Comparison of the performance of mental health, drug and alcohol comorbidities based on ICD-10-AM and medical records for predicting 12-month outcomes in trauma patients

• DOI: 10.1186/s12913-018-3248-x
• Publication Date: 2018-06-05T09:16:19Z
• Authors: Tu Q. Nguyen; Pamela M. Simpson; Sandra C. Braaf; Peter A. Cameron; Rodney Judson; Belinda J. Gabbe
• Keywords: Health administration; Medical record

Many outcome studies capture the presence of mental health, drug and alcohol comorbidities from administrative datasets medical records. How these sources compare as predictors patient outcomes has not been determined. The purpose present study was to based on ICD-10-AM coding record documentation for predicting longer-term in injured patients. A random sample patients (n = 500) captured by Victorian State Trauma Registry selected study. Retrospective reviews were conducted collect data about documented while codes obtained routinely collected hospital data. Outcomes at 12-months post-injury Glasgow Outcome Scale – Extended (GOS-E), European Quality Life Five Dimensions (EQ-5D-3L), return work. Linear logistic regression models, adjusted age gender, using derived comorbidity compared measures calibration (Hosmer-Lemeshow statistic) discrimination (C-statistic R2). There no demonstrable difference predictive performance between models GOS-E, EQ-5D-3L utility sore mobility, self-care, usual activities pain/discomfort items. area under receiver operating characteristic (AUC) (AUC 0.68, 95% CI: 0.63, 0.73) higher than model 0.62, 0.57, 0.67) anxiety/depression item. work with inclusion 0.69, 0.76) 0.59, CL: 0.52, 0.65). Mental information review clearly superior majority assessed when ICD-10-AM. While available records may be more comprehensive ICD-10-AM, there appears little discriminative capacity coded two sources.