Family Practice Vol. 20, No. 6, 682-684
© Oxford University Press 2003, all rights reserved
Article |
Performance of a single screening question for depression in a representative sample of 13 670 people aged 75 and over in the UK: results from the MRC trial of assessment and management of older people in the community
Department of Epidemiology, and a Centre for Ageing and Public Health, London School of Hygiene and Tropical Medicine, Keppel Street, London WC1E 7HT, b Section of Care of the Elderly, Imperial College, Faculty of Medicine, Hammersmith Campus, Du Cane Road, London W12 ONN, c University Department of Geriatric Medicine, University of Wales College of Medicine, Cardiff CF64 2XX and d Unit of Health Care Epidemiology, University of Oxford, Oxford OX3 7LF, UK
Correspondence to Dr DPJ Osborn, Department of Psychiatry and Behavioural Sciences, Royal Free Campus, Royal Free and University College Medical School, Rowland Hill St, London NW3 2PF, UK; E-mail: dosborn{at}rfc.ucl.ac.uk
Received 17 February 2003; Revised 30 June 2003; Accepted 14 July 2003.
Osborn DPJ, Fletcher AE, Smeeth L, Stirling S, Bulpitt CJ, Nunes M, Breeze E, Ng ESW, Jones D and Tulloch A. Performance of a single screening question for depression in a representative sample of 13 670 people aged 75 and over in the UK: results from the MRC trial of assessment and management of older people in the community. Family Practice 2003; 20: 682–684.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Background. A concise, accurate screening question for depression would be an important contribution to the Single Assessment Process for Older People.
Objective. To examine the performance of a previously validated screening question for depression, in a large community sample.
Methods. Both the single screening question, and the Geriatric Depression Scale (GDS-15) were completed by 13 670 people aged 75 and over in the community. Responses to the question were compared with a ‘standard’ of scoring above different thresholds on the Geriatric Depression Scale (GDS-15).
Results. For more severe GDS-15 depression, the best performance of the question was a sensitivity of only 52% and a specificity of 93%.
Conclusion. Even at its best, the question therefore misses almost half the cases. This highlights the problems of such simple approaches to routine screening.
Keywords. Depressive disorder diagnosis, mass screening, psychiatric status rating scales, sensitivity and specificity.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Depression is common in older people. It has wide ranging social, physical and psychological effects, but under-recognition and under-treatment is widespread. The National Service Framework (NSF) for Older People1 acknowledges this and emphasises community-orientated protocols to diagnose and treat depressed patients. The vehicle will be the Single Assessment Process. The Single Assessment Process working party is advising on this process, which must be comprehensive, multidimensional, yet matched to individual circumstances. There is an urgent need for accurate, concise screening questions in each area assessed. These must be brief enough to be realistically employed in over-stretched primary care settings, where depression is only one of many important foci.
Watkins et al. recently validated a single screening question for depression in stroke patients.2 They used a question which had previously compared favourably with the Geriatric Depression Scale (GDS). The MRC trial of assessment and management of older people in the community (MRC Trial) contained a version of this single question, allowing us to assess its performance in a large representative sample of older people in the UK.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The main aims and methods of the MRC Trial have been described elsewhere.3 In the initial ‘brief’ assessment, all trial participants were asked the single question ‘Do you feel sad, depressed or miserable’, with four possible answers ‘always’, ‘often’, ‘occasionally’ and ‘never’. The 15 item GDS (GDS-15),4 was also routinely offered to all participants aged 75 and over, within one arm of the trial (the ‘universal’ arm), allowing comparison with the single question. A recent review of rating scales in old-age psychiatry concluded that the GDS is the standard against which other depression scales should be rated.5 The GDS-15 was administered by a trained research nurse approximately two weeks after the single question. GDS scores in this arm of the MRC trial have already been reported,3 but results from the single question have not been fully analysed before. We examined accuracy of the single screening question against different recognised thresholds on the GDS-15.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
There were 53 GP practices, chosen to be representative of UK mortality and deprivation scores. From these practices, 13 670/21 241 (64.4%) of those eligible provided complete information on the single question and on the GDS-15. The median age of the non-responders at invitation was 81.5 years (interquartile range 78.0–85.7) compared with 80.1 years (interquartile range 77.1–84.0) for responders, P < 0.001. Women (61.9%) were less likely to respond than men (68.6%), P < 0.001. The single question had three possible thresholds. Table 1 shows the prevalence of responses above each threshold, and the performance of each single question threshold compared to different scores on the GDS-15. The standard GDS-15 threshold is
6, but lower cut-offs (3 and 5) have been suggested for screening, and a higher threshold (8) suggested to increase specificity, or to detect depression of greater severity.3
|
We compared the performance of each single question threshold. The highest specificity (more than 99% at each GDS cut-off) is observed with the response ‘always sad’ but at a cost of very poor sensitivity (less than 13%). Conversely, sensitivity is increased (more than 79% at each GDS cut-off) when three categories of single question response are combined (‘always/often/ occasionally’), but specificities are low (less than 60%). For severe depression (GDS
8), the most important condition to identify, the ‘best’ performance comes from the response ‘often/always’, where sensitivity is 52% and specificity is 93%. |
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The single question does not perform well as a routine screening tool in the community. In practice, the only usable threshold would be ‘always or often sad’, which 8.1% reported. The ‘always sad’ response is impractical because it consistently misses most cases. Since 61.3% reported the lowest threshold (‘always/often/occasionally sad’), the burden of false positives generated would be immense and unworkable. Even at the ‘often/always’ cut-off, only half the cases of more severe depression (
8 on GDS-15) are detected, and almost four fifths of these are false positives. Our data highlight the dangers such questions may present in routine assessments. Whilst the single question is certainly brief, it is unlikely to facilitate better care for the majority of depressed older people in the community. It would, however, generate much more work for primary care services and might lead to false assumptions about the lack of depression in those who do not respond ‘often’ or ‘always’.
|
Acknowledgments |
|---|
We would like to thank the following people. The nurses, GPs, the other staff and the patients in the participating practices. Everyone at the MRC General Practice Research Framework co-ordinating centre, particularly Jeannett Martin and Nicky Fasey. Clerical staff at London School of Hygiene & Tropical Medicine (Janbibi Mazar and Rakhi Kabawala) and Hammersmith Hospital (Ruth Peters) for all their work on the trial. The staff of the Research Team for Elderly People, University of Wales College of Medicine, for all their work. The Trial Steering Committee: Professor J Grimley Evans (chair), Professor A Haines (previous chair), Professor K Luker, Professor C Brayne, Dr M Vickers, Professor M Drummond, Dr A Glanz, Dr L Davies.
AF, AT, CB and DJ are the investigators of the MRC Trial of the Assessment and Management of Older People in the Community. AF, LS, DO devised the idea for the paper. DO analysed the data and wrote the paper. LS supervised the analysis. MN and EB were involved in administering the trial and editing data. SS devised the randomisation procedure, monitored data collection and took part in training the nurses and lay interviewers. EN took part in data management and cleaning. All authors commented on drafts of the paper.
The MRC Trial of the Assessment and Management of Older People in the Community was funded by the UK Medical Research Council (Grant number G9223939), the Department of Health and the Scottish Office. DO was funded by a research fellowship from the Medical Research Council. LS was funded by a research fellowship from London NHS Executive.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
1 Department of Health. National service framework for older people. London: Department of Health, 2001.
2 Watkins C, Daniels L, Jack C, Dickinson H, van den Broek M. Accuracy of a single screening question for depression in a cohort of patients after stroke: comparative study. Br Med J 2001; 323: 1159.
3 Osborn DPJ, Fletcher AE, Smeeth LS, Stirling S, Nunes M, Breeze E et al. Geriatric Depression Scale Scores in a representative sample of 14 545 people aged 75 and over in the United Kingdom: results from the MRC trial of assessment and management of older people in the community. Int J Geriatr Psychiatry 2002; 17: 375–382.[CrossRef][ISI][Medline]
4 Sheik JI,Yesavage JA. Geriatric Depression Scale (GDS): Recent evidence and development of a shorter version. Clinical Gerontology 1986; 37: 819–820.
5 Burns A, Lawlor B, Craig S. Rating scales in old age psychiatry. Br J Psychiatry 2002; 180: 161–167.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||