94 Evaluating Quality in Burn Care: Limitations of the Nationwide Readmission Database

Abstract

Listen

Translate article

Hospital readmission data may be helpful in identifying risk factors leading to higher costs of care or undesired outcomes. Few studies have examined readmission rates in the traumatic injury population, with even fewer examining the applicability to the adult burn population. The 2014 Nationwide Readmissions Database (NRD) was queried for 363 burn-specific ICD-9 CM codes and filtered by age and readmission status to capture the adult burn population presenting between January 1 and December 31, 2014. Note that the NRD does not include planned readmissions in the dataset captured. Univariate and multivariate logistical regression models were developed to determine readmission likelihood. A total of 26,719 burn patients were identified in the database with 1,209 all-cause unscheduled readmissions. Initial readmission analysis of demographics and comorbidities identified several readmission correlations in agreement with previous studies. However, 48% of all burn readmissions underwent an excision and grafting procedure on readmission. Additionally the highest readmission likelihood (3.14; 2.45–4.02) was seen in the highest volume centers, those seeing over 350 burn patients per year. The database was then filtered to remove those undergoing excision and grafting on readmission and readmission relationships were found to change considerably. Non-operative management on initial admission was found to have the strongest correlation with readmission (6.06; 4.55–10.0), followed by discharge AMA (3.65; 2.64–5.05). Hospital annual burn patient volume was found to not be significantly correlated with readmission. Non-operative management of burn wounds continues to fail a certain percentage of the time in the absence of definitive ways to assess burn depth. However, increases in readmission rates can be decoupled from hospital volumes when accurate patient characterization within the database is performed. Readmission coding accuracy represents a limitation of the NRD database which should not be overlooked when examining the burn population. Understanding the limitations of national databases with regards to assessing quality in Burn patients.