The February issue of the BJGP devoted space to research into depression and attempted to put this work into context. Here I try to set the scene in a different way – by relating what we do in the consulting room to what is happening in the community at large. I have taken percentages from the literature and converted them to round numbers on a human scale. Suppose a part-time UK GP has 600 adults on their list, sees 200 in a month and discusses a few of them over coffee (see figure).
Each month, 15% of adults report having felt depressed and 4% consult their doctor about this.1 So each month, in an adult population of 600, 100 feel depressed and 27* consult.
The prevalence of depression in general practice is around 10% and its diagnosis by GPs has a sensitivity of 36% and specificity of 84%.2 So of the 200 patients seen in a month, depression is diagnosed in 20 of the 180 who do not have depression and in seven of the 20 who do, giving a total of 27* supposedly depressed patients. Of patients diagnosed with depression 33–43% are given antidepressants.3,4 So, out of 27 patients we prescribe antidepressants for 10.
Antidepressants produce improvement in 28–74% of patients compared with 16–35% on placebo.5 Withdrawal rates from antidepressants are 3–30%.5,6 So, of 10 patients for whom we prescribe antidepressants, two withdraw, three continue without benefit and five continue with benefit, of whom two would have benefited from placebo.
This account omits certain considerations such as compliance and is meant not to be statistically watertight but to give an overview. It indicates the marginal role of antidepressants, the importance of finding out what makes one depressed person consult unlike the other three who don't, and the potential value of depression screening questionnaires.
Footnotes
↵* By my choice of list size and workload I have arranged for these figures to match: that does not mean they represent the same patients.
- © British Journal of General Practice, 2009.