Validating the DIVERT scales, CARS, and EARLI for predicting emergency department visits in home health care in Japan: A retrospective cohort study

Background: The Detection of Indicators and Vulnerabilities for Emergency Room Trips (DIVERT) scale, the Community Assessment Risk Screen (CARS), and the Emergency Admission Risk Likelihood Index (EARLI) are scales that assess the risk of emergency department (ED) visits among home health care patients. This study validated these scales and explored factors that could improve their predictive accuracy among Japanese home health care patients. Methods: This was a single-center retrospective cohort study. The primary outcome of unplanned ED visits was used to assess the validity of the DIVERT scale, CARS, and EARLI. Additionally, we examined whether the addition of patient age and receipt of advance care planning as variables on these assessments could enhance their precision. Results: Altogether, 40 (17.8%) had at least one ED visit during the 6 months study period. In these patients, the DIVERT scale, CARS, and EARLI of the patients with ≥1 ED visit was superior compared with no ED visit (both p < 0.05). The area under the curve (AUC) of the DIVERT scale, CARS, and EARLI were 0.62, 0.59, and 0.60, respectively. Adding patient age and receipt of advance care planning improved the AUC in all three scales. Conclusions: Our findings suggest that these assessment scales could be applicable to home health care patients in Japan. Furthermore, adding age and receipt of advance care planning as variables was found to enhance the predictive accuracy of the scales.