|Home | About | Journals | Submit | Contact Us | Français|
We read with interest the recently published paper on pooled analysis of ultra-short screening tools for depression.1 In our view the issue of importance is the validity rather than the accuracy of the short screening test. A high false-positive rate is not as important an issue in initial screening for depression as it is when screening for conditions that require invasive investigations to establish the diagnosis (for example carcinoma of the breast). A few more questions is all that is required to establish a diagnosis of depression and then offer treatment. The authors refer to our more recent tool, the two questions with help question (TQWHQ).2 A positive response to either depression question plus the help question gave a sensitivity of 96% and specificity of 89%. The likelihood ratio positive increased from 4.4 to 9.1 when considering a positive response to either screening question or the help question, which we consider more than a modest enhancement. The likelihood ratio of a patient who answered yes, wanting help today of having major depression was high (17.5). We agree with Mitchell and Coyne that ultra-short tests need second stage assessment of those who screen positive. In the TQWHQ, the help question acts as a second-stage assessor, significantly reducing the number of false positives.
Allowing our GPs to see questionnaire data is criticised as contaminating our study. This is missing the point that GP diagnosis was one of the outcomes of the study. GPs had a sensitivity of 79% and a specificity of 94% when using the TQWHQ (measured against the CIDI), meaning that patients are unlikely to receive unnecessary treatment.