Preventing 30-day hospital readmissions: a systematic review and meta-analysis of randomized trials

Abstract

Importance

Reducing early (<30 days) hospital readmissions is a policy priority aimed at improving healthcare quality. The Cumulative Complexity Model conceptualizes patient context. It predicts that highly supportive discharge interventions will enhance patient capacity to enact burdensome self-care and avoid readmissions.

Objectives

To synthesize the evidence of the efficacy of interventions to reduce early hospital readmissions and identify intervention features—including their impact on treatment burden and on patients’ capacity to enact post-discharge self-care—that might explain their varying effects.

Data Sources

We searched electronic databases (1990 until April 1^st, 2013), contacted experts, and reviewed bibliographies.

Study Selection

Randomized trials that assessed the effect of interventions on all-cause or unplanned readmissions within 30 days of discharge in adult patients admitted to the hospital for a medical or surgical cause for > 24 hours and discharged to home.

Data extraction and Synthesis

Reviewer pairs extracted trial characteristics and used an activity-based coding strategy to characterize the interventions; fidelity was confirmed with authors. Blinded to trial outcomes, reviewers noted the extent to which interventions placed additional work on patients after discharge or supported their capacity for self-care in accordance with the Cumulative Complexity Model.

Main Outcome

Relative risk of all-cause or unplanned readmission with or without out of hospital deaths at 30 days post-discharge.

Results

In 42 trials, the tested interventions prevented early readmissions [pooled random effects relative risk (RR) 0.82, 95% CI 0.73 to 0.91; p=.03; I²= 32%], a finding that was consistent across patient subgroups. Trials published before 2002 reported interventions that were 1.6 times more effective than those tested later (p_interaction = .01). In exploratory subgroup analyses, interventions with many components (p_interaction <.01), involving more individuals in care delivery (p_interaction = .05), and supporting patient capacity for self-care (p_interaction = .04) were 1.4, 1.3, and 1.3 times more effective than other interventions. A post-hoc regression model showed incremental value in providing comprehensive, post-discharge support to patients and caregivers.

Conclusions

Tested interventions are effective at reducing readmissions, but more effective interventions are complex and support patient capacity for self-care. Interventions tested more recently are less effective.

Registration Number

PROSPERO, CRD42013004773

INTRODUCTION

Early hospital readmissions have been recognized as a common and costly occurrence, particularly among elderly and high risk patients. One in five Medicare beneficiaries is readmitted within 30 days, for example, at a cost of over $26 billion per year.¹ To encourage improvement in the quality of care and a reduction in unnecessary health expense; policymakers, reimbursement strategists, and the United States government have made reducing 30-day hospital readmissions a national priority.^2–4 Achieving this goal, however, requires more complete understanding of the underlying causes of readmission.

The Cumulative Complexity Model (CuCoM)⁵ is a framework developed by our research group that conceptualizes patient context as a balance between workload and capacity (Figure 1). Workload consists of all the work of being a patient and includes efforts to understand and plan for care, to enroll the support of others, and to access and use healthcare services.^6,7 Capacity is determined by the quality and availability of resources that patients can mobilize to carry out this work (physical and mental health, social capital, financial resources, and environmental assets). The CuCoM is novel in its consideration of the effects of treatment burden on patient context and it illustrates how infeasible, unsupported, and context-irreverent care can lead to poor health outcomes and reduced healthcare effectiveness. Because patients recently discharged from the hospital are in a state of extreme physiologic and psychological vulnerability,⁸ their capacity for enacting self-care is low. The CuCoM predicts that, unless sufficient support is given to enhance patient and caregiver capacity to carry out the work of patienthood, placing highly burdensome discharge demands on these patients will lead to poor outcomes and hospital readmission.

Figure 1.

The Cumulative Complexity Model. Patient context is represented as a balance between workload and capacity. This balance must be optimized to ensure care effectiveness and improve outcomes. In turn, the outcomes achieved feedback to affect the workload-capacity balance.

To evaluate the validity of the CuCoM and provide hypothesis-generating work in the understanding of patient context, we chose to synthesize the evidence on the efficacy of interventions to reduce early hospital readmissions. In particular, we sought to determine the degree to which a number of intervention characteristics—including their impact on patient capacity and workload—might account for differences in their effectiveness.

METHODS

A registered protocol (PROSPERO CRD42013004773) guided the conduct of this review,⁹ which we report in adherence to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement.¹⁰

Eligibility Criteria

Eligible studies were randomized trials reported in English or Spanish, since 1990, that assessed the effectiveness of peri-discharge interventions versus any comparator on the risk of early (i.e., within 30 days of discharge) all-cause or unplanned readmissions with or without out-of-hospital deaths. The intervention had to focus its efforts on the hospital to home transition, permit patients across arms to have otherwise similar inpatient experiences, and be generalizable to contexts beyond a single patient diagnosis. Adult patients had to be admitted from the community to an inpatient ward for at least 24 hours with a medical or surgical cause. Studies including obstetric or psychiatric admissions or only including discharges to skilled nursing or rehabilitation facilities were excluded.

Information Sources

In collaboration with an experienced research librarian (P.E.), we searched in April of 2013 the following databases: PubMed, Ovid MEDLINE, Ovid EMBASE, EBSCO CINAHL, and Scopus. Supplement A includes the complete search strategy. Two reviewers (T.S., A.L.) hand searched the bibliographies of included studies and recent reviews. Experts in the field were asked to identify additional references.

Study Selection

Four reviewers (A.L., M.G., J.B., T.S.) worked independently and considered the eligibility of candidate papers by examining their titles and abstracts, and then the full version of papers identified as potentially eligible by at least one reviewer. Conflicts about the eligibility of full papers were resolved by discussion and consensus. Eligibility was delayed for studies reporting outcomes incompletely, pending author contact.

Data Collection

After creating and piloting a standardized form, reviewers (A.L., M.G., J.B.) working independently and in duplicate and using a web-based program (DistillerSR; Ottawa, Canada), abstracted details about the patient population, the interventions compared, and the outcomes reported.

We abstracted details of the interventions tested verbatim from either the trial report or a cited protocol, limiting our focus to the period of hospitalization until 30 days post discharge, and identifying the “net intervention” by selecting out activities that occurred in the intervention arm but not in the control arm. These activities were coded using a taxonomy adapted from Hansen et al¹¹ (Table 1). We also noted the number of meaningfully involved individuals participating in the intervention’s delivery and the number of meaningful interactions these individuals had with patients. “Meaningfully involved individuals” played a structured and requisite function in the delivery of central aspects of the intervention (e.g., a physician that might be contacted only as needed would not be considered meaningfully involved). Similarly, meaningful patient interactions were defined as those that were the proposed sources of the intervention’s effectiveness (e.g., a nurse visiting a patient only to deliver educational materials, but not to actually engage in educational activity, would not be considered a meaningful interaction). Two team members (A.L., M.G.) created summary descriptions of the interventions in a standardized format; these were shared with each author to confirm their fidelity to what happened in the trial.

Table 1.

Activity-based Coding Framework for Discharge Interventions

Label	Activity Observed
Discharge Planning	Simply thinking about and formalizing an approach to prepare for discharge when this did not occur in any way in the control arm
Case Management	Logistical coordination of care and/or resources not specifically focused on self-management and either not occurring in control arm or occurring to lesser degree
Telephone Follow-up	Use of a telephone or videophone for provider-initiated communication after discharge that does not occur in the control arm
Telemonitoring	Use of remote technology designed for the patient to transmit objective measures of health status with or without connected subjective assessment
Patient Education	Patient-directed education related to diagnosis or treatment rationale but not focused on encouraging self-management and not occurring in control arm
Self-Management	Patient-directed education or coaching directly focused on improving patient’s ability to self-manage care needs that does not happen in control arm
Medication Intervention	Medication reconciliation or special education aimed at improving medication understanding or adherence; often conducted by a pharmacist but need not be
Home Visits	Physical visitation by intervention provider to patient’s place of residence when this does not happen in control
Follow-Up Scheduled	Scheduling of a follow-up visit prior to discharge when this is not done in the control arm or is done less reliably
Patient Centered Discharge Instructions	Some difference in the format or usability of discharge materials to make them more accessible or relevant compared to control
Provider Continuity	Increased provider presence on both sides of the hospital to home transition compared to control; may include involvement of PCP in inpatient care or strategic follow up with inpatient provider after discharge or “bridging” provider
Timely Follow-up	Post-discharge follow-up visit or communication with patient when this either does not occur or occurs at a later date in the control arm
Timely PCP Communication	Engagement with PCP in communication about patient status when this either does not occur or occurs at a later date in the control arm
Patient Hotline	Presence of an open line for patient-initiated communication when this either does not exist in the control arm or is more restricted in availability or usefulness
Rehab Intervention	Patient-directed rehabilitation efforts that are not entirely diagnosis-specific but aimed at improving functional status and do not exist in control
Streamlining	A general streamlining of services provided, often with dedicated assignment of responsibility when this does not occur in control
Making Requisite	Increasing the use or quality of services currently available but underutilized compared to the situation in control
Other	Special situations unique to the intervention (caregiver education, peer mentoring, etc)

After calibrating judgments on a pilot sample, two raters familiar with the CuCoM (F.M., K.G.), not involved in data collection and blinded to trial results, evaluated each standardized intervention description on a scale of 1 (substantially decrease) to 4 (no effect) to 7 (substantially increase) to reflect the degree to which the intervention was likely to impact patient workload and patient capacity for self-care. The impact on patient capacity was rated with perfect agreement 50% of the time and within 1 point of difference in 42% of cases (8% differed by 2 points). Because no interventions were rated to decrease patient capacity and all average ratings fell within the range of 4 to 5.5, we elected to dichotomize the variable (threshold of ≥5 for increasing capacity) for analysis. Workload was more difficult to assess reliably: perfect agreement and minor disagreement (+/− 1 point) were seen in 29% and 44% of cases, respectively, with 27% of cases differing by 2 or more points. This variable was divided into 3 categories (increase, decrease, or no change).

For each included trial, we extracted or computed the risk of early readmission for each arm, analyzing patients as randomized (intention to treat analysis). We used the number randomized as the denominator except when the number of patients discharged was reported and differed from the number randomized. We selected the outcome to extract based on an ad-hoc hierarchy of outcomes of interest, with priority given to unplanned readmissions, then to all-cause readmissions, and finally to the composite endpoints of unplanned and all-cause readmissions plus out-of-hospital deaths, respectively. Outcomes were extracted and analyzed at the longest period of follow-up, up to 30 days from discharge. Examination of trials reporting the effect of interventions on more than one of these outcomes revealed that treatment effects were consistent across them (data not shown).

Risk of Bias

Two raters (A.L., M.K.) worked independently and in duplicate to determine the extent to which each trial was at risk of bias using a standardized form based on the Cochrane Collaboration’s tool.¹² The assessment considered the quality of the randomization sequence generation, allocation concealment, blinding of outcome assessors, the potential for missing outcomes (i.e., likelihood of missing readmissions to other hospitals), and the proportion of patients lost to follow-up. For missing outcomes, “high risk of bias” was assigned when the readmissions data came from internal health system records only. To assess for publication bias, we examined a funnel plot for asymmetry and conducted Egger’s asymmetry regression¹³ and determined the associated p-value.

Data Synthesis

We used random effects meta-analyses to estimate pooled risk ratios and 95% confidence intervals for early readmission.^14,15 We tested for heterogeneity of effect on this outcome using the Cochran Q chi-square test¹⁶ and estimated between-trial inconsistency not due to chance using the I² statistic.¹⁷

To explore the effects of patient, intervention, and outcome characteristics on the impact of measured intervention effectiveness, we conducted planned subgroup analyses, testing variables one at a time.

Patient characteristics tested were age (average ≥65 or not), diagnosis (heart failure or other), and hospital ward (general medical or other). Intervention characteristics tested included the number of unique activities involved in the intervention, the number of unique individuals/roles meaningfully involved in its delivery, the minimum number of meaningful patient interactions occurring within 30 days, the location of the intervention activity (i.e. whether it occurred entirely during the inpatient stay, after discharge, or as a combination that “bridged” the transition), whether the intervention was rated to increase or decrease patient workload, and whether the intervention was rated to increase patient capacity (no intervention was found that decreased patient capacity for self-care). Ad-hoc variables tested were year of publication and type of outcome reported (i.e. unplanned readmissions vs. other).

Informed by the findings of the exploratory subgroup analyses and our initial hypotheses, we constructed a post-hoc meta-regression model to test a variable that reflected the degree to which discharge interventions provided comprehensive patient and caregiver support. This “comprehensive support” variable could return values within a range of 0 to 4 “points” based on whether the intervention 1.) was rated to increase patient capacity, 2.) had ≥5 (75^th % of distribution) unique intervention activities, 3.) had ≥5 (75^th % of distribution) meaningful patient contacts, and 4.) had ≥2 (75^th % of distribution) individuals involved in its delivery. We created three categories for this variable: interventions with zero points (Category 1), interventions with one or two points (Category 2), and interventions with three or four points (Category 3). To control for changes in standard care delivery over time, we adjusted based on the year of publication variable.

RESULTS

Study Selection

Our initial database search generated 1128 reports (Figure 2). Through abstract and title screening, 256 reports were identified for full-text review. During full-text screening (agreement 89%), 24 were selected for inclusion, and 39 were set aside for author contact prior to making a decision. Of 7 potentially eligible studies identified from bibliographies and expert consultation, 2 were included, and 1 was set aside for author contact. Of the 40 trials requiring author contact for a final eligibility decision, 21 were deemed eligible. Of the 48 apparently eligible trials, one was found ineligible after the author confirmed that readmission data was only collected for readmissions related to the index diagnosis.¹⁸ The final sample therefore comprised 47 trials from 46 reports.^19–64

Figure 2.

Summary of evidence search and selection. Note that typical papers initially set aside for author inclusion were those that did not report the outcome of interest within 30 days but reported other outcomes within this time, reported the outcome within a survival analysis graph but without information about the number of patients at risk, or reported the outcome as a component of a composite outcome.

Of the 47 eligible trials, 42 contributed data for the primary meta-analysis, and 5 (those that reported numbers of readmissions rather than the number of patients readmitted) were analyzed separately.^{31,45,50,55,61} A complete list of excluded full-text studies and rationale for exclusion is available in Supplement B.

Study Characteristics

Table 2 describes the included trials. Many were single-center trials taking place in academic medical centers, enrolling few patients (e.g., 21 trials enrolled <200 patients), and reporting 30-day readmissions. Most interventions tested took place in both the inpatient and outpatient settings. Supplement C reports the results of the coded activity analysis. In general, interventions included anywhere from 1 to 7 unique activities. Case management, patient education, home visits, and self-management support were commonly present in net activity descriptions (Supplement C). Trial authors responded to confirmation requests for 34 of the 47 net intervention descriptions. Three authors requested minor modifications and one author made major modifications to these descriptions.

Table 2.

Study Characteristics

[Ref] Author, Year (*cluster randomized)	Setting	Population	Intervention	Outcome, Time reported	N	# Activities	# People	# Interaction	Capacity	Workload	Location
19 Melton, 2012	48 different states, USA	Commercially insured for 3 acute DRG’s	Risk-prioritized telephone follow-up	UR, 1 month	3988	1	1	1	4	3.5	Out
20 Marusic, 2013	Academic hospital in Croatia	Elderly pts on two or more meds for chronic disease	Specialized pharmacotherapeutic counseling	UR, 1 month	160	2	1	1	4	3.5	In
21 Altfeld, 2012	USA	Elderly pts with 7+ meds and psychosocial need	Targeted telephone follow-up program	ACR, 1 month	906	2	1	1	5	2	Out
22 Davis, 2012	Academic center, USA	Heart failure pts with mild cognitive impairment	Self-management focused education program	ACR, 1 month	125	5	1	2	5	3.5	Both
23 Bowles, 2011	Urban community, USA	Mostly black HF patients under specific home care agency	Telehomecare substitution of traditional home care	ACR, 1 month	218	1	2	5	4	5.5	Out
24 *Finn, 2011	Academic center, USA	General medical patients of academic center; about 25% discharged to SNF	Embedded nurse practitioner into academic team to improve discharge process	ACR, 1 month	646	2	1	1	4	2.5	In
25 Wong, 2011	Large general hospital, Hong Kong	Elderly general medical population in Hong Kong	Use nurse case managers and trained volunteers to improve transition through health-social partnership	ACR, 1 month	686	4	3	5	4.5	3.5	Both
26 Leventhal, 2011	University hospital in Switzerland	Elderly heart failure patients	Outpatient, interdisciplinary education and support program	ACR, 1 month	34	5	1	4	4.5	4	Out
27 Rytter, 2010	Single center in Denmark	Elderly patients from medical or geriatric ward	Use mandatory home visits to improve follow up from PCP and district nurses	ACR, 1 month	331	4	2	2	4	2.5	Out
28 Koehler, 2009	Academic center, USA	Elderly gen med pts with expected return to home or assisted living	Supplemental care bundle that shifted responsibilities from nurses to care coordinators and added follow-up	UR, 1 month	41	6	2	5	4.5	4	Both
29 Braun, 2009	Medical center, Israel	General medical patients	Use tight telephone follow-up, especially to improve compliance	ACR, 1 month	400	1	1	2	4.5	3	Out
30 Courtney, 2009	Tertiary center, Australia	Elderly general medical patients at high risk	Individualized, exercise-based care plan for elderly	UR, 1 month	128	6	2	5	5	5	Both
31 Jack, 2009	Academic, urban, safety net center, USA	General medical patients; 51% black	Standardized discharge package to minimize failures using discharge planners and pharmacists	ACRE, 1 month	738	6	2	3	5	2.5	Both
32 Wakefield, 2008	VA medical center in Iowa, USA	Males with heart failure; average age 69	Telehealth-facilitated postdischarge support program	ACR, 1 month	148	2	1	5	5	5	Out
33 Balaban, 2008	Small community teaching hospital, USA	Culturally and linguisitically diverse general medical or surgical patients	Program to promptly reconnect patients to medical home through discharge form	ACR, 1 month	96	3	2	2	5	2.5	Both
34 Wong, 2008	Three regional hospitals in Hong Kong	Elderly patients readmitted to department of medicine	Preventive, post-discharge home visits for high risk patients	UR, 1 month	354	2	1	2	4	4	Both
35 Coleman, 2006	Single center, USA	Elderly medical patients that were in capitated delivery system; about 20% discharged to SNF	Use of Transition Coaches™ and personal health record to equip patients and caregivers to be more active in care	UR, 1 month	750	5	1	5	4.5	5	Both
36 Linne, 2006	Multiple community hospitals in Sweden	Heart failure patients; discharge disposition not reported	Use computer-based education session in discharge process	ACR, 1 month	230	2	0	0	4	4.5	Both
37 Casas, 2006	Two tertiary centers; one in Spain and one in Belgium	Elderly COPD patients	Integrated care plan to generate synergy and avoid redundancy between inpatient and outpatient care teams for COPD patients	ACR, 1 month	155	7	3	5	5	4.5	Both
38 Riegel, 2006	Two community hospitals in Southern California, USA	Mexican Americans with HF that were old and ill and poorly accultured	Telephone case management program to improve discharge transition in Mexican Americans	ACR, 1 month	135	5	1	1	5.5	3	Out
39 Koelling, 2005	University hospital, USA	Select HF patients; average age 65	Single pre-discharge education session	ACR, 1 month	223	2	1	1	4.5	4.5	In
40 Mejhert, 2004	University hospital, Sweden	Elderly heart failure patients	Nurse-driven, protocol-based outpatient management program	ACR+D, 1 month	196	3	2	1	5	4.5	Out
41 Kwok, 2004	Two acute hospitals in Hong Kong	Elderly chronic lung disease patients at high-risk	Community nurse-supported program based on weekly home visits	UR, 1 month	157	4	1	5	5.5	4.5	Both
42 *Doughty, 2002	Single center in New Zealand	HF patients; dispositions not reported	Outpatient, integrated management program for heart failure	ACR, 1 month	197	6	3	1	4.5	5	Out
43 Jaarsma, 1999	University Hospital, Netherlands	Elderly heart failure patients; average age 73	Nurse-led education and support program with follow-up home visit	ACR, 1 month	179	3	1	3	4	3.5	Both
44 Naylor, 1999	Two urban academic hospitals	Elderly medical and surgical patients; 45% black	Advanced practice nurse-directed program that stressed continuity, home and telephone follow-up	ACR, 1 month	363	5	1	4	5.5	2.5	Both
45 Stewart, 1998	Tertiary referral center in Australia	General medical and surgical patients; 83% considered high risk	Risk-targeted, home-based intervention by nurse and pharmacist	UR+DE, 1 month	762	5	2	3	5	2.5	Both
46 Dunn, 1995	Geriatric hospital in England	Geriatric ward patients with average age of 83	Single home visit from public health nurse	ACR, 1 month	204	2	1	1	4.5	3	Out
47 Rich, 1995	Single academic center, USA	High risk, elderly heart failure patients	Nurse-directed, multidisciplinary intervention with home visit follow-up	ACR, 1 month	274	6	4	5	5.5	2	Both
48 Naylor (med), 1994	Single University hospital, USA	Elderly patients with or without caregiver for medical cardiac diagnosis	Individualized, comprehensive program directed by clinical nurse specialists, including home follow-up	ACR, 2 weeks	142	5	1	4	4.5	2	Both
48 Naylor (surg), 1994	Single University hospital, USA	Elderly patients with or without caregiver for surgical cardiac diagnosis	Individualized, comprehensive program directed by clinical nurse specialists, including home follow-up	ACR, 2 weeks	134	5	1	4	4.5	2	Both
49 Naylor, 1990	Urban medical center, USA	Elderly general medical or surgical patients	Comprehensive, individualized discharge planning protocol with home follow-up directed by nurse specialists	ACR, 2 weeks	40	4	1	4	4	3	Both
50 Kulshreshtha, 2010	Urban teaching hospital, USA	Heart failure patients; could enter study up to 2 weeks post discharge	Remote monitoring follow-up program for ambulatory patients	ACRE, 1 month	150	2	1	5	5.5	5.5	Out
51 *Graumlich, 2009	Tertiary teaching hospital, USA	General medical patients at high-risk of readmission	Discharge software to improve communication and address deficiencies	ACR, 1 month	631	3	1	0	4	3	In
52 Atienza, 2004	Three tertiary University hospitals, Spain	Heart failure patients; average age 68	Hospital discharge and outpatient disease management program	ACR, 1 month	338	5	2	2	4.5	4	Both
53 Riegel, 2004	Two hospitals in suburban Southwest, USA	Heart failure patients in integrated health system; average age 73	Use of peer mentors to improve self-care in recently-discharged patients	ACR, 1 month	88	1	1	2	4.5	3	Out
54 Stowasser, 2002	Two large hospitals in Australia	General medical and surgical patients	Medication liaison service to improve communication of medication-related issues through discharge process	UR, 1 month	240	3	1	0	4	3	In
55 Li, 2012	Academic center in New York, USA	Elderly patients and their family caregivers	Training of family caregivers to prepare for anticipated post-discharge role	ACRE, 2 weeks	407	1	0	0	4.5	3.5	In
56 Shyu, 2005	Large, single center in Taiwan	Elderly hip fracture patients	Interdisciplinary program of geriatric consultation, rehab, and discharge planning service	ACR, 1 month	137	4	3	5	5.5	3.5	Both
57 Angermann, 2012	Nine centers in Germany	Heart failure patients; average age 69	Nurse-coordinated disease management program that emphasized a “call and care center”	ACR, 1 month	715	4	1	5	4.5	4.5	Both
58 Naylor, 2004	Six academic and community hospitals, USA	Elderly HF patients; 36% black	Advanced practice nurse-directed care program with emphasis on comorbid and chronic condition management; included home follow-up	ACR+D, 1 month	239	6	1	5	5	3	Both
59 Stromberg, 2003	One University and two county hospitals in Sweden	Elderly heart failure patients	Requisite follow-up in specialized, protocol-driven, nurse-led heart failure clinic	ACR+D, 1 month	106	5	1	1	5	3.5	Out
60 Hansen, 1995	University Hospital in Denmark	Highly selected patients from subacute geriatric ward needing home rehab, medical, and social support	Home visit follow-up program for highly-targeted elderly population	ACR, 1 month	193	4	2	2	5	3.5	Out
61 *Maslove, 2009	Single academic center, Canada	General medical patients; about 80% discharged home	Development of more useful and standardized discharge summary	ACRE, 1 month	209	2	1	0	4	3.5	In
62 Forster, 2005	Two campuses of a teaching hospital in Canada	General medical patients; average age 66	Integration of dedicated clinical nurse specialist into care team to facilitate discharge planning process	ACR+D, 1 month	361	3	1	1	4	3	Both
63 Dudas, 2001	Single academic center, USA	General medical service patients	Pharmacy service follow-up call	ACR, 1 month	221	2	1	1	4	3	Out
64 Parry, 2009	Two community hospitals, USA	Fee for service Medicare patients in single health system; inclusion criteria desired to catch patients discharging to SNF—did not report dispositions	Use of Transition Coaches™ and personal health record to equip patients and caregivers to assert more active role in care transition	UR, 1 month	98	5	1	5	5	2.5	Both

#Activities=number of activities in the intervention as evaluated by coding strategy from Table 1

#People=number of individuals meaningfully involved in delivery of the intervention

#Interactions=minimum number of meaningful human interactions in intervention delivery

Capacity=rated likelihood of intervention to affect patient capacity for self-care on scale of 1 to 7, where 4 is no change

Workload=rated likelihood of intervention to impose work or burden on patient on scale of 1 to 7, where 4 is no change

Location=setting (inpatient, outpatient, or both) where intervention activity occurred

ACR=all cause readmission rate; UR=unplanned readmission rate; ACR+D=all cause readmission and out of hospital deaths rate; ACRE=all cause readmission event count; UR+DE=unplanned readmission and out of hospital deaths event count; N=total number of analyzed patients from both arms (number discharged); In=all activity occurred in inpatient environment; Out=all activity occurred in outpatient environment; Both=activity occurred in both inpatient and outpatient environment

Most studies were at low risk of bias (Supplement D). The most common methodological limitation of these trials was the lack of a reliable method for dealing with missing data.

Meta-analysis

In the 42 trials reporting readmission rates, the overall pooled relative risk (RR) of readmission within 30 days was 0.82, 95% CI 0.73 to 0.91; p<.001 (Figure 3). Inconsistency across trials was low (I²=31%). Funnel plot examination showed asymmetry suggestive of publication bias in the context of smaller studies (see Supplement E) and Egger’s test was significant (p=0.02). The five trials reporting number of readmissions (rather than number of patients with readmissions) had a pooled relative risk of readmission of 0.93, 95% CI 0.72 to 1.20, I²=23%, p=0.59. While this result was consistent with the risk found in trials reporting readmission rates (p_interaction= 0.38), we opted not to include these trials in subgroup analyses.

Figure 3.

Results of primary meta-analysis. RR=relative risk.

Subgroup analyses failed to find an interaction between trial results and patient characteristics or outcome measured (Table 3). A number of intervention characteristics, however, did interact with measured effectiveness. These include whether the intervention was rated to augment patient capacity for self-care [RR 0.68, 95% CI 0.53 to 0.86 when it was and RR 0.88, 95% CI 0.80 to 0.97 when it was not; p_interaction=0.04], whether the intervention had at least five unique, component activities [RR 0.63, 95% CI 0.53 to 0.76 when it did and RR 0.91, 95% CI 0.81 to 1.01 when it did not; p_interaction<0.01], and whether the intervention had at least two individuals involved in delivery [RR 0.69, 95% CI 0.57 to 0.84 when it did and RR 0.87, 95% CI 0.77 to 0.98 when it did not; p_interaction=0.05]. Studies testing interventions more recently were associated with reduced effectiveness [RR 0.89, 95% CI 0.81 to 0.97 when published in 2002 or later and RR 0.56, 95% CI 0.40 to 0.79 when published prior to 2002; p_interaction=0.01] Other characteristics of the interventions, such as their rated effect on patient workload, and the site of delivery had no significant interaction with the intervention effect.

Table 3.

Subgroup Analyses

Category	Study Subgroup Characteristic	Number of Studies in Subgroup*	Ratio of Relative Risks of Readmission+ (95% CI)	P Value for Interaction
Patient Characteristics	Patients Enrolled had CHF	16	1.02 (0.82 to 1.27)	0.83
	Patients Enrolled were Aged >65	36	0.87 (0.69 to 1.10)	0.24
	Patients Enrolled were from General Medical Ward	18	0.97 (0.78 to 1.21)	0.79
Intervention Characteristics	Intervention was Rated to Increase Capacity	16	0.76 (0.59 to 0.99)	0.04
	Intervention was Rated to Increase Workload	5#	0.93 (0.67 to 1.29)	0.68
	Intervention was Rated to Decrease Workload	19#	0.99 (0.78 to 1.25)	0.90
	b Study was Published in 2001 or Later	24	1.59 (1.05 to 2.40)	0.03
	a Two or More Individuals Delivered the Intervention	13	0.79 (0.63 to 1.00)	0.05
	a Five or More Meaningful Patient Interactions in the Intervention	13	0.91 (0.73 to 1.14)	0.43
	a Five or More Unique Activities in the Intervention	16	0.70 (0.56 to 0.86)	<0.01
	Intervention had Both an Inpatient and Outpatient Component	22	0.92 (0.74 to 1.15)	0.46
Outcome Characteristics	Outcome Measured was Unplanned Readmissions	9^	0.95 (0.76 to 1.21)	0.70

⁺This represents the ratio of pooled relative risks of early hospital readmission among studies that had a certain characteristic as compared to those that did not have this characteristic.

^*Compared to the remainder of analyzed studies (e.g. 42-N) unless otherwise noted.

^#Compared to “No Change,” (N=18).

^{^}Compared to “All-Cause,” (N=30).

^aThese cut-offs were chosen as they represented the 75^th percentile of distribution of each variable.

^bThis cut-off was chosen as it represented the mid-point of study eligibility for this review.

Post-hoc Meta-Regression Analysis

Despite potential colinearity of the contributing variables, meta-regression showed a significant and incremental effect of “comprehensive support” on reducing readmissions (Table 4). Seven interventions comprised Category 3^{28,30,37,47,56,58,64}. Compared to Category 1 interventions, these were associated with a relative risk of readmission of 0.63, 95% CI 0.43 to 0.93; p=0.02. Category 3 interventions used a consistent and complex strategy that emphasized the assessment and addressing of factors related to patient context and capacity for self-care (including the impact of comorbidities, functional status, caregiver capabilities, socioeconomic factors, potential for self-management, and patient and caregiver goals for care). These interventions coordinated care across the inpatient to outpatient transition and involved multiple patient interactions; all but one²⁸ involved patient home visits.

Table 4.

Effects of “Comprehensive Support” in Meta-Regression Analysis

Study Characteristic	Number of Studies	Relative Risk of Readmission Compared to Reference^	95% Confidence Interval	P value
Comprehensive Support Category 1 (0 points)	15	1 (reference)
Comprehensive Support Category 2 (1 or 2 points)	20	0.82	0.66 to 1.02	0.07
Comprehensive Support Category 3 (3 or 4 points)	7	0.63	0.43 to 0.93	0.02
Publication in 2002 or After	33	1.47	1.10 to 1.96	0.01

The “Comprehensive Support” variable returned one point each for interventions that 1.) were rated to increase patient capacity, 2.) had five or more unique intervention activities, 3.) had five or more meaningful patient interactions, and 4.) had two or more individuals involved in its delivery.

^{^}This represents the adjusted effect of each characteristic on early readmission

DISCUSSION

Our Findings

The body of randomized trial evidence shows a consistent and beneficial effect of tested interventions on the risk of 30-day readmissions. Exploratory subgroup analyses suggest that effective interventions are more complex and seek to enhance patient capacity to reliably access and enact post-discharge care. Additionally, interventions tested more recently are, in general, less efficacious when compared to controls.

Our findings are consistent with the CuCoM in their suggestion that providing comprehensive and context-sensitive support to patients reduces the risk of early hospital readmission; however, we could not identify an effect of rated intervention workload on this risk.

Limitations and strengths of this review

Many studies in this review were conducted in single, academic centers; this raises questions about applicability. Also, the scales we used to evaluate intervention effects on patient workload and capacity relied on global judgments (rather than criterion-based judgments) and are original to this work. To our knowledge, no validated scale exists to assess the potential of an intervention to impose patient workload or treatment burden and/or affect a patient’s capacity for self-care. Although our raters were consistent in their assessments of interventions’ effect on patient capacity, their judgment of impact on patient workload was less reliable. Particularly, raters felt that some burdensome interventions could be beneficial if the patient had the capacity and resources to access and enact the care. Because the experience of treatment burden is not constant between patients, an ideal analysis of its effects would be based on patient-reported assessments of intervention workload. Indeed, many eligible patients declined enrollment into some studies^{23,28,44,50,53}, often because they were not agreeable to the perceived burden of the intervention; evaluating the effect of intervention-imposed workload in such samples is of limited applicability. In general, these assessments should be regarded as hypothesis-generating and the inferences made based on subgroup analyses must be viewed as tentative (given the borderline significance of the analyses, the potential for chance findings from testing multiple hypotheses, and the possibility that some variables are correlated). Finally, despite robust efforts to obtain unpublished data, there was evidence of publication bias. The overall effect of this on our findings is not known.

This review also has many strengths. First, it provides the largest and most comprehensive assessment of discharge interventions and their effect on 30-day readmissions, including 47 randomized trials at low risk of bias. This is a stronger and less heterogeneous body of evidence than previously assembled^11,65 and it includes unpublished data from 18 trials. Our study used an activity-based coding method designed to ensure appropriate characterization of each intervention and their net activities when comparing the intervention and control arms. This method contributes to the field and can be applied to future assessments of complex interventions. To our knowledge, this is also the first use of the CuCoM⁵ to analyze the impact of healthcare delivery interventions on patients as an explanation for their relative efficacy.

Comparison with other studies

We identified 31 more RCT’s than were accumulated in the most recent review of discharge intervention effects on 30 day readmission rates¹¹ and we provide the first meta-analysis on this topic. Although previous studies and reviews have suggested that “bundled” interventions are of greater value,^11,65 this meta-analysis provides objective support for this claim. Additionally, our study adds to and enhances the body of evidence related to the importance of patient contextual factors in affecting health outcomes.⁶⁶

Implications for policy and practice

In this analysis, interventions that used a complex and supportive strategy to assess and address contextual issues and limitations in patient capacity were most effective at reducing early hospital readmissions. Many of these contacted the patient frequently, used home visits, and reported cost savings. This information can be used to guide the design and testing of future interventions. The CuCoM may also have value in helping to conceptualize the effects of healthcare interventions across diverse patient contexts, but we were unable to characterize a consistent effect of rated intervention workload on outcomes. Finally, we found that more recently tested interventions were less effective. We hypothesize that this may represent 1.) a general improvement over time in the standard of care that was not fully appreciated in control descriptions, 2.) an increased effort over time to test simpler and less comprehensive interventions, 3.) a higher likelihood over time of more diverse interventions to measure and report 30 day readmission rates (e.g. including those less focused on reducing early readmissions) and/or, 4.) a general shift away from interventions stressing human interaction toward those more “high tech” in nature. Further study is needed to determine the implications of this finding.

Conclusions

Our results suggest that most interventions tested are effective in reducing the risk of early readmissions. Some features, however, may enhance the effect of these programs. In particular, we found value in interventions that supported patients’ capacity for self-care in their transition from hospital to home. Future work intended to improve the effectiveness of healthcare delivery may benefit from consideration of the demands that healthcare interventions place on recently-discharged patients and their caregivers, and the extent to which these demands are offset by comprehensive support for implementation.