What is the aim of this review?
To find out whether short stays in hospital in short-stay units improve outcomes in adults with internal medicine diseases and conditions compared to usual care.
Key messages
We are unsure about the effect of short-stay unit hospitalisation for adults with internal medicine diseases and conditions compared to usual care. The evidence was uncertain for several important reasons; including not having enough data, differences among participants and co-interventions, and problems with the methods that the trials used that could have led to false results. We need more high-quality trials to test the impact of short-stay unit hospitalisation on individual patients and costs.
What was studied in the review?
Short-stay units are hospital units that provide short-term care in selected patients. Their services are often defined by the type of patient, the unit's function, and a time frame. Studies have indicated that short-stay units may reduce the number of people admitted to hospital, the length of time they spend in hospital, the number of people who have to go back into hospital (readmission), and costs, without losing any quality of care, but a thorough evaluation of effects of short-stay unit hospitalisation compared with usual care (mainly hospitalisation in a traditional hospital ward) was lacking before we conducted the present review. The review focused on short-stay unit hospitalisation for internal medicine diseases and conditions, such as pneumonia or chest pain. We compared the effect of short-stay unit hospitalisation with usual care by looking at deaths (mortality), serious problems (serious adverse events), quality of life, activities of daily living (such as managing housework or medications ), hospital readmission, non-serious adverse events, and costs.
What are the main results of the review?
The review authors found 14 relevant trials with a total of 2872 participants. All trials were randomised trials, i.e. people participating in the trials had been assigned by chance alone to either hospitalisation in a short-stay unit or a control group that received usual care. Randomised trials are considered to be the most reliable trial design to test effects of an intervention.
Thirteen trials evaluated short-stay unit hospitalisation for six specific conditions (asthma, atrial fibrillation (irregular heartbeat), chest pain, decompensated (worsening of the signs of) heart failure, syncope (losing consciousness due to a fall in blood pressure), and transient ischaemic attack (mini stroke)) and one trial did not specify which condition its participants had. We identified one ongoing trial. The components of short-stay unit hospitalisation differed among the trials depending on the trial participants' conditions. All of the included trials had problems with their methods that potentially could have led to false results. We were uncertain whether there was any difference between short-stay unit hospitalisation and usual care for reducing mortality, serious adverse events, and hospital readmissions. We were not able to combine and examine results for any other outcomes, because the trials used different ways of measuring (e.g. using different scales), or did not give enough data, or reported their results in a way that meant we could not use them. Individual trials said that short-stay unit hospitalisation led to higher quality of life, fewer non-serious adverse events, and lower costs. However, we cannot be certain about this evidence and need to be careful about interpreting the trials' results; all included trials were at high risk of errors, which questions the validity of these results and we cannot exclude that the findings were merely due to the play of chance. It is crucial to validate the findings in larger, well-conducted trials.
How up-to-date is this review?
The review authors searched for trials that had been published before 13 December 2017.
Overall, the quantity and the certainty of the evidence was very low. Consequently, it is uncertain whether there are any beneficial or harmful effects of short-stay unit hospitalisation for adults with internal medicine diseases and conditions - more trials comparing the effects of short-stay units with usual care are needed. Such trials ought to be conducted with low risk of bias and low risks of random errors to improve the overall confidence in the evidence.
Short-stay units are hospital units that provide short-term care for selected patients. Studies have indicated that short-stay units might reduce admission rates, time of hospital stays, hospital readmissions and expenditure without compromising the quality of care. Short-stay units are often defined by a target patient category, a target function, and a target time frame. Hypothetically, short-stay units could be established as part of any department, but this review focuses on short-stay units that provide care for participants with internal medicine diseases and conditions.
To assess beneficial and harmful effects of short-stay unit hospitalisation compared with usual care in people with internal medicine diseases and conditions.
We searched CENTRAL, MEDLINE, Embase, three other databases and two trials registers up to 13 December 2017 together with reference checking, citation searching and contact with study authors to identify additional studies. We also searched several grey literature sources and performed a forward citation search for included studies.
We included randomised trials and cluster-randomised trials, comparing hospitalisation in a short-stay unit with usual care (hospitalisation in a traditional hospital ward or other services). We defined a short-stay unit to be a hospital ward where the targeted length of stay in hospital for patients was five days or less. We included both multipurpose and specialised short-stay units. Participants were adults admitted to hospital with an internal medicine disease or condition. We excluded surgical, obstetric, psychiatric, gynaecological, and ambulatory participants. Trials were included irrespective of publication status, date, and language.
We used standard methodological procedures expected by Cochrane. Two review authors independently extracted data and assessed the risk of bias of each included trial. We measured intervention effect sizes by meta-analyses for two primary outcomes, mortality and serious adverse events, and one secondary outcome, hospital readmission. We narratively reported the following important outcomes: quality of life, activities of daily living, non-serious adverse events, and costs. We used risk ratio differences of 15% for mortality and of 20% for serious adverse events for minimal relevant clinical consideration. We rated the certainty of the evidence and the strength of recommendations of the outcomes using the GRADE approach.
We included 19 records reporting on 14 randomised trials with a total of 2872 participants. One trial was ongoing. Thirteen trials evaluated short-stay unit hospitalisation for six specific conditions (acute decompensated heart failure (one trial), asthma (one trial), atrial fibrillation (one trial), chest pain (seven trials), syncope (two trials), and transient ischaemic attack (one trial)) and one trial investigated participants presenting with miscellaneous internal medicine disease and conditions. The components of the intervention differed among the trials as dictated by the trial participants' condition. All included trials were at high risk of bias.
The certainty of the evidence for all outcomes was very low. Consequently, it is uncertain whether hospitalisation in short-stay units compared with usual care reduces mortality (risk ratio (RR) 0.73, 95% confidence interval (CI) 0.47 to 1.15) 5 trials (seven additional trials reporting on 1299 participants reported no deaths in either group)); serious adverse events (RR 0.95, 95% CI 0.59 to 1.54; 7 trials (one additional trial with 108 participants reported no serious adverse events in either group)), and hospital readmission (RR 0.80, 95% CI 0.54 to 1.19, 8 trials (one additional trial with 424 participants did not report results for participants)). There was not enough information to confirm or refute that short-stay unit hospitalisation had relevant effects on quality of life, activities of daily living, non-serious adverse events, and costs.