Sunday, October 25, 2009
Since psychiatric epidemiologic surveys began being conducted in the 1980s a strange pattern has been seen in one of the commonly estimated parameters: lifetime prevalence (LTP). LTP refers to the proportion of a sample who report having major depression during their prior lifetime. Since older people have had longer lives and a greater chance of becoming depressed, intuition suggests that LTP should increase with age. In contrast, most studies have found that it increases in young adult life and then declines subsequently with age (see September posting to this blog). There are a variety of possible explanations: (1) that the first onset of of depression occurs commonly in young people but that this incidence declines with age - this could explain the flattening of an age specific prevalence curve - but not a decline, (2) higher mortality in people with depression, (3) recall bias- ie. people forgetting about, or reframing, their experience of depression when they were young as they get older. These explanations have largely been ignored, however, and most researchers have instead claimed that more recent birth cohorts (those of younger age in surveys) have a higher risk of depression. We have developed some simulations to assess these explanations - and (as noted below) have confirmed that a cohort effect is not a necessary or likely explanation. The other factors can explain the observed pattern, the simulations may be accessed here. If the pattern is due to a cohort effect, however, this will lead to a 10-fold increase in depression lifetime prevalence in elderly people in upcoming decades, a dynamic depicted here.