Mitchell A.J., Meader N., Bird V. et al.
British Journal of Psychiatry: 2012, 201, p. 93–100.
Unable to obtain a copy by clicking title? Try asking the author for a reprint by adapting this prepared e-mail or by writing to Dr Mitchell at firstname.lastname@example.org.
The policy emphasis on systematic screening to identify risky drinkers seems justified by this review, which found that without this GPs and other non-specialist doctors and nurses missed about half the risky drinkers they saw. However, that is better than in many screening programmes, prompting the reviewers to query whether these really do improve on clinical judgement.
Summary The featured review amalgamated results from relevant studies to estimate how well medical staff identify alcohol problems in their patients in routine practice, using normal procedures and clinical judgement rather than set screening methods. Typically their judgements were compared against assessments based on diagnostic interviews or patients’ responses to standard assessment questionnaires. The review separately considered findings from primary care, general hospitals and psychiatric units, and divided the nature of the drink problems in to specifically diagnosed dependence, intoxication, and alcohol use problems – a broad range from hazardous drinking to ‘alcoholism’, which accounted for 39 of the 48 studies. Just over half the studies were from the USA; none were from the UK.
Across the 12 relevant studies, primary care doctors correctly identified 42% of problem drinkers and documented this in the medical records of 27%. Only two studies calculated what proportion of patients without such problems were wrongly identified as problem drinkers – about 7%. No studies specifically assessed accuracy of identification of dependence or intoxication.
Across the 23 studies in hospitals, health professionals correctly identified 52% of problem drinkers and documented this in the case notes of 37%. Across nine studies, about 12% of patients without such problems were wrongly identified as problem drinkers. In respect of dependence in particular, 42% of dependent drinkers were documented as such in case notes. In trauma centres and against a benchmark of blood alcohol levels, across four studies health professionals correctly identified 90% of cases of intoxication and recorded this in 76% of case notes. However, they also falsely identified about 39% of patients as drunk when blood tests did not indicate they were.
Across three studies, using their clinical judgement mental health professionals correctly identified 55% of problem drinkers, but falsely identified drinking problems in 16% of patients. In one study, 28% of problem drinkers had been documented as such in case notes. In another, 83% of dependent drinkers had been correctly identified.
Where it was possible to assess this, figures did not substantially differ depending on whether the diagnosis against which clinical judgement was benchmarked had been reached by interviewing the patient or their self-completing a questionnaire.
Broadly the findings indicate that all health care professionals have considerable difficulty identifying problem drinking in routine practice. Relying on clinical judgement, they correctly identified about half those with drinking problems and documented just one in three. However, this was based on a single encounter; from the results of one study, we can expect diagnosis to improve after several consultations. Most successful was the identification of intoxication by emergency department specialists, who correctly spotted nine out of ten cases.
A four-in-ten detection rate in primary care is in line with previous work suggesting that most problem drinkers are not routinely detected, depriving patient and doctor of a chance to deal with the problem. Hospital and psychiatric staff missed about half the problem drinkers they saw, important because drinking can exacerbate physical and mental illness. The low detection rate of mental health professionals may seem surprising, but from other work we know that even in mental health settings, often alcohol problems are not discussed.
Patients generally do not object to being asked about their drinking and will disclose if asked in a sensitive manner. This leaves the willingness of the clinician to appropriately ask about drinking as the most important factor in whether problems get identified. Training can improve recognition rates. Whether clinical judgement or formal screening is the best recognition method remains to be addressed by research.
commentary It seems likely that the detection rates found in the reviewed studies exceed those to be expected in normal practice; during a study doctors and nurses may be sensitive to the fact that their judgements are being ‘checked’ by researchers. Nevertheless, the review exposes the room for improvement in identifying alcohol problems in generic medical settings, one reason why systematic universal or selective screening using research-validated methods has been seen as the way to improve performance.
However, in practice screening is usually far from universal or systematic; the willingness or ability of doctors and nurses to screen patients for risky drinking is the Achilles’ heel of attempts to generate public health gains from screening and brief intervention programmes. A review has documented screening rates above 26% only in (self-)selected practices, special populations (patients already known to have high blood pressure), or when powerful levers were available to incentivise screening. It is not unusual for very small proportions to be screened in studies which rely on doctors and nurses to do the screening rather than researchers. Such findings seem to justify the featured review’s comment that “the added value of screening over and above clinical diagnosis remains unclear”.
Incentives can substantially raise screening rates, though not always with the accompanying quality needed to identify most problem drinkers. Scotland set a national policy target for brief alcohol interventions in medical settings, among which the environment was most conducive in primary care practices, which were incentivised by specific funding available through local enhanced contracts. In three areas investigated in greater depth, across one an estimated 41% of hazardous or harmful drinkers visiting practices were screened, and in the two others areas, about 30%. This was despite the incentive system in those areas being focused on promoting opportunities for screening.
Last revised 18 August 2014. First uploaded 11 August 2014
DOCUMENT 2015 Alcohol-use disorders