practice
Wound closure in patients with DFU: a cost-effectiveness analysis of two cellular/tissue-derived products Objective: Determine the cost-effectiveness of extracellular matrix (ECM) relative to human fibroblast-derived dermal substitute (HFDS) on diabetic foot ulcer (DFU) wound closure. l Method: Outcomes data were obtained from a 12-week, randomised, clinical trial of adults aged 18 years or older diagnosed with type 1 or type 2 diabetes with a DFU. Patients were treated with either ECM or HFDS treatment. A two-state Markov model (healed and unhealed) with a 1-week cycle length was developed using wound-closure rates from the trial to estimate the number of closed-wound weeks and the expected DFU cost per patient. Results were recorded over 12 weeks to estimate the number of closed-wound weeks per treatment and the average cost to achieve epithelialisation (primary outcome). The perspective of the analysis was that of the payer, specifically the Centers for Medicare and Medicaid Services. No cost discounting was performed because of the short duration of the study. l Results: The study consisted of 26 patients, with13 in each group. In the ECM group, 10 wounds closed (77%), with an average closure time of 36 days; 11 wounds closed in the HFDS group (85%), with an average closure time of 41 days. There was no significant difference between these results (p=0.73). Over 12 weeks, the expected cost per DFU was $2522 (£1634) for ECM and $3889 (£2524) for HFDS. Patients treated with HFDS incurred total treatment costs that were approximately 54% higher than those treated with ECM. Sensitivity analyses revealed that the total cost of care for two applications of HFDS was more costly than eight applications of ECM by approximately $500 (£325). l Conclusion: In patients with DFU, ECM yielded similar clinical outcomes to HFDS but at a lower cost. Health-care providers should consider ECM as a cost-saving alternative to HFDS. l Declaration of interest: A.M. Gilligan, and C.R. Waycaster, are employees of Smith & Nephew Inc.. This study was funded by Smith & Nephew Inc.. A.L. Landsman, reports no conflicts of interest. l
diabetic foot ulcer; extracellular matrix; human fibroblast-derived dermal substitute; cost-effectiveness
J O U R N A L O F WO U N D C A R E V O L 2 4 , N O 3 , M A R C H 2 0 1 5
tion of advanced biological therapies to standard care for diabetic foot ulcers (DFU) results in improved healing compared with standard care only.10–13 Treatment algorithms for difficult-to-heal DFU acknowledge the usefulness of biological therapies in combination with standard care.14 Two commonly used CTPs are extracellular matrix (ECM) and human fibroblast-derived dermal substitute (HFDS). ECM (OASIS, Cook Biotech Inc., West Lafayette, USA) is a porcine-derived small-intestine submucosa biomaterial that consists primarily of a collagen-based ECM that provides strength, structural support, stability, and cellular attachments to the constantly regenerating mucosal cell layer.15,16 HFDS (Dermagraft, Organogenesis Inc., Canton, USA) is a cryopreserved fibroblast-derived dermal substitute composed of fibroblasts, extracellular matrix, and a bioabsorbable scaffold.17,18 Clinical efficacy for the treatment of DFU has been demonstrated in multiple clinical trials evaluating both therapies.13,19–26 However, many advanced therapies for the treatment of DFU have greater product costs, including initial costs, compared with standard
A.M. Gilligan,1,2 PhD, Manager, Health Economics and Outcomes Research, Adjunct Assistant Professor; C.R. Waycaster, 1,2 PhD, Director, Health Economics and Outcomes Research, Adjunct Assistant Professor; A.L. Landsman,3 DPM, PhD, Assistant Professor of Surgery; 1 Smith & Nephew Inc., Fort Worth, TX 2 University of North Texas Health Sciences Center, Department of Pharmacotherapy, Fort Worth, TX 3 Harvard Medical School, Division of Podiatric Surgery, Cambridge Health Alliance, Cambridge, MA Email: Adrienne.Gilligan@ smith-nephew.com
s
© 2015 MA Healthcare
ltd
C
hronic wounds can enter a state of pathologic inflammation or proliferation.1 The most common feature of chronic wounds is the failure to re-epithelialise, which is due to a failure of keratinocyte migration.2 This may occur because of the lack of a functional extracellular matrix, when fibronectin and collagen molecules needed for cell attachment and migration are deficient.3–6 Diabetes directly affects the normal healing process at the molecular level.7 Multiple key factors contribute to impaired wound healing in patients with diabetes and include inadequate blood supply, infection, impaired fibroblast function, inhibited keratinocyte migration, reduction of growth factors, and absence of normal protein matrix in the dermis.8 Care of difficult-to-heal wounds has improved as a result of advances in science and associated techniques, such as autologous growth factors and the development of recombinant growth factors and cellular/tissue-derived products (CTPs).9 When standard care is inadequate, additional advanced therapies, such as CTPs, may be required. The addi-
149
ournal of Wound Care. Downloaded from magonlinelibrary.com by 130.179.016.201 on September 4, 2015. For personal use only. No other uses without permission. . All rights reserved
practice Wounds that closed in week 12 were rechecked at an additional visit at week 13.
Fig 1. A two-state Markov model
Economic analysis
State 1
State 2
therapy. These costs may be justified if they improve ulcer healing, reduce morbidity, and improve patient functional status. The primary objective of this study was to determine the cost-effectiveness of ECM relative to HFDS on wound closure for the treatment of DFU.
Methods Study participants, design and interventions Data for this economic analysis were derived from a 12-week randomised clinical trial of adults aged 18 years or older with a diagnosis of type 1 or 2 diabetes who had DFU.13 Eligible patients were required to have a neuropathic foot ulcer, from 1cm2 to 16cm2 in area, for a minimum of 4 weeks. Additional inclusion and exclusion criteria are described in Table 1. Patients were examined once weekly for the first 8 weeks and then every other week until wound closure was achieved or for up to 12 weeks. The trial was conducted at outpatient sites in Illinois (1), Virginia (1), New Jersey (1), and Florida (1). All patients provided written informed consent. Due to having four different sites across the US, training of the study investigators was provided in a uniform manner at each location to ensure consistent treatment and application techniques. Data collection was monitored and analysed by Clinical Science Corporation (Skokie, IL, USA) to eliminate potential bias.13 Patients were randomly assigned to either ECM or HFDS for the 12-week treatment phase. Randomisation was performed when the investigative site identified a qualified candidate and contacted an independent site (MED Institute Inc., IN) that randomly assigned patients to one of the two treatment arms. Randomisation was conducted across all sites and the result was a balanced pool of subjects between treatment groups. Each product was applied in accordance with the manufacturer’s recommendations. A maximum of 8 ECM applications in total were permitted. HFDS could be reapplied twice more (total of 3 applications) from baseline at weeks 2 and 4, only if closure had not been achieved in the HFDS treatment arm. Wounds were cleaned, measured, and evaluated at weeks 1, 2, 3, 4, 6, 8, 10, and 12. 150
Model inputs Time horizon A 12-week time horizon was chosen because this was the length of the randomised clinical trial. Costs associated with outpatient treatment continued to accrue until epithelialisation was achieved. l Two-state Markov model and transition probabilities A two-state Markov model with a cycle length of 1 week was chosen to follow the healed (epithelialised) and unhealed stages of DFU (Fig 1). State 1, the unhealed state, represents healing DFU and, consequently, the total costs associated with treatment in the outpatient setting (for example, clinic visits, application of skin substitutes, and facility and physician reimbursement). State 2, the epithelialised phase, represents a closed wound and requires no further dressing or treatment and, consequently, incurs no additional costs. The transition probabilities from the unhealed phase to the healed phase were determined using wound-closure rates.13 These weekly transition rates were used to populate the Markov model and to assess clinical and economic outcomes. Because adverse events were not reported in the trial, additional complications such as infection or oedema were not considered for the Markov analysis. l
ltd
Healed
© 2015 MA Healthcare
Unhealed
Markov models are designed to aid decision-making in clinical situations for events and cost transition over time. The patient(s) are considered to be in one of a finite number of discrete health states that incorporate all clinically important events into transition probabilities from one state to another.27 A Markov model (TreeAge Pro version 2014, TreeAge Software, Inc., Williamstown, Massachusetts, USA) was developed to compare the costs and outcomes of ECM versus HFDS using the wound closure rates to estimate the number of closed-wound weeks and the expected DFU cost per patient. Outputs from the Markov model were then used to derive a costeffectiveness ratio for each treatment group, defined as the expected cost per epithelialised-wound week. Using this approach, results were derived over 12 weeks to estimate the number of closed-wound weeks per treatment, as well as the average cost to achieve epithelialisation (primary outcome). Resource use was based on the treatment regimen used in the trial. Costs were derived from standard cost references and medical supply wholesalers. The number of ECM applications and HFDS grafts were based on the treatment regimen used in the trial. A sensitivity analysis of various applications for each skin substitute was performed to assess changes in expected total costs over 12 weeks.
J O U R N A L O F WO U N D C A R E V O L 2 4 , N O 3 , M A R C H 2 0 1 5
ournal of Wound Care. Downloaded from magonlinelibrary.com by 130.179.016.201 on September 4, 2015. For personal use only. No other uses without permission. . All rights reserved
practice Table 1. Inclusion/exclusion criteria Inclusion
Exclusion
• •
• • • • • • • • • • •
• • • • • •
18 years or older Diagnosed as having type 1 or type 2 diabetes Ulcers that extend through the epidermis and dermis but no exposed tendon or bone Diagnosed with chronic diabetic ulcer Have viable wound bed with granulation tissue as determined by bleeding following debridement Ulcer size ≥1cm2 and ≤16cm2 Signed an informed consent form Wound present for at least 4 weeks at the time of screening
• • • • • • •
Malnourished (defined by albumin 3 seconds at tips of toes) Sickle cell disease Active Charcot’s neuroarthropathy as determined by clinical and/or radiographic examination Having a religious or cultural objection to use of porcine products Enrolled in a clinical evaluation of another investigational device or drug, or have received and investigational treatment for DFU in the last 30 days Unable to comply with the procedures described in the protocol
DFU - diabetic foot ulcer; EDTA - ethylenediaminetetraacetic acid; HFDS - human fibroblast-derived dermal substitute.
Clinical outcomes definition The clinical benefit for the Markov model was defined as ‘epithelialised weeks’ and represents the expected number of weeks that the wound closed over the 12-week trial. Epithelialisation was defined as healing by the growth of epithelium over a denuded surface without any evidence of drainage or bleeding.13 These results were presented as closed-wound weeks to effectively demonstrate the differences in the wound healing trajectories between the two CTP treatment groups. Closed-wound weeks represent the average expected time, in weeks, that DFU remain closed in the two comparative cohorts given their respective transition probabilities from the unhealed state to the healed state. Closedwound weeks are the mathematical complement of open-wound weeks and represent a positive measure of clinical outcomes. l
Economic outcomes definition The analysis was from the perspective of a third-party payer. The Centres for Medicare and Medicaid Services’ maximum allowable costs were used as proxies for assessing the total cumulative cost of care. Only the direct medical costs of care were considered in the economic analysis. All costs were reported in 2014 dollars, and no discounting of costs was conducted because of the 12-week duration of the model. A cost-effectiveness analysis was performed assessing cost per epithelialised-wound week on a per-patient basis. Derivation of costs is displayed in Table 2. Given the recent changes to the Hospital Outpatient Prospective Payment System (HOPPS) rates,28 ECM was considered a low-cost skin substitute and HFDS a high-cost skin substitute. It is important to note that the new HOPPS rates represent a bundled payment of the facility reimbursement and product l
Quantity
Cost
OPPS rate low-cost skin substitute (C5271)*
1 visit/week
$409.41
OPPS rate high-cost skin substitute (CPT 15275)
1 visit/week
$1371.19
OPPS rate hospital outpatient established clinic visit (G0463)*
1 visit/week
$92.53
Physician rate skin substitute application (CPT 15275)
1 visit/week
$102.81
Physician rate evaluation and management visit level 2 (CPT 99212)
1 visit/week
$25.43
*Healthcare Common Procedure Coding System (HCPCS) stated CPT - current procedural terminology; OPPS - Outpatient Prospective Payment System.
152
© 2015 MA Healthcare
Resource
ltd
Table 2. Unit cost table
J O U R N A L O F WO U N D C A R E V O L 2 4 , N O 3 , M A R C H 2 0 1 5
ournal of Wound Care. Downloaded from magonlinelibrary.com by 130.179.016.201 on September 4, 2015. For personal use only. No other uses without permission. . All rights reserved
practice Table 3. Population demographics Characteristic
ECM
HFDS
p-value
Male (%)
76.9
61.5
0.36
Age years, mean (SD)
62.2 (12.2)
63.4 (9.8)
0.78
Average wound size cm , mean (SD)
1.9 (1.8)
1.9 (1.4)
0.99
Average number of dressings, mean (SD)
6.5 (1.4)
2.5 (0.8)
-
2
ECM - extracellular matrix; HFDS - human fibroblast-derived dermal substitute; SD - standard deviation.
At the start of the study, 40 patients were screened. In total, 26 of the 31 patients who met the inclusion criteria completed the study, with 13 in the ECM group and 13 in the HFDS group. Five were lost to follow-up. Demographic and wound characteristics for both treatment groups are described in Table 3. The majority of patients were male (69.2%), and the mean age of the overall patient population was 62.8 years (standard deviation (SD)=11). The average initial wound surface area was 1.9 ± 1.6cm2 (mean ± SD). There was no statistically significant difference in initial average wound size between groups (p=0.94). The average number of ECM applications applied was 6.5±1.4, while the average number of HFDS grafts applied was 2.5±0.8. Fig 2 displays the wound closure rates over the 12-week trial. In the ECM group, 10 wounds closed (77%), with an average closure time of 35.7±41.5 days (95% confidence interval (CI)=13.14–58.21). In the HFDS group, 11 wounds closed (85%), with an average closure J O U R N A L O F WO U N D C A R E V O L 2 4 , N O 3 , M A R C H 2 0 1 5
Economic results Primary outcome On the basis of the transition rates of the prospective clinical trial data, the expected number of epithelialised weeks accumulated over 12 weeks was similar across the two treatment arms (5.3 weeks for ECM versus 6.1 weeks for HFDS) (Fig 3). To provide another perspective, the clinical complement to epithelialised weeks (closedwound weeks) is open-wound weeks. Consequently, the expected number of open-wound weeks for the ECM and HFDS cohorts was estimated at 6.7 and 5.9 weeks, respectively. Therefore, it is important to note that there was an insignificant difference in wound closure (approximately 0.8 weeks) between ECM and HFDS. These results indicate that both treatments provide similar clinical benefits for the management of DFU. The expected costs per ulcer at the end of the 12-week phase were $2522 (£1634) and l
Fig 2. Wound closure rates for ECM and HFDS treatments 90
85%
80
77%
70 60 50 ECM HFDS
40 30 20 10 0 1
2
3
4
5
6
7
8
9
10
11
12
13
Study duration (weeks) ECM - extracellular matrix; HFDS - human fibroblast-derived dermal substitute.
s
© 2015 MA Healthcare
ltd
Results
time of 40.9 ± 32.3 days (95% CI=23.34–58.46). The average difference in time to closure between ECM and HFDS was 5.2 days. No significant difference was observed in the percentages of wound closure or average closure time across the two groups (p=0.73).
Wounds closed (%)
cost. Costs for low- and high-cost skin substitutes, physician reimbursement for application of a skin substitute (CPT 15275), facility reimbursement for an outpatient established clinic visit, and physician evaluation and management visits (level 2), were used per the design of the clinical trial. l Sensitivity analysis Sensitivity analysis is the process of changing the value of an input parameter to assess the magnitude of its effect on the final results of the analysis. One-way sensitivity analyses test the robustness of the model’s assumptions on the results. Deterministic sensitivity analysis assists with justification for the choice of variables included in the model, along with an explanation for the source of ranges used. Variables with uncertainty or expectation of sensitivity were selected for these analyses. In this study, the probability of healing for ECM and HFDS, the costs for the bundled facility and product payment, and the physician reimbursement, were set at ±50% compared with the base case. In addition, the number of CTP applications was varied from 1 to 8 applications because HFDS has not been studied in patients receiving more than 8 applications.29
153
ournal of Wound Care. Downloaded from magonlinelibrary.com by 130.179.016.201 on September 4, 2015. For personal use only. No other uses without permission. . All rights reserved
practice Fig 3. Closed-wound weeks among patients treated with ECM and HFDS 7 6.1 5.3
5 4 ECM HFDS
3 2 1 1
2
3
4
5
6
7
8
9
10
11
12
Study duration (weeks)
Discussion
ECM - extracellular matrix; HFDS - human fibroblast-derived dermal substitute.
$3889 (£2524) for the ECM and HFDS groups, respectively, which equates to a cost savings of $1367 (£887) in favour of ECM (Fig 4). The differences in the product costs between ECM and HFDS led to differences in cost between the two therapies because both treatments yielded similar clinical outcomes. When the total costs of DFU care were estimated over the course of 12 weeks, ECM was more cost-effective than HFDS, indicating that ECM provides similar outcomes as HFDS but at a lower cost. Patients treated with HFDS incurred total treatment costs that were approximately 54% higher than those receiving ECM. The estimated cost per closedwound week was 33% higher for HFDS than for ECM ($633 versus $475, respectively). l Sensitivity analyses The one-way deterministic sensitivity analyses revealed no thresholds where ECM
Cumulative cost in US dollars ($)
Fig 4. Cumulative costs for ECM and HFDS treatments 4000
$3889
3500 3000 $2522
2500 2000
ECM HFDS
1500 1000 500 0
1
2
3
4
5
6
7
8
9
10
Study duration (weeks) ECM - extracellular matrix; HFDS - human fibroblast-derived dermal substitute.
154
11
12
DFU often require substantial healing time and are associated with increased risk for infection, hospitalisation and other events that can result in severe and costly outcomes.30 The economic burden of DFU is composed of several factors, including late management of diabetes, high recurrence and amputation rates, complexity of treatment regimens in patients with osteomyelitis, and high morbidity and mortality rates after amputation.31 The results of this Markov analysis demonstrate that ECM provides similar outcomes to HFDS but at a substantially lower cost, thus reducing the overall cost of DFU therapy. No significant difference was found in time to closure or in the percentage of wound closures between the two treatment groups. There was less than a 1-week difference in wound closure time between HFDS and ECM. The average closure time of approximately 7 weeks found in the clinical data represents a substantial improvement over standard care with saline-moistened gauze, where wound healing has been reported at approximately 50% at 12 weeks.32 Patients treated with HFDS incurred total treatment costs that were approximately 54% higher than those treated with ECM. The overall cost of care for the ECM cohort provided a cost savings of more than $1,360, which translates to $158/closedwound week difference between ECM and HFDS. The substantial difference in the product cost was the primary driver in the cost difference between ECM and HFDS. The sensitivity analysis that varied the number of skin substitute applications revealed that the total cost of care for 2 applications of HFDS was more costly than 8 applications of ECM by approximately $500. These results demonstrate the economic value of ECM for the treatment of DFU in speciality wound care clinics. Per the inclusion/exclusion criteria of this study, the majority of wounds fell into a specific size range,
ltd
0
© 2015 MA Healthcare
Closed-wound weeks
6
lost cost-effectiveness. The most influential variables were the weekly cost of ECM, the probability of healing with ECM, and the weekly cost of HFDS. One of the largest varying costs in advanced DFU therapy is the cost/frequency of skin substitutes. The assumption is that the clinical outcomes remain the same with costs varying because of the number of skin substitutes applied. Fig 5 illustrates the relationship between the number of skin substitutes used and the expected costs for ECM and HFDS groups. Regardless of the number of skin substitutes applied, the total costs for ECM remained lower than the total cost of care for HFDS. For example, the cost of 4 ECM applications was similar to that of 1 HFDS application, even when controlling for additional costs such as physician and facility reimbursement for outpatient clinic visits.
J O U R N A L O F WO U N D C A R E V O L 2 4 , N O 3 , M A R C H 2 0 1 5
ournal of Wound Care. Downloaded from magonlinelibrary.com by 130.179.016.201 on September 4, 2015. For personal use only. No other uses without permission. . All rights reserved
practice Fig 5. Sensitivity analysis: influence of the number of skin substitute applications on cumulative cost of care for wound therapy 7000 Expected cumulative cost of DFU in US dollars ($)
with a wound surface area of ≥1cm2 and ≤16cm2. In a large retrospective study of 26,599 patients with DFU, more than 60% had a mean wound surface area of >0.5cm2 and
Wound closure in patients with DFU: a cost-effectiveness analysis of two cellular/tissue-derived products Objective: Determine the cost-effectiveness of extracellular matrix (ECM) relative to human fibroblast-derived dermal substitute (HFDS) on diabetic foot ulcer (DFU) wound closure. l Method: Outcomes data were obtained from a 12-week, randomised, clinical trial of adults aged 18 years or older diagnosed with type 1 or type 2 diabetes with a DFU. Patients were treated with either ECM or HFDS treatment. A two-state Markov model (healed and unhealed) with a 1-week cycle length was developed using wound-closure rates from the trial to estimate the number of closed-wound weeks and the expected DFU cost per patient. Results were recorded over 12 weeks to estimate the number of closed-wound weeks per treatment and the average cost to achieve epithelialisation (primary outcome). The perspective of the analysis was that of the payer, specifically the Centers for Medicare and Medicaid Services. No cost discounting was performed because of the short duration of the study. l Results: The study consisted of 26 patients, with13 in each group. In the ECM group, 10 wounds closed (77%), with an average closure time of 36 days; 11 wounds closed in the HFDS group (85%), with an average closure time of 41 days. There was no significant difference between these results (p=0.73). Over 12 weeks, the expected cost per DFU was $2522 (£1634) for ECM and $3889 (£2524) for HFDS. Patients treated with HFDS incurred total treatment costs that were approximately 54% higher than those treated with ECM. Sensitivity analyses revealed that the total cost of care for two applications of HFDS was more costly than eight applications of ECM by approximately $500 (£325). l Conclusion: In patients with DFU, ECM yielded similar clinical outcomes to HFDS but at a lower cost. Health-care providers should consider ECM as a cost-saving alternative to HFDS. l Declaration of interest: A.M. Gilligan, and C.R. Waycaster, are employees of Smith & Nephew Inc.. This study was funded by Smith & Nephew Inc.. A.L. Landsman, reports no conflicts of interest. l
diabetic foot ulcer; extracellular matrix; human fibroblast-derived dermal substitute; cost-effectiveness
J O U R N A L O F WO U N D C A R E V O L 2 4 , N O 3 , M A R C H 2 0 1 5
tion of advanced biological therapies to standard care for diabetic foot ulcers (DFU) results in improved healing compared with standard care only.10–13 Treatment algorithms for difficult-to-heal DFU acknowledge the usefulness of biological therapies in combination with standard care.14 Two commonly used CTPs are extracellular matrix (ECM) and human fibroblast-derived dermal substitute (HFDS). ECM (OASIS, Cook Biotech Inc., West Lafayette, USA) is a porcine-derived small-intestine submucosa biomaterial that consists primarily of a collagen-based ECM that provides strength, structural support, stability, and cellular attachments to the constantly regenerating mucosal cell layer.15,16 HFDS (Dermagraft, Organogenesis Inc., Canton, USA) is a cryopreserved fibroblast-derived dermal substitute composed of fibroblasts, extracellular matrix, and a bioabsorbable scaffold.17,18 Clinical efficacy for the treatment of DFU has been demonstrated in multiple clinical trials evaluating both therapies.13,19–26 However, many advanced therapies for the treatment of DFU have greater product costs, including initial costs, compared with standard
A.M. Gilligan,1,2 PhD, Manager, Health Economics and Outcomes Research, Adjunct Assistant Professor; C.R. Waycaster, 1,2 PhD, Director, Health Economics and Outcomes Research, Adjunct Assistant Professor; A.L. Landsman,3 DPM, PhD, Assistant Professor of Surgery; 1 Smith & Nephew Inc., Fort Worth, TX 2 University of North Texas Health Sciences Center, Department of Pharmacotherapy, Fort Worth, TX 3 Harvard Medical School, Division of Podiatric Surgery, Cambridge Health Alliance, Cambridge, MA Email: Adrienne.Gilligan@ smith-nephew.com
s
© 2015 MA Healthcare
ltd
C
hronic wounds can enter a state of pathologic inflammation or proliferation.1 The most common feature of chronic wounds is the failure to re-epithelialise, which is due to a failure of keratinocyte migration.2 This may occur because of the lack of a functional extracellular matrix, when fibronectin and collagen molecules needed for cell attachment and migration are deficient.3–6 Diabetes directly affects the normal healing process at the molecular level.7 Multiple key factors contribute to impaired wound healing in patients with diabetes and include inadequate blood supply, infection, impaired fibroblast function, inhibited keratinocyte migration, reduction of growth factors, and absence of normal protein matrix in the dermis.8 Care of difficult-to-heal wounds has improved as a result of advances in science and associated techniques, such as autologous growth factors and the development of recombinant growth factors and cellular/tissue-derived products (CTPs).9 When standard care is inadequate, additional advanced therapies, such as CTPs, may be required. The addi-
149
ournal of Wound Care. Downloaded from magonlinelibrary.com by 130.179.016.201 on September 4, 2015. For personal use only. No other uses without permission. . All rights reserved
practice Wounds that closed in week 12 were rechecked at an additional visit at week 13.
Fig 1. A two-state Markov model
Economic analysis
State 1
State 2
therapy. These costs may be justified if they improve ulcer healing, reduce morbidity, and improve patient functional status. The primary objective of this study was to determine the cost-effectiveness of ECM relative to HFDS on wound closure for the treatment of DFU.
Methods Study participants, design and interventions Data for this economic analysis were derived from a 12-week randomised clinical trial of adults aged 18 years or older with a diagnosis of type 1 or 2 diabetes who had DFU.13 Eligible patients were required to have a neuropathic foot ulcer, from 1cm2 to 16cm2 in area, for a minimum of 4 weeks. Additional inclusion and exclusion criteria are described in Table 1. Patients were examined once weekly for the first 8 weeks and then every other week until wound closure was achieved or for up to 12 weeks. The trial was conducted at outpatient sites in Illinois (1), Virginia (1), New Jersey (1), and Florida (1). All patients provided written informed consent. Due to having four different sites across the US, training of the study investigators was provided in a uniform manner at each location to ensure consistent treatment and application techniques. Data collection was monitored and analysed by Clinical Science Corporation (Skokie, IL, USA) to eliminate potential bias.13 Patients were randomly assigned to either ECM or HFDS for the 12-week treatment phase. Randomisation was performed when the investigative site identified a qualified candidate and contacted an independent site (MED Institute Inc., IN) that randomly assigned patients to one of the two treatment arms. Randomisation was conducted across all sites and the result was a balanced pool of subjects between treatment groups. Each product was applied in accordance with the manufacturer’s recommendations. A maximum of 8 ECM applications in total were permitted. HFDS could be reapplied twice more (total of 3 applications) from baseline at weeks 2 and 4, only if closure had not been achieved in the HFDS treatment arm. Wounds were cleaned, measured, and evaluated at weeks 1, 2, 3, 4, 6, 8, 10, and 12. 150
Model inputs Time horizon A 12-week time horizon was chosen because this was the length of the randomised clinical trial. Costs associated with outpatient treatment continued to accrue until epithelialisation was achieved. l Two-state Markov model and transition probabilities A two-state Markov model with a cycle length of 1 week was chosen to follow the healed (epithelialised) and unhealed stages of DFU (Fig 1). State 1, the unhealed state, represents healing DFU and, consequently, the total costs associated with treatment in the outpatient setting (for example, clinic visits, application of skin substitutes, and facility and physician reimbursement). State 2, the epithelialised phase, represents a closed wound and requires no further dressing or treatment and, consequently, incurs no additional costs. The transition probabilities from the unhealed phase to the healed phase were determined using wound-closure rates.13 These weekly transition rates were used to populate the Markov model and to assess clinical and economic outcomes. Because adverse events were not reported in the trial, additional complications such as infection or oedema were not considered for the Markov analysis. l
ltd
Healed
© 2015 MA Healthcare
Unhealed
Markov models are designed to aid decision-making in clinical situations for events and cost transition over time. The patient(s) are considered to be in one of a finite number of discrete health states that incorporate all clinically important events into transition probabilities from one state to another.27 A Markov model (TreeAge Pro version 2014, TreeAge Software, Inc., Williamstown, Massachusetts, USA) was developed to compare the costs and outcomes of ECM versus HFDS using the wound closure rates to estimate the number of closed-wound weeks and the expected DFU cost per patient. Outputs from the Markov model were then used to derive a costeffectiveness ratio for each treatment group, defined as the expected cost per epithelialised-wound week. Using this approach, results were derived over 12 weeks to estimate the number of closed-wound weeks per treatment, as well as the average cost to achieve epithelialisation (primary outcome). Resource use was based on the treatment regimen used in the trial. Costs were derived from standard cost references and medical supply wholesalers. The number of ECM applications and HFDS grafts were based on the treatment regimen used in the trial. A sensitivity analysis of various applications for each skin substitute was performed to assess changes in expected total costs over 12 weeks.
J O U R N A L O F WO U N D C A R E V O L 2 4 , N O 3 , M A R C H 2 0 1 5
ournal of Wound Care. Downloaded from magonlinelibrary.com by 130.179.016.201 on September 4, 2015. For personal use only. No other uses without permission. . All rights reserved
practice Table 1. Inclusion/exclusion criteria Inclusion
Exclusion
• •
• • • • • • • • • • •
• • • • • •
18 years or older Diagnosed as having type 1 or type 2 diabetes Ulcers that extend through the epidermis and dermis but no exposed tendon or bone Diagnosed with chronic diabetic ulcer Have viable wound bed with granulation tissue as determined by bleeding following debridement Ulcer size ≥1cm2 and ≤16cm2 Signed an informed consent form Wound present for at least 4 weeks at the time of screening
• • • • • • •
Malnourished (defined by albumin 3 seconds at tips of toes) Sickle cell disease Active Charcot’s neuroarthropathy as determined by clinical and/or radiographic examination Having a religious or cultural objection to use of porcine products Enrolled in a clinical evaluation of another investigational device or drug, or have received and investigational treatment for DFU in the last 30 days Unable to comply with the procedures described in the protocol
DFU - diabetic foot ulcer; EDTA - ethylenediaminetetraacetic acid; HFDS - human fibroblast-derived dermal substitute.
Clinical outcomes definition The clinical benefit for the Markov model was defined as ‘epithelialised weeks’ and represents the expected number of weeks that the wound closed over the 12-week trial. Epithelialisation was defined as healing by the growth of epithelium over a denuded surface without any evidence of drainage or bleeding.13 These results were presented as closed-wound weeks to effectively demonstrate the differences in the wound healing trajectories between the two CTP treatment groups. Closed-wound weeks represent the average expected time, in weeks, that DFU remain closed in the two comparative cohorts given their respective transition probabilities from the unhealed state to the healed state. Closedwound weeks are the mathematical complement of open-wound weeks and represent a positive measure of clinical outcomes. l
Economic outcomes definition The analysis was from the perspective of a third-party payer. The Centres for Medicare and Medicaid Services’ maximum allowable costs were used as proxies for assessing the total cumulative cost of care. Only the direct medical costs of care were considered in the economic analysis. All costs were reported in 2014 dollars, and no discounting of costs was conducted because of the 12-week duration of the model. A cost-effectiveness analysis was performed assessing cost per epithelialised-wound week on a per-patient basis. Derivation of costs is displayed in Table 2. Given the recent changes to the Hospital Outpatient Prospective Payment System (HOPPS) rates,28 ECM was considered a low-cost skin substitute and HFDS a high-cost skin substitute. It is important to note that the new HOPPS rates represent a bundled payment of the facility reimbursement and product l
Quantity
Cost
OPPS rate low-cost skin substitute (C5271)*
1 visit/week
$409.41
OPPS rate high-cost skin substitute (CPT 15275)
1 visit/week
$1371.19
OPPS rate hospital outpatient established clinic visit (G0463)*
1 visit/week
$92.53
Physician rate skin substitute application (CPT 15275)
1 visit/week
$102.81
Physician rate evaluation and management visit level 2 (CPT 99212)
1 visit/week
$25.43
*Healthcare Common Procedure Coding System (HCPCS) stated CPT - current procedural terminology; OPPS - Outpatient Prospective Payment System.
152
© 2015 MA Healthcare
Resource
ltd
Table 2. Unit cost table
J O U R N A L O F WO U N D C A R E V O L 2 4 , N O 3 , M A R C H 2 0 1 5
ournal of Wound Care. Downloaded from magonlinelibrary.com by 130.179.016.201 on September 4, 2015. For personal use only. No other uses without permission. . All rights reserved
practice Table 3. Population demographics Characteristic
ECM
HFDS
p-value
Male (%)
76.9
61.5
0.36
Age years, mean (SD)
62.2 (12.2)
63.4 (9.8)
0.78
Average wound size cm , mean (SD)
1.9 (1.8)
1.9 (1.4)
0.99
Average number of dressings, mean (SD)
6.5 (1.4)
2.5 (0.8)
-
2
ECM - extracellular matrix; HFDS - human fibroblast-derived dermal substitute; SD - standard deviation.
At the start of the study, 40 patients were screened. In total, 26 of the 31 patients who met the inclusion criteria completed the study, with 13 in the ECM group and 13 in the HFDS group. Five were lost to follow-up. Demographic and wound characteristics for both treatment groups are described in Table 3. The majority of patients were male (69.2%), and the mean age of the overall patient population was 62.8 years (standard deviation (SD)=11). The average initial wound surface area was 1.9 ± 1.6cm2 (mean ± SD). There was no statistically significant difference in initial average wound size between groups (p=0.94). The average number of ECM applications applied was 6.5±1.4, while the average number of HFDS grafts applied was 2.5±0.8. Fig 2 displays the wound closure rates over the 12-week trial. In the ECM group, 10 wounds closed (77%), with an average closure time of 35.7±41.5 days (95% confidence interval (CI)=13.14–58.21). In the HFDS group, 11 wounds closed (85%), with an average closure J O U R N A L O F WO U N D C A R E V O L 2 4 , N O 3 , M A R C H 2 0 1 5
Economic results Primary outcome On the basis of the transition rates of the prospective clinical trial data, the expected number of epithelialised weeks accumulated over 12 weeks was similar across the two treatment arms (5.3 weeks for ECM versus 6.1 weeks for HFDS) (Fig 3). To provide another perspective, the clinical complement to epithelialised weeks (closedwound weeks) is open-wound weeks. Consequently, the expected number of open-wound weeks for the ECM and HFDS cohorts was estimated at 6.7 and 5.9 weeks, respectively. Therefore, it is important to note that there was an insignificant difference in wound closure (approximately 0.8 weeks) between ECM and HFDS. These results indicate that both treatments provide similar clinical benefits for the management of DFU. The expected costs per ulcer at the end of the 12-week phase were $2522 (£1634) and l
Fig 2. Wound closure rates for ECM and HFDS treatments 90
85%
80
77%
70 60 50 ECM HFDS
40 30 20 10 0 1
2
3
4
5
6
7
8
9
10
11
12
13
Study duration (weeks) ECM - extracellular matrix; HFDS - human fibroblast-derived dermal substitute.
s
© 2015 MA Healthcare
ltd
Results
time of 40.9 ± 32.3 days (95% CI=23.34–58.46). The average difference in time to closure between ECM and HFDS was 5.2 days. No significant difference was observed in the percentages of wound closure or average closure time across the two groups (p=0.73).
Wounds closed (%)
cost. Costs for low- and high-cost skin substitutes, physician reimbursement for application of a skin substitute (CPT 15275), facility reimbursement for an outpatient established clinic visit, and physician evaluation and management visits (level 2), were used per the design of the clinical trial. l Sensitivity analysis Sensitivity analysis is the process of changing the value of an input parameter to assess the magnitude of its effect on the final results of the analysis. One-way sensitivity analyses test the robustness of the model’s assumptions on the results. Deterministic sensitivity analysis assists with justification for the choice of variables included in the model, along with an explanation for the source of ranges used. Variables with uncertainty or expectation of sensitivity were selected for these analyses. In this study, the probability of healing for ECM and HFDS, the costs for the bundled facility and product payment, and the physician reimbursement, were set at ±50% compared with the base case. In addition, the number of CTP applications was varied from 1 to 8 applications because HFDS has not been studied in patients receiving more than 8 applications.29
153
ournal of Wound Care. Downloaded from magonlinelibrary.com by 130.179.016.201 on September 4, 2015. For personal use only. No other uses without permission. . All rights reserved
practice Fig 3. Closed-wound weeks among patients treated with ECM and HFDS 7 6.1 5.3
5 4 ECM HFDS
3 2 1 1
2
3
4
5
6
7
8
9
10
11
12
Study duration (weeks)
Discussion
ECM - extracellular matrix; HFDS - human fibroblast-derived dermal substitute.
$3889 (£2524) for the ECM and HFDS groups, respectively, which equates to a cost savings of $1367 (£887) in favour of ECM (Fig 4). The differences in the product costs between ECM and HFDS led to differences in cost between the two therapies because both treatments yielded similar clinical outcomes. When the total costs of DFU care were estimated over the course of 12 weeks, ECM was more cost-effective than HFDS, indicating that ECM provides similar outcomes as HFDS but at a lower cost. Patients treated with HFDS incurred total treatment costs that were approximately 54% higher than those receiving ECM. The estimated cost per closedwound week was 33% higher for HFDS than for ECM ($633 versus $475, respectively). l Sensitivity analyses The one-way deterministic sensitivity analyses revealed no thresholds where ECM
Cumulative cost in US dollars ($)
Fig 4. Cumulative costs for ECM and HFDS treatments 4000
$3889
3500 3000 $2522
2500 2000
ECM HFDS
1500 1000 500 0
1
2
3
4
5
6
7
8
9
10
Study duration (weeks) ECM - extracellular matrix; HFDS - human fibroblast-derived dermal substitute.
154
11
12
DFU often require substantial healing time and are associated with increased risk for infection, hospitalisation and other events that can result in severe and costly outcomes.30 The economic burden of DFU is composed of several factors, including late management of diabetes, high recurrence and amputation rates, complexity of treatment regimens in patients with osteomyelitis, and high morbidity and mortality rates after amputation.31 The results of this Markov analysis demonstrate that ECM provides similar outcomes to HFDS but at a substantially lower cost, thus reducing the overall cost of DFU therapy. No significant difference was found in time to closure or in the percentage of wound closures between the two treatment groups. There was less than a 1-week difference in wound closure time between HFDS and ECM. The average closure time of approximately 7 weeks found in the clinical data represents a substantial improvement over standard care with saline-moistened gauze, where wound healing has been reported at approximately 50% at 12 weeks.32 Patients treated with HFDS incurred total treatment costs that were approximately 54% higher than those treated with ECM. The overall cost of care for the ECM cohort provided a cost savings of more than $1,360, which translates to $158/closedwound week difference between ECM and HFDS. The substantial difference in the product cost was the primary driver in the cost difference between ECM and HFDS. The sensitivity analysis that varied the number of skin substitute applications revealed that the total cost of care for 2 applications of HFDS was more costly than 8 applications of ECM by approximately $500. These results demonstrate the economic value of ECM for the treatment of DFU in speciality wound care clinics. Per the inclusion/exclusion criteria of this study, the majority of wounds fell into a specific size range,
ltd
0
© 2015 MA Healthcare
Closed-wound weeks
6
lost cost-effectiveness. The most influential variables were the weekly cost of ECM, the probability of healing with ECM, and the weekly cost of HFDS. One of the largest varying costs in advanced DFU therapy is the cost/frequency of skin substitutes. The assumption is that the clinical outcomes remain the same with costs varying because of the number of skin substitutes applied. Fig 5 illustrates the relationship between the number of skin substitutes used and the expected costs for ECM and HFDS groups. Regardless of the number of skin substitutes applied, the total costs for ECM remained lower than the total cost of care for HFDS. For example, the cost of 4 ECM applications was similar to that of 1 HFDS application, even when controlling for additional costs such as physician and facility reimbursement for outpatient clinic visits.
J O U R N A L O F WO U N D C A R E V O L 2 4 , N O 3 , M A R C H 2 0 1 5
ournal of Wound Care. Downloaded from magonlinelibrary.com by 130.179.016.201 on September 4, 2015. For personal use only. No other uses without permission. . All rights reserved
practice Fig 5. Sensitivity analysis: influence of the number of skin substitute applications on cumulative cost of care for wound therapy 7000 Expected cumulative cost of DFU in US dollars ($)
with a wound surface area of ≥1cm2 and ≤16cm2. In a large retrospective study of 26,599 patients with DFU, more than 60% had a mean wound surface area of >0.5cm2 and
