PIT1 Negative Pressure Wound Therapy for Pressure Ulcer: Effect on Outcomes and Utilization Using Real-World DATA

Abstract

Clinical trials evaluating negative pressure wound therapy (NPWT) as a treatment for pressure ulcer (PU) have been inconclusive. The objective of this research was to evaluate the effect of NPWT on length of stay (LOS), charges, and mortality for PU patients. Clinical data collected over a 4-year period from a large medical institution in the OMNY Health System Database were used to identify adult PU patients. Patients were excluded if also diagnosed with diabetic, chronic skin non-pressure, or varicose ulcers. Current Procedural Terminology codes were used to divide patients into NPWT and comparison cohorts. Crude estimates for mortality (percent and 95% confidence interval [95% CI]) and for LOS and charges (median, quartiles 1 and 3 [Q1, Q3]) were generated for each outcome. Propensity score methods were used to account for underlying differences between cohorts. After exclusions, 749 patients were divided into NPWT (N=135) and comparison (N=614) cohorts. Crude mortality was slightly greater in the NPWT cohort (28.9%; 95% CI: 21.2%-36.5%) than in the comparison cohort (22.2%; 95% CI: 18.9%-25.4%). NPWT patients incurred notably higher charges (median [Q1-Q3]: $65,197 [$32,232-$143,607] vs. $13,850 [$3,056-$35,929]) and LOS (median [Q1-Q3]: 11 days [7, 22] vs. 5 days [3, 9]) than comparison patients. After application of propensity score methods, the adjusted mortality relative risk was 1.20 (95% CI: 0.88-1.65), and the adjusted median differences in LOS and charges were 5.8 days (Q1-Q3: 3.9-9.2) and $30,900 (Q1-Q3: $15,434-$58,868), respectively. No notable positive effects of NPWT on PU outcomes or hospitalizations were observed even after adjusting for underlying differences. Residual confounding by indication could be a possible explanation for the observed associations between NPWT and PU outcomes. Further research is needed to examine longer term differences in outcomes such as readmission rates and to evaluate drivers of cost difference.