Systematic review of contemporary improvement interventions on care coordination, discharge support and transitional care from the patient experience perspective

Background

The patient experience with care is an integral component of healthcare quality and a construct used to assess the person-centeredness of healthcare delivery, informing quality-improvement (QI) activities.^1–3 Specifically, patient experience refers to how patients have experienced valuable aspects of healthcare delivery, such as physician communication, involvement in decision-making, getting timely appointments, or supports to care coordination and discharge planning in the transitions across settings to address care fragmentation issues.¹

As synthesized in several systematic reviews, better patient experiences with care have been associated with improved treatment adherence and to improved patient and health system outcomes (e.g., reduced healthcare utilization or malpractice litigation).^4–7 For example, a recent systematic review found that better scores on standardized patient experience measures were associated with greater self-reported physical and mental health, lower frequency and length of hospitalizations, and fewer emergency room visits.⁴ Furthermore, improving the patient experience of care has been recognized as a QI aim in itself, under the goal of improving the person-centeredness of care.^8,9

Many health systems increasingly reimburse providers for the value of care, including patient experience scores. The US Agency for Healthcare Research and Quality (AHRQ) has developed the Consumer Assessment of Healthcare Providers and Systems (CAHPS) surveys and quality measures, which have been widely used to track, publicly report, and compare providers’ patient experience performance.¹⁰ Within selected pay-for-performance programs, the US Centers for Medicare & Medicaid Services uses CAHPS scores, reflecting patient experience of care, when computing payment incentives.¹¹ Several other countries require providers to monitor and improve patient experience.^12,13 Overall, healthcare systems, organizations, and frontline practitioners are increasingly subject to quality mandates, value-based incentives, or market pressures (e.g., customer loyalty) to develop systematic activities to improve their patient experience performance,¹⁴ including for care coordination and discharge support.^15,16

Reducing care fragmentation, improving continuity of care, and preventing medication errors or rehospitalizations can be achieved by interventions such as supported discharge, case management programs, transitional care services, use of care coordinators as well as models of care such as patient-centered medical homes (PCMHs). Systematic reviews have addressed the effectiveness of these health service interventions, and many of these have been shown to be effective for a set of health, quality of life, and health system outcomes (e.g., reduced rehospitalizations).^17–20 For example, interventions to facilitate the transition from hospital to home when started in the hospital and continued in the community through telephone or other telehealth follow-up or scheduled home visits can be effective in reducing hospital readmission, especially when tailored to each patient.^18,19,21

However, we found no systematic review focused on the impact that care coordination, discharge support, or transitional care activities have on the patient’s experience with care. A recent overview of systematic reviews for care coordination interventions found patient experiences were rarely addressed or synthesized as outcomes of these interventions.²² Here, we aim to synthesis the contemporary literature (2015-2022) on the impact of care coordination, discharge support, and transitional care activities on patient experience as assessed by formal patient-reported experience measures.

Methods

Results

Figure 1 provides the PRISMA flowchart.

• Download figure
• Open in new tab

Figure 1: PRISMA flowchart of the systematic review

Among the 1087 unique papers screened, 138 full texts were assessed for eligibility and 15 were included.^{15,16,26–38} The absence of impact assessment with inferential statistics was the main reason for exclusion at the full-text assessment. Among papers included, seven focused on enhanced care coordination supports (e.g., over and beyond models of care based on a PCMH), whereas eight focused on enhanced discharge supports, specifically for enhanced patient education, reengineered discharge, telephone follow-up including for transitional care, or a mix of those interventions.

• a) Enhanced care coordination supports – primary care level (n= 7)

Table 1 summarizes each study’s methods (e.g., design, context) and risks of bias; Table 2 summarizes the intervention and study findings.

Context-wise, all studies were multi-site PCMHs or other primary care practices, all from a network or integrated health system in the US: Veterans’ Health Administration (VHA),^32–34 a statewide PCMH network,³⁵ Federally Qualified Health Centers,³⁶ a Safety Net Medical Homes’ Initiative,³⁸ or academically-affiliated primary care practices.³⁷

Three studies were RCTs for facilitating site-based improvement projects. A recent RCT at the VHA, with the fewest appraised risks of bias (Table 1), used QI facilitation (e.g., coaching) on the site’s use of a care-coordination improvement toolkit – beyond its internal dissemination (control); both groups improved, with no significant benefit for the more resource-intensive arm.³⁴

Another RCT focused on the impact of augmented care coordination for patients with high risk of hospitalization (beyond standard PCMH; control group) and found improvements in 2 of 6 care coordination items.³³ The remaining RCT encouraged the selection of high-value care-coordination QI goals (beyond standard QI facilitation: control group) and found mixed results.³⁵ These RCTs had some appraised risks of bias (e.g., potential contamination;³³ selection bias³⁵).

Among other study types, one non-randomized trial focused on nurses taking on additional care coordination and navigation support for socially vulnerable patients, beyond a PCMH model (control), and found close but no significant differences between groups (p= 0.07).³⁶ Finally, three observational studies (tailoring PCMH model to homeless patients;³² practice transformation tailored to the safety nets;³⁸ learning collaborative³⁷) showed mixed results and had substantive risks of bias (Table 1).

• b) Enhanced discharge or follow-up supports - after hospital or surgical care (n= 8)

Table 3 summarizes the methods and Table 4 the interventions and results for the studies on enhanced discharge support, including patient education or follow-up. All studies were based in the US.

Among the three RCTs, two focused on patient discharge education using the teach back method, after a health literacy assessment²⁶ or as part of a multi-modal language-concordant discharge support and telephone follow-up.²⁷ None of the study interventions led to better experience results than standard care. The other RCT was on a web-based, surgical patient education tool for pre-operative & follow-up instructions, and found improvements in the overall experience score.²⁸ However, several risks of bias were appraised in all the RCTs such single-center study,^26–28 limited follow-up,^27,28 or no correction for multiple comparisons (Table 3).^26-28

Two non-randomized trials were found: one on the implementation of ReEngineered Discharge for three patients a day that nurses selected as complex cases,¹⁶ the other on teach back discharge education by nurses.³¹ Both achieved improvements in selected patient experience items (e.g., information provision). In turn, an observational study with historical controls focused on a telephone follow-up after hospital discharge with possible remediation plans or rescue resources (e.g., same-day clinic appointment);¹⁵ it achieved higher percentages of top-box scores (i.e., highest score on the given patient experience items) compared to controls for 5 of the 11 patient experience items. Several risks of bias applied to the three studies (Table 3), including risks of contamination and unadjusted analyses.

Finally, two pre-post studies were included, one using teach back and other discharge redesigns,²⁹ the other a pharmacist-led discharge education and follow-up intervention.³⁰ These provided either positive or mixed results (improvement in top box percentages but not in the percentile ranks with peer providers). In addition to the uncontrolled study design, these studies had multiple substantive risks of bias (Table 3), limiting the ability to draw conclusions.

Discussion

This knowledge synthesis addresses a gap in the literature regarding the impact of care coordination or discharge support improvement on standardized patient experience measures. Health-service improvement activities, including enhanced supports or tailoring for vulnerable populations, achieved mixed to null improvements in the patient experience with the care coordination or discharge processes, compared to standard models of care or simpler quality-improvement supports. This was found by seven studies (including three multi-site RCTs) on enhanced QI support provided to PCHMs or other primary care practices. It was also found by eight studies on enhanced discharge support such as patient education, reengineered discharge, and/or telephone follow-up or transitional support. These latter studies, as a body of literature, had a greater risk of bias.

Several risks of bias were identified in the reviewed literature across the study designs, limiting the internal and external validity of the findings. The studies on enhanced discharge support were typically single center. Even the multi-site studies on enhanced care coordination were performed within a network or organization. Hence, it is unknown how the findings apply elsewhere. For instance, the study with the lowest appraised risks of bias found that the more resource-intensive type of QI support provided to PCMH units was not superior to the dissemination of the QI tool alone for driving improvements in patient experience of care coordination activities. But this finding directly applies to the VHA. The VHA has been subject to systematic QI programs, studies, or activities,^39,40 which may have contributed to building the capacity for effective QI activities across their units as merely supported by an online QI tool, even without a centralized and more costly QI facilitation. The same may not apply to other healthcare organizations and development circumstances. Because context is a vital variable in QI science,⁴¹ future studies may consider generating evidence with validity across organizations and networks and their varied sets of circumstances. QI collaboratives or hubs, including those externally funded, might equitably foster the capacity to develop and assess QI activities across service delivery organizations, say within a jurisdiction within the US.⁴²

Varied study designs were included here. While RCTs can produce the most solid evidence when well conducted, a wider range of designs were included to reflect real-world improvement activities and evaluations, including pragmatic constraints.⁴³ Study limitations within each design were reflected in the appraised risks of bias, especially among the single-center discharge support studies. Moreover, studies’ delivery of the tested intervention was sometimes pragmatically constrained, too. For instance, a pharmacist-led patient education intervention was performed for patients discharged only during particular hours.³⁰ Similarly, a reengineered discharge support program was capped at a number of patients per day as selected by nurses as “complex” cases.¹⁶ These limitations may affect the fidelity of the improvement interventions and accuracy of the results and may often reflect budget or other limitations (e.g., human resources, in-house improvement or research expertise) for the conduct or evaluation of QI activities.⁴³ Percentile ranks (i.e., provider ranks in patient experience performance relative to peers) were used as an additional metric only in one included paper (with a pre-post design), showing mixed results while the percentage of top-box scores had positive results.²⁹ Reporting results in more than one metric of a patient experience measure can be important, especially for percentile ranks in pre-post studies to partly overcome the lack of a control group or the use of historical ones. Overall, the inclusion of varied study designs, interventions, and patient experience measures and metrics may inform further improvement activities and its study, at varying levels of available resources. Across study designs, some common intervention approaches were found, such as enhanced supports for QI facilitation on care coordination^34,35 or tailored approaches for at-risk populations (e.g., socially vulnerable, homeless),^32,33,36,38 often showing mixed patient experience results relative to standard supports. In turn, discharge support approaches that used a reengineered discharge or teach back strategies were also common but with less favorable results when compared to studies and reviews on other outcomes, such as reducing readmissions.^44–46 For example, one recent systematic review of teach back strategies applied to discharge education identified a 45% reduction in 30-day readmission.⁴⁶ Here, studies using the teach back method had mixed results overall^26,27,29,31 but neutral results in the RCTs,^26,27 either as the main strategy²⁶ or as part of a multi-modal strategy that also included a reengineered discharge, motivational interviewing, follow-up and transitional support, and overall language-concordant supports.²⁷ When found to be effective compared to a control intervention, the teach back method was used in isolation and specifically improved the patient experience with receiving information about new medicines.³¹ For transitional care post discharge, which was rarely addressed in the body of literature reviewed, the multi-modal intervention that included this component found neutral results,²⁷ whereas the intervention specifically focused on this component found positive results in five patient experience items.³⁰ Further research could compare the cost-effectiveness of multi-modal and single-method improvement approaches as well as determine whether multi-modal interventions, due to their complexity, are delivered with fidelity to the main active ingredients of the intervention.^47,48

To be impactful on the patient experience, it may be key for studies to assess how patient-centered (e.g., how attentive, respectful, responsive, with genuine interest) was the delivery of a teach back, telephonic transitional support, and other supportive patient-provider communications intervention. This might include how tailored the communication was to the individual patient¹⁸ to avoid the risks of being perceived as just another box-ticking exercise.⁴⁹ Finally, in alignment with growing focus on coproduction and on experience-based codesign for health-service improvement activities,^50,51 the patient experience with care can be further engaged all the way through from the intervention codesign to its evaluation. The literature reviewed here had no focus on engaging the patient perspective in the codesign of the improvement interventions.

All the included studies were conducted in the US, which may result from multiple factors. First, the CAHPS program in the US provides established, widely used standardized patient experience measures, which was a criterion for inclusion. Second, providers’ performance on CAHPS measures is publicly reported and included in some value-based purchasing programs,¹¹ which may have driven the interest of providers to develop and report these improvement. Third, CAHPS has been a primary quality target of a “person-centered” medical home, with the US National Committee for Quality Assurance recommending administration of the CAHPS PCMH survey for PCMH transformation, and a related study found that CAHPS surveys were considered actionable for PCMH transformation.⁵² Fourth, the VHA in the US has its own patient experience measurement system, which was used a standardized assessment measure in studies externally reported and included here.^33,34 Fifth, improvement activities otherwise fitting the study eligibility criteria from other countries may have been conducted but not reported in English or through peer-reviewed journals.

This review has several limitations. First, it excluded articles published in languages other than English or in the grey literature, which affects representativeness but reflects peer-reviewed knowledge readily available to an international audience. Second, representativeness is also affected by the exclusion of articles published before 2015, which allowed for a synthesis of contemporary approaches in an ever-changing healthcare delivery landscape. Third, for the data extraction, the second reviewer performed only confirmation tasks while two independent reviewer roles were used for all other tasks. Fourth, we did not perform a formal grading of the evidence or of the risks of bias within or across study types, within heterogenic health service delivery contexts, interventions, and study methods. Fifth, we did not include studies that used patient experience as a secondary study outcome. While including these studies may have added to the pool of evidence, it would have risked including studies in which improving the patient experience of care was not a primary intention.

Conclusion

Enhanced support for improving care coordination and discharge support had mixed or neutral results for improving the patient experience with care beyond standard care or simpler supports. Substantial risk of bias and lack of comparison between improvement approaches impede firmer conclusions, especially for the studies on enhanced discharge support, which were also typically single center. The multi-site studies on enhanced care coordination were also limited by being performed within an integrated delivery system, network, or organization, limiting the generalizability of the results across organizational and service delivery contexts. Improving the person-centeredness of care may require strengthening the capacity to fund, design, conduct, and evaluate improvement studies focused on patient experience across providers and their networks. Improving the patient experience with care coordination and discharge may require the strengthening of the improvement studies, and further engaging the patient perspective from the intervention development to its evaluation.