Detecting DSM-5 somatic symptom disorder in general hospitals in China: B-criteria instrument has better accuracy-A secondary analysis

ObjectiveThis study investigates the diagnostic accuracy of the PHQ-15, SSS-8, SSD-12 and Whitley 8 and their combination in detecting DSM-5 somatic symptom disorder in general hospitals. MethodsIn our former multicenter cross-sectional study enrolling 699 outpatients from different departments in five cities in China, SCID-5 for SSD was administered to diagnose SSD and instruments including PHQ-15, SSS-8, SSD-12 and WI-8 were used to evaluate the SSD A and B criteria. In this secondary analysis study, we investigate which instrument or combination of instrument has best accuracy for detecting SSD in outpatients. Receiver operator curves were created, and area under the curve (AUC) analyses were assessed. The sensitivity and specificity were calculated for the optimal individual cut points. ResultsData from n = 694 patients [38.6% male, mean age: 42.89 years (SD = 14.24)] were analyzed. A total of 33.9% of patients fulfilled the SSD criteria. Diagnostic accuracy was moderate or good for each questionnaire (PHQ-15: AUC = 0.72; 95% CI = 0.68-0.75; SSS-8: AUC = 0.73; 95% CI = 0.69-0.76; SSD-12: AUC = 0.84; 95% CI = 0.81-0.86; WI-8: AUC = 0.81; 95% CI = 0.78-0.84). SSD-12 and WI-8 were significantly better at predicting SSD diagnoses. Combining PHQ-15 or SSS-8 with SSD-12 or WI-8 showed similar diagnostic accuracy to SSD-12 or WI-8 alone (PHQ-15 + SSD-12: AUC = 0.84; 95% CI = 0.81-0.87; PHQ-15 + WI-8: AUC = 0.82; 95% CI = 0.79-0.85; SSS-8 + SSD-12: AUC = 0.84; 95% CI = 0.81-0.87; SSS-8 + WI-8: AUC = 0.82; 95% CI = 0.79-0.84). In the efficiency analysis, both SSD-12 and WI-8 showed good efficiency, SSD-12 slightly more efficient than WI-8; however, within the range of good sensitivity, the PHQ-15 and SSS-8 delivered rather poor specificity. For a priority of sensitivity over specificity, the cutoff points of >= 13 for SSD-12 (sensitivity and specificity = 80 and 72%) and >= 17 for WI-8 (sensitivity and specificity = 80 and 67%) are recommended. ConclusionsIn general hospital settings, SSD-12 or WI-8 alone may be sufficient for detecting somatic symptom disorder, as effective as when combined with the PHQ-15 or SSS-8 for evaluating physical burden.