Background: Treatment non-response among high-risk, psychiatric patients exposes those suffering to suicidal risk as well as persistent social and occupational difficulties. Strategies for identification of treatment non-response are limited.
Aims: Diagnostic efficiency of a self-report, cross-cutting symptom measure was assessed as a marker of treatment non-response.
Method: 835 inpatients at a specialist psychiatric hospital completed the Patient Health Questionnaire - Depression (PHQ-9) at admission and every two weeks during hospitalization.
Results: For patients admitted with severe depression (PHQ-9 ≥ 20), results indicated good accuracy of 2-week PHQ-9 change score in identifying treatment non-response (AUC = 0.80, SE = 0.04, p < .0001; sensitivity = 85%; specificity = 73%; OR = 14.91).
Conclusions: The search for predictors of non-response to psychiatric treatment has a long and generally unfulfilled history. The PHQ-9 change score holds promise as a cost-effective test with comparable diagnostic characteristics to other medical tests.
Keywords: Diagnostic efficiency; PHQ-9; Prediction; Treatment non-response.
Copyright © 2015 Elsevier Ltd. All rights reserved.