Depression in chronic pain patients: prevalence and measurement

This study aimed to: (1) determine prevalence of depression in patients referred to specialist pain services using the Structured Clinical Interview (SCID) diagnostic interview, (2) compare results on the Beck Depression Inventory II (BDI-II) with the SCID to determine the utility of the BDI-II as a screening tool in this population. Thirty-six participants were recruited, mainly women, with a mean age = 47.83 years (standard deviation = 12.85 years), who were heterogeneous with regard to their pain. All completed the BDI-II and SCID. The SCID diagnosed 26 (72%) cases of depression. BDI-II scores showed 31 (86%) that reported at least mild depression. Agreement between BDI-II scores over threshold for mild depression and SCID diagnosis were assessed by Cohen's kappa (= 0.6). ROC analysis for BDI-II scores against SCID diagnosis gave a large area under the curve (0.97, 95% confidence interval 0.93 to 1.02), suggesting BDI-II is an excellent screen for this population, although the curve was unusual in that sensitivity was high even when the false positive rate was zero. ROC analysis suggested 22 or above as an optimum cut-off score for depression on the BDI-II-higher than for a general population sample. It has been suggested that the BDI overestimates incidence of depression in pain patients, but this study confirmed through diagnostic interview the very high incidence of depression in this population. It is therefore questionable whether there is value in screening referrals for depression. When using BDI-II for screening, audit or evaluation purposes with a pain clinic population, we suggest a cut-off of 22 or above.