Background
Readmissions to hospital within 30 days of discharge are generally considered an unplanned or potentially avoidable event [
1,
2]. In the United States (US), 1 in 5 Medicare fee-for-service patients are readmitted to hospital within 30 days of discharge and it is estimated that up to 90 % of readmissions within 30 days are unplanned [
3]. Reported estimations of annual health system costs due to readmission range from $12 billion to $17.4 billion in the US [
3,
4] and £1.6 billion in the United Kingdom (UK) [
5]. Readmissions are also associated with human costs such as feelings of frustration and time lost from an individual’s usual role within the workplace and family [
6].
Readmissions are highest amongst those with chronic diseases, in particular amongst patients with cardiovascular disease, chronic respiratory disease and diabetes [
3,
7‐
10]. Patients with chronic heart failure have been reported to be at the highest risk of readmission to hospital within 30 days [
7,
8,
11] with reported rates of 26.9 % amongst Medicare fee-for-service patients [
3]. Individuals with chronic obstructive pulmonary disease (COPD) and diabetes also have high reported readmission rates (22.6–20 % respectively [
3]) [
7,
8,
12]. Patients with chronic disease discharged from hospital often have complex health care needs and treatment plans, which means the early discharge period is a challenging time for the patient and their carer [
9,
13].
Inadequate discharge planning, poor follow up from community health care services, and a lack of patient and carer education in chronic disease self-management skills are believed to contribute to unplanned readmission [
14,
15]. Healthcare guidelines in the UK and the US penalise hospitals by restricting government payments for excess unplanned readmissions within 30 days of discharge, based on the rationale that readmissions result from suboptimal care and are preventable [
4,
16‐
18]. This has led to increased motivation to find effective strategies to reduce unplanned readmissions [
1,
14,
19].
The effectiveness of a number of intervention strategies, including discharge planning, patient education, telephone follow up (TFU), home visits, and transition coaching, have been explored to reduce readmissions. Research to date has found no consistent evidence of a singular or multicomponent intervention in reducing readmission [
14]. However previous systematic reviews have highlighted that TFU is a common component of successful randomised trials of multi-component interventions in reducing readmissions [
14,
20]. Therefore it is a potentially promising intervention amongst patients with chronic disease. TFU, where a hospital or community health worker calls a recently discharged patient at home, is used to provide ongoing education, management of symptoms and prescribed medication, recognition of complications and reassurance to patients with the aim of facilitating a smooth transition into community or specialist health care [
21,
22]. TFU is considered easy to implement and low cost [
2,
21]. Telephone contact has been linked to increased patient satisfaction [
23].
Several reviews to date have examined the effectiveness of TFU [
2,
14,
21,
24]. Hansen and colleagues examined the effectiveness of 43 studies which used different types of singular and multi-component interventions in reducing 30 day readmissions in both surgical and medical patients [
14]. Following assessment of included studies against Cochrane Effective Practice and Organisation of Care (EPOC) criteria, they found most were observational studies and there was extensive heterogeneity in content and context. They concluded there was no intervention, including TFU, which was consistently effective in reducing readmissions [
14]. A Cochrane systematic review examined the effectiveness of TFU delivered by hospital-based staff on health outcomes in 33 studies involving 5110 surgical and medical patients [
21]. While the main focus of the review was on psychosocial and physical outcomes, four studies reporting readmission outcomes amongst patients with cardiac disease were pooled together and no effect was found at three months. Again applying EPOC criteria, they found studies were of low methodological quality. Readmission outcomes at 30 days were not assessed. Another review by Bahr and colleagues focused on hospital based TFU as a singular intervention amongst medical and surgical patients, with no impact on readmissions within 30 days [
2]. However they included descriptive studies and no formal assessment of methodological quality was performed. Crocker and colleagues in their review of three included studies also concluded that TFU alone is ineffective in reducing readmissions amongst general medical patients [
24]. Risk of bias in study design was assessed but no formal scoring was reported. They did not assess 30 day outcomes and focussed solely on TFU delivered by a primary care team member, and therefore the results are not generalizable to more common hospital based models of TFU where calls are made by the discharge nurse.
While overall, these reviews suggest that the evidence for TFU in reducing readmissions is inconclusive, none have focussed specifically on hospitalised chronic disease patients, and therefore it is unclear to which results are generalizable to this population. Given the increasing prevalence and healthcare burden of chronic diseases, its disease complexity, and the development of government chronic disease strategies [
25,
26], it is pertinent to examine the effectiveness of TFU in patients with one or more chronic disease separately from general medical and surgical patients. Therefore, the aim of this review is to assess the methodological quality and effectiveness of interventions using TFU in reducing readmission within 30 days amongst patients with cardiovascular disease, chronic respiratory disease and diabetes.
Discussion
This systematic review examined the effectiveness of TFU in reducing readmission within 30 days of discharge among patients with cardiovascular disease, chronic respiratory disease and diabetes. Of the ten intervention studies which met the EPOC research design criteria, five were effective in reducing readmissions within 30 days. However the methodological quality of studies was poor. Apart from one low risk study, most had similar limitations, which weakens the overall strength of evidence. There was a lack of uniformity in how readmission was measured which highlights the need for consistency and precision in the measurements used in studies aiming to reduce readmission. Most studies identified were single site interventions and thus findings may have limited generalisability. In addition, the studies presented wide variation in standard care provided to control groups. Some studies included very little information on what constituted standard care. This made it difficult to interpret study results in relation to the circumstances under which the interventions were likely to be effective or ineffective.
All identified studies combined TFU with other intervention components. All three studies evaluating TFU with pre-discharge interventions showed effectiveness, however in two studies readmission was significantly reduced in only one of the two intervention groups, i.e. in the craniotomy group and not the stroke group [
37]; and in the medical group and not the surgical group [
36]. Two of four studies evaluating TFU with both pre- and post-discharge components were effective [
29,
34]. There was no evidence that TFU and telemedicine or TFU and post-discharge interventions was effective, however, only one to two studies examined each of these types of interventions. On balance, the evidence for TFU is equivocal. There is some suggestion however that combining TFU with pre-discharge intervention components may be promising but further interventions are needed to confirm whether this is the case for both medical and surgical patients with chronic disease. Although the effective studies all offered some form of continuity or bridging for the patient from the hospital to the community setting, none included components distinctive from the ineffective studies. This equivocal finding aligns with that of Hansen and colleagues, who also found no conclusive evidence for a multi-component intervention in reliably reducing readmissions amongst general and surgical patients [
14].
Questions also still remain as to whether TFU itself is the effective component or not. The outcomes of TFU may be masked by many factors such as individual professional and patient actions and behaviour, social interactions and environmental settings [
21]. Further randomised trials of high methodological quality examining the effectiveness of TFU in a standardised way are needed. In particular, given the lack of detail given in many included studies with regards to TFU, it may be warranted to examine the intensity, content and length of calls needed to achieve a significant effect for such patients. TFU is a popular feature of interventions in reducing readmissions, however given limited health resources, the specific details surrounding the effectiveness of TFU for patients with chronic disease still needs to be tested.
Seven of the ten included studies focused on patients with heart failure. Although chronic diseases share common features in terms of intermittent exacerbation of disease, persistence over time and are rarely cured [
38], there are differences with respect to the type and intensity of treatment, symptoms and the professional care needed. Therefore, study results derived from one chronic disease population cannot necessarily be generalised to other chronic disease groups. Given this, there is a need for more intervention research on reducing 30 day readmissions for patients with other prevalent chronic disease such as diabetes and chronic respiratory disease.
Patient-centred care requires communication between hospital and community based physicians; ensuring patients do not experience a gap in care and understanding. The roles of these health professionals are critical to preventing readmission [
39]. One included study focused on training hospital doctors in patient-centred transitional care through telephoning community physicians, home visits to the patient and conducting TFU which resulted in a significant reduction in 30 day readmissions [
34]. However, most studies focused on patient-level interventions rather than provider-level change. Record and colleague’s study points to the potential importance of enhancing provider skills in patient education, transitional care and conducting TFU calls.
This review had a number of limitations. Firstly, a meta-analysis was not possible due to the wide variation in interventions between studies and readmission measures used. Secondly, many included studies were of low methodological quality and lacked detail making it difficult to determine the content or effectiveness of the interventions and to draw firm conclusions applicable to other hospitals and communities. Lastly, it is acknowledged that data on rates of readmissions will inevitably include some readmissions which are appropriate and unavoidable, for example, when a readmission is medically necessary due to an unavoidable change in chronic condition [
40,
41]. Although two of the included studies measured unplanned readmissions, no studies measured avoidable readmissions. This is mainly due to the fact there is no agreed method of measuring avoidable readmissions [
40]. Therefore data on rates of readmissions included in this review may be overestimated in terms of true avoidability.
Conclusions
Although there is increasing priority being placed on reducing readmissions within 30 days, the evidence for the effectiveness of TFU alone or in combination with other intervention components in reducing readmissions in patients with chronic disease remains inconclusive. However despite the equivocal findings, there remain important implications for practice. Due to a lack of studies, there is no well-controlled evidence to suggest that TFU in isolation is an effective strategy. TFU combined with pre-discharge interventions show some promise, however, results are not consistent across patient groups. This may suggest the importance of ensuring that the pre-discharge and / or TFU intervention components are carefully tailored to the needs of the patient group. There is also potential importance in focusing interventions on enhancing provider skills in patient education, transitional care and conducting TFU. In generating good research evidence in this area, priority should be given to conducting studies of high methodological quality. Where possible, studies should be multi-site in order to enhance generalisability, and measurements of readmission need to be consistent across studies. In order to build upon the existing evidence-base, there is merit in focussing research efforts on the evaluation of delivery of standardised TFU in combination with pre-discharge interventions.
Acknowledgements
All persons contributing to the manuscript met the criteria of authorship. There was no external funding for this paper.