[abstract] Factors affecting patient cost of HBO2T: An analysis using data from the U.S. Hyperbaric Registry

Abstract

Background: The Federal Coordinating Counsel for Comparative Effectiveness Research (CER) concluded that longitudinally linked electronic health record (EHR) databases are useful tools for CER research. We report on the first use of such a registry to analyze the cost of hyperbaric oxygen therapy (HBO2T) among wound care patients and the factors which drive HBO2T costs. Methods: Clinics using the Intellicure electronic health record (EHR) allowed data to be deidentified and pooled. A retrospective analysis was performed of prospectively collected data from 1,023 patients with 1,538 wounds that received hyperbaric oxygen therapy (HBO2T) over a period of two years. Results: Patients undergoing HBO2T had an average of 2.3 co-morbidities. Total cost per wound undergoing HBO2T averaged $6,025 and differed significantly by wound type with amputations being the most costly wounds ($7,838). Cost to heal a wound on a per-patient basis increased when patients had more co-morbidities, from a median of $15,572 with no co-morbidities to $23,245 per patient. The likelihood of HBO2T being performed was determined by ICD-9 diagnosis with the most common diagnoses being jeopardized flap, soft tissue radionecrosis and diabetic foot ulcer. Independent factors significantly associated with an increased likelihood of HBO2T use were larger wound area or deeper wounds, being on oral steroids, being paralyzed, having heart failure or coronary artery disease. Patients who did not have renal failure or limitations in activities of daily living were more likely to get HBO2T Lower costs per patient (fewer treatment numbers) were associated with younger patient age, not having negative wound pressure therapy in addition to HBO2T, not being a diabetic, or not being prescribed pain medications, antibiotics or oral steroids. Conclusion: Cost of care is the first step in the process of determining cost effectiveness and the use of deidentified pooled EHR data is feasible for such projects.