M Hickman1, H Sutcliffe, A Sondhi, G V Stimson. 1. Centre for Research on Drugs and Health Behaviour, Social Science and Medicine, Imperial College School of Medicine, London.
Abstract
BACKGROUND: We report detailed findings of the first systematic validation of a Regional Drug Misuse Database (RDMD); such databases constitute the main investment in routine drug statistics in the UK by the Department of Health. METHODS: A retrospective case-finding study in a stratified random sample of one in three specialist drug agencies was carried out. Agency records of clients attending during 1994 were matched with reports (episodes) to the North Thames RDMD to assess the level of under-reporting, and the relationship between RDMD reports (episodes) and the number of problem drug users in contact with agencies. Under-ascertainment of cases was estimated using two-sample capture-recapture. RESULTS: Under-reporting was associated with agency records missing full date of birth or initials (attributers), and agency type. Compared with drug dependency units (DDU) the odds of under-reporting were 3-18 times higher by the other specialist drug agencies. Even after excluding episodes with missing attributers the odds ratio (OR) of not being reported was significantly higher among needle exchanges (OR 2.7), non-statutory community based drug teams (OR 3.2), statutory community based drug teams (OR 4.9) and residential rehabilitation units (OR 8.7) compared with DDUs. Overall database episodes represented 60 per cent of the number of clients attending specialist agencies as a result of a mixture of under-reporting and the proportion of clients retained in treatment, which also varied by agency type. A total of 727 individuals (16 per cent) had never been reported. CONCLUSIONS: Surveillance of drug misuse through RDMDs does not yet fulfil its objectives. It is essential that a system of following up reports is introduced to improve their utility, and to contribute to the monitoring of the UK Government's new drugs strategy, and wider European surveillance.
BACKGROUND: We report detailed findings of the first systematic validation of a Regional Drug Misuse Database (RDMD); such databases constitute the main investment in routine drug statistics in the UK by the Department of Health. METHODS: A retrospective case-finding study in a stratified random sample of one in three specialist drug agencies was carried out. Agency records of clients attending during 1994 were matched with reports (episodes) to the North Thames RDMD to assess the level of under-reporting, and the relationship between RDMD reports (episodes) and the number of problem drug users in contact with agencies. Under-ascertainment of cases was estimated using two-sample capture-recapture. RESULTS: Under-reporting was associated with agency records missing full date of birth or initials (attributers), and agency type. Compared with drug dependency units (DDU) the odds of under-reporting were 3-18 times higher by the other specialist drug agencies. Even after excluding episodes with missing attributers the odds ratio (OR) of not being reported was significantly higher among needle exchanges (OR 2.7), non-statutory community based drug teams (OR 3.2), statutory community based drug teams (OR 4.9) and residential rehabilitation units (OR 8.7) compared with DDUs. Overall database episodes represented 60 per cent of the number of clients attending specialist agencies as a result of a mixture of under-reporting and the proportion of clients retained in treatment, which also varied by agency type. A total of 727 individuals (16 per cent) had never been reported. CONCLUSIONS: Surveillance of drug misuse through RDMDs does not yet fulfil its objectives. It is essential that a system of following up reports is introduced to improve their utility, and to contribute to the monitoring of the UK Government's new drugs strategy, and wider European surveillance.
Authors: Matthew Hickman; Zenobia Carnwath; Peter Madden; Michael Farrell; Cleone Rooney; Richard Ashcroft; Ali Judd; Gerry Stimson Journal: J Urban Health Date: 2003-06 Impact factor: 3.671