This trial demonstrated that elderly patients with acute diseases or deterioration of a chronic disease initially handled at a general hospital and subsequently offered intermediate care, had lower readmission rates (p = 0.03), and had a higher number of patients independent of community care (p = 0.02) than patients given traditional prolonged care at a general hospital. The differences in total number of days with inward care were minor. The differences in number of deaths and need of home care were in favour of the intervention group, however, statistically insignificant.
All patients were transferred immediately after randomisation to the community hospital except the eight patients with a severe and acute deterioration of their disease. These patients could have been treated as readmissions for the same diseases in the intention-to-treat analyses. However, the decisions not to transfer these patients were undertaken by the physicians at the general hospital and not by the physicians at the community hospital. Treated as readmissions in the statistical analyses resulted in an insignificant reduction of the number of readmissions (p = 0.14, adjusted p = 0.11) and an insignificant difference in number of days readmitted in favour of the intervention group; 4.4 (95% CI 2.6–6.9) days versus 7.6 (95% CI 3.6–11.6) days.
The present study appears to be the first randomised controlled trial where included patients have been an unselected general hospital population above 60 years of age. Another strength of this trial was that all patients received the same optimal care in the initial stage of their illness before randomisation.
As one of the authors, blinded for which group the patients belonged to, collected all information from medical records and from the patient administrative systems, information bias by collection was possible. As all data was objective measures as readmissions for the same disease, use of home care and number of deaths, the registration was considered to be accurate.
Several efforts have been developed to reduce number of days of inward care and to facilitate discharge from general hospitals including discharge planning, nurse led inpatient care, hospital at home regimes, general practitioners hospitals, community hospitals and patients hotels . Some studies have found a better functional outcome and reduced mortality when older patients were treated at specialised geriatric wards [17–19], whilst the benefit of early supported discharge of stroke patients was ascribed a structured collaboration between primary and secondary health care [20, 21].
Several community hospitals in Norway are comparable with community hospitals in England [7, 8] and general practitioners hospitals in Holland  where some studies have explored their appropriateness [11, 12, 22–25]. In Norway the use of nursing homes and community hospitals may have been overlooked as appropriate alternatives, and research on such models has been sparse both nationally and internationally [7, 22].
A limitation of performing intermediate care is the lack of possibility to identify which of the components that are working so well. However, some of the main components in the intervention were assessments of ADL and a consecutive and closely communicating and cooperating with each patient and his social and professional networks to identify the best supportive solutions. This communication, including the continuous dialogue with the rest of the primary health care in the municipality, was probably the central element of the care that seems to be efficient in reducing the number of readmissions for the same disease, the need of community care and allowing the professional teams to optimise the follow-up after discharge.
The communication process is always complex. Older people are a more heterogeneous group than younger people, and maybe they have experienced several more or less successful diagnosing and treatment procedures. Health personal and older people can have different perception of what are illness and the consequences of illness. As a consequence, unclear communication can cause the whole medical encounter to fall apart.
Intermediate care at a community hospital seems to be highly effective.
In a modern health care system care is more and more specialised, fragmented and organ-focused. In addition to the expansion of further sub-specialising in modern medicine, the results from this study underscore the additional need of better step-down care systems at an intermediate level. It is indeed relevant to question the appropriateness of prolonged traditional general hospital care for this rapidly increasing group of patients.
There are little existing scientific evidence of the benefits of intermediate care  and more randomised controlled trials are necessary to test different models for intermediate care at community hospitals as alternatives to general hospital admissions and as alternatives to prolonged general hospital care to confirm any benefits of intermediate care. Additionally, the economic consequences have to be explored.