Systematic review
Discharging patients with chronic disease: effect of various forms of cooperation between hospital and the community health service
Systematic review
|Updated
The report review available research on effect of various forms of coordination between hospital and the community health service when discharging patients with chronic illnesses.
Key message
The Norwegian Knowledge Center for Health Services was asked by Diakonhjemmet Hospital, Oslo, to review available research on effect of various forms of coordination between hospital and the community health service when discharging patients with chronic illnesses. This review is thought to inform about choices of interventions and consideration of how to design a coordination intervention.
Key results
We included 45 studies about coordination interventions, performed both with and without other interventions at the same time. This is a subject of substantial interest, and the studies also imply a large number of different way of operationalisating the concept. Regardless of whether the coordination was studied as the only element in an intervention, or alongside other interventions, there were no significant differences between the intervention and control group for the outcomes readmissions and deaths. Of four studies performed without other simultaneous interventions, only one was of such quality that we trust the documentation of effect of the intervention: Inviting the GP to the hospital before discharge may possibly increase the proportion of geriatric patients receiving community services after 6 months, than if the GP is not invited. In the remaining studies, the coordination was a part of a multifaceted intervention where also other interventions were performed simultaneously.
It is possible that: Coordinators provided with considerably more time than usually available, to provide case management, including telephone follow-up, liaison with local councils and nursing agencies, coordination of service provision instead of usual hospital discharge planning may result in a higher proportion living at home after one month, a more extensive use of community services, personal care and higher costs within 6 months. Comprehensive geriatric assessment in addition to the usual emergency department care, and a referral faxed from the emergency department to the community agency to expedite home care services may result in lower use of nursing home after 30 days, but no significant differences in costs after 30 days and 6 months or in use of nursing home after 6 months. We evaluated the quality of the documentation for all the reported outcomes as low or very low. This means that we expect further research to influence our conclusions.
Summary
Background
The Norwegian Knowledge Center for the Health Services was asked by Diakonhjemmet Hospital in Oslo to review the effect of coordination between hospital and the community health services while discharging patients with chronic illnesses. They wanted a contribution to the discussion on how to choose between various interventions and to know if there was any documentation that one coordination at discharge lead to better and safer treatment of patient, less days in hospital and less use of various resources for either the hospital or the community health service.
The outcomes of interest were patient related such as quality of life, quality of treatment or health related outcomes. The outcomes could also be related to the hospital or community health service like use of resources, quality of coordination, development of skills or readmissions. To be classified as a coordination intervention in this project, the study had to describe how the coordination took place. We included studies where at least one representative from the hospital cooperated with at least one person from the community health services.
Methods
The systematic literature search was conducted August 16, 2011, and updated August 9th and 10th 2012. We searched in these databases: Medline, EMBASE, Cinahl, Cochrane Database of Systematic Reviews, Cochrane Central Register of Controlled Trials (CENTRAL), DARE and HTA. Inclusion criteria: Population People with chronic diseases, including psychiatric diseases, children and adults with congenital injury that need medical treatment Intervention All specific interventions involving both the hospital and the community health services cooperating in relation the discharge of patients with chronic disease
Comparison
Discharge without specific coordination between the hospital and the community health services. Standard discharge where coordination is not described. Other types of coordination than the intervention
Outcomes
- Patient outcomes: functional level, treatment outcomes, undesired events, satisfaction, quality of life, mortality and disease specific indicators.
- Organizational outcomes: length of stay, readmissions, resources used, quality of coordination, efficiency, work environment, improvement of skills, and knowledge.
Design Systematic reviews of high quality.Primary studies of the following designs: randomized controlled trials, clinical controlled trials, controlled before and after studies, and interrupted time series.
The results from the search for literature were evaluated by two persons independently. Risk of bias in the studies was assessed by checklists.
The quality of the documentation for each outcome was evaluated with GRADE.
Results
The literature search resulted in 11304 references. A total of 504 articles were read in full text and we finally included 51 articles from 45 studies. Four studies included a coordination intervention without other elements. These comprised letters from the hospital together with various documents, an invitation to the GP to visit the hospital before discharge, and contact between the pharmacist at the hospital and the pharmacy in the community. We sorted the 41 complex interventions into three groups: coordination; planning; and transmission. We looked for similarities between the interventions. However, none were similar enough that we could combine them in a metaanalysis.
Some elements in the interventions appeared more often in some groups of patients, i.e. patient education was more often studied for other groups of patients than geriatric patients. Seven of the 45 included studies were performed with methodology that reduce risk of bias, thus we have greater confidence in these effect estimates than we have for the estimates from the other 38 studies. One of these studies included just a coordination intervention, and is the only study that we have some confidence in the estimate the effect of cooperation: Inviting the GP to the hospital before discharge may possibly increase the proportion of geriatric patients receiving community services after 6 months, than if the GP is not invited. The other six studies that evaluated a coordination intervention as part of a multifaceted intervention, found no significant differences between the intervention and control group with regard to the proportion of readmissions or deaths. For other outcomes the effect estimates varied.
The studies with the highest quality of the documentation show that: Coordinators with considerably more time than usually available, providing case management, including telephone follow-up, liaison with local councils and nursing agencies, coordination of service provision instead of usual hospital discharge planning may possibly result in a higher proportion living at home after one month, a more extensive use of community services, personal care and costs after 6 months. Comprehensive geriatric assessment in addition to the usual emergency department care, and a referral was faxed from the emergency department to the community agency to expedite home care services may possibly result in lower use of nursing home after 30 days, but possibly no significant differences in costs after 30 days and 6 months or in use of nursing home after 6 months. In addition to usual care guidance from the study pharmacist, home visits by the pharmacist after 1-2 weeks and that the hospital pharmacist contacted the pharmacy of the patient, may possibly contribute to patients with coronary artery diseases taking beta-blockers according to guidelines after six months when this was measured by the pharmacy, not self-reported by the patient.
Discussion
Seven studies were performed with methods that implied low risk of bias, while the other 38 had unclear or high risk of bias. Six of these seven studies were multifaceted interventions, where the authors have reported the effect of the multifaceted intervention. All studies have different interventions, and we should be careful of trusting the effect estimates from single studies as the context of the study could be essential for the result. By studying the intervention in other settings it may be possible to do a metaanalysis and increase confidence in the results.
Conclusion
There are many studies evaluating coordination when discharging patients with chronic illness from the hospital to the community health service, however, identifying an effect was difficult. There were several coordination interventions and they differed, both in terms of content and in terms of which interventions they were combined with.
One study with low quality of the documentation studied one coordination intervention without other elements. Inviting the GP to the hospital before discharge, may possibly increase the proportion of geriatric patients receiving community services after 6 months, than if the GP was not invited. A large number of studies on this topic are published every year.
Hopefully, in the future, it will be possible to have a more solid documentation than at present. With the great attention on coordination in all parts of the Norwegian health system, it is a unique opportunity to conduct good and important studies to evaluate the effect of such interventions.