Multiple Sclerosis, 1 June 2003, vol. 9, no. 3, pp. 307-310(4)
Avasarala J.R.; Cross A.H.; Trinkaus K.
 Departments of Neurology, Washington University School of Medicine, Campus Box 8111, 660 S. Euclid Ave., St Louis, MO 63110, USA  Departments of Neurology, Washington University School of Medicine, St Louis, MO 63110, USA  Departments of Biostatistics, Washington University School of Medicine, St Louis, MO 63110, USA
To examine if depression in multiple sclerosis (MS) can be accurately recognized using the Yale Single Question (YSQ) screen as compared with the Beck Depression Inventory (BDI), a 21-item self-report rating scale for depression.
In addition, we sought to assess the sensitivity, specificity, the positive predictive value (PPV) and the negative predictive value (NPV) of the YSQ.
Depression associated with MS is a major contributor to morbidity.
Screening for depression in patients with MS currently includes the BD1, which measures characteristic attitudes and symptoms of depression.
However, in a busy outpatient clinic, the BDI might not be the most appropriate instrument, particularly if depression screening can be assessed accurately using simpler techniques that are easy to administer and consume less time.
We compared the accuracy of the YSQ screen response against the BDI to screen for depression in MS patients, in an outpatient setting.
This is a comparative outcome study of two 'instruments' used for screening depression in MS patients in an academic outpatient setting.
All patients were initially screened for depression by asking patients the YSQ – 'Do you frequently feel sad or depressed?', followed by BDI administration.
Depression was defined as a score of 13 on the BDI.
One hundred and twenty successive patients who presented to the MS clinic at Washington University School of Medicine and met the criteria for diagnosis of MS were screened for depression.
All patients diagnosed as having MS, regardless of type, were included in the study.
Of the 120 patients studied, a total of 49 of 120 were clinically depressed as defined by a BDI cut-off of 13; 71 of 120 were not.
The sensitivity of the YSQ was 32 of 49=65.3% with a 95% confidence interval (0.50, 0.78), specificity was 62 of 71=87.3% (0.77, 0.94), PPV was 32 of 41=78.0% (0.62, 0.89) and NPV was 62 of 79=78.5% (0.68, 0.87).
Of the 49 patients depressed by BDI criteria, 17 responded 'no' to the YSQ, yielding a false-negative rate of 34.7% (0.22, 0.50).
The Wilcoxon–Mann–Whitney test for difference in age among those on antidepressants compared with those who were not showed no statistical difference (P=0.35).
For patients on antidepressants, the mean BDI score was 16.0±8.9 (mean±SD) and 9.5±8.7 for those not on antidepressants.
Differences in BDI scores among patients on antidepressants versus those who were not were statistically significant (P<0.0001).
Patients on antidepressants had significantly higher BDI scores.
Our results show that the YSQ cannot replace the BDI as a screening instrument for depression in MS.
The YSQ could not identify 34.7% of patients who were depressed by BDI criteria.
However,as reported in a published study, BDI missed 30% of cases with early depression in MS when a cut-off of 13 was used.
The YSQ appears to be specific in ruling out depression when a patient is not depressed.
MS is a chronic disease and since prevalence of depression varies, it is important to screen patients repeatedly over time so as not to miss the diagnosis.
That BDI scores were higher among those taking antidepressants underscores the fact that this subset of patients need to be on medication, but the higher scores could also represent a sampling error since the duration of antidepressant use was not studied.