Hospital at home: from red to amber?

Hospital at home schemes providing care in the patient’s home that is traditionally provided in hospital have grown in importance in health services in both Europe and North America and are seen as a possible substitute for inpatient care in the National Health Service.¹ The limited experience of hospital at home developments in the United Kingdom suggests that savings can be made when such services are substituted for usual hospital care, at least for some patient groups,² but a recent systematic review of the English language literature provides little evidence to support this innovatory approach to acute care.³ Few trials of hospital at home services have, however, been done, and most have been small, with no consistency in outcome measures and little attempt at economic evaluation.

The many purchasers and providers planning hospital at home schemes⁴ have had little guidance on how to proceed, with the result that many recent attempts to create this type of service have been unplanned experiments. Economic evaluation of some new hospital at home services in London found them to be more expensive than usual care,⁵ and several services have been decommissioned as a result. The separate trials reported in this issue (pp 1786, 1796)⁶,⁷ are therefore important contributions to our understanding of the potential for developing hospital at home schemes in the NHS, not least because they incorporated economic evaluations in their designs (pp 1791, 1802).⁸,⁹

The results of these trials will reassure advocates of hospital at home schemes without satisfying sceptics. In both centres hospital at home seems as effective and as acceptable to patients as routine hospital care, although in the Northamptonshire study the trial did not have the power to detect differences in mortality and morbidity.⁶ The exceptions to effectiveness and acceptability seem to be patients who had under- gone knee replacements and those with chronic obstructive airways disease—and will not surprise clinicians. Hospital at home patients had more days of care than their inpatient counterparts, but this finding is difficult to interpret. Were hospital at home teams having difficulty discharging patients, perhaps with perverse incentives to hold on to them during periods of underutilisation? Or is discharge from inpatient care sometimes premature, so that recipients of hospital at home services get longer, but more appropriate, care?

The economic evaluations of the two services come to opposing conclusions, with reduced costs for hospital at home patients in Bristol,⁹ despite their greater lengths of stay, and higher costs in Northamptonshire for elderly medical patients and those with obstructive airways disease.⁸ However, the sensitivity analyses are crucial, because hospital at home costs in Bristol would exceed usual inpatient care costs only if the latter were reduced by 50%, while in the Northamptonshire study a reduction in hospital at home care of only one or two days could alter the study’s conclusion, at least for some patient groups.

Where does this leave providers and commissioners hopeful that hospital at home services could solve some of their service delivery problems? Paediatric hospital at home schemes are well established, as are some forms of highly focused, high technology medical care,¹ but new services aimed at older patients with a wide range of medical and surgical problems remain problematic. These two trials do not and could not answer fundamental questions about the value of hospital at home as a substitute for usual inpatient care. They are too small, and (despite the efforts of a research group hosted by North Thames region to coordinate trial development) they are difficult to compare and combine because they use different outcome measures.

More importantly, the results of these studies seem to be contingent on characteristics of local services that may have influenced the application of eligibility criteria, recruitment to the study, and length of stay in and timing of discharge from hospital at home. Previously untried features of an innovative service, like the special payments to general practitioners for caring for hospital at home patients in the Northamptonshire study,⁶ may have unforeseen effects on care pathways. Descriptive studies of the organisational culture and practice of innovative services are needed to place their findings in context and might be useful components of future studies, since pragmatic randomised controlled trials alone seem to be necessary but insufficient guides for service development.

Nevertheless, the two trials reported in this issue do provide useful pointers for service developers. It seems that hospital at home can substitute for usual hospital care for some diagnostic groups, without adverse effects on patients, and potentially with release of resources. Resources are not always released, however, and the outcome may be supplementation of existing hospital services, at overall greater cost to the local health service. Supplementary services may be desirable for those commissioners who can afford them, but knowing whether an innovation will supplement or substitute for existing services matters greatly. The size of hospital at home schemes and their case mix are clearly important factors that influence the impact of these schemes on other services.

More trials are about to report, and in a year or so the picture may be very different. Commissioners and providers impatient with academic conservatism and the quest for evidence may scorn suggestions that “more research is needed,” but they might be wise to wait just a little longer before giving hospital at home the green light.

General practice pp 1786, 1791, 1796, 1802