Effect of health information technology (HIT)-based discharge transition interventions on patient readmissions and emergency room visits: a systematic review

Copyright © The Author(s) 2022. Published by Oxford University Press on behalf of the American Medical Informatics Association. All rights reserved. For permissions, please email: journals.permissions@oup.com

This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Associated Data

The data extracted from the original research articles underlying this systematic review are available in the article and in its online supplementary material.

Abstract

Objective

To systematically synthesize and appraise the evidence on the effectiveness of health information technology (HIT)-based discharge care transition interventions (CTIs) on readmissions and emergency room visits.

Materials and Methods

We conducted a systematic search on multiple databases (MEDLINE, CINAHL, EMBASE, and CENTRAL) on June 29, 2020, targeting readmissions and emergency room visits. Prospective studies evaluating HIT-based CTIs published as original research articles in English language peer-reviewed journals were eligible for inclusion. Outcomes were pooled for narrative analysis.

Results

Eleven studies were included for review. Most studies (n = 6) were non-RCTs. Several studies (n = 9) assessed bridging interventions comprised of at least 1 pre- and 1 post-discharge component. The narrative analysis found improvements in patient experience and perceptions of discharge care.

Discussion

Given the statistical and clinical heterogeneity among studies, we could not ascertain the cumulative effect of CTIs on clinical outcomes. Nevertheless, we found gaps in current research and its implications for future work, including the need for a HIT-based care transition model for guiding theory-driven design and evaluation of HIT-based discharge CTIs.

Conclusions

We appraised and aggregated empirical evidence on the cumulative effectiveness of HIT-based interventions to support discharge transitions from hospital to home, and we highlighted the implications for evidence-based practice and informatics research.

Keywords: discharges, adult, pediatric, electronic care transitions, care continuity, emergency department

INTRODUCTION

Over 35 million hospital discharges occur annually in the United States. 1 Effective hospital-to-home discharges involve patient engagement and successful transfer of information, responsibility, and authority between inpatient and outpatient community providers, teams, and/or health systems. 2 However, failures in information transfer, and communication between the inpatient and outpatient teams pose serious threats to care coordination and patient safety. As such, poorly conducted discharge care transitions are vulnerable to poor patient outcomes, including adverse events, 3 , 4 unplanned emergency room (ER) and hospital readmissions, 5 , 6 increased mortality, 7 impaired quality of life, 8 physical and emotional stress for patients and families, 9 lower patient satisfaction, 10 and higher economic burden. 4 , 11 More importantly, unplanned hospital readmissions are a marker of poor care quality affecting clinical outcomes and increasing hospital costs. 12 Factors contributing to readmissions include poor care coordination within and between care teams (eg, primary care and subspecialty clinicians) and lack of patient follow-up. 13 The Centers for Medicare and Medicaid Services reported that nearly 20% of Medicare patients were readmitted within 30 days, 14 , 15 costing nearly $26 billion, of which $17 billion was considered avoidable expenses. 16

Several multi-component discharge care transition interventions 17–22 (eg, patient education, standardized discharge summaries) have been implemented to support the movement of patients from hospital to home settings as their conditions and care needs change during the course of a chronic or acute illness. 23 Such interventions facilitate the coordination and continuity of care activities and bridge the gaps between providers, services, and settings. 24 The discharge care transition interventions fall under 3 categories: 25 pre-discharge (prior to patient discharge); post-discharge (after patient discharge); and bridging (at least 1 pre-discharge and 1 post-discharge component) interventions. 26

Reviews have aggregated the empirical evidence about the effectiveness of discharge care transition interventions on clinical and patient outcomes, including adverse events, ER visits, readmissions, healthcare utilization, and mortality rates, with mixed results. 19 , 25 , 27–32 A recent meta-review concluded there is limited cumulative evidence that discharge planning and support interventions positively impact patient status at hospital discharge, healthcare utilization, patient quality of life, or costs. 33 Furthermore, given the recent focus on the use of health information technologies (HIT) 34 , 35 to improve patient safety, several electronically integrated discharge processes have been evaluated; some of these have shown improvements in communication at discharge, 31 timeliness of discharge, and patient satisfaction. 32 Although recent reviews have included HIT tools as part of their interventions, they report on overarching topics that encompassed HIT and non-HIT discharge transitional programs 27 , 28 and effect of the discharge interventions on readmission outcomes regardless of intervention type. 19 , 29 Despite the widespread implementation and use of HIT, there are no systematic reviews or sub-group analyses assessing the role and impact of HIT-based discharge care transition interventions.

Toward this end, we performed a systematic review to determine the effectiveness of HIT-based discharge care transition interventions (hereafter, CTI) on the process and clinical outcomes, including readmissions and ER visits. This review provides insights on (1) CTI types (eg, discharge plan, medication reconciled lists, discharge summary, discharge checklist), functions, components, phases, and elements supported; (2) impact of CTIs on various reported outcomes including hospital readmissions and ER visits; (3) CTI design and implementation strategies; (4) quality of evidence on CTI effectiveness; (5) gaps in research on evaluation studies on CTI; and (6) future directions for CTI research, and implications for practice.

MATERIALS AND METHODS

Search strategy

An informatics researcher (SK) and a medical librarian conducted a systematic search of MEDLINE, CINAHL, EMBASE, and the Cochrane Central Register of Controlled Trials (CENTRAL) on June 29, 2020, to identify relevant articles on HIT-based discharge CTIs for inclusion. Two reviewers (JA, TK) conducted an initial scoping review of the literature on discharge care transitions to finalize the specific problem area and related search terms. A combination of standardized, controlled vocabulary (eg, MeSH terms) and keywords (eg, electronic medical record (EMR), discharge tool, patient discharge) was used. Manual screening of relevant reference lists supplemented the search. Full search strategy details are provided in Supplementary Appendix S1.

Study screening and selection

Three reviewers (AS, AM, JA) independently screened article titles and abstracts for eligibility and retrieved relevant articles for full-text review. Two reviewers (AM, AS) independently assessed full-text articles for inclusion criteria. Disagreements were discussed and resolved with a third reviewer (JA). Three reviewers (ST, AS, JA) additionally screened references from included articles to identify relevant articles for eligibility. Eligibility criteria included prospective studies that evaluate the effect of HIT-based CTIs that explicitly support patients’ transitions to home on ER visits and hospital readmissions. HIT was defined as “applications of information processing involving computer hardware and software that deals with the storage, retrieval, sharing, and use of health care information, data, and knowledge for communication and decision-making.” 36 Only English language original research articles published in peer-reviewed journals were included (see inclusion criteria in Supplementary Appendix S2).

We excluded articles that: did not study hospital-to-home discharges; did not use a prospective study design; did not report on readmission outcomes or ER visits; used telehealth interventions for longitudinal remote monitoring without explicitly supporting care transitions; were not HIT based; or reported only on intervention design and usability. Although we acknowledge that clinicians utilize HIT to support routine activities within their clinical workflows (eg, use of EHR for reviewing patient information before discharge), we only included HIT-based interventions specifically designed to support discharge activities for a patient (eg, an EHR-integrated discharge summary). Full citations and reasons for exclusion are provided in Supplementary Appendix S3. The review protocol was registered and publicly available under the International Prospective Register of Systematic Reviews (PROSPERO ID #: CRD42021232216).

Data abstraction and management

Three reviewers developed a data extraction form (JA, AM, TK), iteratively tested it, and finalized it to organize relevant study data (Supplementary Appendix S4). Two reviewers (AM, ST) independently extracted and recorded data on the study population, study design, setting, intervention details, comparison group, and outcomes characteristics. A third reviewer (JA) reviewed all the extraction forms for accuracy and completeness. Discrepancies on approximately 15% of the data required team discussion and 100% consensus was reached.

Risk of bias assessment

Two reviewers (AM, ST) independently assessed study risk of bias (ROB) using Cochrane Collaboration criteria for randomized control trial (RCT) studies 37 and Cochrane Risk-of-Bias in Non-Randomized Studies of Interventions (ROBINS-I) for non-RCTs. 38 RCTs were assessed for low, high, or unclear risk across 5 domains: selection bias, performance bias, detection bias, attrition bias, and reporting bias (see the template in Supplementary Appendix S5). Non-RCTs were assessed for low, moderate, serious, or critical risk across seven domains related to confounding bias, selection bias, intervention classification bias, intervention deviation bias, missing data bias, outcome measurement bias, and reporting bias (see the template in Supplementary Appendix S6).

A third reviewer (JA) reviewed ROB scores for accuracy, and disagreements were resolved through team discussion. RCT ROB across categories was reported using Review Manager 5.4 software.

Data coding, synthesis, and analysis

Three reviewers (JA, AM, ST) coded study characteristics based on source country, type of site, the number of sites, types of participants (eg, patients, clinicians, caregivers), and patient population types (eg, adult/pediatric/all, patients in specific units or with certain health conditions). Intervention characteristics were analyzed using content analysis and classified by the number of intervention components, CTI elements, type of care continuity function supported, and the type of CTI ( Table 1 ).

Table 1.

Intervention characteristics coding framework

CTI: care transition intervention; HIT: health information technology.

Comparison groups and study design were classified using Kaushal et al’s 39 four levels of study design. All outcomes across studies were retrieved for analysis and organized as a matrix to avoid reporting bias. All codes and descriptions are presented in Supplementary Appendix S7. Two reviewers (AM, ST) independently coded the data, and a third reviewer (JA) reviewed all codes for accuracy and completeness. Disagreements were resolved through discussion until 100% consensus was reached.

Narrative analysis

Narrative analysis was conducted on all studies. 42 We first identified the direction of outcome effect through vote counting based on statistical significance (ie, by counting the studies and effect ascertained using a majority). 43 Next, we confirmed whether the direction was meaningful based on whether the effect was found across at least 3 studies. 44 By counting results, we derived overall statements according to the outcome of interest. For example, if 3 or more studies concluded that a specific outcome favored the intervention and met the majority rule, the outcome was recorded as “significantly improved.” If most studies did not favor the intervention, then the outcome was recorded as “significantly worsened.” If most studies did not note a significant difference, a “no significant difference” was recorded. Outcomes reaching a verdict with fewer than 3 studies or outcomes with no majority (ie, tie) were reported as “no conclusive evidence.”

RESULTS

The search identified 6832 articles. After removing 1185 duplicates, 5647 articles were selected for the initial title and abstract screening. Twenty-nine full-text articles were reviewed, and 10 met the inclusion criteria. We manually screened citations from included articles to capture any missing publications and retrieved 34 additional references for full-text review; 2 of these extra references were included. Eleven studies (12 references in total; 2 references were related to the same study) met the inclusion criteria and were included in the review ( Figure 1 ).

Study selection process using the Preferred Reporting Items for Systematic Reviews and Meta-Analysis. 45

Study characteristics

Studies were published between 2009 and 2020, with all but 1 study 46 , 47 published after 2013. Of the 11 studies, 4 were multi-site; 48–51 2 spanned nonacademic and academic settings, 48 , 49 with the remaining conducted in academic settings ( Table 2 ).

Table 2.

Included study characteristics

Note: ROB and ROBINS have been assessed based on risk categories described in the “Risk of bias assessment” section.

CPOE: computerized physician order entry; CTI: care transition intervention; HIT: health information technology; ePED: engaging parents in education for discharge; ER: emergency room; HF: heart failure; QoL: quality of life; ROB: risk of bias; ROBINS: Cochrane Risk-of-Bias in Non-Randomized Studies of Interventions; THEM: Telehealth electronic monitoring

Population

The study population mainly included patients, except one with caregivers 53 ; 3 studies also included clinicians. 46 , 47 , 50 , 55 Nine studies focused on adult patients (≥18), with one exclusively on geriatric patients (65 years or older). 5 One study focused on pediatric patients. 53 Seven studies targeted medical illnesses (major chronic conditions; heart failure; conditions within an internal medicine unit; or all conditions). Three studies targeted surgical patients undergoing vascular surgery, hip or knee replacements, or nonspecified surgery. Sample sizes varied between 16 and 19 043 patients in the intervention group (median = 358) and between 14 and 18 975 patients in the control group (median = 698).

Intervention

Table 3 highlights intervention characteristics and functions.

Table 3.

Intervention characteristics and functions

Design: transitions of care committee; development of patient criterion list; pilot study; project expansion

Implementation: Institute for Safe Medication Practices guidelines/recommendations; systematic review

Implementation: Reflective Practitioner Theory 57 ; Individual and Family Self-Management Theory 58 ; teach-back method

CMS: Centers for Medicare & Medicaid Services; CPOE: computerized physician order entry; CTI: care transition intervention; ePED: engaging parents in education for discharge; HIT: health information technology.

Types of CTIs

One study 53 focused solely on a pre-discharge intervention: a mobile/tablet-based application for a parent discharge preparation. Nine studies 5 , 46–52 , 55 , 56 assessed bridging interventions. One study 54 evaluated a post-discharge intervention: a tablet-based application for home monitoring and assessment, along with patient support and education.

CTI components and elements

Lerret et al 53 evaluated a single-component electronic tablet discharge application. In contrast, the remaining studies reported on multi-component interventions with a combination of the following components: designated discharge support responsibilities (eg, new discharge team roles, added tasks to discharge workflow); paper-based tools; and HIT-based tools. Paper-based tools included physical copies of discharge documents such as patient education and instructions and discharge summaries. HIT-based tools included EMR-integrated tools, 5 , 48 , 49 , 52 computerized physician order entry, 47 mobile and tablet-based applications, 53 , 56 telehealth remote monitoring to support care transitions, 51 , 54 electronic discharge summaries, 55 and electronic patient instruction notes. 50

Each component consisted of at least one of the following elements: discharge planning, discharge summary, follow-up scheduling, medication reconciliation, and patient education and instructions. Care team follow-up scheduling was the most common CTI element across all studies, during or after discharge, followed by patient education and instructions, and then medication reconciliation ( Figure 2 ). Follow-ups were conducted virtually across 5 studies (eg, telephone, video call, web portal). 48 , 49 , 51 , 54 , 56

HIT-based discharge CTI elements. Please note: none of the studies included a discharge checklist. CTI: care transition intervention; HIT: health information technology.

We further characterized the intervention elements to align with the discharge transition process and identified 2 phases: pre-discharge preparation and post-discharge follow-up ( Figure 3 ).

Phases of discharge and order of discharge steps across studies. Numbers across elements indicate the order of steps followed in the discharge process (eg, Zheng et al 56 intervention included 2 steps: coordination of follow-up virtual contact; inpatient nurse and inpatient provider communication about patient care updates). Empty cells indicate the step was not included in the intervention.

All studies included steps across both discharge phases, except Lerret et al, 53 which focused on the pre-discharge preparation.

Care continuity functions supported by CTIs

Two CTIs supported the care continuity functions of information transfer, coordination, and communication. 5 , 49 Three others supported information transfer and communication using at least one of the following: follow-up scheduling, discharge planning, patient education and instructions, medication reconciliation, and discharge summaries. 51 , 53 , 54 Three CTIs supported coordination and information transfer by utilizing 2 or more of the following: patient instructions, medication reconciliation, discharge summary, discharge planning, and follow-up scheduling. 46 , 47 , 52 , 55 One CTI only supported information transfer 50 but used several elements including discharge planning, patient education and instructions, and discharge summaries.

Comparison

Five studies were Level I, 5 , 46 , 47 , 51 , 54 , 55 4 were Level II, 48 , 49 , 53 , 56 and 2 were Level IV. 50 , 52 Control conditions, where applicable, commonly involved usual care without HIT (eg, dictated summaries, standard discharge instructions) (see Table 1 and additional details in Supplementary Appendix S8).

Outcomes

All outcomes from Table 2 were analyzed.

Narrative analysis

The impact of CTIs on the various outcomes was summarized using the vote-counting method in Table 4 (see Supplementary Appendix S9 for outcomes reported only in a single study).

Table 4.

Study outcomes and significance of findings

Note: The number of follow-up consultations and completion of intervention use were reported separately. No outcomes were significantly worsened.

ER: emergency room; NR: not reported.

Hospital readmissions

Among studies reporting 30-day readmissions, 5 , 48–55 no significant differences were found compared to control groups overall. However, in addition to 3 studies finding significant improvement, Bailey et al 48 identified that Medicaid populations had significantly reduced readmissions when using CTIs.

There was no significant difference between intervention and control groups across studies examining ≤6-month hospital readmissions. 46 , 47 , 51 , 55 Both studies examining unspecified time frames of hospital readmissions 52 , 56 noted significant improvements compared to control groups, but with only 2 studies, no conclusive evidence could be drawn.

ER visits

There was no significant difference between intervention and control groups across studies examining ≤6-month ER visits. 46–48 , 50 , 56

Length of stay

With only 2 studies examining the length of stay during readmissions, 50 , 55 no conclusive evidence could be drawn.

Adverse events

Across 3 studies 46 , 47 , 54 , 55 reporting on specific adverse events, there was no significant difference between intervention and control groups in adverse events, such as 30-day infection rates, adverse drug events, and procedure-related injuries.

Outpatient follow-up

No conclusive evidence could be drawn with only 2 studies examining 14-day 5 , 48 and overall 54 , 56 outpatient follow-up rates.

30-Day mortality

No conclusive evidence could be drawn with only 2 studies examining 30-day patient mortality. 51 , 55

Medical expenditures

No conclusive evidence could be drawn with only 2 studies examining overall medical expenditures. 48 , 52 However, in the sub-group analysis, we found that Bailey et al 48 identified a significant reduction in spending within the Medicaid population.

Patient experiences and perceptions of discharge care

Three 46 , 47 , 52 , 53 of 4 studies found significant improvements in perceived intervention quality, discharge quality, and patient satisfaction with discharge areas such as overall care and medication information.

Compliance to intervention

Two studies examined different forms of intervention compliance. Moy et al 50 found that over 98% of patient documentation was completed, and there was an increased number of handoffs within the intervention group. Ong et al 51 reported 82.7% compliance to intervention, which decreased over time.

Risk of bias assessment

Five RCT studies were assessed using ROB, and all were determined to have low overall risk. 5 , 46 , 47 , 51 , 54 , 55 The domain judged to have the lowest ROB was selective outcome reporting. In contrast, the domain with the highest ROB was blinding of participants and personnel, as this is not always feasible in studies on care transition interventions. Six non-RCT studies using ROBINS-I: 3 studies 48 , 49 , 56 were classified as serious; 2 studies 50 , 53 had a moderate ROBINS-I ROB score, and 1 study 52 had a low ROBINS-I ROB score. Most serious risks were attributed to bias due to missing data, and moderate risks were attributed to bias in outcome measurements. Supplementary Appendix S10 presents a risk of bias summary with review authors' judgments about each quality item for each included study and risk of bias graph with percentages of high, low, and unclear risk for each quality item across all studies.

DISCUSSION

Poorly coordinated hospital-to-home discharge transitions can lead to adverse events and complications in ∼20% of patients within 3 weeks of discharge, 3 , 4 resulting in preventable readmissions and unnecessary financial burden. 19 , 27 , 59–62 Given the widespread use of HIT in healthcare, this review systematically synthesizes the studies evaluating HIT-based discharge CTIs to mitigate care transition failures leading to poor patient outcomes. Based on this review, we described the “lay of the land,” highlighting the underlying characteristics and functions of current HIT-based discharge CTIs. In addition, we appraised and aggregated the empirical evidence on the cumulative effect of HIT-based interventions to support discharge transitions from hospital to home, potentially contributing to implications for evidence-based practice and informatics research.

We report on the key findings from this review: First, most of the included studies were non-RCTs conducted at a single academic center and were predominantly focused on adult patients. Only 3 studies focused on surgical patients and one on pediatric patients. Second, all studies assessed the impact of these interventions on clinical outcomes, but only a few investigated patient experiences 46 , 47 , 52–54 and clinician compliance to intervention outcomes. 46 , 47 , 50 Third, since the HITECH Act of 2011, 63 there has been an increase in HIT-based interventions 64 to support coordination during discharges, as reflected by the majority of our included studies (10 of 11 studies; 10 out of 12 references) being published after 2013. Fourth, most included studies evaluated multi-component CTIs supporting pre-discharge preparation and post-discharge follow-up phases and consisting of designated discharge support responsibilities with HIT being used for follow-up scheduling, patient education, and medication reconciliation. Tools were also frequently utilized for follow-up virtual contact and care coordination with outpatient providers.

We attempted a meta-analysis on outcomes with greater than 2 studies (30-day readmissions, ER visits, 3- and 6-month quality of life measures) using a random-effects model. However, the statistical heterogeneity was too high for meaningful analysis [eg, 30-day readmissions (2 studies), mean difference = −1.54, 95% confidence interval −4.20 to 1.20, I2 = 95%]. In addition, several studies lacked sufficient data for pooled meta-analysis; 7 study authors 5 , 46 , 47 , 50 , 51 , 54–56 were contacted for missing data, of which only 2 responded, but stated that the missing data was unavailable.

Our narrative analysis suggested that CTIs demonstrated a significant improvement in patient experiences and perceptions regarding discharge care. 46 , 47 , 52 , 53 Across the studies with improved patient experiences and perceptions regarding discharge care, 2 bridging interventions 46 , 47 , 52 supported coordination and information functions of care continuity, whereas the pre-discharge intervention 53 supported communication and information functions. These 3 interventions also encapsulated 2–4 discharge elements. Alternatively, studies demonstrated no statistically significant differences across hospital readmissions, ER visits, length of stay, 30-day adverse events, and 14-day and overall outpatient follow-up. In addition, no conclusive evidence regarding the impact of these interventions on 30-day mortality, outpatient follow-ups, and medical expenditures could be drawn. With such varied interventions and inconclusive findings related to clinically relevant outcomes, we are unable to pinpoint the HIT characteristics that had a meaningful impact. Nevertheless, there was no evidence from the included studies of any harm or unintended consequences associated with the use of HIT-based interventions.

This systematic review has several limitations. First, there was considerable heterogeneity across the healthcare systems and resources, varying nature of discharge components and type of HIT-based intervention, and nonstandard outcome measurement and reporting methods. For example, Bailey et al 48 measured average readmission rates per participant using mean and SD; Farford et al 49 measured readmission rates using percentages. Second, our analysis was limited by the English language and peer-reviewed journal publications of included articles. Third, we did not include retrospective studies; the choice of excluding retrospective observational studies was made to reduce risk of bias in the evidence synthesis process and also because such studies are not well-suited for determining the efficacy of interventions. Fourth, some improvements in outcomes could be attributed to reliance on smaller sample sizes, detailed pre-implementation strategies, or other factors.

Potential implications for practice and research in informatics

Our review findings have several implications for both practice and research. There are several care transitions models, such as Better Outcomes for Older Adults Through Safe Transitions and Project Re-Engineered Discharge, which have been shown to reduce readmissions. 65–68 However, several limitations to sustainability exist, including limited resources (eg, funding, training, and new staff) and clinician buy-in. 69 , 70 Current CTIs consist of numerous components (eg, tools for health professionals, in-hospital visits, follow-up phone calls) but are not fully leveraged to support the early identification of patient risks and problems after discharge and patient-centered care. 71 Furthermore, given that the elements of prior care models are resource-intensive, we can potentially leverage HIT components to optimize the existing interventions using CPOE, mobile or web-based applications, and EHR- or system-integrated discharge support tools. For example, machine learning (ML) models have been harnessed to enhance the functionalities of CDS, predicting patient health status (ie, lactate levels 72 ) risks to complications, and other outcomes within postoperative settings. 73 Equipped with ML-based technology, remote monitoring via telehealth can provide a unique opportunity for remote care teams to assess and predict patient risks while accurately incorporating contextual stressors to develop care plans accordingly.

In addition, further work is essential to develop a HIT-based care transition model that can guide the design and evaluation of effective and efficient HIT discharge CTIs. The lack of information on the cost of intervention implementation is particularly problematic for healthcare organizations planning strategic approaches for reducing readmissions. Further research on intervention costs are needed to assess the return on investment and value added by HIT-based discharge CTIs.

Out of the 11 studies, only 6 were assessed as low risk of bias and 5 were RCTs. Allocation concealment and incomplete outcome data were concerns highlighting the high risk of bias in RCTs; similarly, bias in the measurement of outcomes was evident across 5 out of the 6 non-RCTs, and bias due to missing data was a serious risk in the 3 remaining studies. 48 , 49 , 56 Combined with the lack of standardized and validated outcome measures, these concerns suggest a strong need for standardized data collection and analysis in future studies. Following a standard template for better reporting of intervention description is recommended to better ascertain the intervention characteristics. For example, the TIDieR (Template for Intervention Description and Replication checklist) 74 guides authors to specify details including intervention name; rationale; theory or goal of the elements essential for the intervention; any information or physical materials used in the intervention; procedure; activities and processes used in the intervention; intervention provider; models of intervention delivery; intervention venue; intervention intensity; planned intervention modifications/tailoring; actual intervention modifications; and intervention fidelity and adherence.

Additional research gaps identified include the need for concerted efforts toward (1) studying hospital discharges from nonacademic and community hospitals to home settings; (2) examining surgical patient discharges; (3) using standardized outcome measures for outcome measurement and reporting; (4) incorporating both clinical and process outcomes using a mixed-method approach to study perspectives of patients/caregivers and clinicians on the use and delivery of discharge intervention and its impact on clinical outcomes and lastly; and (5) ascertaining the impact of a HIT-based care transition model on clinical and patient outcomes.

CONCLUSION

High-quality discharge care transitions are crucial for hospital-to-home care continuity and patient safety. Multiple CTIs consisting of designated discharge support responsibilities, paper-based tools, and HIT-based elements were implemented and evaluated to support discharge functions and increase the safety and quality of discharges. Our systematic review synthesized and appraised the empirical evidence on the characteristics of these interventions, assessing effectiveness on hospital readmissions and ER visits along with other clinical and process/satisfaction outcomes using narrative analysis methods. Evidence demonstrated that overall hospital readmissions, patient experiences, and care perceptions improved significantly with the use of CTIs, but all other outcomes had no significant improvement or were inconclusive.

AUTHOR CONTRIBUTIONS

JA, TK, and SK conceived the study. JA, TK, and SK were involved in developing the search strategy; JA, TK, AM, and ST were involved in piloting and conducting data extraction and organization. JA, AM, and ST were involved in the analysis and quality assessment. All authors were involved in the writing of the manuscript and revising it for publication.

SUPPLEMENTARY MATERIAL

Supplementary material is available at Journal of the American Medical Informatics Association online.