Cost Effectiveness of Becaplermin Gel on Wound Closure for the Treatment of Pressure Injuries

Objective. This study aims to determine the cost effectiveness of becaplermin gel on wound healing for the treatment of stage 3 and stage 4 pressure injuries (PIs). Materials and Methods. A 2-stage Markov model was used to predict expected costs and outcomes of wound healing for becaplermin gel once daily plus good wound care (BGWC) compared with a placebo gel plus good wound care (control) over 1 year; good wound care consisted of debridement, infection management, and moisture balance. Patients in both arms received dressing changes and gel applications twice daily. Outcome data used in the analysis were derived from a 16-week randomized clinical trial. The primary outcome of interest was PI-free weeks. Transition probabilities for the Markov states were estimated from the clinical trial. Pressure injury recurrence rates were derived from PI literature. Utilization for becaplermin was calculated using the manufacturer’s recommended dosing algorithm. Costs were derived from standard cost references and medical supply wholesalers; economic perspective taken was that of the long-term care facility. Results. A total of 62 patients completed the study: 31 for BGWC and 31 for control. Over 1 year, patients treated with BGWC had substantially higher PI-free weeks compared with control patients (11.6 vs. 3.1, respectively). Patients treated with BGWC incurred higher total costs than those receiving the control treatment. Expected annual direct costs for PI were $3827 for BGWC and $1279 for the control. The incremental cost-effectiveness ratio was $298 (about $43/day), indicating that patients would have to pay an extra $298 to gain 1 additional PI-free week. Conclusions. Becaplermin gel plus good wound care was cost effective over standard of care, yielding better outcomes at a slightly higher cost and should be considered for management of PIs.

Pressure injuries (PIs), recently superseding the term pressure ulcer,¹ pose significant clinical and economic challenges to health systems, patients, providers, and society. The prevalence of PIs in the United States in long-term care (LTC) settings approaches nearly 30%.^2,3 Pressure injuries are associated with higher rates of morbidity and mortality, particularly in the geriatric population.⁴ Patients who develop PIs often have an increased length of stay in hospitals and use significantly more resources than patients without PIs.^5,6 The estimated annual expenditures for PIs are about $11 billion per year.^7,8 

Good wound bed preparation is essential for the management of chronic and nonhealing PIs.⁹ Normal wound healing is a complex process between the extracellular matrix, controlled angiogenesis, epidermal and dermal cells, and soluble mediators.¹⁰ After hemostasis, wound healing involves 3 phases: inflammation, proliferation, and remodeling. Without complications, these phases follow a specific time sequence which yields restoration of anatomic and functional integrity.^11,12 Chronic wounds do not follow this normal sequence and are often prolonged and incomplete, resulting in poor clinical outcomes. These wounds require LTC and are characterized by inflammatory stimuli such as relative ischemia, repeat trauma, and bacterial contamination.¹³ 

Platelet-derived growth factor (PDGF) is a dimeric protein composed of 2 disulfide-linked polypeptide chains.¹⁴ It exists in 3 isoforms: the heterodimer PDGF-AB (consisting of an A and B chain) and 2 homodimers, consisting of 2 A or 2 B chains (PDGF-AA and PDGF-BB, respectively). The homodimer PDGF-BB has been shown in preclinical and clinical studies to promote the formation of granulation tissue at the wound site and to promote wound healing.^15-18

Becaplermin gel (REGRANEX Gel; Smith & Nephew, Fort Worth, TX) is formulated in a preserved sodium carboxymethylcellulose-based gel for topical administration, which may provide the additional benefit of a moist wound healing environment.^18,19 This gel is produced using recombinant DNA technology by insertion of the gene for the B chain of PDGF into the yeast Saccharomyces cerevisiae. Compared with naturally occurring PDGF, becaplermin gel contains similar biological activity. Becaplermin gel was one of the first advanced therapies approved by the US Food and Drug Administration as an adjunct to standard of care for the treatment of lower extremity diabetic neuropathic ulcers that extend into the subcutaneous tissue or beyond and have adequate blood supply.^16,17 Pressure injuries are another type of chronic full-thickness wound, similar to diabetic foot ulcers,²⁰ and represent a potential application for becaplermin gel treatment.

A study by Robson et al²¹ found that becaplermin treatment using doses < 100 µg/mL on chronic stage 3 or 4 PIs have little effect on healing compared with a placebo. However, a phase II clinical trial of 41 patients²² showed that, when expressed as a percentage of initial wound volume, once-daily application of becpalermin gel 100 µg/mL or 300 µg/mL resulted in a decreased wound volume at endpoint, a trend that approached significance when compared with the placebo. In addition, a phase II randomized, double-blind, placebo-controlled study²³ resulted in significantly increased incidence of complete and ≥ 90% healing and significantly reduced the median relative wound volume at endpoint for PIs treated with becaplermin gel 100 µg/mL or 300 µg/mL compared with that of the placebo gel. 

Limited research has been done to assess the impact of becaplermin gel for the management of chronic wounds. It is important for health care providers to assess the efficacy and effectiveness of other methods of treatment such as becaplermin gel to establish how it may be best used in various health care settings. Given the clinical challenges and economic implications surrounding stage 3 and 4 PIs, it is important to assess the cost effectiveness of advanced therapies. The primary objective of this investigation was to assess the cost effectiveness of becaplermin gel compared with a placebo gel for the management of stage 3 and 4 PIs.

Study participants, design, and interventions

Data for this economic analysis were derived from a 16-week randomized clinical trial²³ of adults aged 18 years or older who had at least 1 but no more than 3 chronic full-thickness (stage 3 or 4 as defined by the National Pressure Ulcer Advisory Panel²⁴) PIs (primary or recurrent) without involvement of bone tissue. To be eligible for the study, patients had to have a target PI with a volume between 10 mL and 150 mL, inclusive, following debridement at the baseline visit. Target PIs had to present for at least 4 weeks despite previous treatment and be anatomically located where pressure could be offloaded for the duration of the study.

Patients were randomly assigned to 1 of 2 treatment groups: becaplermin gel 100 µg/g plus good wound care (BGWC) or sodium carboxymethylcellulose gel (placebo gel) plus good wound care (control) twice daily.²³ A thin layer of the study drug (becaplermin gel or placebo gel) was placed on the entire exposed wound surface, then the wound was packed with saline-moistened gauze. The second daily dressing was applied in a similar method, after rinsing the wound surface with saline or water, based on recommended application guidelines.²⁵ Debridement of PIs to remove fibrin and necrotic tissue is an essential component of good wound care and was performed at clinic visits if necessary. Standard of care also included culture of tissue biopsy specimens, systemic treatment of wound infections, offloading pressure from the affected area, maintenance of a moist wound environment, and nutritional support as needed.

Economic analysis

Markov models are designed to aid health care decision-making in clinical situations when health-related events occur and as costs transition over time; they also are a convenient and effective way to model clinical problems for patient populations with ongoing risk, such as patients with PIs. Markov models consider patients to be in one of a finite number of discrete health states. All clinically important health-related events are represented by transitions from one heath state to another. Calculating probabilities (ie, the likelihood they are in a certain health state) to these transitions and including their associated costs allow for a more accurate representation of health states over time.²⁶

A Markov model (TreeAge Pro version 2016; TreeAge Software Inc, Williamstown, MA) was developed to compare the costs and clinical outcomes of becaplermin gel versus placebo using the wound closure and ≥ 90% closure rates from the clinical trial²³ to estimate the number of closed wound weeks, wounds with ≥ 90% closure, and the expected PI cost per patient (primary and secondary outcomes). Outputs from the Markov model were used to determine the expected cost per closed wound week and expected costs for wounds with ≥ 90% closure for each treatment group. Using this approach, results were extrapolated to 1 year to estimate the primary and secondary clinical outcomes per treatment group as well as the average costs over time.

Utilization for becaplermin gel was calculated using the dosing algorithm in the product label.²⁵ Based upon the average wound surface volume (mL) in the becaplermin treatment arm and the recommended dosing formula,²⁵ it was estimated that 3 tubes would be utilized across the time horizon. Costs for becaplermin gel were derived from standard cost references and medical supply wholesalers. Total weekly costs per episode of PI care was estimated from the clinical trial protocol.²³ Because this study was retrospective in nature and used de-identified data from a previously published trial²³ where institutional review board (IRB) approval was obtained at all sites and all patients provided written informed consent, IRB approval for this economic analysis was not required.

Model inputs

Time horizon

Because costs associated with PI treatment continue to accrue until epithelialization or ≥ 90% closure was achieved, a 1-year time horizon was selected to capture the additional costs that occur after completion of the 16-week clinical trial.²³ This time period allowed for adequate modeling of total costs across the entire episode of care for both treatment groups.

Three-state Markov model and transition probabilities

A 3-state Markov model with a cycle length of 1 week was chosen to follow the unhealed, healed, and death stages of a PI. Since results were extrapolated out to 1 year, death was included in order to capture the overall mortality rate of the US population for individuals aged 45 to 54 years, as this was the average age in both treatment groups. State 1, the unhealed state, represents a healing PI and, consequently, all the costs associated with treatment in the LTC setting (ie, debridement, nursing time, and cost of therapies). State 2, the closed state, represents a closed wound requiring no further dressing of treatment and incurring no further costs. State 3, the death state, was defined as a probability of death (per week) set to 0.0000781 on the basis of Centers for Disease Control data for the 2013 annual mortality rate (all causes) for persons aged 45 to 54 years.²⁷

The transition probabilities from the unhealed phase to the closed phase were determined using wound closure rates from the clinical trial.²³ At the end of 16 weeks, about 23% of patients in the BGWC group were considered healed compared with 0% in the control group. Furthermore, at the end of the trial, approximately 58% of patients in the BGWC group had ≥ 90% healing compared with only 29% in the control. These probabilities were extrapolated to 52 weeks, assuming cumulative probability rates over time using equations described by Briggs et al.^28-30 Given that the closure rate at 16 weeks for the control group was 0%, a closure rate of 5% was assumed for modeling purposes. Using this time-dependent Markov model provided a robust method to modeling a chronic illness, since the assumption of constant transition probabilities is considered too conservative for applications in health care.³⁰ These weekly transition rates were used to populate the Markov model and to measure the clinical and economic outcomes. Recurrence was not measured in the clinical trial and therefore not captured. A total of 4 patients (2 in each treatment group) experienced adverse events that were not related to the therapies administered.²³ The majority of events were mild to moderate in severity and were generally consistent with the underlying disease state and age of the patient; therefore, costs associated with severe complications resulting in hospitalization were omitted from the model input.

Clinical outcomes definition

The clinical benefit for the Markov model was defined as “closed wound weeks” or “wounds with ≥ 90% closure” and represents the expected number of weeks that the wound was closed or achieved ≥ 90% closure over the 1-year time horizon. This was presented as closed wound weeks to effectively demonstrate the differences in healing trajectories between the 2 treatment groups. Closed wound weeks represent the average expected time that a PI remains closed in the 2 comparative cohorts given their respective transition probabilities from the unhealed state to the closed (epithelialized) state. Closed wound weeks are the mathematical complement of open wound weeks and represent a positive measure of clinical outcomes. In addition, ≥ 90% healing was measured based on reports in the literature indicating that patients benefit substantially from achieving considerable but incomplete healing.^31-33 Healing to 90%, which largely reflects granulation tissue formation and contraction of the wound, is a relatively rapid process and is followed by a slower progression to epithelialization and resurfacing.³³ The model evaluated 4 separate comparative pathways (2 for healed and 2 for ≥ 90% healing): BGWC healed versus control healed and BGWC ≥ 90% healing versus control ≥ 90% healing.

Economic outcomes definition

The perspective of the analysis was that of the LTC facility. Only the direct medical costs associated with PIs were considered. Results of the 1-year model were reported in US dollars (2016); therefore, no discounting was necessary. A cost-effectiveness analysis was conducted assessing cost per closed wound week and wound with ≥ 90% closure on a per-patient basis. Becaplermin and placebo gels’ cost were based on the 2016 wholesale acquisition cost ($920/tube and $8/tube, respectively). Resource utilization included nursing time, becaplermin gel, placebo gel, saline-moistened gauze, and facility and physician reimbursement (Table 1). The Centers for Medicare and Medicaid Services (CMS) maximum-allowable costs were used as proxies for assessing debridement costs. Resource utilization estimates were derived from the clinical trial.²³ For this analysis, it was assumed that wounds were cleaned and dressed twice daily as they were in the clinical trial and based on recommended application guidelines of twice daily dressing changes for becaplermin.²⁵ 

Sensitivity analysis

Sensitivity analyses test the robustness of the model’s assumption (ie, variables chosen for model input) on the results. Changing the value of an input parameter helps assess the magnitude of its effect on the final results of the analysis. Deterministic sensitivity analysis provides an explanation for the source of ranges used, along with justification for choice of the variables included. In this analysis, the probability of healing for BGWC and the control, the probability of ≥ 90% closure, the probability of death, the costs of becaplermin and placebo gels, LTC clinic visits, and nursing time were included in the sensitivity analysis. Given the uncertainty that exists in rates of healing with wounds of various sizes and shapes,³⁴ a ±50% compared with the base case was utilized to increase the robustness of the model’s results. In addition, the number of tubes to becaplermin gel varied from 1 to 3.³⁵

A probabilistic sensitivity analysis was performed to evaluate parameter uncertainty by using second-order Monte-Carlo simulations of 10 000 trials in which all model inputs were varied simultaneously. This method is comprised of generating a dummy data set by resampling with replacement (ie, randomly selecting 1 patient at a time) from the original data set and repeating this random patient selection until the dummy data set reaches the same size as the original.²⁶

A total of 62 patients were assigned randomly to the treatment groups. No clinically important differences between treatment groups were observed for any of the demographic or baseline efficacy variables (Table 2). Mean patient age was 50 years (standard deviation [SD] = 13.6) for patients treated with the placebo and 48 years (SD = 13.1) for patients treated with BGWC. About 81% and 84% of patients were male in both treatment groups, respectively. The mean wound volume was 19.6 mL (SD = 21.9) for PIs treated with the placebo and 16.6 mL (SD = 15.1) for PIs treated with BGWC. The average duration of the PI was 30 weeks (SD = 43) for the control and 22 weeks (SD = 32) for the BGWC group. 

In the BGWC group, the incidence of complete healing was significantly greater than that of the placebo gel (49.4% vs. 9.7%, respectively; P < .01) over 1 year (Figure 1). A similar difference was observed in the incidence of ≥ 90% healing (82.0% vs. 49.4%, respectively). 

Economic Results

The primary clinical outcome for this economic analysis was closed wound weeks. Based on the transition rate of the prospective clinical trial data, the number of closed wound weeks accumulated over 1 year was higher for the BGWC-treated arm (11.6 weeks for BGWC vs. 3.1 weeks for control, respectively; Figure 2). To provide another perspective, the clinical complement to closed wound weeks is open wound weeks. Consequently, the expected number of open wound weeks for the BGWC and placebo cohorts was estimated at 40.4 weeks and 48.9 weeks, respectively. It is important to note that there is almost a 9-week difference (about 3 months) in wound closure between the BGWC and control groups. Patients receiving the placebo would have, on average, 3 additional open wound months compared with patients treated with BGWC. These results indicate that BGWC is more clinically effective than standard of care alone for the management of chronic stage 3 and 4 PIs. 

Expected costs per PI at the end of the 52-week phase were $3827 and $1279 for the BGWC and control groups, respectively (Table 3), which equates to an incremental cost-effectiveness ratio (ICER) of $298 per closed wound week. If an individual is willing to pay an additional $298 per additional closed wound week (about $43/day), that would provide the patient an additional week of clinical benefit. When analyzing ≥ 90% closure at the end of the 52-week phase, costs were $3152 and $1022 for the BGWC and control groups, respectively, which equates to an ICER of $150 per ≥ 90% healed wound week (about $21/day). While the cost of BGWC is higher regardless of the number of becaplermin tubes used (Figure 3), it is important to note that the clinical benefit of BGWC surpasses that of standard of care alone by providing, on average, an additional 3 months of wound closure.

Sensitivity analyses

One-way deterministic sensitivity analysis revealed no thresholds where the primary findings changed. All model inputs were considered in the sensitivity analysis. The most influential variables were the Outpatient Prospective Payment System reimbursement for a hospital outpatient clinic debridement and the physician reimbursement for debridement. The values of these inputs were derived from Common Procedural Terminology (CPT) codes for 2016 Medicare-allowable total expected costs. 

Probabilistic sensitivity analysis indicated that BGWC had the highest average effectiveness of 11.6 ± 1.6 closed wound weeks whereas the control had the lowest at 3.1 ± 0.9 closed wound weeks. The cost-effectiveness acceptability curve illustrates the probability that any 1 strategy is cost effective as a function of willingness to pay. Given a maximum acceptable ceiling ratio of $880 per closed wound week, the probability that BGWC is cost effective compared with the control is 100%. If a payer’s willingness to pay is $320, the probability that BGWC is cost effective compared with the control is 61.6% (Figure 4). Similar results were observed (at a lower cost) for patients who achieved ≥ 90% closure. 

Pressure injury treatment remains a clinical and economic challenge. There is a critical need for effective treatments that reduce the burden of care and increase patient quality of life. Once-daily treatment of stage 3 and 4 PIs with becaplermin gel significantly increased the incidences of completed (23% vs. 0%) and ≥ 90% healing (56% vs. 27%) at 16 weeks. In contrast, none of the patients treated with placebo gel achieved complete healing by the end of the clinical trial or the extrapolated economic model. The results of this study show that the use of becaplermin gel in difficult-to-heal PIs can accelerate wound closure and that the total estimated 1-year cost of BGWC therapy was slightly higher than the cost of placebo gel. If the payer is willing to pay an additional $400 per additional closed wound week (about $57/day), then the likelihood that becaplermin gel is cost effective increases to more than 83%. Furthermore, if a patient achieves ≥ 90% healing and the payer is willing to spend an additional $280 per additional closed wound week (about $40/day), then the likelihood that becaplermin gel is cost effective increases to more than 95%. Together, these results indicate that BGWC therapy is more clinically efficacious; given the little to no success of closure using placebo gel, BGWC could be considered a cost-effective alternative to standard of care for the management of late-stage PIs.

Consistent with results of clinical studies of becaplermin for the treatment of lower extremity diabetic ulcers,^16,17 treatment of patients with chronic PIs was well tolerated and there were no safety concerns associated with the use of becaplermin gel.²³ Although the cost of adding a 15-g tube of becaplermin as an adjunct to good wound care was approximately 6 times higher than the cost of good wound care alone, the therapeutic effect of BGWC was approximately 70% greater than for the control when wound closure was measured across the entire episode of care. Treatment with BGWC provides 11.6 closed wound weeks compared with 3.1 weeks in the placebo cohort. These 9 additional open wound weeks can influence clinical, economic, and humanistic outcomes. The longer a wound remains open and unhealed, the greater the risk of infection, decreased quality of life, and higher health care utilization and cost. Results from this study demonstrate the economic value of becaplermin gel for the treatment of PIs in LTC settings.

There is limited evidence available to guide clinicians for optimal treatment approaches for slow-to-heal PIs. A health care provider’s decisions on appropriate treatment regimens for PIs are contingent on the patient’s goals and condition, the wound characteristics, the care setting and talents of other team members, and most importantly, evidence of the efficacy, effectiveness, and limitations of the various modalities available. These findings may assist US health providers (including those in LTC facilities), health plans, and patients in assessing the economic effectiveness among patients with stage 3 and 4 PIs who are treated with becaplermin gel to reduce utilization and costs. 

Outside of its labeled indication, becaplermin gel has been demonstrated to be efficacious for the treatment of chronic PIs.^21,23,36 In addition, several analyses and systematic reviews have demonstrated the cost effectiveness of becaplermin both in the United States and Europe for the treatment and management of diabetic neuropathic foot ulcers.^37-43 Differences in health care systems, including reimbursement, relative prices, and treatment practices, vary among countries and affect cost savings globally.^37,44 Health care providers and coverage decision makers should take both the high initial cost of growth factors and the clinical benefit into account when deciding the appropriate allocation of resources.⁴³ At this point, most studies examining the cost effectiveness of advanced therapies have occurred only in Western countries. The effectiveness of advanced biologic therapies (particularly PDGFs) should be evaluated in developing countries, given that the high cost of advanced therapies and the financial constraints in developing nations are not directly generalizable on a global scale.⁴⁵

Certain limitations should be considered when interpreting findings from this study. First, the specific amount of becaplermin use was unknown and was estimated using the recommended dosing formula from the product labeling.²⁵ Second, the results of this study cannot be applied to other standard care methods, health care settings, or treatments for wounds of other etiologies. Third, only direct costs of care were included in the analysis; indirect and intangible costs were not measured. Decision makers in other health care settings should investigate the cost effectiveness of advanced treatment modalities in their own care setting. Fourth, modeling and parameter uncertainty occurs with most economic modeling. A limitation of the Markov model is that the probability of moving out of a health state is not dependent on prior patient states before entering that specific health state.⁴⁶ However, with the use of time-dependent transition probabilities, this limitation is minimized. In addition, parameter uncertainty was reduced by conducting deterministic and probabilistic sensitivity analyses. Fifth, because the investigators used results derived from Rees et al²³ to populate the economic model, clinical practices and outcomes (ie, definition of good wound care) were limited to the practice patterns that occurred at the time of the clinical trial. Sixth, the model did not incorporate any PI recurrence or adverse events that are associated with PIs, although these events would not be expected to be high over the specified time period. Finally, even though the probability of mortality secondary to malignancy was not measured in the economic model, it should be noted that becaplermin gel contains a black box warning.

The results of this analysis suggest that once-daily treatment with becaplermin gel 100 µg/g increased the incidence of complete healing relative to a placebo gel. Treatment of PIs with becaplermin gel may improve clinical outcomes at a slightly higher cost. Health care providers and decision makers should consider both the acquisition cost of becaplermin and the total cost of care when making decisions about the allocation of financial resources. Future investigations of the cost effectiveness of becaplermin for the treatment and management of PIs in routine clinical practice and using real-world data are warranted. 

The authors are grateful to Dr. Rees and colleagues for their previously published work, “Becaplermin gel in the treatment of pressure ulcers: a phase II randomized, double-blind, placebo-controlled study,” that allowed them to conduct this comparative cost-effectiveness study.

Affiliations: Truven Health Analytics, an IBM Company, Houston, TX; University of North Texas Health Sciences Center, Department of Pharmacotherapy, Fort Worth, TX; Smith & Nephew Inc, Fort Worth, TX; and Connecticut Clinical Nursing Associates, LLC, Plymouth, CT

Correspondence: Adrienne M. Gilligan, PhD, Truven Health Analytics, an IBM Company, 2 Riverway, Houston, TX 77056; agilliga@us.ibm.com

Disclosure: This study, medical writing, and editorial support were funded by Smith & Nephew, Inc (Forth Worth, TX). Curtis Waycaster is an employee of Smith & Nephew. The product used in this study is not indicated for the treatment of pressure injuries.