Single-item patient-rated helpfulness and improvement as an alternative to standardized questionnaires for establishing anxiety and depression treatment efficacy.

Abstract

Evidence-based psychological treatments for anxiety and depression are widely used, yet roughly half of those treated do not respond. Treatment response prediction could help to optimize patient outcomes and use of clinical resources. However, existing longitudinal studies with potentially valuable predictors are unlikely to include comprehensive, prospective measures of symptoms throughout therapy. Single-item patient ratings of helpfulness and improvement are a potentially cost-effective and efficient alternative, but their relationship with typically used change score measures is unknown. Data were analyzed from 135 participants (124 female sex; 120 female gender; 127 White) who received cognitive-behavioral therapy. Anxiety symptoms (Generalized Anxiety Disorder-7), depression symptoms (Patient Health Questionnaire-9), and impairment (Work and Social Adjustment Scale) questionnaire scores were obtained before therapy (assessment), before each session, and 1-month posttreatment (follow-up). Helpfulness (binary) and improvement (continuous) ratings were collected at follow-up. Linear regression models assessed the relationship between helpfulness and improvement ratings and questionnaire change scores from the first to the last session. Logistic regressions modeled the relationships between single-item measures and National Health Service Talking Therapies outcomes, derived from questionnaire change scores. Helpfulness and improvement showed significant associations with questionnaire change scores as well as National Health Service Talking Therapies outcomes. In a joint model, improvement retained significant associations while helpfulness became nonsignificant. Improvement, and to a lesser extent helpfulness, patient ratings may be a cost-effective alternative for establishing treatment efficacy and outcome. The items' wording and response scales may underlie observed differences. While not equivalent to change score-based measures, they may be adequate for studies requiring large sample sizes. (PsycInfo Database Record (c) 2025 APA, all rights reserved).