Skip navigation

Caring for older people after acute illness


Over recent years the time patients spend in hospital has been reducing as pressure on resources and beds has grown and developments in medical technology have improved recovery times. Support can now be provided in a variety of different settings. Between 1982 and 1993 the average length of stay for geriatric patients has fallen by over 60%. At the same time there has been considerable innovation in alternative patterns of care for older people. Where is the most effective and appropriate care provided for older people after acute illness? This was the question that prompted a systematic review to establish the strength of existing research on the costs, quality and effectiveness of different locations of acute, post and sub-acute rehabilitation care for older (65+ years) people.


1. Parker G and colleagues. Best place of care for older people after acute and during subacute illness: a systematic review. J Health Serv Res Policy Vol 5, No 3: July 2000

2. Allen KR, Counsell SR. Josephson RA et al. Improving clinical and functional outcomes of hospitalised older patients with cardiovascular illness, Cardiovascular Reviews and reports 1996; 17:43-57

How was the question tackled?

A strategy was devised to guide the initial search of 25 databases using approaches relevant to the particular database: searching was not restricted by research design, country of origin or language. After the initial scoping work, the main search was focused on papers published since 1988. 'Procite' an electronic reference system was used to support the process although most of the work was undertaken by hand.

Study selection was iterative with the focus on randomised or pseudo-randomised studies. Two quality assessment tools were used to guide the selection of the papers for inclusion in the review. Papers were selected on the basis of study design and comparisons of the location of care. Analysis was predominantly qualitative. The outcomes covered were mortality, length of stay, readmission, physical function, cognitive function, costs, patient satisfaction, quality of life and impact on carers and destination. 84 papers, covering 45 separate trials, were included in the review.

What did the review find?

Services that adjust skill mix , including those that care for patients in settings managed by nurses (nurse-led beds), or attempt to discharge them 'early' or divert them from acute hospitals.

Earlier reviews had been equivocal. With one review suggesting that home care showed a small to moderate benefit in terms of subsequent hospital use. However another recent (Cochrane) review concluded that there was insufficient evidence to determine the effect on patient outcomes or costs to the health service.

The results of this review are clearer. There is no evidence that care given in any such setting adversely affected mortality. All but one of these studies reported higher levels of return home. Trials on the use of nurse-led beds were relatively small, with no firm conclusions on mortality but return home did seem more likely for patients cared for here.

Early discharge schemes do not seem to disadvantage patients in terms of mortality or readmission and indeed may offer benefits in terms of patients returning home. Such schemes do not seem to increase costs. Early discharge must mean significantly early discharge if it is to have an economic effect. Schemes to avoid admission may also have potential for reducing costs - although further study here is needed.

Increased condition-specific expertise in hospital settings such as stroke units, hip units, geriatric assessment units and acute care for elders units. All the units were in specific physical locations where care was provided care, informed by specialist expertise, rather than a team approach.

The results of earlier reviews were variable. A Cochrane review of stroke units concluded that they decrease mortality, improve physical function and improve destinational outcome at discharge. Other reviews reached similar conclusions in respect of geriatric assessment units. However, another Cochrane review of the effectiveness of co-ordinated multi-disciplinary rehabilitation for patients with proximal femoral fracture was more equivocal but nevertheless concluded that there was 'some rationale for the adoption of the model of care'.

This review show that although physically distinct stroke units significantly improve patients' chances of survival, the other types of unit do not have the same effect - although this may need to be countered because stroke units generally deal with more severely ill or older patients. Readmission from all units was generally lower.

Stroke units deliver better outcomes in terms of mortality and return home - but this review suggest these benefits weaken over time. The length of stay was generally lower with a greater likelihood of patients being discharged home. Information about patient's quality of life, satisfaction with care, how long they spend in the place they wish to live and costs are important to future debate about service developments - but are lacking at present.

There are unresolved questions about what aspects of care in a stroke unit, or in an inpatient geriatric assessment unit, actually improves care. The review unearthed no firm evidence about the impact of hip units. Geriatric assessment units appear to improve outcomes in terms of destination at discharge and this benefit appears to endure in terms of follow-up with reduced readmission rates.

Rehabilitation provided as part of inpatient services, in community services or at day hospitals.

Earlier reviews of stroke units and hip units that deliver rehabilitation care suggested that increased survival, improved physical function and return home are achieved in the shorter term. A review of day hospitals concluded that their patents had less functional deterioration and fewer admissions to institutional care although studies that used active (community based) controls showed that day hospitals offered little advantage over other forms of care.

This review suggests that inpatient rehabilitation (largely condition specific) reduces mortality when compared to usual care. This effect may reduce over time: a situation evident in most of the trials. There is no evidence of a difference in mortality rates in community based rehabilitation of day hospitals (but the numbers are small). Several of the trials addressed issues about readmission and length of stay. The inpatient rehabilitation service reported substantially lower readmission. There was little evidence of benefit from community based rehabilitation. Day hospital based rehabilitation has similarly been inadequately evaluated.

Few trials reported on destinational outcomes and fewer still destination beyond discharge - although again inpatient rehabilitation showed significant advantage. The amount of cost data varied. Three papers on inpatient rehabilitation and day hospitals suggest some significant additional costs for the health service - another paper reports the opposite! The remainder are broadly neutral in terms of costs, although the basis of the comparisons was varied.

What has the review told us?

This is an excellent review, which has tackled an important topic. It offers a useful baseline and there are helpful reminders about the need to be systematic and make sure that future well-intended service developments are suitably evaluated. Many recent service developments have not been approached with the right rigour so are under-evaluated. There are many substantial gaps in our knowledge about what works and why in this field. The exceptions are stroke units and inpatient rehabilitation where much is now known about how to make them 'work'.

Some of the more important pointers from the work are however about research methodology and need to be heeded.

Three fundamental issues are raised about the usefulness of systematic reviews in areas of service delivery and organisation.

First , there are problems about the lack of consistency about the way services are described by those involved and where they are provided, for example definitions of stroke units vary.

Second , there are no standards terms to describe different models of care; for example the term intermediate care is used in different locations to describe different approaches.

Third , there are problems about distinguishing between clinical interventions and locations of care.

The team also experienced problems in applying and interpreting methodological quality criteria to trials of service effectiveness, particularly about how outcomes should be assessed. Other problems were encountered which were more to do with the trials and in particular the generalisability of results when the trials has themselves used comprehensive exclusion criteria.

Tackling these conceptual problems added complexity to the management of the search process - it took much longer than planned and meant that much irrelevant material had to be screened.

The review has mirrored findings in a parallel survey [2] that demonstrated that there has been considerable recent service development in areas where this review showed that evidence is weakest. It cannot be right to promote services based on assumption about what will work any more than it is not right develop clinical interventions on assumptions rather than evidence. More work to develop appropriate primary and secondary methods of research valid in these fields should be a priority.