Effectiveness of transitional care program in older adults with heart failure: a systematic review and meta-analysis of randomised controlled trials

Abstract

Background

Heart failure leads to adverse clinical and patient-reported outcomes. Its prevalence has increased markedly among adults aged 60 years and over. Transitional care interventions are recommended to address these issues; however, their effectiveness on health outcomes, particularly in older adults, remains limited and inconclusive.

Methods

This study aimed to systematically review and synthesise the existing evidence from randomised controlled trials on the effectiveness of transitional care interventions on health outcomes among older adults with heart failure. Only randomised controlled trials were included. This systematic review and meta-analysis were conducted according to the Cochrane Collaboration methodology and were reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Studies from 2001 to the present were identified through searches of PubMed, the Cochrane Library, Web of Science Core Collection, and Cumulative Index to Nursing and Allied Health Literature Plus with full text. Risk ratios, mean differences, and standardised mean differences were calculated. Heterogeneity was evaluated with the I² statistic. The certainty of the evidence was evaluated with the Grading of Recommendations Assessment, Development and Evaluation criteria.

Results

Eight studies with 967 subjects (483 in the intervention group and 484 in the control group) were included in this systematic review and meta-analysis. Transitional care interventions were associated with improvements in self-care confidence and reductions in heart failure-specific readmission. There were no statistically significant effects on self-care maintenance, self-care management, or heart failure knowledge. Findings for health-related quality of life, functional status, and event-free survival varied across studies. The certainty of the evidence ranged from very low to moderate.

Conclusions

Transitional care interventions were associated with improved self-care confidence and reduced heart failure-specific readmissions in older adults with heart failure. However, variations in the certainty of the evidence create uncertainty about the intervention’s effect. The limited number of trials included in this review demonstrates an evidence gap in this area. Further high-quality studies with transparent reporting through prospective trial registration should be conducted to determine the optimal content of the transitional care intervention.

Registration

The protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO) (CRD42021229464).

Clinical trials number

Not applicable.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12877-026-07642-0.

Introduction

Heart failure (HF) has been defined as a global pandemic, and approximately 64.3 million people are estimated to live with HF worldwide [1]. Trends in HF incidence vary by age, sex, and race or ethnicity. Globally, the incidence of HF was shown to increase markedly among adults aged 60 and over. The risk of developing HF is 20 times higher among this age group compared with those under 60 years of age [2–5]. Older adults with HF are more likely to experience frequent hospital readmissions and emergency department visits [6, 7]. Approximately half of patients are hospitalised at least once a year [8]. After hospital discharge, 7%-13% of patients are readmitted within the first 30 days [6, 7, 9, 10], and 14%-25% within the first 9 months [11, 12]. Some of the most important predictors for long-term mortality, hospitalisation, and emergency department visits are insufficient self-care behaviours such as the regular use of prescription medications, adherence to diet, evaluation and management of HF symptoms in older patients with HF [6, 7, 13–15]. Studies have shown that most HF patients have insufficient self-care behaviours, which worsen with age [6, 15–18]. This has been linked to lack of HF knowledge, low self-confidence, and poor self-care management after hospital discharge [19]. To address these issues, transitional care is recommended to support continuity of care across settings, from hospital to outpatient clinic to home [9, 10]. After discharge, patients may be monitored through various methods, including home or clinic visits, telephone follow-up, or web-based approaches, as part of transitional care. Education and counselling can also be provided to support self-care and self-management for patients [10, 20, 21]. The content of transitional care intervention (TCI) differs among the countries and studies; consequently, it is referred to as a home-based program, a nurse-led visiting program, a clinical-based follow-up program, or a tele-monitoring program in the literature [20, 22–26]. However, TCIs generally include key components such as education, counselling and monitoring practices to support especially older adults in bridging the gaps between care settings or their discharge from the hospital to home, to reduce readmissions, ensure communication and care coordination, conduct timely patient follow-up, provide support for patients and families and ultimately improve the quality of patient care [27–29].

Several meta-analyses have been conducted to determine the effectiveness of TCIs for people with HF, reporting reductions in hospitalisation and mortality, as well as improved quality of life. However, much of the existing review literature on TCI mixes designs and age ranges [23, 26, 30–35]. As a result, there is still uncertainty about whether TCIs for older adults with HF, who are often underrepresented in studies, are effective on health outcomes. This review aims to address this gap by systematically synthesising and, where appropriate, pooling evidence from randomised controlled trials (RCTs). The findings of this review may help inform the development or redesign of older-age-specific TCI.

Aim

This study aimed to systematically review and synthesise the existing evidence from RCTs on the effectiveness of transitional care interventions on health outcomes among older adults with HF.

Research question

• What is the effectiveness of transitional care interventions on health outcomes among older adults with HF?

Methods

This systematic review and meta-analysis were conducted according to the Cochrane Collaboration methodology [36] and reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [37]. The protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO) (CRD42021229464).

Data sources and search strategy

The first literature search was conducted from December 2001 (The retrieval period for one of the first reviews of transitional care was limited to 1985–2001 [38], and structured TCI for HF began to be more clearly conceptualised and empirically evaluated in the early 2000s [39, 40]) to December 2021 (the year when the initial search was conducted), in PubMed, Cochrane Library, Web of Science Core Collections, and Cumulative Index to Nursing and Allied Health Literature (CINAHL) Plus with full text. To ensure currency, a re-run search covering the period from January 2021 to 2025 was performed in the included databases, except CINAHL, which was no longer accessible at the time of the updated search. The search strategy was systematically assessed using the Peer Review of Electronic Search Strategies (PRESS) by a statistician, two experts, and an experienced information specialist, as recommended in the literature [41]. A combination of MeSH and free-text terms that included “aged”, “heart failure” and “transitional care” was used in both the first and re-run searches. Cochrane Sensitivity-maximising Randomised Controlled Trial filter [42] was applied to PubMed and adjusted for the other databases. Details on the search strategies for included databases are shown in Supplementary Material 1. The reference lists of the eligible studies were hand-searched to identify other relevant articles.

Eligibility criteria

We developed inclusion and exclusion criteria concerning the “PICO(S)” framework (population, intervention, comparison, outcome, and study). The included studies should meet the following criteria:

• (i)    Population: aged more than 60 years, hospitalised with primary diagnosis of HF, discharged from hospital to home, and enrolled in the study before discharge. One of the population criteria in this study was that patients be aged 60 years or older, since HF is more prevalent among adults over 60 years of age, and the United Nations defines an older adult as a person aged 60 years or older [2–4, 43].

• (ii)   Intervention: defined as structured interventions initiated before discharge and continued after discharge, designed to support the transition from hospital to home. These interventions included one or more components such as discharge planning, education, counselling, or monitoring, and involved a follow-up process delivered through different modes (e.g., home-based or clinic-based care).

• (iii)  Comparison: usual care or routine care.

• (iv)  Outcomes: all health outcomes that were investigated for the effect of TCI.

• (v)  Studies: RCTs reported as full texts were included, as the cause-and-effect can be based on RCTs' findings with far more confidence than almost any other type of study [36].

Studies focusing only on pharmacotherapy, exercise, invasive monitoring, or end-of-life care were excluded. Editorials, reviews, conference abstracts, letters, notes, and trials without full-text access were also excluded. In addition, grey literature and non-English studies were not considered.

Study selection and data extraction

Based on inclusion and exclusion criteria, titles and abstracts of each reference after duplicates were excluded were independently assessed by two authors (AY and ONE) for potential relevance. Any disagreement regarding eligibility was resolved through discussion between the two authors (AY, ONE), with a third author (EK) consulted when a consensus could not be reached. The full texts of all eligible papers were retrieved and coded as “retrieve” or “exclude” by two authors (AY, ONE). Conflicts in the eligibility of any report were resolved by consensus or reconciled by a third reviewer (EK). “Data Extraction Form” has been developed by adopting and customising the “Data collection form for intervention reviews for RCTs only-template” of The Cochrane Collaboration. The following data were extracted from each study: first author, publication year, country, participant characteristics, delivery personnel, delivery modes, core components, duration of intervention, and outcomes. One review author (AY) manually extracted study characteristics. A second author (OYE) checked the extraction’s accuracy. Any disagreement was resolved by consensus or by consulting a third reviewer (EK). The study authors were contacted to request missing data or verify key study characteristics where these were unclear. For trials that did not explicitly restrict enrolment by age but reported a mean age ≥ 60 years and met al.l other inclusion criteria for this review, the study authors were contacted to request age-stratified data for participants aged 60 or above to increase the number of included trials. The data obtained through this process were included in this review. Using this approach, the data on people aged 60 or over from the study by Chang, Wang [44] were obtained from the study authors. Missing data that could not be obtained were excluded. We have not made any assumptions about missing data. The re-run search used the same procedures, adding one more eligible study for a total of 8 trials.

Risk of bias assessment

The risk of bias assessment of included studies by using the Cochrane Risk of Bias Tool for Randomised Trials-2 (Rob 2.0), which assessed the five bias domains covering the randomisation process, deviations from intended intervention, missing outcome data, measurement of the outcome, and selection of the reported result and then came up with the overall risk of bias [45]. Two review authors (AY, OYE) independently classified each trial’s risk of bias as low risk, some concerns, or high risk, and disagreements were resolved by discussion or by consulting a third reviewer (EK). Studies coded as high risk of bias were not excluded, and the results were discussed with consideration of the risk of bias.

Statistical analysis

The meta-analysis was conducted using statistical software R with the meta package [46, 47]. Risk ratio (RR) with 95% confidence intervals (CIs) was calculated for HF–specific readmissions, and hazard ratio (HR) with 95% CIs for event-free survival. Mean difference (MD) with 95% CIs was calculated to estimate the intervention effects on quality of life, self-care behaviours, and HF knowledge, as these outcomes were measured using the same scales. The standardised mean difference (SMD) with a 95% CI was calculated to estimate the effect sizes of functional status, as this outcome was assessed using different measurement tools. A forest plot was used to depict the MDs and CIs. When outcomes were reported at multiple time points, data from the longest time point were extracted. Substantial heterogeneity was anticipated due to differences in intervention characteristics and study design. Therefore, meta-analyses of each outcome were performed using a random-effects approach [48]. When three or more studies were available, the Hartung–Knapp method was preferred to reflect better uncertainty around the between-study variance [49, 50]. Heterogeneity across studies was evaluated using I² statistics. I² values of 0%-40%, 30%-60%, 50%-90%, or 75%-100% indicated “not important”, “moderate”, “substantial”, or “considerable” heterogeneity [36]. Where substantial heterogeneity was observed, studies were not pooled, and the findings were summarised narratively. With respect to event-free survival, the included trials were broadly comparable, as all defined this outcome as a composite of all-cause mortality and HF–related rehospitalisation. Hazard ratios were extracted and standardised so that values below 1 consistently favoured the intervention across studies. Due to the limited number of trials included in each meta-analysis, subgroup analyses or meta-regression based on intervention characteristics were not conducted. Assessment of publication bias using funnel plots was not performed, as fewer than 10 studies were available [51, 52]. Sensitivity analyses were conducted using a leave-one-out approach that involved iteratively rerunning the meta-analysis and deleting studies to identify influential studies and explore potential sources of heterogeneity.

Certainty of the evidence

The Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach was followed to assess the certainty of evidence in the current review [36] and created a Summary of Findings table (SoF). Two review authors (AY, OYE) independently evaluated the certainty of the evidence, and disagreements were resolved by discussion or consultation with a third reviewer (EK). All outcomes with findings from included studies are reported in the SoF table (Supplementary Material 2).

Results

Search results

A total of 2761 records were identified from the four databases. After deduplication, 1860 records remained for screening. The remaining 1708 records based on title and abstract screening were excluded. After evaluating the 152 full-text papers, 145 ineligible reports were excluded because they did not meet the inclusion criteria. Ultimately, 7 studies were selected for inclusion from the first search. The subsequent re-run search identified 857 potential papers. The entire process, which was deduplication, title and abstract screening, and full-text evaluation, was repeated in the re-run search. This resulted in one included study from the re-run search, bringing the new total to 8. The PRISMA flow diagram of study selection is presented in Fig. 1.

Fig. 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) study selection diagram

Study characteristics

The major characteristics of the included studies are summarised in Table 1. The included 8 articles were published between 2004 and 2025. Studies were conducted in the United States of America [53, 54], China [55–57], Italy [58, 59], and Taiwan [44]. A total of 967 people were included in this systematic review (483 in the intervention group and 484 in the control group). The sample size ranged from 40 [53] to 239 [54]. The mean ages of participants in the intervention group ranged from 70.82 years (SD: 6.34) [57] to 79.5 years (SD: 6.6) [55], whereas those in the control group ranged from 70.53 years (SD: 7.68) [44] to 78.7 years (SD: 6.7) [56]. The severity of HF is based on the New York Heart Association classification, and 4 of the 7 studies reported that most participants were in classes II-III. Seven trials reported reduced ejection fraction, with a mean below 40%.

Table 1.

Major characteristics of included studies

First Author (Year)Country	Participants (IG/CG)		Intervention				Outcomes
	Sample Size	Mean Age	Delivery Personnel	Delivery Mode	Key Components	Duration
Kwok (2008),China [55]	49/56	79.5±6.6 / 76.8±7.0	Nurse-led (Community nurse)	Home-based(with supportive telephone contact)	-    Education (Standardised) -    Monitoring (home visits)	6 months	No significant differences in re-admission rates, functional status, total public health care, and personal care costs; lower handicap
Naylor (2004), USA [54]	118/121	76.4±6.9 / 75.6±6.5	Nurse-led (Advanced practice nurse)	Home-based (with supportive telephone contact)	-    Discharge planning -    Education (Individualised) -    Monitoring (home visits) -    Counselling	1 year***	Lower mean total healthcare costs; increased the length of time between hospital discharge and readmission/death; fewer rehospitalisations; greater improvements in HRQoL and satisfaction; no significant differences in functional status, event-free survival, or unscheduled visits to physicians, clinics, and emergency departments.
Yu (2015),China [56]	90/86	78.6±7.1 / 78.7±6.7	Nurse-led (Cardiac nurse)	Hybrid (home-based and structured telephone support)	-    Discharge planning -    Education (Individualised) -    Monitoring (home visits, structured telephone follow-up) -    Counselling	9 months	No significant differences in event-free survival, hospital readmission, or mortality; fewer hospital stays; greater improvements in self-care behaviour and HRQoL
Del Sindaco (2007),Italy [59]	86/87	77.4 ± 5.9 / 77.5±5.7	Multidisciplinary (Nurses, cardiologists, primary care physicians)	Hybrid (clinic-based, home-based, and structured telephone support)	-    Discharge planning -    Education (Standardised) -    Monitoring (clinic/outpatient follow-up, home visits, structured telephone follow-up) -    Counselling	2 years****	Fewer all-cause deaths and heart failure hospital admissions; shorter length of hospital stay; greater improvements in functional status, HRQoL; lower mean healthcare costs
Chang (2020),Taiwan* [44]	37/28	71.83±8.08 / 70.53±7.68	Nurse-led (Research nurse)	Structured telephone support	-    Education (Individualised) -    Monitoring (structured telephone follow-up)	3 months	No significant differences in the illness representation, dimension, or self-care maintenance; greater improvements in the self-care confidence
Leavitt (2020), USA [53]	19/21	82.7±8.27**	Nurse-led (Home health nurse)	Home-based	-    Education (Individualised) -    Monitoring (home visits) -    Counselling	1 month	No significant differences in readmissions; greater improvements in HF knowledge, self-care confidence, and HRQoL
Villani (2014), Italy [58]	40/40	71±4 / 73±5	Multidisciplinary (Nurse, cardiologist, psychologist)	Telemonitoring	-    Monitoring (telemonitoring) -    Counselling	1 year***	Lower mortality and fewer hospital readmissions for congestive heart failure; greater improvements in the perceived well-being; lower anxiety and depression levels; lower NYHA class
Shan (2025),China [57]	44/43	70.82±6.34 / 71.67±6.82	Nurse-led (Nurse, graduate nursing students)	Clinic-based	-    Discharge planning -    Education (Standardised) -    Monitoring (clinic/outpatient follow-up)	3 months	Greater improvements in HRQoL and functional status; fewer unplanned readmissions

APN Advanced Practice Nurse, HRQoL Health Related Quality of Life, CHF Chronic Heart Failure, NYHA New York Heart Association, HF Heart Failure, COPD Chronic Obstructive Pulmonary Disease, PDA Personal Digital Assistant, ECG Electrocardiography

*The results of people aged 60 or over were obtained from the study authors and include a meta-analysis

**The mean age was not separately reported for the intervention and control groups in trials

***52 weeks and 12 months were accepted as 1 year

****24 months were accepted as 2 years

Due to variations in intervention content, TCIs were descriptively classified by delivery mode, informed by the American Heart Association (AHA) taxonomy [60], and extant literature, including systematic reviews [23, 26, 61–63]. Across the eight trials, interventions were conducted during pre-discharge and post-discharge periods as home-based, clinic-based, structured telephone support, telemonitoring, or hybrid approaches combining two or more delivery modes. Home-based interventions, often supported by telephone contact, were the most commonly applied approach [53–55]. These interventions were generally delivered by nurses, relied on follow-up through home visits, and focused on self-care education, symptom monitoring, and treatment adherence. Structured telephone support, primarily consisted of post-discharge telephone follow-up following pre-discharge education, was used in one study [44]. Telemonitoring was applied in another study to enable remote monitoring of patients’ clinical status [58]. A clinic-based approach was used in one study, focusing on discharge management and outpatient follow-up [57]. In two trials, hybrid interventions combining home-based care with structured telephone follow-up and/or clinic-based services were reported, providing a more comprehensive approach to care [56, 59].

Interventions included key components: discharge planning, education, monitoring, and counselling, although the combination of these components varied. The discharge planning was clearly reported in four trials [54, 56, 57, 59] and involved a comprehensive pre-discharge assessment and structured pre-discharge planning. These activities were generally aimed at assessing care needs, planning post-discharge care, and, often, providing patient education. One trial reported only limited, non-structured pre-discharge contact rather than a clearly defined planning process [55]. The education, whether standardised [55, 57, 59] or individualised [44, 53, 54, 56], generally covered HF diagnosis, symptoms, treatment, and self-care practices. The monitoring was conducted through different approaches, including home visits [53–55], structured telephone follow-up [44], telemonitoring [58], clinic/outpatient follow-up [57] and combination of these [56, 59], primarily focusing on evaluating clinical status, adherence to treatment, and possible adverse reactions, and was carried out at multiple time points with variable durations ranging from 4 weeks to 2 years. The counselling was reported in five studies, most often delivered by nurses, and was generally aimed at supporting self-management and adherence [53, 54, 56, 58, 59].

Interventions were predominantly delivered by nurses with advanced competencies (cardiac nurses, home care nurses, community nurses, or specially trained advanced practice nurses) [44, 53–56], although multidisciplinary teams were also involved in some studies (cardiologists, primary care physicians, psychologists, and specialised nurses or registered nurses) [58, 59].

Comparison

In the included studies, TCI was compared with usual care, which was variably described across studies. Five of the eight studies included discharge education focused on HF management in accordance with country-specific health policies [44, 53, 54, 57, 58]. In addition, patients in the usual care groups received routine or limited follow-up, most commonly through telephone contact or clinic visits [44, 55–57, 59]. The frequency and purpose of follow-up varied across studies, ranging from scheduled outpatient visits to brief or infrequent telephone contacts. Routine home healthcare was provided as part of usual care in one study [53] and upon referral in one study [54].

Measuring outcomes

Regarding outcomes and measures, self-care behaviours in three studies were evaluated using the “Self-Care of Heart Failure Index” (SCHFI) [44, 53, 56], with higher scores indicating better self-care. In five studies, quality of life was assessed using the “Minnesota Living with Heart Failure Questionnaire” (MLHFQ) [53, 54, 56, 57, 59]. On this scale, a lower score indicates a better quality of life. In two articles, HF knowledge was assessed using the “Dutch Heart Failure Knowledge Scale” [53, 56], a higher score reflects better HF–related knowledge. Functional status was assessed using the “six-minute walk test” in two studies [55, 57] and in one study using the “Enforced Social Dependency Scale” [54].

Assessing the risk of bias

The risk of bias evaluation for the 8 included studies, assessed using RoB 2.0, is summarised in Fig. 2. Overall, three studies were judged as high risk of bias [44, 55, 58], three studies were judged as low risk of bias [54, 56, 57], and two studies were assessed to have some concerns about the risk of bias [53, 59]. Two studies were judged to have some concerns about risk of bias because of insufficient information about achieving allocation sequence concealment until participants were enrolled and assigned to interventions [44, 58]. All studies were at low risk of performance bias, but it is unlikely, given the nature of this type of intervention. Moreover, there was a low risk of attrition bias as no significant differences in demographic or clinical profiles were found between subjects who dropped out of the intervention and control groups. Since it was unclear whether outcome assessors were aware of the intervention received, two studies were evaluated as high risk of bias in this domain [44, 58]. We could not identify published protocols or trial registrations for seven trials. One study, which published the trial registration in advance, was assessed as high risk for reporting bias, as there were differences in the listed outcomes between the registered and published versions [44]. One trial was assessed as a high risk of bias because it reported primary outcomes only for the intervention group and not for the control group [59]. In one trial, many results were reported for all participants rather than separately for the intervention and control groups [58]; therefore, it was evaluated as high risk of reporting bias. Five of the included studies were assessed as being at low risk of reporting bias.

Fig. 2.

Risk of bias assessment summary for each study

Intervention effects

Self-care behaviours

Self-care maintenance

Three trials, including 283 participants, evaluated the effect of TCIs on self-care maintenance [44, 53, 56] (Fig. 3). The duration of the intervention ranged from 1 to 9 months. The meta-analysis did not demonstrate a statistically significant improvement in self-care maintenance between the intervention and usual care groups (MD = 5.76, 95% CI -2.01–13.54; I² = 48.8%, p = .014; low certainty of evidence). Leave-one-out analysis showed a stable and statistically significant pooled effect. Heterogeneity was eliminated after exclusion of Yu, Lee [56] and Chang, Wang [44], indicating that between-study heterogeneity was mainly driven by these studies (Supplementary Material 3).

Fig. 3.

Forest plot showing the transitional care interventions on self-care maintenance

Self-care management

Two trials, including 218 participants, evaluated the effect of TCIs on self-care management [53, 56] (Fig. 4). The duration of the intervention ranged from 1 to 9 months. The meta-analysis did not demonstrate a statistically significant improvement in self-care management between the intervention and usual care groups (MD = -3.62, 95% CI -65.94–73.18; I² = 67.1%, p = .08; very low certainty of evidence).

Fig. 4.

Forest plot showing the transitional care interventions on self-care management

Self-care confidence

Three trials, including 283 participants, evaluated the effect of TCIs on self-care confidence [44, 53, 56] (Fig. 5). The duration of the intervention ranged from 1 to 9 months. The meta-analysis demonstrated a statistically significant improvement in self-care confidence between the intervention and usual care groups (MD = 16.35, 95% CI 10.18–22.51; I² = 31%, p = .23; low certainty of evidence). Leave-one-out analysis showed that the pooled effect remained statistically significant. The observed heterogeneity was mainly attributable to Leavitt, Hain [53], as removal of this study led to a substantial reduction in I² (Supplementary Material 3).

Fig. 5.

Forest plot showing the transitional care interventions on self-care confidence

Heart failure knowledge

Two trials, including 218 participants, evaluated the effect of TCIs on HF knowledge [53, 56] (Fig. 6). The duration of the intervention ranged from 1 to 9 months. The meta-analysis did not demonstrate a statistically significant improvement in HF knowledge between the intervention and usual care groups (MD = 1.40, 95% CI -0.81–3.62; I² = 0%, p = .50; moderate certainty of evidence).

Fig. 6.

Forest plot showing the transitional care interventions on heart failure knowledge

Heart failure-specific readmissions

Three trials, including 492 participants, evaluated the effect of TCIs on HF-specific readmissions [54, 58, 59] (Fig. 7). The duration of the intervention ranged from 1 to 2 years. The meta-analysis demonstrated a statistically significant difference in HF–specific readmissions between the intervention and usual care groups (RR = 0.56, 95% CI 0.46–0.70; I² = 0%, p = .95; moderate certainty of evidence).

Fig. 7.

Forest plot showing the transitional care interventions on heart failure-specific readmissions

Health-related quality of life

Five trials evaluated the effect of TCIs on health-related quality of life (HRQoL); however, one trial reported outcomes only for the intervention group and was therefore excluded from the analysis [59]. Four trials, including 544 participants, reported HRQoL [53, 54, 56, 57], with intervention duration ranging from 1 month to 1 year. Due to substantial heterogeneity across studies, a meta-analysis was not performed (I² = 95.6%, p < .0001; low certainty of evidence). Across the included studies, HRQoL improved following TCIs, although the effect size and statistical significance varied between studies.

Functional status

Three trials, including 431 participants, reported functional status [54, 55, 57]. The duration of the intervention ranged from 3 months to 1 year. Due to substantial heterogeneity across studies, a meta-analysis was not performed (I² = 93%, p < .0001; very low certainty of evidence). Findings varied across studies; one reported improvement [57], whereas the others found no significant differences compared with usual care.

Event-free survival

Three trials evaluated the effect of TCIs on event-free survival; however, one trial was excluded due to missing data [59]. Two trials, including 405 participants, reported event-free survival [54, 56]. The duration of the intervention ranged from 9 months to 1 year. Due to substantial heterogeneity across studies, a meta-analysis was not performed (I² = 80%, p = .03; low certainty of evidence). Findings varied across studies; one reported improvement [54], whereas the others found no significant differences compared with usual care.

Discussion

The main aim of this study, which summarised data from eight trials that randomised 967 participants, was to synthesise and assess the effects of TCIs on all health outcomes in older adults with HF. Overall, the findings suggest that TCIs were associated with improvements in self-care confidence and reductions in HF-specific readmissions. Reducing HF-specific readmissions may be clinically relevant because it is a key goal in the management of older adults with HF. Interestingly, studies showing reductions in HF-specific readmissions generally had longer follow-up periods, suggesting that these effects may emerge over time. An improvement in self-care confidence may also be clinically significant, as it could help patients manage their condition more effectively over time, potentially leading to better clinical outcomes. In contrast, there was little or no improvement in self-care management, self-care maintenance, or HF knowledge compared with usual care. Findings for HRQoL, functional status, and event-free survival differed. While improvements in HRQoL were observed, the size of the effect varied, and the certainty of the evidence was low. In contrast, results for functional status and event-free survival were inconsistent: some studies reported improvement, while others did not, with very low certainty of evidence. In several studies, usual care included discharge planning, discharge education, and routine follow-up contacts such as clinic visits or telephone calls. These were standard or less structured practices and lacked continuity. As a result, some overlap between groups may have occurred, which could explain the limited effects.

The included studies were most commonly delivered by nurses; however, they varied substantially in delivery mode, intervention components, and duration. Considering the AHA disease management taxonomy and relevant literature helped to describe and compare intervention characteristics more consistently. Most interventions were combined different components, delivery modes, and durations, which limited the ability to draw clear conclusions about the specific contributions of these factors and to compare intervention types meaningfully. No single intervention combination was consistently associated with positive outcomes. These findings suggest that intervention effectiveness may be influenced not by a single component but by multiple components combined and implemented within each intervention. In addition, the limited number of included trials may limit the applicability of these findings to clinical practice. As the studies were conducted in four different high- or middle-income countries, the effectiveness of TCIs may have been influenced by socioeconomic and healthcare system contexts, including TCIs’ implementation, hospitalisation criteria, patients’ health literacy, caregiver support, and differences in cultural factors. While this contextual diversity suggests that the intervention can be broadly diversified and adapted to different contexts, it also means that an effective TCI for improving health outcomes in older adults with HF may not have been clearly established. This limits the generalisability of the findings.

Heterogeneity among pooled outcomes varied, and outcomes with substantial heterogeneity were presented narratively. Variability in heterogeneity likely reflects differences in intervention content and in usual care. Given the observed heterogeneity, the results should be interpreted with caution. Leave-one-out sensitivity analyses showed that, for some outcomes, heterogeneity decreased when individual studies were removed, indicating that variability was partly driven by specific trials. However, although sensitivity analyses identified specific studies contributing to heterogeneity, the underlying reasons could not be determined because studies could not be grouped by their characteristics. Small-study effects should also be considered a potential source of heterogeneity. Several of the included trials had relatively small sample sizes, increasing the likelihood of imprecise effect estimates and exaggerated intervention effects. A formal statistical assessment of small-study effects was not performed because of the limited number of studies per outcome. Therefore, the possibility of small-study effects cannot be completely ruled out. Selective outcome reporting is another potential concern. As a trial registry was unavailable for 7 of the included trials, the completeness of reporting for prespecified outcomes could not be verified. Consequently, selective reporting bias cannot be ruled out, although no clear evidence of systematic selective reporting was identified in the published reports. Five of the included studies were assessed as having “high” risk of bias or “some concerns”. Factors related to risk of bias, such as blinding and selective outcome reporting, could have influenced the reported outcomes and potentially led to the interventions’ effectiveness being over- or underestimated. Given these methodological considerations, the certainty of the evidence ranged from very low to moderate, and the true effects of TCIs in older adults with HF remain unclear.

To our knowledge, no previous review has specifically addressed this research question in the systematic review literature; there was variability across populations, study aims and reported outcomes. Despite these differences, several previous reviews may still be useful for providing context when comparing and interpreting the findings of the present review [31–35, 64]. A prior review, which included a wider age range, reported that the TCIs were associated with reductions in the HF-specific readmission rate [32]. This finding is consistent with our study’s results; however, its applicability to older adults is limited due to the moderate certainty of evidence. Evidence on patient-reported outcomes remains inconsistent. Although Lee, Yang [34] found no effect of TCIs on quality of life in frail older adults, other reviews [31, 35] reported improvements in quality of life among adults with HF. In the present review, improvements in quality of life were observed in older adults; however, the effect size varied across studies, and there was low certainty of evidence, suggesting that the evidence in this age group remains limited. Similarly, while self-care interventions have been shown to improve self-care maintenance and self-care management in broader age groups with HF [64], these effects were not consistently observed in studies limited to older adults. This review did not indicate improvement, with low and very low certainty of evidence for both outcomes. As a result, the effectiveness of these interventions in older adults remains unclear. A recent review in populations with serious illness also reported no differences in functional status [33]. In the present review, findings for functional status in older adults with HF were inconsistent across studies, with very low and low certainty of evidence. Consequently, these results in the literature indicate substantial uncertainty. The variability of the evidence in the literature may also reflect heterogeneity in the content of TCIs across the included studies in the review. Similarly, the present review identified marked differences in the delivery mode, components, and duration of interventions. These differences highlight the complexity of HF management, but also limit comparability and generalisability.

Although such comparisons are informative, the present review is distinct from earlier syntheses in several important respects. HF is largely considered a disease of older adults, and the outcomes examined in our review are closely age-related health outcomes. Consequently, findings derived from studies including mixed-age populations may not be directly applicable to older adults. Moreover, previous reviews mainly focused on clinical indicators such as rehospitalisation, mortality, length of hospital stay, and emergency department use. It was necessary to evaluate the TCIs that indirectly influence clinical outcomes by underrepresented personal outcomes, including self-care behaviour and HF knowledge. By including both clinical and personal outcomes, the present review addresses an important gap in the literature and provides evidence that has not been captured in earlier reviews. In addition, HF–specific readmissions were poorly reported in previous literature syntheses of the topic, despite HF being the most common cause of hospital readmissions among older adults [11]. Previous systematic reviews in the literature synthesised not only RCT but also grey literature. We may not have been able to include results from grey literature and obtain unreported results for other trials. There is some controversy as to whether unpublished studies should be included in meta-analyses because they may be incomplete and their methodological quality can be difficult to assess [36]. Additionally, the previous evidence suggests that published trials tend to be larger, show an overall greater effect than trials found only in the grey literature, and are more likely to have statistically significant results compared to grey trials [65]. However, neither strategy is infallible on its own; the literature on TCIs for older adults with HF is limited, and data from the grey literature may explain the conflicting and inconclusive results. Non-English publications were also excluded due to the lack of multilingual reviewers in this review. Consequently, the potential for bias due to missing evidence cannot be fully eliminated; therefore, the risk of bias should be assessed in the current review. In order to overcome some of the problems associated with publication bias, it is suggested that scholars pay closer attention to screening grey literature in the future.

Limitation

This review has several limitations that need to be addressed when interpreting the current results. First, the current article was based on 8 trials focused on the effect of TCIs in older adults with HF, which did not allow us to conduct subgroup analyses by intervention components, durations, and delivery personnel. Second, the small number of studies included in each meta-analysis may have affected the statistical significance and may reduce the generalisability of the findings. Thirdly, one included trial [44] enrolling broader adult populations, and age-stratified data for participants aged ≥ 60 52s was obtained from the authors. This approach enabled the inclusion of relevant evidence for older adults. A leave-one-out sensitivity analysis showed that the observed heterogeneity was not primarily attributable to this study. As the randomisation was performed at the full-sample level, balance within the ≥ 60-year subgroup may not have been fully ensured, which could affect the validity of the subgroup. Finally, it cannot be ruled out that studies were missed in the selection process due to the exclusion of grey literature and non-English publications, which may pose a risk of reporting bias in this systematic review and meta-analysis.

Conclusion

In line with the stated review objective, this review indicates that TCIs were associated with improvements in self-care confidence and HF–specific readmissions. These findings may be clinically relevant. In contrast, little or no consistent effects were observed for self-care management, self-care maintenance, and HF knowledge. The certainty of the evidence ranged from very low to moderate, and the findings should therefore be interpreted with caution regarding their clinical relevance. Overall, the findings indicate that the effectiveness of TCIs may depend on the combination of delivery mode, intervention components, and duration. In light of current studies, future well-designed RCTs that clearly define intervention components, ensure comparable durations, and are transparently reported through prospective trial registration are needed to identify which TCIs are most effective for older adults with HF. In addition, future studies should prioritise examining TCIs in low-income settings, where evidence remains limited, and the burden of HF is expected to rise. Such efforts would facilitate more robust comparisons, reduce heterogeneity and strengthen the evidence base.

Abbreviations

CI

Confidence Interval

GRADE

Grading of Recommendations Assessment, Development, and Evaluation

HF

Heart Failure

HR

Hazard Ratio

HRQoL

Health-Related Quality of Life

MD

Mean Difference

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

PROSPERO

Prospective Register of Systematic Reviews

RCT

Randomised Controlled Trial

RR

Risk Ratio

SMD

Standardized Mean Difference

Rob 2.0

Risk of Bias Tool for Randomized Trials-2

TCI

Transitional Care Intervention

Authors’ contributions

All authors read and approved the final manuscript. AY: Conceptualization (lead); Data curation (equal); Investigation (lead); Methodology (lead); Project administration (lead); Resources (equal); Validation (supporting); Visualization (lead); Writing – original draft (lead). EK: Data curation (equal); Formal analysis (lead); Investigation (supporting); Methodology (supporting); Resources (equal); Validation (lead); Writing – review & editing (equal). ONE: Conceptualization (supporting); Investigation (supporting); Methodology (supporting); Resources (equal); Writing – review & editing (equal).

Funding

No funding has been obtained for this study.

Data availability

All data generated or analysed during this study are included in this published article [and its supplementary information files].

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.