of compression therapy with early endovenous ablation, we
found early ablation to be a cost-effective alternative to delayed ablation with compression
therapy for Medicare patients diagnosed with VLU. Medicare should consider innovative
payment models that increase incentives for health systems to deploy early endovenous ablation
to all eligible VLU patients. Doing so will save on the excessive costs of chronic wound care and
improve clinical benefits for patients that currently face long durations of follow-up and extreme
pain caused by VLU.
83
Table 4.1 Transition Probability Base Case Inputs and Range for Sensitivity Analyses
Parameter Cycle transition probability
(± 20% range) Reference
Compression with Early Ablation
Probability of healing - Month 1 to 6 0.133 (0.106-0.160) Gohel et al,12 2020.
Probability of healing - Month 6 to 12 0.090 (0.072-0.108) Gohel et al,12 2020.
Probability of healing - Month 12+ 0.045 (0.036-0.054) Gohel et al,12 2020.
Probability of recurrence - Month 1-12 0.011 (0.009-0.013) Gohel et al,12 2020.
Probability of recurrence - Month 12-24 0.004 (0.003-0.004) Gohel et al,12 2020.
Probability of recurrence - Month 24-36 0.007 (0.006-0.009) Gohel et al,12 2020.
Compression with Deferred Ablation
Probability of healing - Month 1 to 6 0.119 (0.096-0.143) Gohel et al,12 2020.
Probability of healing - Month 6 to 12 0.065 (0.052-0.078) Gohel et al,12 2020.
Probability of healing - Month 12+ 0.032 (0.026-0.039) Gohel et al,12 2020.
Probability of recurrence - Month 1-12 0.016 (0.013-0.019) Gohel et al,12 2020.
Probability of recurrence - Month 12-24 0.005 (0.004-0.006) Gohel et al,12 2020.
Probability of recurrence - Month 24-36 0.007 (0.005-0.008) Gohel et al,12 2020.
All-Cause Mortality
Mortality - Month 1-12 0.001 (0.0008-0.0012) CDC,16 2017.
Mortality - Month 12-24 0.001 (0.0008-0.0012) CDC,16 2017.
Mortality - Month 24-36 0.001 (0.0008-0.0012) CDC,16 2017.
Notes: CDC = Centers for Disease Control and Prevention
84
Table 4.2 Direct Costs of Venous Leg Ulceration Treatment
Cost parameter CPT/DRG
code
Medicare costs
(± 20% range), $. Reference
Intervention Costs
Endovenous radiofrequency 36475 1,323 (1,059-1,588) CMS, 18 2021.
Radiofrequency added on with
multiple veins treatment 36476 314 (251-377) CMS, 18 2021.
Endovenous laser 36478 1,215 (972-1,458) CMS, 18 2021.
Laser added on with multiple
veins treatment 36479 138 (111-166) CMS, 18 2021.
Mechanochemical ablation 36473 1,448 (1,158-1,737) CMS, 18 2021.
Mechanochemical ablation
added on with multiple veins
treatment 36474 296 (237-356) CMS, 18 2021.
Physician Payment, Facility
Physician, evaluation, initial visit 99203 85 (68-102) CMS, 18 2021.
Physician, debridement, initial
visit 11042 63 (51-76) CMS, 18 2021.
Physician, debridement,
established visit 97597 36 (29-44) CMS, 18 2021.
Physician, compression only 99212 36 (29-44) CMS, 18 2021.
Facility Reimbursement
Facility, initial visit 99213 86 (69-104) Nherera et al, 21 2016.
Facility, debridement, initial visit 11042 220 (176-264) Nherera et al, 21 2016.
Facility, debridement,
established visit 97597 114 (91-137) Nherera et al, 21 2016.
Facility, compression only 29581 83 (66-99) Nherera et al, 21 2016.
Home Health
Home health (60-day episode) C2F2S1 2,808 (2,246-3,370) Carter et al, 20 2014
Compression
Compression stocking (per pair
for 6 months) A6532 72 (58-86) Carter et al, 20 2014
Hospitalization costs
Skin debridement with
complication 571 10,832 (8,665-12,998) CMS, 19 2019
Skin ulcer with complication 593 8,882 (7,105-10,658) CMS, 19 2019
Cellulitis (No major
complication) 603 5,562 (4,449-6,674) CMS, 19 2019
Cellulitis (major complication) 602 9,872 (7,898-11,847) CMS, 19 2019
Pain medications (prescription
drugs)
85
Amitriptyline, calculated
monthly cost 43 (35-52) Carter et al, 20 2014
Gabapentin, calculated monthly
cost 124 (99-149) Carter et al, 20 2014
Hydrocodone, calculated
monthly cost 22 (17-26) Carter et al, 20 2014
Notes: CMS = Centers for Medicare & Medicaid Services; CPT = Current Procedural Terminology; DRG =
Diagnosis Related Groups.
86
Table 4.3 Base Case Results
Time
Horizon Treatment Costs, $ Costs
Difference, $. QALYs Incremental
QALYs ICER INMB, $
1 Years Deferred
ablation 8,423 0.699
Early ablation 7,787 -636 0.703 0.004 Dominate 981
3 Years Deferred
ablation 15,208 1.985
Early ablation 12,527 -2,681 2.011 0.026 Dominate 5,226
Notes: ICER = incremental cost-effectiveness ratio; INMB = incremental net monetary benefits; QALYs = qualityadjusted life years.
87
Figure 4.1 Markov Model Simulating Disease Progression of Venous Leg Ulcers
Notes: VLU = venous leg ulcers.
88
Figure 4.2 Cost-Effectiveness Acceptability Curve of Compression Therapy with Early vs.
Deferred Ablation
Notes: QALY = quality adjusted life year.
89
Chapter 5: Conclusions
Hanke Zheng
90
5.1 Aims and Hypothesis Revisited
In this dissertation, we sought to identify and evaluate the defects in value for limb
preservation care in chronic leg ulcerations (CLU), particularly focusing on diabetic foot ulcers
(DFU) and venous leg ulcers (VLU). Due to systematic inefficiency in quality care delivery and
resources allocation, defects in value are common and have caused unnecessary financial waste
in the downstream costs for preventable adverse outcomes, such as major lower extremity
amputation (LEA) and severe infections.1 Thus, there is an urgent need to identify these value
defects to inform key stakeholders and healthcare system to take proactive actions to eliminate
them. Leveraging the novel value defects framework, the dissertation work deepened the
understanding of defects in the field of CLU through the lens of real-world evidence and
economic modeling evaluation, which provided valuable insights that would assist the healthcare
decision making in the efforts of eliminating the suboptimal treatment behaviors.
In Chapter 2, we used the OPTUM’s de-identified Clinformatics® Data Mart Database to
explore the utilization of guideline-consistent limb preservation procedures, including vascular
testing and/or revascularization, among patients with DFU in the year before major LEA. We
revealed a huge gap in utilization of vascular testing and revascularization in real-world setting,
even within the insured population. More than 27% of the DFU patients who underwent a major
LEA did not receive any forms of vascular testing in the year leading up to the amputation.
About 36.6% of identified patients had no attempt of revascularization in the year preceding
major LEA. Our finding is alarming as it represents missed opportunities of preserving their
limbs if patients’ conditions were timely assessed, constantly monitored, and properly treated.
We analyzed the variability by socio-demographics and clinical factors using descriptive and
adjusted methods. Hispanic patients had 46% greater possibility of receiving vascular testing
91
(p<0.01), which may be attributable to the more advanced disease severity in the minority group
that was not captured in the claims database.2 Other factors predicting the pre-LEA vascular
testing include age, diagnosis of gangrene, and baseline comorbidities. The observation
regarding predictors of pre-LEA revascularization was similar to that of pre-LEA vascular
testing. However, we identified an income disparity in the use of revascularization, reflecting the
potential impact of economic barriers on low-income individuals. Unlike vascular testing,
revascularization is not an inexpensive procedure that could still be an unaffordable out-ofpocket cost even in an insured population.3
In Chapter 3, we assessed the value of alternative initial revascularization strategies, open
bypass surgery against endovascular therapy (EDVT), in DFU patients by comparing their longterm clinical outcomes after initial revascularization. The majority of identified patients (85%)
underwent EDVT as the first revascularization. After adjusting for observable confounders
through regression models and inverse probability treatment weighting (IPTW), we found open
bypass surgery-first approach was associated with greater risk of major LEA than EDVT-first
approach, but there was no difference in all-cause mortality by choice of initial revascularization.
However, patients who received EDVT first were more likely to get a reintervention.
Interestingly, the association between revascularization procedures and outcomes varied by
subgroups. The elevated risk of major LEA associated with open bypass surgery became smaller
and insignificant for patients younger than 65. Open bypass generated survival benefits to
patients with gangrene by decreasing the mortality by 15% (p<0.001). While the limb
preservation benefit of EDVT-first became greater for patients without gangrene, nothing was
observed for severe patients with gangrene.
92
In Chapter 4, we sought to evaluate the cost-effectiveness and budget impact of
compression therapy with early intervention of endovenous ablation as opposed to deferred
ablation to treat VLU from the U.S. Medicare perspective, respectively. We concluded that early
ablation was the dominant one by generating more quality-adjusted life years (QALYs) at a
lower cost than deferred ablation. In the base case, the total cost of early intervention was
$12,527 and the total QALYs gained were 2.011 per person at a 3-year time horizon. In contrast,
compression therapy with deferred ablation yielded a total cost of $15,208 and 1.985 QALYs per
person. Thus, at the cost-effectiveness threshold of $100,000/QALY, the incremental NMB of
early ablation was $5,226 per person at 3 years. Our analysis also demonstrated significant cost
savings generated by early ablation for payers (Medicare). By improving the healing trajectory
and preventing VLU recurrence, endovenous ablation generated a total cost saving of $2,680,246
at 3 years, which was equivalent to a per-member-per-month (PMPM) difference of $0.075. Our
findings were robust in the univariate and probability sensitivity analyses.
We expand on the interpretation and policy implication of our findings as final remarks in
the next section of this chapter.
5.2 Final Remarks
The overarching theme of this dissertation work surrounds the defects in value in the
management of CLU, which addresses the phenomenon of healthcare system spending extra
amount of money and healthcare resources on no-value and low-value care, rather than highvalue preventive care. We have been in the vicious cycle in which we pay for unnecessary care
and adverse outcomes (recurrence, major LEA, infections, severe complications, etc) that can be
averted by implementation of patient-centered high-value care. Two thirds of healthcare
expenditure have been wasted on downstream costs derived from defects in value.1
93
We addressed our first aim in Chapter 2, which revealed the underutilization of guidelineconsistent limb salvage care before above-ankle LEA of DFU patients, and the variability in the
utilization by patients’ social determinants of health and their clinical characteristics. It
highlighted the unmet need and underscored the importance for decision makers to continue
making efforts to facilitate the use of guideline-consistent vascular care, particularly for the
disadvantaged population. Healthcare system should enhance education and training programs to
strengthen the awareness of importance of timely vascular testing and appropriate management
for both healthcare professionals and patients. We also encourage the hospitals and vascular care
centers to adopt a proactive approach for timely vascular testing to identify high-risk patients and
intervene with patient-centered care. Healthcare providers should enhance a shared decision
making to ensure patients understand their options and importance of timely vascular testing and
management in preventing major adverse limb events. Payers should revisit the coverage policy
to remove access barriers for high-risk DFU patients to necessary limb preservation care. A payfor-performance (P4P) payment model may be a solution that would move the clinical behaviors
towards high-value care and enhance the adherence of guideline-recommended limb preservation
care, which in the long run would not only benefit patients’ health but also reduce financial
wastes on unnecessary care and suboptimal treatments.
The second aim, addressed in Chapter 3, compared the clinical effectiveness of two
mainstay revascularization strategies in DFU patients at high risk of above-ankle LEA, using
propensity score approach to control for the observable confounding. The findings supplemented
the two recent randomized trials with opposite conclusions by incorporating the key endpoints
from both, thus helping inform the decision making of clinicians, payers, and policy makers from
the real-world evidence perspective. Providers must weigh the reduced major LEA risk
94
associated with EDVT against the increased probability of reintervention it presents, compared
to an open bypass surgery-first strategy. In the decision-making process, patients’ individual
characteristics should also be considered as critical components given the variation in the
association between initial revascularization and outcomes by age, DFU severity, and potentially
other factors. Instead of prioritizing one treatment over the other for all, payers should adopt a
more granular case-by-case reimbursement strategy to incentivize providers to make the most
appropriate and clinically sound treatment recommendation for individual patient. A P4P model,
coupled with pre-defined value indicators such as avoiding major LEA, hospitalizations due to
severe infections, etc., may be viable to improve the clinical practices that achieve more
favorable patient outcomes.
The third aim was an economic evaluation that shed light on the cost-effectiveness of
early intervention of endovenous therapy for Medicare beneficiaries with VLU. It demonstrated
the value of timely intervention of appropriate vascular intervention, such as endovenous
therapy, to prevent reduced quality of life and the downstream healthcare resources utilization
due to disease recurrence and escalation from the Medicare perspective. These findings imply
that this spending could be conserved if more interventional procedures such as early
endovenous ablation are undertaken decisively for VLU patients who are clinically significant to
mitigate downstream costs of chronic would care, not to mention gains in clinical benefits for
patients through intact skin. Health systems that implement these types of early interventions
need to be properly incentivized by Medicare payments and performance measures to be
sustainable. Thus, Medicare should work with health systems to increase the reimbursement rate
for early endovenous ablation while maintain its dominance as cost-effective to increase
incentives for its use and compensate losses from less chronic wound care in the long run.
95
In conclusion, instead of wasting money and resources, we encourage payers and clinical
decision makers to invest upfront on high value care and eliminate suboptimal treatment
behaviors that lead to value defects. As an initial effort, the dissertation recommended a set of
high-value limb salvage procedures to guide the decision making of various stakeholders, which
included but not limited to (1) timely vascular testing and assessment, (2) patient-centered
decision making of revascularization for DFU patients, and (3) early intervention of endovenous
ablation for VLU patients who are clinically eligible. Future research and policy efforts should
aim to eliminate defects in value not only in DFU and VLU but also across broader disease areas.
96
References
Chapter 1 references
1. Agale SV. Chronic Leg Ulcers: Epidemiology, Aetiopathogenesis, and Management.
Ulcers. 2013;2013(1):413604.
2. Cheng CF, Sahu D, Tsen F, et al. A fragment of secreted Hsp90α carries properties that
enable it to accelerate effectively both acute and diabetic wound healing in mice. J Clin
Invest. 2011;121(11):4348-4361.
3. Nunan R, Harding KG, Martin P. Clinical challenges of chronic wounds: searching for an
optimal animal model to recapitulate their complexity. Dis Model Mech.
2014;7(11):1205-1213.
4. Boyko EJ, Zelnick LR, Braffett BH, et al. Risk of Foot Ulcer and Lower-Extremity
Amputation Among Participants in the Diabetes Control and Complications
Trial/Epidemiology of Diabetes Interventions and Complications Study. Diabetes Care.
2022;45(2):357-364.
5. Armstrong DG, Swerdlow MA, Armstrong AA, Conte MS, Padula WV, Bus SA. Five
year mortality and direct costs of care for people with diabetic foot complications are
comparable to cancer. J Foot Ankle Res. 2020;13(1):16.
6. Hinchliffe RJ, Forsythe RO, Apelqvist J, et al. Guidelines on diagnosis, prognosis, and
management of peripheral artery disease in patients with foot ulcers and diabetes
(IWGDF 2019 update). Diabetes Metab Res Rev. 2020;36 Suppl 1:e3276.
7. Meloni M, Morosetti D, Giurato L, et al. Foot Revascularization Avoids Major
Amputation in Persons with Diabetes and Ischaemic Foot Ulcers. J Clin Med.
2021;10(17).
8. Beckman JA, Schneider PA, Conte MS. Advances in Revascularization for Peripheral
Artery Disease: Revascularization in PAD. Circ Res. 2021;128(12):1885-1912.
9. Edwards H, Courtney M, Finlayson K, Shuter P, Lindsay E. A randomised controlled
trial of a community nursing intervention: improved quality of life and healing for clients
with chronic leg ulcers. J Clin Nurs. 2009;18(11):1541-1549.
10. O'Meara S, Al-Kurdi D, Ologun Y, Ovington LG, Martyn-St James M, Richardson R.
Antibiotics and antiseptics for venous leg ulcers. Cochrane Database Syst Rev.
2014;2014(1):Cd003557.
11. Rahman GA, Adigun IA, Fadeyi A. Epidemiology, etiology, and treatment of chronic leg
ulcer: experience with sixty patients. Ann Afr Med. 2010;9(1):1-4.
97
12. Gohel MS, Heatley F, Liu X, et al. A Randomized Trial of Early Endovenous Ablation in
Venous Ulceration. N Engl J Med. 2018;378(22):2105-2114.
13. Nussbaum SR, Carter MJ, Fife CE, et al. An Economic Evaluation of the Impact, Cost,
and Medicare Policy Implications of Chronic Nonhealing Wounds. Value Health.
2018;21(1):27-32.
14. Rice JB, Desai U, Cummings AK, Birnbaum HG, Skornicki M, Parsons N. Burden of
venous leg ulcers in the United States. J Med Econ. 2014;17(5):347-356.
15. Dietz DW, Padula WV, Zheng H, Pronovost PJ. Costs of Defects in Surgical Care: A
Call to Eliminate Defects in Value. NEJM Catalyst Innovations in Care Delivery.
2021;2(6).
16. Tang L, Paravastu SCV, Thomas SD, Tan E, Farmer E, Varcoe RL. Cost Analysis of
Initial Treatment With Endovascular Revascularization, Open Surgery, or Primary Major
Amputation in Patients With Peripheral Artery Disease. J Endovasc Ther.
2018;25(4):504-511.
17. Pearson SD. The Patient Not in the Room. Ann Intern Med. 2021;174(1):109-110.
18. Elwyn G, Fisher E. Higher integrity health care: evidence-based shared decision making.
Circ Cardiovasc Qual Outcomes. 2014;7(6):975-980.
19. Bradbury AW, Moakes CA, Popplewell M, et al. A vein bypass first versus a best
endovascular treatment first revascularisation strategy for patients with chronic limb
threatening ischaemia who required an infra-popliteal, with or without an additional more
proximal infra-inguinal revascularisation procedure to restore limb perfusion (BASIL-2):
an open-label, randomised, multicentre, phase 3 trial. Lancet. 2023;401(10390):1798-
1809.
20. Farber A, Menard MT, Conte MS, et al. Surgery or Endovascular Therapy for Chronic
Limb-Threatening Ischemia. N Engl J Med. 2022;387(25):2305-2316.
21. Forsythe RO, Apelqvist J, Boyko EJ, et al. Effectiveness of revascularisation of the
ulcerated foot in patients with diabetes and peripheral artery disease: A systematic
review. Diabetes Metab Res Rev. 2020;36 Suppl 1:e3279.
22. Darling JD, Bodewes TCF, Deery SE, et al. Outcomes after first-time lower extremity
revascularization for chronic limb-threatening ischemia between patients with and
without diabetes. J Vasc Surg. 2018;67(4):1159-1169.
23. Chang CH, Lin JW, Hsu J, Wu LC, Lai MS. Stent revascularization versus bypass
surgery for peripheral artery disease in type 2 diabetic patients - an instrumental variable
analysis. Sci Rep. 2016;6:37177.
98
24. Lindsay E, Renyi R, Wilkie P, et al. Patient-centred care: a call to action for wound
management. J Wound Care. 2017;26(11):662-677.
25. Wennberg JE. The Dartmouth Atlas of Health Care. In: Supply-Sensitive Care: A
Dartmouth Atlas Project Topic Brief. Lebanon (NH): The Dartmouth Institute for Health
Policy and Clinical Practice © The Trustees of Dartmouth College.; 2007.
26. Wennberg JE, Fisher ES, Skinner JS. Geography and the debate over Medicare reform.
Health Aff (Millwood). 2002;Suppl Web Exclusives:W96-114.
27. Berwick DM, Hackbarth AD. Eliminating waste in US health care. Jama.
2012;307(14):1513-1516.
28. Gray TA, Rhodes S, Atkinson RA, et al. Opportunities for better value wound care: a
multiservice, cross-sectional survey of complex wounds and their care in a UK
community population. BMJ Open. 2018;8(3):e019440.
29. Woods TJ, Tesfay F, Speck P, Kaambwa B. Economic evaluations considering costs and
outcomes of diabetic foot ulcer infections: A systematic review. PLoS One.
2020;15(4):e0232395.
30. Cheng Q, Lazzarini PA, Gibb M, et al. A cost-effectiveness analysis of optimal care for
diabetic foot ulcers in Australia. Int Wound J. 2017;14(4):616-628.
31. Keyhani S, Falk R, Howell EA, Bishop T, Korenstein D. Overuse and systems of care: a
systematic review. Med Care. 2013;51(6):503-508.
32. Ikegami N. Fee-for-service payment - an evil practice that must be stamped out? Int J
Health Policy Manag. 2015;4(2):57-59.
33. Lyu H, Xu T, Brotman D, et al. Overtreatment in the United States. PLoS One.
2017;12(9):e0181970.
34. Zgierska A, Rabago D, Miller MM. Impact of patient satisfaction ratings on physicians
and clinical care. Patient Prefer Adherence. 2014;8:437-446.
Chapter 2 references
1. Boyko EJ, Zelnick LR, Braffett BH, et al. Risk of Foot Ulcer and Lower-Extremity
Amputation Among Participants in the Diabetes Control and Complications
Trial/Epidemiology of Diabetes Interventions and Complications Study. Diabetes Care.
2022;45(2):357-364.
2. Hinchliffe RJ, Forsythe RO, Apelqvist J, et al. Guidelines on diagnosis, prognosis, and
management of peripheral artery disease in patients with foot ulcers and diabetes
(IWGDF 2019 update). Diabetes Metab Res Rev. 2020;36 Suppl 1:e3276.
99
3. Curry SJ, Krist AH, Owens DK, et al. Screening for Peripheral Artery Disease and
Cardiovascular Disease Risk Assessment With the Ankle-Brachial Index: US Preventive
Services Task Force Recommendation Statement. Jama. 2018;320(2):177-183.
4. Holman KH, Henke PK, Dimick JB, Birkmeyer JD. Racial disparities in the use of
revascularization before leg amputation in Medicare patients. J Vasc Surg.
2011;54(2):420-426, 426.e421.
5. Alvidrez J, Castille D, Laude-Sharp M, Rosario A, Tabor D. The National Institute on
Minority Health and Health Disparities Research Framework. Am J Public Health.
2019;109(S1):S16-s20.
6. Vemulapalli S, Greiner MA, Jones WS, Patel MR, Hernandez AF, Curtis LH. Peripheral
arterial testing before lower extremity amputation among Medicare beneficiaries, 2000 to
2010. Circ Cardiovasc Qual Outcomes. 2014;7(1):142-150.
7. Alabi O, Beriwal S, Gallini JW, et al. Association of Health Care Utilization and Access
to Care With Vascular Assessment Before Major Lower Extremity Amputation Among
US Veterans. JAMA Surg. 2023;158(6):e230479.
8. Secemsky EA, Kirksey L, Quiroga E, et al. Impact of Intensity of Vascular Care
Preceding Major Amputation Among Patients With Chronic Limb-Threatening Ischemia.
Circ Cardiovasc Interv. 2024;17(1):e012798.
9. Barnes JA, Eid MA, Creager MA, Goodney PP. Epidemiology and Risk of Amputation
in Patients With Diabetes Mellitus and Peripheral Artery Disease. Arterioscler Thromb
Vasc Biol. 2020;40(8):1808-1817.
10. Gorina Y, Kramarow EA. Identifying chronic conditions in Medicare claims data:
evaluating the Chronic Condition Data Warehouse algorithm. Health Serv Res.
2011;46(5):1610-1627.
11. Fowler XP, Eid MA, Barnes JA, et al. Deriving International Classification of Diseases,
9th and 10th revision, codes for identifying and following up patients with diabetic lower
extremity ulcers. J Vasc Surg Cases Innov Tech. 2022;8(4):877-884.
12. Lee DC, Liang H, Shi L. The convergence of racial and income disparities in health
insurance coverage in the United States. Int J Equity Health. 2021;20(1):96.
13. Goodney PP, Holman K, Henke PK, et al. Regional intensity of vascular care and lower
extremity amputation rates. J Vasc Surg. 2013;57(6):1471-1479, 1480.e1471-1473;
discussion 1479-1480.
14. Tan TW, Calhoun EA, Knapp SM, et al. Rates of Diabetes-Related Major Amputations
Among Racial and Ethnic Minority Adults Following Medicaid Expansion Under the
Patient Protection and Affordable Care Act. JAMA Netw Open. 2022;5(3):e223991.
100
15. Clayton EO, Njoku-Austin C, Scott DM, Cain JD, Hogan MV. Racial and Ethnic
Disparities in the Management of Diabetic Feet. Curr Rev Musculoskelet Med.
2023;16(11):550-556.
16. Perlander A, Nordanstig J, Österberg K, Svensson M. Cost-Effectiveness of
Endovascular Intervention Versus Bypass Surgery in Patients With Chronic LimbThreatening Ischemia and Principal Target Lesion in the Femoropopliteal Segment. J
CRIT LIMB ISCHEM. 2022;2(1):E19-E26.
17. Alabi O, Hunt KJ, Patzer RE, Henry Akintobi T, Massarweh NN. Racial Differences in
Vascular Assessment Prior to Amputation in the Veterans Health Administration. Health
Equity. 2023;7(1):346-350.
18. Varghese JJ, Estes BA, Martinsen BJ, et al. Utilization Rates of Diagnostic and
Therapeutic Vascular Procedures Among Patients Undergoing Lower Extremity
Amputations in a Rural Community Hospital: A Clinicopathological Correlation. Vasc
Endovascular Surg. 2021;55(4):325-331.
19. Kalbaugh CA, Witrick B, Sivaraj LB, et al. Non-Hispanic Black and Hispanic Patients
Have Worse Outcomes Than White Patients Within Similar Stages of Peripheral Artery
Disease. J Am Heart Assoc. 2022;11(1):e023396.
20. Brennan MB, Powell WR, Kaiksow F, et al. Association of Race, Ethnicity, and Rurality
With Major Leg Amputation or Death Among Medicare Beneficiaries Hospitalized With
Diabetic Foot Ulcers. JAMA Netw Open. 2022;5(4):e228399.
21. Mathes T, Pieper D, Morche J, Polus S, Jaschinski T, Eikermann M. Pay for performance
for hospitals. Cochrane Database Syst Rev. 2019;7(7):Cd011156.
Chapter 3 references
1. Boyko EJ, Zelnick LR, Braffett BH, et al. Risk of Foot Ulcer and Lower-Extremity
Amputation Among Participants in the Diabetes Control and Complications
Trial/Epidemiology of Diabetes Interventions and Complications Study. Diabetes Care.
2022;45(2):357-364.
2. Armstrong DG, Swerdlow MA, Armstrong AA, Conte MS, Padula WV, Bus SA. Five
year mortality and direct costs of care for people with diabetic foot complications are
comparable to cancer. J Foot Ankle Res. 2020;13(1):16.
3. Jude EB, Eleftheriadou I, Tentolouris N. Peripheral arterial disease in diabetes--a review.
Diabet Med. 2010;27(1):4-14.
4. Meloni M, Morosetti D, Giurato L, et al. Foot Revascularization Avoids Major
Amputation in Persons with Diabetes and Ischaemic Foot Ulcers. J Clin Med.
2021;10(17).
101
5. Hinchliffe RJ, Forsythe RO, Apelqvist J, et al. Guidelines on diagnosis, prognosis, and
management of peripheral artery disease in patients with foot ulcers and diabetes
(IWGDF 2019 update). Diabetes Metab Res Rev. 2020;36 Suppl 1:e3276.
6. Chang CH, Lin JW, Hsu J, Wu LC, Lai MS. Stent revascularization versus bypass
surgery for peripheral artery disease in type 2 diabetic patients - an instrumental variable
analysis. Sci Rep. 2016;6:37177.
7. Bradbury AW, Moakes CA, Popplewell M, et al. A vein bypass first versus a best
endovascular treatment first revascularisation strategy for patients with chronic limb
threatening ischaemia who required an infra-popliteal, with or without an additional more
proximal infra-inguinal revascularisation procedure to restore limb perfusion (BASIL-2):
an open-label, randomised, multicentre, phase 3 trial. Lancet. 2023;401(10390):1798-
1809.
8. Farber A, Menard MT, Conte MS, et al. Surgery or Endovascular Therapy for Chronic
Limb-Threatening Ischemia. N Engl J Med. 2022;387(25):2305-2316.
9. Conte MS, O'Banion LA. Apples and oranges? A comparison of BEST-CLI to BASIL-2.
J Vasc Surg. 2023;78(2):265-268.
10. Abuhalimeh B, Abumoawad A, Behrooz L, Dominguez YC, Hurtado CM, Regan CJ.
Lessons Learned From the BEST-CLI and BASIL-2 Trials. Published 2024. Accessed
June, 2024.
11. Hinchliffe RJ, Andros G, Apelqvist J, et al. A systematic review of the effectiveness of
revascularization of the ulcerated foot in patients with diabetes and peripheral arterial
disease. Diabetes Metab Res Rev. 2012;28 Suppl 1:179-217.
12. Forsythe RO, Apelqvist J, Boyko EJ, et al. Effectiveness of revascularisation of the
ulcerated foot in patients with diabetes and peripheral artery disease: A systematic
review. Diabetes Metab Res Rev. 2020;36 Suppl 1:e3279.
13. Darling JD, Bodewes TCF, Deery SE, et al. Outcomes after first-time lower extremity
revascularization for chronic limb-threatening ischemia between patients with and
without diabetes. J Vasc Surg. 2018;67(4):1159-1169.
14. Majmundar M, Patel KN, Doshi R, et al. Comparison of 6-Month Outcomes of
Endovascular vs Surgical Revascularization for Patients With Critical Limb Ischemia.
JAMA Netw Open. 2022;5(8):e2227746.
15. Gorina Y, Kramarow EA. Identifying chronic conditions in Medicare claims data:
evaluating the Chronic Condition Data Warehouse algorithm. Health Serv Res.
2011;46(5):1610-1627.
102
16. Fowler XP, Eid MA, Barnes JA, et al. Deriving International Classification of Diseases,
9th and 10th revision, codes for identifying and following up patients with diabetic lower
extremity ulcers. J Vasc Surg Cases Innov Tech. 2022;8(4):877-884.
17. Tan TW, Calhoun EA, Knapp SM, et al. Rates of Diabetes-Related Major Amputations
Among Racial and Ethnic Minority Adults Following Medicaid Expansion Under the
Patient Protection and Affordable Care Act. JAMA Netw Open. 2022;5(3):e223991.
18. Chesnaye NC, Stel VS, Tripepi G, et al. An introduction to inverse probability of
treatment weighting in observational research. Clin Kidney J. 2022;15(1):14-20.
19. Greifer N, Stuart EA. Matching Methods for Confounder Adjustment: An Addition to the
Epidemiologist's Toolbox. Epidemiol Rev. 2022;43(1):118-129.
20. Austin PC, Stuart EA. Moving towards best practice when using inverse probability of
treatment weighting (IPTW) using the propensity score to estimate causal treatment
effects in observational studies. Stat Med. 2015;34(28):3661-3679.
21. Lee BK, Lessler J, Stuart EA. Weight trimming and propensity score weighting. PLoS
One. 2011;6(3):e18174.
22. Stuart EA, Lee BK, Leacy FP. Prognostic score-based balance measures can be a useful
diagnostic for propensity score methods in comparative effectiveness research. J Clin
Epidemiol. 2013;66(8 Suppl):S84-S90.e81.
23. He P, Eriksson F, Scheike TH, Zhang MJ. A Proportional Hazards Regression Model for
the Sub-distribution with Covariates Adjusted Censoring Weight for Competing Risks
Data. Scand Stat Theory Appl. 2016;43(1):103-122.
24. Baubeta Fridh E, Andersson M, Thuresson M, et al. Editor's Choice - Impact of
Comorbidity, Medication, and Gender on Amputation Rate Following Revascularisation
for Chronic Limb Threatening Ischaemia. Eur J Vasc Endovasc Surg. 2018;56(5):681-
688.
25. Butt T, Lilja E, Örneholm H, et al. Amputation-Free Survival in Patients With Diabetes
Mellitus and Peripheral Arterial Disease With Heel Ulcer: Open Versus Endovascular
Surgery. Vasc Endovascular Surg. 2019;53(2):118-125.
26. Meloni M, Izzo V, Giurato L, et al. Recurrence of Critical Limb Ischemia After
Endovascular Intervention in Patients with Diabetic Foot Ulcers. Adv Wound Care (New
Rochelle). 2018;7(6):171-176.
27. Butt T, Lilja E, Elgzyri T, et al. Amputation-free survival in patients with diabetic foot
ulcer and peripheral arterial disease: Endovascular versus open surgery in a propensity
score adjusted analysis. J Diabetes Complications. 2020;34(5):107551.
103
28. Schechter MC, Fayfman M, Khan L, et al. Evaluation of a comprehensive diabetic foot
ulcer care quality model. J Diabetes Complications. 2020;34(4):107516.
29. Kyeremanteng K, Robidoux R, D'Egidio G, Fernando SM, Neilipovitz D. An Analysis of
Pay-for-Performance Schemes and Their Potential Impacts on Health Systems and
Outcomes for Patients. Crit Care Res Pract. 2019;2019:8943972.
30. Kim KM, Max W, White JS, Chapman SA, Muench U. Do penalty-based pay-forperformance programs improve surgical care more effectively than other payment
strategies? A systematic review. Ann Med Surg (Lond). 2020;60:623-630.
Chapter 4 references
1. Edwards H, Courtney M, Finlayson K, Shuter P, Lindsay E. A randomised controlled
trial of a community nursing intervention: improved quality of life and healing for clients
with chronic leg ulcers. J Clin Nurs. 2009;18(11):1541-1549.
2. O'Meara S, Al-Kurdi D, Ologun Y, Ovington LG, Martyn-St James M, Richardson R.
Antibiotics and antiseptics for venous leg ulcers. Cochrane Database Syst Rev.
2014(1):Cd003557.
3. Nelson EA, Adderley U. Venous leg ulcers. BMJ Clin Evid. 2016;2016.
4. Probst S, Weller CD, Bobbink P, et al. Prevalence and incidence of venous leg ulcers-a
protocol for a systematic review. Syst Rev. 2021;10(1):148.
5. Sen CK, Gordillo GM, Roy S, et al. Human skin wounds: a major and snowballing threat
to public health and the economy. Wound Repair Regen. 2009;17(6):763-771.
6. de Araujo T, Valencia I, Federman DG, Kirsner RS. Managing the patient with venous
ulcers. Ann Intern Med. 2003;138(4):326-334.
7. Amsler F, Willenberg T, Blättler W. In search of optimal compression therapy for venous
leg ulcers: a meta-analysis of studies comparing diverse [corrected] bandages with
specifically designed stockings. J Vasc Surg. 2009;50(3):668-674.
8. Dissemond J, Assenheimer B, Bültemann A, et al. Compression therapy in patients with
venous leg ulcers. J Dtsch Dermatol Ges. 2016;14(11):1072-1087.
9. Epstein D, Gohel M, Heatley F, Davies AH. Cost-effectiveness of treatments for
superficial venous reflux in patients with chronic venous ulceration. BJS Open.
2018;2(4):203-212.
10. Vicente-Jiménez S, Lopez-Valcarcel B, Maynar M, et al. Clinical results and costeffectiveness of radiofrequency and cyanoacrylate ablation compared with traditional
surgical stripping for treating varicose veins. J Vasc Surg Venous Lymphat Disord.
2022;10(4):846-854.e842.
104
11. Gohel MS, Heatley F, Liu X, et al. A Randomized Trial of Early Endovenous Ablation in
Venous Ulceration. N Engl J Med. 2018;378(22):2105-2114.
12. Gohel MS, Mora MJ, Szigeti M, et al. Long-term Clinical and Cost-effectiveness of Early
Endovenous Ablation in Venous Ulceration: A Randomized Clinical Trial. JAMA Surg.
2020;155(12):1113-1121.
13. Sanders GD, Neumann PJ, Basu A, et al. Recommendations for Conduct, Methodological
Practices, and Reporting of Cost-effectiveness Analyses: Second Panel on CostEffectiveness in Health and Medicine. Jama. 2016;316(10):1093-1103.
14. Husereau D, Drummond M, Petrou S, et al. Consolidated Health Economic Evaluation
Reporting Standards (CHEERS)--explanation and elaboration: a report of the ISPOR
Health Economic Evaluation Publication Guidelines Good Reporting Practices Task
Force. Value Health. 2013;16(2):231-250.
15. Centers for Disease Control and Prevention. Changes in Life Expectancy at Birth, 2010–
2018. https://www.cdc.gov/nchs/data/hestat/life-expectancy/life-expectancy-2018.htm.
Accessed July 31, 2021.
16. Centers for Disease Control and Prevention. U.S. Life Tables, 2017.
https://www.cdc.gov/nchs/nvss/life-expectancy.htm. Accessed July 30, 2021.
17. Beck JR, Kassirer JP, Pauker SG. A convenient approximation of life expectancy (the
"DEALE"). I. Validation of the method. Am J Med. 1982;73(6):883-888.
18. Centers for Medicare & Medicaid Services. Physician Fee Schedule.
https://www.cms.gov/medicare/physician-feeschedule/search?Y=0&T=4&HT=0&CT=3&H1=99203&M=5. Published 2021.
Accessed Mar 13, 2022.
19. Centers for Medicare & Medicaid Services. The Medicare Inpatient Hospitals by
Geography and Service dataset. https://data.cms.gov/provider-summary-by-type-ofservice/medicare-inpatient-hospitals/medicare-inpatient-hospitals-by-geography-andservice. Published 2019. Accessed July 30, 2021.
20. Carter MJ, Waycaster C, Schaum K, Gilligan AM. Cost-effectiveness of three adjunct
cellular/tissue-derived products used in the management of chronic venous leg ulcers.
Value Health. 2014;17(8):801-813.
21. Nherera LM, Woodmansey E, Trueman P, Gibbons GW. Estimating the Clinical
Outcomes and Cost Differences Between Standard Care With and Without Cadexomer
Iodine in the Management of Chronic Venous Leg Ulcers Using a Markov Model.
Ostomy Wound Manage. 2016;62(6):26-40.
105
22. Center for Medicaid & Medicare Services. Billing and Coding: Treatment of Varicose
Veins of the Lower Extremity. https://www.cms.gov/medicare-coveragedatabase/view/article.aspx?articleId=52870&ver=31. Accessed July 31, 2021.
23. Guest JF, Ruiz FJ, Mihai A, Lehman A. Cost effectiveness of using
carboxymethylcellulose dressing compared with gauze in the management of exuding
venous leg ulcers in Germany and the USA. Curr Med Res Opin. 2005;21(1):81-92.
24. Cox NH. Oedema as a risk factor for multiple episodes of cellulitis/erysipelas of the
lower leg: a series with community follow-up. Br J Dermatol. 2006;155(5):947-950.
25. Pugliese DJ. Infection in Venous Leg Ulcers: Considerations for Optimal Management in
the Elderly. Drugs Aging. 2016;33(2):87-96.
26. Melikian R, O'Donnell TF, Jr., Iafrati M. The economic impact of infection requiring
hospitalization on venous leg ulcers. J Vasc Surg Venous Lymphat Disord.
2022;10(1):96-101.
27. Briggs M, Bennett MI, Closs SJ, Cocks K. Painful leg ulceration: a prospective,
longitudinal cohort study. Wound Repair Regen. 2007;15(2):186-191.
28. Chen WY, Rogers AA. Recent insights into the causes of chronic leg ulceration in venous
diseases and implications on other types of chronic wounds. Wound Repair Regen.
2007;15(4):434-449.
29. Sullivan PW, Ghushchyan V. Preference-Based EQ-5D index scores for chronic
conditions in the United States. Med Decis Making. 2006;26(4):410-420.
30. Iglesias CP, Birks Y, Nelson EA, Scanlon E, Cullum NA. Quality of life of people with
venous leg ulcers: a comparison of the discriminative and responsive characteristics of
two generic and a disease specific instruments. Qual Life Res. 2005;14(7):1705-1718.
31. Jiang R, Janssen MFB, Pickard AS. US population norms for the EQ-5D-5L and
comparison of norms from face-to-face and online samples. Qual Life Res.
2021;30(3):803-816.
32. Nussbaum SR, Carter MJ, Fife CE, et al. An Economic Evaluation of the Impact, Cost,
and Medicare Policy Implications of Chronic Nonhealing Wounds. Value Health.
2018;21(1):27-32.
33. Sen CK. Human Wounds and Its Burden: An Updated Compendium of Estimates. Adv
Wound Care (New Rochelle). 2019;8(2):39-48.
34. Spivack SB, Bernheim SM, Forman HP, Drye EE, Krumholz HM. Hospital
cardiovascular outcome measures in federal pay-for-reporting and pay-for-performance
programs: a brief overview of current efforts. Circ Cardiovasc Qual Outcomes.
2014;7(5):627-633.
106
35. Wennberg JE, O'Connor AM, Collins ED, Weinstein JN. Extending the P4P agenda, part
1: how Medicare can improve patient decision making and reduce unnecessary care.
Health Aff (Millwood). 2007;26(6):1564-1574.
36. White C, Chan C, Huckfeldt PJ, et al. Specialty Payment Model Opportunities and
Assessment: Oncology Simulation Report. Rand Health Q. 2015;5(1):12.
37. Dietz DW, Padula WV, Zheng H, Pronovost PJ. Costs of Defects in Surgical Care: A
Call to Eliminate Defects in Value. NEJM Catalyst Innovations in Care Delivery.
2021;2(6).
38. The Kaiser Family Foudnation. Profile of Medicare Beneficiaries by Race and Ethnicity:
A Chartpack. Published 2016. Accessed July 31, 2021.
39. Sen CK, Roy S. Sociogenomic Approach to Wound Care: A New Patient-Centered
Paradigm. Adv Wound Care (New Rochelle). 2019;8(11):523-526.
Chapter 5 references
1. Dietz DW, Padula WV, Zheng H, Pronovost PJ. Costs of Defects in Surgical Care: A
Call to Eliminate Defects in Value. NEJM Catalyst Innovations in Care Delivery.
2021;2(6).
2. Clayton EO, Njoku-Austin C, Scott DM, Cain JD, Hogan MV. Racial and Ethnic
Disparities in the Management of Diabetic Feet. Curr Rev Musculoskelet Med.
2023;16(11):550-556.
3. Perlander A, Nordanstig J, Österberg K, Svensson M. Cost-Effectiveness of
Endovascular Intervention Versus Bypass Surgery in Patients With Chronic LimbThreatening Ischemia and Principal Target Lesion in the Femoropopliteal Segment. J
CRIT LIMB ISCHEM. 2022;2(1):E19-E26.
Abstract (if available)
Abstract
Value defects in health care are the suboptimal treatment behaviors that needlessly reduce quality of care or lead to unnecessary financial waste to the healthcare system or society. Leveraging an innovative Value Defect Framework on a cost-effectiveness scale, these three papers comprehensively identify the unmet need and assess the value in the field of limb preservation care, focusing on diabetic foot ulcer (DFU) and venous leg ulcer (VLU), which are two leading foot ulcerations that impact millions of lives and result in significant financial toll in the U.S healthcare system and society. The first paper highlights the unmet need and underscores the importance for decision makers to continue making efforts to facilitate the use of guideline-consistent vascular care, particularly for the disadvantaged population with DFU. The second paper compares the clinical effectiveness of two mainstay revascularization strategies in DFU patients who are at high risk of above-ankle amputation, supplementing the two recent randomized trials with opposite conclusions by incorporating the key endpoints from both, thus helping inform the decision making of clinicians, payers, and policy makers from the real-world evidence perspective. The third paper is an economic evaluation that sheds light on the cost-effectiveness of early intervention of endovenous therapy for Medicare beneficiaries with VLU. It demonstrates the value of timely intervention of appropriate vascular intervention, which is endovenous therapy, to prevent reduced quality of life and the downstream healthcare resources utilization due to disease recurrence and escalation from the payers perspective.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Physician behavior and low-value care
PDF
Value in health in the era of vertical integration
PDF
Advances and applications for economic evaluation methods in health technology assessment (HTA)
PDF
Value in oncology care and opportunities for improvement
PDF
The disease burden among aging people living with HIV/AIDS
PDF
Evaluating treatment options for metastatic, castration-resistant prostate cancer: a comprehensive value assessment
PDF
Crohn’s disease: health outcomes and resource utilization in the biologic era
PDF
Disease modifying treatments for Alzeimer's disease: modeling, value assessment and eligible patients
PDF
Three essays on emerging issues in hemophilia care
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
Essays in health economics
PDF
Essays in pharmaceutical and health economics
PDF
Health care utilization and spending of the U.S. aging population
PDF
Assessment and analysis of direct care community worker training: addressing social determinants of health in the home care setting
PDF
Exploring end-of-life care experiences and disparities among Hispanic populations in the United States
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
Expansion of emergency medical service regional routing for stroke patient care in the United States from 2000-2010
PDF
Wearables in healthcare
PDF
Exploring mental health care utilization and academic persistence among college students: evaluating racial and stigma-related disparities
Asset Metadata
Creator
Zheng, Hanke
(author)
Core Title
Assessing value defects in limb preservation care
School
School of Pharmacy
Degree
Doctor of Philosophy
Degree Program
Health Economics
Degree Conferral Date
2024-08
Publication Date
08/13/2024
Defense Date
08/12/2024
Publisher
Los Angeles, California
(original),
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
Health Economics,limb preservation care,value assessment,value defect
Format
theses
(aat)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Padula, William (
committee chair
), Seabury, Seth (
committee member
), Tan, Tze-Woei (
committee member
)
Creator Email
hankezhe@usc.edu,hankezheng1996@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-oUC113998TG0
Unique identifier
UC113998TG0
Identifier
etd-ZhengHanke-13378.pdf (filename)
Legacy Identifier
etd-ZhengHanke-13378
Document Type
Dissertation
Format
theses (aat)
Rights
Zheng, Hanke
Internet Media Type
application/pdf
Type
texts
Source
20240813-usctheses-batch-1197
(batch),
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 author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.
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
Repository Email
cisadmin@lib.usc.edu
Tags
limb preservation care
value assessment
value defect