Comparison of Health Care Costs and Hospital Readmission Rates Associated With Negative Pressure Wound Therapies

Objective. A retrospective national claims database analysis evaluated total and wound-related costs (eg, hospital readmission rates) for patients with chronic wounds treated with negative pressure wound therapy (NPWT), comparing NPWT-V (V.A.C. Therapy, KCI, an Acelity company, San Antonio, TX) and NPWT-O (other non-KCI models of NPWT, the brands of which were not known to the researchers). Methods. Patients with ≥ 1 NPWT claim from January 2009-June 2012 in outpatient settings in the United States were included, if they had continuous medical and pharmacy benefits for 12 months before the initial index date of their NPWT claim and at least 3 months post index. Mean total health care costs were assessed at 3 months and 12 months; wound-related hospital readmission rates were assessed at 3 months and 6 months. Cost differences between cohorts were analyzed by t test and readmission rates were analyzed by chi-square test. Results. At 3 months, the cohort of NPWT-V patients was significantly younger (59.2 vs 63.6 years, P < 0.01). The same patients were followed at 3, 6, and 12 months, although some fell out as time progressed. At the 3-month assessment, mean comorbidity scores did not differ between the NPWT-V group and the NPWT-O groups (3.38 vs 3.66). Total costs were lower for NPWT-V vs NPWT-O at 3 months ($35,498 vs $39,722, respectively; P = 0.08) and 12 months ($80,768 vs $111,212; P = 0.03). Significantly lower inpatient (P = 0.01), emergency room (P < 0.01), and home (P = 0.05) costs, despite higher (P = 0.04) NPWT costs, accounted for lower 12-month NPWT-V total costs. Wound-related readmission rates were significantly lower for NPWT-V at 3 months (5% vs 8%; P ≤ 0.01) and 6 months (6% vs 11%; P ≤ 0.01). For all wound types, NPWT-O patients had a 17-fold higher rate of switching to alternate NPWT models compared with NPWT-V patients. Conclusion. In this retrospective analysis, NPWT-V patients had lower total costs, lower wound-related costs, and lower hospital readmission rates than NPWT-O patients at all time points assessed.

  Chronic wound care continues to pose a significant challenge to the US health care system, as the prevalence of diabetes and obesity increases and the population ages.^1,2 Lazarus and colleagues³ define chronic wounds as “those that have failed to proceed through an orderly and timely reparative process to produce anatomic and functional integrity” of the injured site. Patients who are not healthy have a higher risk of developing nonhealing wounds.³

  Diabetic foot ulcers (DFUs), venous leg ulcers (VLUs), and pressure ulcers (PrUs) are common chronic wounds often treated in the outpatient setting. Among patients with diabetes, DFU prevalence has been reported as ranging from 4%-10% with an estimated lifetime incidence rate of 10%-25%.⁴ Diabetic foot ulcers are associated with 20% of hospital admissions and 40%-70% of nontraumatic amputations in the lower extremities occur in patients with diabetes.⁴ The prevalence of chronic venous ulcers in the general population is 1% compared to 4% in the population > 65 years of age.⁵ As many as 33% of patients treated for VLUs are estimated to have ≥ 4 recurrences.⁶ In US health care institutions, an estimated 2.5 million patients will develop PrUs, with an associated mortality rate of 60,000 deaths.⁷ During the period of 1993 to 2003, PrUs in hospitalized patients increased by 79%.⁸ According to the Agency for Healthcare Research and Quality, 72% of the patients who had PrUs during hospitalization in 2006 were ≥ 65 years old.⁹

  The increase in chronic wound patients has been reflected in rising health care costs. In 2009, Sen and coauthors² reported 6.5 million patients in the United States were living with chronic wounds. Moreover, in 2007 the annual cost associated with chronic wounds was estimated at $25 billion.² Estimated US costs related to treatment of a DFU ranged from $4,595 per ulcer episode to nearly $28,000 (in 1995 US dollars) for the 2-year period following diagnosis and more than $30,000 during the life of the ulcer.¹⁰ Many DFUs lead to minor or major amputations. In 2001 the average cost of a DFU-related amputation was $38,077 and annual cost for DFU care and related amputations was estimated to be $10.9 billion.¹¹ In 1998, the annual US cost of treating VLUs was estimated at $2.5–$3.5 billion.¹² In 2008, the average cost to treat a PrU as a secondary diagnosis was estimated at $43,180 per hospital stay,¹³ and $11 billion was the projected annual cost to treat PrUs.⁷

  Negative pressure wound therapy (NPWT) is an advanced adjunctive therapy used to facilitate healing in chronic wounds.^14-19 For example, in a randomized controlled trial (RCT), Blume and colleagues14 compared NPWT (n = 169) to advanced moist wound therapy (AMWT) (n = 166) for treatment of DFUs. Significantly more DFUs achieved 100% wound closure with NPWT (73 of 169, 43.2%) than with AMWT (48 of 166, 28.9%) within the 112-day active treatment phase (P = 0.007). Patients treated with NPWT also had significantly (P = 0.035) fewer secondary amputations (7 of 169, 4.1% vs 17 of 166, 10.2%).¹⁴ While the majority of literature reports outcomes using 1 model of NPWT with reticulated open-cell foam dressings (V.A.C. Therapy, KCI, an Acelity company, San Antonio, TX), the number of NPWT alternatives has grown significantly in the past few years. Because material costs of NPWT may be higher than other therapies, the cost effectiveness of NPWT in terms of total costs and wound-related costs is an important consideration. Additionally, the varying costs across the range of NPWT models raise the issue of comparative effectiveness.

  To better understand the comparative effectiveness of NPWT as a treatment for chronic wounds, this study analyzed de-identified insurance claim data for patients receiving any model of NPWT for costs and hospital readmission rates for the period following the initial NPWT claim in an outpatient setting.

  Design and data source. A retrospective analysis was conducted using the large, longitudinal, closed national administrative claims database of a major US insurance company (Optum Life Sciences, Eden Prairie, MN). All data were de-identified and accessed in compliance with Health Insurance Portability and Accountability Act privacy guidelines.²⁰

  Study participants. Patients were included if they had ≥ 1 diagnosis claim with an NPWT HCPCS code (E2402) in the outpatient setting during the period of January 2009-June 2012. Patients were grouped according to the initial NPWT therapy received: NPWT-V or NPWT-O (other non-KCI models of NPWT). Claims data were analyzed at 3 months, 6 months, and 12 months after the index date of the first NPWT claim. The number of patients analyzed in each period varied as members left the database. Patients also had to have continuous medical and pharmacy benefits (ie, commercial or Medicare Advantage) for 12 months preindex date and for at least 3 months, and up to 12 months, post index. Patient demographics were presented at 3 months only as age and comorbidities were not anticipated to materially change over the 12 month period.

  Total cost data was presented at 3 months and 12 months only, as inclusion of interim 6-month data suggested no deviation from the 3 to 12 month cost trend and was not anticipated to add to the discussion. Mean per patient wound-related inpatient stays, ER visits, and re-admission rates were not reported at 12 months, as events beyond 6-months could easily be due to another cause (eg, a new wound).

  Measures. Patient demographic and clinical characteristics were assessed on the index date. The Charlson Comorbidity Index score²¹ was computed for the 12 month preindex period. Wound types were classified according to the ICD-9-CM code on the claim.

  Total health care costs at 3 months and 12 months were calculated as the sum of all costs for wound-related and nonwound-related services based on standardized costs, which removed variations in the cost of service based upon specific contractual arrangements between payers and providers.

  Hospital readmission rates at 3 months and 6 months were defined as the proportion of patients with hospital readmissions. Readmissions were categorized as total, wound-related, and nonwound-related. All claims with a wound diagnosis within the top 3 diagnoses on the claim were categorized as wound-related.

  The numbers of wound-related emergency room visits and inpatient stays were also calculated for 3 months and 6 months. Switching to alternate NPWT was measured as the proportion of patients switching from the initial NPWT model to a different NPWT device.

  Statistical analysis. Differences between cohorts (NPWT-V vs NPWT-O) were assessed by t tests for costs and chi-square tests for readmission rates. Significance was defined as P ≤ 0.05. Data analysis was generated using SAS/STAT Software, version 9.2 (SAS Institute, Inc, Cary, NC).

  At 3 months, the analysis included 12,843 NPWT-V and 713 NPWT-O patients. Mean age for the NPWT-V cohort was 59.2 years and 48.3% of these patients were men (Table 1). Patients in the NPWT-V cohort were significantly younger than NPWT-O patients (59.2 vs 63.6 years, P < 0.01); mean Charlson Comorbidity Index scores between groups did not differ (3.38 vs 3.66). The most common comorbid conditions were diabetes without complications, peripheral vascular disease, and chronic pulmonary disease (Table 1).

  Based on ICD-9-CM codes, chronic wound types (ie, PrUs, DFUs [with and without amputation], VLUs, and nonhealing surgical wounds) comprised the majority (82.1%) of the wounds. Open wounds, cellulitis, and necrotizing fasciitis were some of the acute wound groups, although it was difficult to determine whether open wounds were chronic or acute wounds.

  Mean total health care costs per patient were calculated and compared for 3 months and 12 months after the initial NPWT claim in an outpatient setting (Figure 1). At 3 months, there were 12,843 patients in the NPWT-V cohort and 713 patients in the NPWT-O cohort. The difference between the mean total costs per patient at 3 months for NPWT-V ($35,498) vs NPWT-O ($39,722) trended lower by $4,224 (11%) per patient for NPWT-V (P = 0.08). At 12 months, there were 7,860 patients in the NPWT-V cohort and 378 patients in the NPWT-O cohort. The 12-month mean total costs (NPWT-V, $80,768 vs NPWT-O, $111,212) (Figure 1) were significantly (P = 0.03) lower for NPWT-V patients by $30,444 (27%).

  At 3 months, wound-related costs represented 37.9% of total costs for NPWT-V and 37.6% of total costs for NPWT-O; at 12 months, wound-related costs represented 25.7% of total costs for both NPWT-V and NPWT-O. Wound-related costs (NPWT-V, $20,801 vs NPWT-O, $28,647) at 12 months were significantly (P = 0.01) lower for NPWT-V by $7,846 (27%). At 12 months, cost savings for NPWT-V were driven by $14,598 (35%) lower inpatient costs (P = 0.01), $4,188 (65%) lower emergency room costs (P < 0.01), and $4,810 (35%) lower home costs (P = 0.05), despite $982 higher NPWT-related costs (P = 0.04.)

  Across all wound types, hospital readmission rates for patients treated with NPWT-V vs patients treated with NPWT-O were significantly lower at 3 months (23% vs 31%, respectively; P ≤ 0.01) and at 6 months (31% vs 43%, P ≤ 0.01) (Figure 2). Wound-related readmission rates were also significantly lower for NPWT-V vs NPWT-O patients at 3 months (5% vs 8%, respectively; P ≤ 0.01) and 6 months (6% vs 11%; P ≤ 0.01) (Figure 2).

  Mean counts of wound-related inpatient stays and emergency room visits were 53% and 83% lower (P < 0.0001) at 3 months and 6 months for patients in the NPWT-V cohort compared to patients in the NPWT-O cohort (Table 2). For each wound category examined, a similar pattern of lower or significantly lower mean number of inpatient stays and emergency room visits was evident for patients treated with NPWT-V (Table 2). Patients treated with NPWT-V also had lower home costs than those treated with NPWT-O. At 3 months, NPWT-V vs NPWT-O home costs were 32% lower ($4,462 vs $6,588, respectively) and at 12 months, home costs were significantly lower ($8,828 vs $13,638, P = 0.05) (Figure 1).

  For all wound types, patients treated with NPWT-O had a 17-fold higher rate of switching to alternate NPWT compared to patients treated with NPWT-V (Figure 3). At 6 months after the initial NPWT claim, rates of switching among NPWT-O patients were 9-fold higher for pressure ulcers, 15-fold higher for nonhealing surgical wounds, and 32-fold higher for open wounds compared to NPWT-V patients (Figure 3).

  In the interest of conciseness, the authors have presented data at the most appropriate time points.

  Studies have shown NPWT to be effective in facilitating closure of chronic wounds.^{14,15,17,18,22} In the context of increasing numbers of chronic wounds being treated in the outpatient setting, evaluating the cost effectiveness of using an advanced wound care therapy like NPWT is particularly relevant. Since multiple NPWT models have become available in recent years, comparative effectiveness is also an important consideration.

  In a retrospective analysis of a large national administrative claims database, the authors compared total costs and total wound-related costs associated with use of NPWT-V and NPWT-O to treat wounds in an outpatient setting. Compared to the NPWT-O group, the NPWT-V group had 11% lower mean total costs at 3 months after the index claim and significantly lower mean total costs at 12 months. The NPWT-V group had significantly lower wound-related costs because of significantly lower readmission rates across all wound types, fewer wound-related inpatient stays and emergency room visits, and lower home costs compared to the NPWT-O group. These reductions in wound-related resource use drove the significantly lower 12-month total costs for the NPWT-V group.

 The authors speculate that training as well as technical and clinical support may represent important differentiating factors when evaluating the comparative effectiveness of various brands of NPWT in the home setting and may have played a role in the results of this study. The authors also recognize that user familiarity with NPWT-V, which has been applied in inpatient, outpatient, and home settings for almost 15 years, may be a hidden confounder in the data. It is possible that users’ lack of familiarity and expertise with the various devices in the NPWT-O group had a negative influence, and outcomes and costs for the NPWT-O devices may change over time as user experience increases.

  While retrospective analysis of large patient databases can provide a real world overview of costs across care settings, there are also inherent limitations. Claims could be missing or miscoded and might not provide proof positive of the presence of an actual condition. In this study, costs in the 6-month period prior to NPWT initiation were compared between patients treated with NPWT-V and NPWT-O to ensure there was no significant difference in the groups; however, study results represented actual, observed values and were not adjusted for potential differences in patient characteristics between cohorts. Consequently, adjusted results might differ. Patients in this study were required to have continuous medical and pharmacy coverage for at least 15 months; thus, results could be primarily applicable to patients in a stable, managed care setting. Also, as seen in the retrospective patient database study by Schwien and colleagues,²² the comparison of new vs established treatments can result in disproportionate group sizes. The relatively small size of the NPWT-O group, compared to the NPWT-V group, could have introduced some unanticipated bias, even allowing for the statistical significance shown, and there could have been other factors and/or comorbidities that were not measured. For example, wound age can have a significant impact on healing time^24-27 and may have differed between the cohorts. Finally, the general nature of the data could demonstrate trends but could not provide information as to their specific causes.

  In this retrospective claims database study, patients treated with NPWT-V had lower total costs and total wound-related costs based on lower overall resource use. Compared with patients treated with NPWT-O, total health care cost savings trended lower for patients treated with NPWT-V at 3 months and were significantly lower at 12 months. The primary drivers of cost savings for the NPWT-V group were significantly lower inpatient, emergency room, and home costs despite higher NPWT costs. Due to the importance of understanding how to manage outcomes and costs for patients with chronic wounds, further comparative studies on total cost to treat these patients with NPWT are warranted.

  The authors would like to thank Ricardo Martinez, MS, Director of Medical Information and Publications, KCI, and Alice Goodwin, BA, Medical Editor, KCI, for medical writing support.

Amy Law, MBA is from KCI, an Acelity company, San Antonio, TX. Anissa Cyhaniuk, MA; and Blake Krebs, MBA are from Optum Life Sciences, Eden Prairie, MN.

Address correspondence to:
Amy Law, MBA
KCI, an Acelity company
12930 Interstate Highway 10 West
San Antonio, TX 78249
amy.law@acelity.com

Disclosure: The authors disclose that Amy Law is an employee of KCI, an Acelity Company (San Antonio, TX); Anissa Cyhaniuk and Blake Krebs, with Optum Life Sciences (Eden Prairie, MN), were paid consultants on this study. Claims data and analytics were provided by Optum Life Sciences.