Intended for healthcare professionals

Views And Reviews Acute Perspective

David Oliver: Avoiding hospital admission—are we really falling short?

BMJ 2019; 364 doi: (Published 26 February 2019) Cite this as: BMJ 2019;364:l747
  1. David Oliver, consultant in geriatrics and acute general medicine
  1. Berkshire
  1. davidoliver372{at}
    Follow David on Twitter: @mancunianmedic

Last week I wrote about the huge policy push to keep patients away from acute hospitals. The focus isn’t surprising, with full hospitals, rising attendance and admission rates, “exit block” in patients who could technically leave,1 potential distress or risks to patients resulting from admission,23 and patients whose problems might have been prevented or dealt with upstream of hospital and closer to home.45

Here, I want to look at whether the evidence supports such efforts and how we might reframe the issue rather than reflexively labelling acute hospital activity as a problem.

No consistent evidence

The empirical evidence is problematic. Many small scale pockets of innovation have modestly cut hospital attendance or admissions. This has been shown in recent reports on primary care home projects6 or NHS England’s “enhanced health in care homes” vanguards,7 as well as local examples for certain patient groups in NHS England’s Next Steps on the NHS Five Year Forward View.8 But, in a series of systematic reviews, researchers at Bristol and Cardiff found no consistent evidence that interventions to reduce hospital attendance or admissions succeeded.9 A further systematic review of studies focusing on people over 65 with a predicted high risk of admission also found no compelling evidence.10

In 2013 the Nuffield Trust also reviewed a variety of models of integrated care out of hospital and found no consistent evidence that acute admission or bed use could be prevented or that this would save money.11 In 2019 its evaluation of integrated care models reached similar conclusions,12 and Cochrane reviews on “hospital at home,” even for selected patient groups, showed mixed results.13

Lower growth (NB—not reductions) in emergency admission rates was seen from 2014-15 to 2017-18 in NHS England’s “new care models” vanguard sites than in the rest of the NHS.14 But these sites received targeted investment support, and the rates are summarised as “NHS England analysis” and “statistically significant” without having been released for full peer review.

Then we have lessons from recent history. In 2007 an evaluation of case management by community matrons showed that the big cuts in admissions promised by the Department of Health did not materialise.15 National targets for 3.4% reductions in urgent activity were set in 2012 but never met,16 and the National Audit Office has highlighted repeated failed attempts and missed targets to reduce demand, let alone release savings.16

Let’s make acute hospitals fit for the types of patients who now mostly use them

The flagship government policy of extending GPs’ opening hours has not been shown to reduce acute hospital attendance,17 and NHS 111 seems effective at steering more, not fewer, people to hospital despite plans to roll it out further.18 Overall admissions have risen far more quickly than those classified as ambulatory or urgent care sensitive (and hence theoretically more preventable).19 Meanwhile, the case mix of acute hospitals has grown older, more medically complex, and more acute20 despite considerable success in sending more patients home within the first 24 hours or to ambulatory emergency care.2122

Maybe here the tail is wagging the “dog” of bed capacity, population health demographics, and demand, with a near obsessional focus on reducing acute care demand and activity at all human costs—and promising reductions in financial costs that just can’t be delivered.

Necessary and appropriate

Let’s perhaps acknowledge that hospital is often necessary and appropriate, even if patients end up staying too long. Let’s make acute hospitals fit for the types of patients who now mostly use them—often old, frail, or with multiple long term conditions. The evidence for interventions among people actually in hospital, such as acute stroke units or ward based comprehensive geriatric assessments, is stronger than the evidence for preventing admissions.2324 Patients often freely choose to attend hospital despite awareness of alternatives—and they’re often assured by it.252627

Finally, perhaps we should stop prematurely judging every community or integrated care model solely on whether it reduces hospital use or saves money, instead of the range of other benefits it might deliver for patients, their families, and practitioners.



View Abstract