Quality Assessment of Studies of Diagnostic Accuracy (QUADAS).

aSee Supplementary Information S5 for QUADAS items. Items are rated ‘yes’, ‘no’ and ‘unclear’ based on the user's guide23 and reflecting the likelihood of being free of bias. Items #3 (reference standard appropriate), #10 (blind interpretation of test results) and #12 (same clinical data available as in practice) were not evaluated because an appropriate reference standard was a criterion for review eligibility and because scoring of all self-report depression screening tools is fully automated and does not require judgment.bItem #1 scored ‘no’ if patients with already diagnosed or treated depression were not excluded from study sample as they would not constitute newly identified cases in clinical practice. Studies were not downgraded for only sampling one type or stage of cancer.cItem #4 scored ‘yes’ if index test and reference standard were administered within 1 week of each other, ‘no’ if longer, and ‘unclear’ if not specified. Studies in which a significant number of patients received assessments more than 2 weeks apart were not included in the systematic review.dItem #13 originally was “Were uninterpretable, indeterminate or intermediate test results reported?” This item was adapted as “Were missing data on the index test handled correctly?”.eAuthors clarified that most patients received the index test and reference standard on the same day and all within 5 days.fAuthors clarified that 67% of interviews were conducted within one week and 93% within 2 weeks.