EFFECT OF WOUND EXUDATE SEVERITY ON TIME TO COMPLETE HEALING IN DIABETIC FOOT ULCERS: A REAL-WORLD LONGITUDINAL ANALYSIS (EXACT STUDY)
Author(s)
David G. M. Armstrong, PhD, MD1, Grace Gahlon, PhD2, Vladica M. Velickovic, PhD, MD3, Yun Han, PhD2, Niuosha M. Sanaeifar, PhD3, Shelby Corman, PhD2, Kenneth Rothman, PhD4;
1Keck School of Medicine of University of Southern California, Los Angeles, CA, USA, 2Precision AQ, Bethesda, MD, USA, 3HARTMANN GROUP, Evidence Generation Department, Heidenheim, Germany, 4Boston University School of Public Health, Boston, MA, USA
1Keck School of Medicine of University of Southern California, Los Angeles, CA, USA, 2Precision AQ, Bethesda, MD, USA, 3HARTMANN GROUP, Evidence Generation Department, Heidenheim, Germany, 4Boston University School of Public Health, Boston, MA, USA
OBJECTIVES: To estimate the effect of wound exudate severity on time to complete healing in diabetic foot ulcers (DFUs) using longitudinal real-world data.
METHODS: We conducted a retrospective cohort study of DFUs documented in the Net Health WoundExpert electronic medical record from U.S. outpatient wound centers (2013-2019). Eligible wounds required ≥2 visits with exudate documentation and baseline area 2-100 cm². Follow-up began with the index visit, with the index date defined as the first visit for each eligible wound, continuing until healing or censoring (loss to follow-up or initiation of advanced therapies). Exudate was modeled as a time-updated exposure (none/mild/moderate/heavy). Confounders were selected using a directed acyclic graph confirmed for clinical plausibility by wound-care experts. We used marginal structural models with stabilized inverse-probability of treatment and censoring weights. Time-to-event effects were estimated using Royston-Parmar flexible parametric survival models with restricted cubic splines and time-varying coefficients. Confidence intervals (CI) were calculated via a patient-level clustered bootstrap procedure. The parametric g-formula generated counterfactual probabilities and restricted mean survival time (RMST) for healing through 52 weeks.
RESULTS: A total of 8,324 patients contributed 19,750 DFUs. Kaplan-Meier healing probabilities were 57.8% at 12 weeks and 75.8% at 24 weeks. Relative to no exudate, hazards of healing were significantly lower for mild (HR 0.14; 95% CI 0.12-0.17), moderate (HR 0.05; 0.04-0.08), and heavy exudate (HR 0.03; 0.00-1.01). At 52 weeks, counterfactual healing probabilities were 0.493 (none) versus 0.086 (mild), 0.043 (moderate), and 0.008 (heavy). RMST at 52 weeks was 235.5 days unhealed (none) versus 342.9 (mild), 353.8 (moderate), and 361.9 (heavy), corresponding to 107.4, 118.3, and 126.4 additional unhealed days, respectively, compared to wounds with no exudate.
CONCLUSIONS: Increasing exudate burden delayed DFU healing. These clinically interpretable absolute-effect estimates provide inputs for prognostic assessment and economic evaluation of exudate-management strategies and may ultimately inform actionable interventions in clinical practice.
METHODS: We conducted a retrospective cohort study of DFUs documented in the Net Health WoundExpert electronic medical record from U.S. outpatient wound centers (2013-2019). Eligible wounds required ≥2 visits with exudate documentation and baseline area 2-100 cm². Follow-up began with the index visit, with the index date defined as the first visit for each eligible wound, continuing until healing or censoring (loss to follow-up or initiation of advanced therapies). Exudate was modeled as a time-updated exposure (none/mild/moderate/heavy). Confounders were selected using a directed acyclic graph confirmed for clinical plausibility by wound-care experts. We used marginal structural models with stabilized inverse-probability of treatment and censoring weights. Time-to-event effects were estimated using Royston-Parmar flexible parametric survival models with restricted cubic splines and time-varying coefficients. Confidence intervals (CI) were calculated via a patient-level clustered bootstrap procedure. The parametric g-formula generated counterfactual probabilities and restricted mean survival time (RMST) for healing through 52 weeks.
RESULTS: A total of 8,324 patients contributed 19,750 DFUs. Kaplan-Meier healing probabilities were 57.8% at 12 weeks and 75.8% at 24 weeks. Relative to no exudate, hazards of healing were significantly lower for mild (HR 0.14; 95% CI 0.12-0.17), moderate (HR 0.05; 0.04-0.08), and heavy exudate (HR 0.03; 0.00-1.01). At 52 weeks, counterfactual healing probabilities were 0.493 (none) versus 0.086 (mild), 0.043 (moderate), and 0.008 (heavy). RMST at 52 weeks was 235.5 days unhealed (none) versus 342.9 (mild), 353.8 (moderate), and 361.9 (heavy), corresponding to 107.4, 118.3, and 126.4 additional unhealed days, respectively, compared to wounds with no exudate.
CONCLUSIONS: Increasing exudate burden delayed DFU healing. These clinically interpretable absolute-effect estimates provide inputs for prognostic assessment and economic evaluation of exudate-management strategies and may ultimately inform actionable interventions in clinical practice.
Conference/Value in Health Info
2026-05, ISPOR 2026, Philadelphia, PA, USA
Value in Health, Volume 29, Issue S6
Code
CO52
Topic
Clinical Outcomes
Topic Subcategory
Relating Intermediate to Long-term Outcomes
Disease
SDC: Diabetes/Endocrine/Metabolic Disorders (including obesity), SDC: Injury & Trauma