Identifying people at high risk of emergency hospital admission

BMJ 2005; 330 doi: https://doi.org/10.1136/bmj.330.7486.266 (Published 03 February 2005) Cite this as: BMJ 2005;330:266
  1. Jill Morrison, professor of general practice (jmm4y{at}clinmed.gla.ac.uk)
  1. General Practice and Primary Care, Division of Community Based Sciences, University of Glasgow, Glasgow G12 9LX

    Simply measuring previous hospital admission rates would be misleading

    Predicting who is at high risk of becoming acutely ill and requiring emergency admission to hospital has become an important issue for the NHS in England. The Department of Health intends that people with complex long term conditions, who are frequent users of unplanned secondary care, will be identified.1 After identification, a community matron or other health professional will care for these people by using case management, which has previously been defined as “a programme to intensively manage individual patients with uncontrolled or high cost conditions.”2 Community matrons could come from any branch of nursing, but district nurses are considered to be the group that will find it easiest to adapt to the role. This approach could reduce the number of avoidable emergency hospital admissions and help to meet the public services agreement target to reduce inpatient emergency bed days by 5% by March 2008. But how are people at high risk of emergency admission to be identified?

    The Department of Health reports that health and social care partners can agree the criteria now for selecting people for case management by “drawing on the good practice already established in different parts of the country.”1 It suggests that the criteria are likely to take account of the frequency and length of previous stays in hospital, the number of medical and other problems, the number of medicines taken, general practitioners' consultation rates, and other high risk factors such as the death of a carer.

    Some evidence about the utility of one of these criteria is reported in this issue (p 289).3 Roland et al used administrative data to track subsequent patterns of emergency admissions of individual patients aged 65 or older who had two or more emergency admissions to hospital in one year: the same criteria used to identify patients for enrolment in the NHS Evercare pilots.2 4 The high risk cohort was identified in 1997-8 and tracked for five years by using hospital episode statistics, which allow separate inpatient admissions to be identified. The researchers found that total admissions and bed days for the cohort decreased sharply in subsequent years. In the 65-74 age group the cohort had an admission rate 20 times greater than the general population of the same age in 1997-8. This ratio fell to 5.2 in 1998-9 and by 2002-3 it was 1.7. The entire cohort, comprising 2.9% of the total population of England aged over 65 in mid-1997, accounted for 38% of emergency admissions in 1997-8 but only 3.2% in 2002-3.

    The authors assume that the progressive reduction in admissions and bed days is likely to be due to deaths among the cohort and to regression towards the mean among surviving patients but they were not able to assess the relative contributions of these explanations. Caution is required when assuming that patients identified as being at high risk, based on their previous rate of emergency admission to hospital, will continue to be at high risk of emergency admission.

    The implication of this research is that assessing an intervention such as case management in a cohort of older people by simply measuring its effect on the hospital admission rates of the cohort would be very misleading because their admissions would tend to decline regardless of the intervention. Roland et al show that, in evaluations of interventions designed to reduce emergency admissions to hospital, comparing the intervention group with a carefully selected control population is essential.

    When evaluating case management by community matrons, methods and outcome measures will need to be chosen with care to ensure that the true effect of this initiative on emergency admissions is measured. The wider impact of the plan to transfer 3000 experienced district nurses or other health professionals from their current roles to undertake this new role should also be considered because this might leave gaps in the delivery of other community based services.

    Primary carep 289


    • Competing interests None declared


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    View Abstract

    Sign in

    Log in through your institution