BACKGROUND: Suicidal ideation (SI) is a major health concern in children, adolescents, and young adults (CAYA) population. Inaccurate estimates of SI-related hospital readmission rates may contribute to inappropriate allocation of resources for the prevention of future readmissions. The estimation of these readmission rates using claims data may be sensitive to the diagnosis code position used to establish analytic cohorts.
OBJECTIVE: To examine the prevalence and effects of SI diagnosis code position in claims on 30-day readmission rates using the Nationwide Readmissions Database (NRD).
METHODS: This was a cross-sectional study using the NRD (2010-2015). We established six cohorts of hospitalized CAYA (5-24 years old) with a diagnosis of SI based on different combinations of SI diagnosis code (ICD-9 code V62.84) positions in claims. Thirty-day hospital readmission rates following an index SI discharge were estimated for each cohort. We tested the null hypothesis that hospital readmission rates following an index SI discharge are not sensitive to diagnosis code positions using a test for equality of proportions between the predefined SI cohorts.
RESULTS: The prevalence of SI diagnosis codes increased yearly from 2.9% in 2010 to 5.8% in 2015. SI hospital readmission rates ranged from 0 to 41.1 per 1000 index events based on cohort definitions (i.e. diagnosis code positions). We rejected the null hypothesis that SI-related readmission rates are not sensitive to diagnosis code positions.
CONCLUSION: SI-related readmission rate estimates are sensitive to SI diagnosis code positions. Determining appropriate diagnostic positions can further improve readmission analyses for SI and its applications in healthcare policies.