Size and burden of depressive disorders in Europe

Abstract

We review epidemiological studies of depression in Europe. Community surveys are essential. Methodological differences in survey methods, instruments, nuances in language and translation limit comparability, but consistent findings are emerging. Western European countries show 1 year prevalence of major depression of around 5%, with two-fold variation, probably methodological, and higher prevalences in women, the middle-aged, less privileged groups, and those experiencing social adversity. There is high comorbidity with other psychiatric and physical disorders. Depression is a major cause of disability. Incidence has been less studied and lifetime incidence is not clear, with longitudinal studies required. There is pressing need for prevalence studies from Eastern Europe. The considerable differences in health care systems among European countries may impact on proportions of depressives receiving treatment and its adequacy, particularly in the key area of primary care, and require further study. There is a need for public health programmes aimed at improving treatment, reducing rates and consequences of depressive disorders.

Introduction

The term depression covers a spectrum, ranging from a normal mood of unhappiness to an abnormal, severe and incapacitating disorder. These phenomena appear to range along a continuum without clear cut points (Melzer et al., 2002). As depression becomes more severe it also becomes more pervasive, including additional symptoms. Modern definitions for disorder, as in ICD-10 Depressive Episode (World Health Organisation, 1992) and DSM-IV Major Depression (American Psychiatric Association, 1994), depend upon the presence of a specific number of symptoms to define a severity threshold, lasting at least 2 weeks, with distress or disability, and without the qualifying features of a number of other disorders. Dysthymia, a chronic or lifelong disorder, sets a milder severity threshold, but a minimum length of 2 years.

Statistics on treatment contacts miss the many cases not reaching health care, mandating community surveys. Within a continuum of severity, cut-off points, precise diagnostic definitions and instruments inevitably influence the results. Public health information on treatment needs should be based on thresholds for benefit in treatment trials but severity measures from these are rarely used in surveys. Below ICD-10/DSM IV depression thresholds are two phenomena which are more frequent: depressive symptoms, for which rates are very high (Boyd and Weissman, 1982), and so-called ‘common mental disorder’, mild mixed disorder usually including depressive and anxiety symptoms, but still with only a proportion fitting criteria for specific disorders.

This review primarily concerns defined disorder at ICD-10/DSM IV level, specifically depressive episode/major depression. Dysthymia adds a small amount to this, but most dysthymics ultimately develop depressive episodes (Keller, 1994). The main instruments in current use, the CIDI (Robins et al., 1988) and the SCAN (Wing et al., 1990) and their predecessors from the 1970s and 1980s, the DIS (Robins et al., 1981, Wittchen, 1994), and the PSE (Wing et al., 1978) use somewhat different case definitions and algorithms. Instruments employ either skilled clinician interviewers (SCAN), or fully structured interviews given by lay interviewers (CIDI). Direct comparisons consistently show that the two methods give different findings at symptom and case-diagnosis levels in general populations and also on co-morbidity (Brugha et al., 1999, Brugha et al., 2001), but better agreement for patients attending services. In primary care attenders, using lay interviewers against the psychiatrist administered SCAN, Jordanova et al. (2004) found the CIDI highly valid, and the CIS-R moderately valid.

In comparisons among European countries, with different languages and cultures, even with careful translation and testing, nuances still occur in language, and in readiness of response. There is little evidence available on the effects of this and its influence on rates reported from different European countries, which could be substantial. In order to compare regions or countries it is crucial that identical procedures and methods are used throughout. The most consistent rates found are for 12 months and 6 months.