Article type
Year
Abstract
Background: The Depression subscale of the Hospital Anxiety and Depression Scale (HADS-D) is recommended for depression screening in people with physical health problems. A score of ≥ 11 is considered a clinically significant depressive disorder, whereas a score between 8 and 10 suggests a mild disorder. Existing conventional meta-analyses, however, were conducted 10 years ago and limited by excluding studies that did not report the results for standard cutoffs (≥ 8 or ≥ 11), incomplete cutoff reporting within included studies, and by not examining accuracy by reference standard.
Objectives: To evaluate HADS-D accuracy for detecting major depression in medically ill patients using individual participant data meta-analysis.
Methods: Eligible studies compared PHQ-9 scores with major depression diagnoses from validated diagnostic interviews. Primary study data and study level data extracted from primary reports were synthesized. For HADS-D cut-off scores 5-15, bivariate random effects meta-analysis was used to estimate pooled sensitivity and specificity, separately, among studies that used semi-structured diagnostic interviews, which are designed for administration by clinicians; fully structured interviews, which are designed for lay administration; and the Mini International Neuropsychiatric (MINI) diagnostic interviews, a brief fully structured interview. Sensitivity and specificity were examined using one-stage meta-regression by reference standard categories, considering all participant characteristics in a single model.
Results: Individual participant data were obtained from 82 of 141 eligible studies (17,176 participants, 2,100 cases) (See Table 1). Combined sensitivity and specificity was maximized at cutoff ≥ 7 for semi-structured interviews, and fully structured interviews, and maximized at cutoff ≥ 6 for the MINI (See Table 1). Among studies with a semi-structured interview (44 studies, 6,614 participants, 754 cases), sensitivity and specificity (95% CI) were 0.83 (0.77, 0.88) and 0.78 (0.73, 0.82) for a cutoff of ≥ 7, 0.74 (0.68, 0.80) and 0.84 (0.80, 0.87) for a cutoff of >= 8, and 0.45 (0.38, 0.53) and 0.95 (0.93, 0.96) for a cutoff of ≥ 11. Accuracy was similar across reference standards and subgroups, including for participants in different recruiting settings.
Conclusions: A HADS-D cutoff of ≥ 7 maximized combined sensitivity and specificity; but the standard cutoffs of ≥ 8 and ≥ 11 were less sensitive but more specific.
Patient or healthcare consumer involvement: There was no direct patient or healthcare consumer involvement in this study. However, we will develop a web-based knowledge translation tool which will help clinicians considering screening for depression with the HADS-D to estimate the expected numbers of positive screens and the true and false screening outcomes based on results from the present IPDMA.
Objectives: To evaluate HADS-D accuracy for detecting major depression in medically ill patients using individual participant data meta-analysis.
Methods: Eligible studies compared PHQ-9 scores with major depression diagnoses from validated diagnostic interviews. Primary study data and study level data extracted from primary reports were synthesized. For HADS-D cut-off scores 5-15, bivariate random effects meta-analysis was used to estimate pooled sensitivity and specificity, separately, among studies that used semi-structured diagnostic interviews, which are designed for administration by clinicians; fully structured interviews, which are designed for lay administration; and the Mini International Neuropsychiatric (MINI) diagnostic interviews, a brief fully structured interview. Sensitivity and specificity were examined using one-stage meta-regression by reference standard categories, considering all participant characteristics in a single model.
Results: Individual participant data were obtained from 82 of 141 eligible studies (17,176 participants, 2,100 cases) (See Table 1). Combined sensitivity and specificity was maximized at cutoff ≥ 7 for semi-structured interviews, and fully structured interviews, and maximized at cutoff ≥ 6 for the MINI (See Table 1). Among studies with a semi-structured interview (44 studies, 6,614 participants, 754 cases), sensitivity and specificity (95% CI) were 0.83 (0.77, 0.88) and 0.78 (0.73, 0.82) for a cutoff of ≥ 7, 0.74 (0.68, 0.80) and 0.84 (0.80, 0.87) for a cutoff of >= 8, and 0.45 (0.38, 0.53) and 0.95 (0.93, 0.96) for a cutoff of ≥ 11. Accuracy was similar across reference standards and subgroups, including for participants in different recruiting settings.
Conclusions: A HADS-D cutoff of ≥ 7 maximized combined sensitivity and specificity; but the standard cutoffs of ≥ 8 and ≥ 11 were less sensitive but more specific.
Patient or healthcare consumer involvement: There was no direct patient or healthcare consumer involvement in this study. However, we will develop a web-based knowledge translation tool which will help clinicians considering screening for depression with the HADS-D to estimate the expected numbers of positive screens and the true and false screening outcomes based on results from the present IPDMA.