Studying the incidence of depression: An 'interval' effect

Abstract

A review of studies about the incidence of depression suggested that the length of the 'interval' of follow up may influence the findings. Exploration of these issues is carried out using data from the Stirling County Study, an investigation of psychiatric epidemiology in a general population. The study's customary method of diagnosis, DePression and AnXiety (DPAX), and the Diagnostic Interview Schedule (DIS) were used in an incidence investigation whose 'interval' was less than three years. Average annual incidence rates of depression for both DPAX and DIS were about 15 per 1000. Where longer intervals were used in the Stirling Study, rates were close to four per 1000. Projected lifetime risk based on the lower rates was more congruent with reported lifetime prevalence than that based on the higher rates. Irrespective of method, 90% or more of the incident cases gave an onset that predated the initial interview, suggesting poor reliability. This was often due to the fact that information given in the first interview met some but not all of the criteria for diagnosis. Being in the 'borderline' category at the beginning of the study significantly increased incidence. Thus, evidence from the Stirling County Study replicated findings that suggest an 'interval effect' and pointed to the need in incidence studies for distinguishing between the onset of the prodrome and the onset of diagnosable depression.