In order to capture the differences in burden between the subtypes of depression, the Global Burden of Disease 2010 Study for the first time estimated the burden of dysthymia and major depressive disorder separately from the previously used umbrella term ‘unipolar depression’. A global summary of epidemiological parameters are necessary inputs in burden of disease calculations for 21 world regions, males and females and for the year 1990, 2005 and 2010. This paper reports findings from a systematic review of global epidemiological data and the subsequent development of an internally consistent epidemiological model of dysthymia.

A systematic search was conducted to identify data sources for the prevalence, incidence, remission and excess-mortality of dysthymia using Medline, PsycINFO and EMBASE electronic databases and grey literature. DisMod-MR, a Bayesian meta-regression tool, was used to check the epidemiological parameters for internal consistency and to predict estimates for world regions with no or few data.

The systematic review identified 38 studies meeting inclusion criteria which provided 147 data points for 30 countries in 13 of 21 world regions. Prevalence increases in the early ages, peaking at around 50 years. Females have higher prevalence of dysthymia than males. Global pooled prevalence remained constant across time points at 1.55% (95%CI 1.50–1.60). There was very little regional variation in prevalence estimates.

There were eight GBD world regions for which we found no data for which DisMod-MR had to impute estimates.

The addition of internally consistent epidemiological estimates by world region, age, sex and year for dysthymia contributed to a more comprehensive estimate of mental health burden in GBD 2010.