Author
DEWEY Michael E.; CHEN Chih-Mei;
Title
Neurosis and mortality in persons aged 65 and over living in the community: a systematic review of the literature.
Journal citation/publication details
International Journal of Geriatric Psychiatry, 19(6), June 2004, pp.554-557.
Summary
This review synthesises research findings on associations between mortality in people over 65 and three types of mental condition: anxiety disorders (however defined); phobic disorders; and obsessive-compulsive disorders. Eligible studies used samples drawn from the community rather than from health care facilities, and seven reports met the inclusion criteria for the review. Statistical synthesis of this limited evidence base indicates a non-significant increase in mortality associated with these conditions.Context
Although two recent systematic reviews (referenced) have surveyed the international literature on dementia and mortality, no previous attempt has been made to examine what is known about mortality among older people and ‘those illnesses loosely defined as neurosis'. This review aims to fill the gap.
Methods
What sources were searched?
Electronic searches of EMBASE (Excerpta Medica), 1980-2002, and Medline (1966-2002), combined with searches of personal files and checking of references to papers identified in the searches.
What search terms/strategies were used?
The search strategy combined ‘mortality' with ‘anxiety' or ‘phobic disorders' or ‘obsessive-compulsive disorders', limited by ‘aged 65+ years'. Journal papers only were considered, published in English, French, German, Italian, Portuguese or Spanish.
What criteria were used to decide on which studies to include?
Eligible studies followed a cohort selected at random from a community, or representing the whole of a geographically limited cohort in the community. In addition, the majority of participants at baseline had to be 65 or over and living in the community (or this data could be extracted from wider study reports). Studies dealing with populations in which illness or contact with a health care facility was a criterion for participation were explicitly excluded, as were studies dealing with neurotic depression and mortality in older people.
Who decided on their relevance and quality?
No details are given.
How many studies were included and where were they from?
Seven studies, all in English from England (2), Sweden , Australia , Germany , Spain and the Netherlands .
How were the study findings combined?
The synthesis is statistical, and the language of the paper assumes statistical knowledge. The analysis is based entirely on information in the papers themselves (authors were not contacted for any missing data) and the variety of methods used to report results means that the reviewers were ultimately ‘forced to use a vote counting method (Fisher's method)' to synthesise the findings.
Findings of the review
The statistical synthesis of results from the six papers using a diagnostic approach and reporting an effect size produced an overall p -value of 0.08, which is not significant. The remaining study used a symptom score, and found no significant relationship between anxiety and mortality.
The reviewers note the disadvantages associated with using vote-counting approaches, and the likely implications for the results of their synthesis: ‘we may have gained a different result if we had been able to use ORs [odds ratios] or relative risks directly'. In addition, they note ‘the issue of combining studies which have used different diagnostic entities', and the fact that the small number of studies precludes the investigation of publication bias or formal testing for heterogeneity.
Authors' conclusions
‘Individual studies in this area have not been conclusive' and there is ‘a dearth of knowledge' about the possible excess mortality associated with the neurotic disorders covered by the review. This may be because the prevalence of dementia and depression has overshadowed these less frequent, though serious, conditions.Implications for policy or practice
None are discussed.