Impact of multicomponent integrated care on mortality and hospitalization after acute coronary syndrome: a systematic review and meta-analysis

Abstract Aims Multicomponent integrated care is associated with sustained control of multiple cardiometabolic risk factors among patients with type 2 diabetes. There is a lack of data in patients with acute coronary syndrome (ACS). We aimed to examine its efficacy on mortality and hospitalization outcomes among patients with ACS in outpatient settings. Methods and results A literature search was conducted on PubMed, EMBASE, Ovid, and Cochrane library databases for randomized controlled trials, published in English language between January 1980 and November 2020. Multicomponent integrated care defined as two or more quality improvement strategies targeting different domains (the healthcare system, healthcare providers, and patients) for one month or more. The study outcomes were all-cause and cardiovascular-related mortality, hospitalization, and emergency department visits. We pooled the risk ratio (RR) with 95% confidence interval (CI) for the association between multicomponent integrated care and study outcomes using the Mantel–Haenszel test. 74 trials (n = 93 278 patients with ACS) were eligible. The most common quality improvement strategies were team change (83.8%), patient education (62.2%), and facilitated patient-provider relay (54.1%). Compared with usual care, multicomponent integrated care was associated with reduced risks for all-cause mortality (RR 0.83, 95% CI 0.77–0.90; P < 0.001; I2 = 0%), cardiovascular mortality (RR 0.81, 95% CI 0.73–0.89; P < 0.001; I2 = 24%) and all-cause hospitalization (RR 0.88, 95 % CI, 0.78–0.99; P = 0.040; I2 = 58%). The associations of multicomponent integrated care with cardiovascular-related hospitalization, emergency department visits and unplanned outpatient visits were not statistically significant. Conclusion In outpatient settings, multicomponent integrated care can reduce risks for mortality and hospitalization in patients with ACS.


Introduction
Acute coronary syndrome (ACS) is typically presented with three main subtypes namely non-ST-elevation myocardial infarction (NSTEMI), ST-elevation myocardial infarction (STEMI), and unstable angina. 1 According to the World Health Organization, the leading cause of death was coronary heart disease which accounted for 16% of total mortality in 2019. 2 Its disease burden rises with ageing, resulting in increasing disability, functional decline, and healthcare costs. 3 Despite the availability of organ-protective medications and percutaneous coronary intervention (PCI) in past few decades, 1,4 the clinical outcomes of patients with coronary heart disease in low-and middle-income countries have been suboptimal. [5][6][7] In the Malaysian general population, the leading cause of death was coronary heart disease (17%), followed by pneumonia (11.4%), and cerebrovascular disease (8.3%) based on the 2020 Statistics on Causes of Death. 8 The attained levels of modifiable cardiometabolic risk factors can influence the onset and progression of ACS. [9][10][11][12] Apart from the overburdened health care system, fragmented care may reduce the service quality among patients after the onset of ACS with increased health care expenditure. 13,14 In the 2014-15 Malaysian Acute Coronary Syndrome Registry, the proportions of patients with ACS who reported having diabetes, hypertension, and dyslipidaemia were 43.2%, 60.8%, and 35.9%, respectively. 15 This highlights the challenges in translating evidence into real-world practice with significant gaps in diagnosis, treatment, and monitoring.
Multicomponent integrated care can be the key to relieve the overburdened healthcare systems. It is defined as the implementation of two or more quality improvement (QI) strategies from different domains targeting the healthcare system, healthcare providers and patients. 16,17 Our previous meta-analysis involving 181 randomized clinical trials (RCTs) confirmed the sustained improvements in multiple cardiometabolic risk factors in patients with type 2 diabetes with the implementation of multicomponent integrated care. 16 Herein, we performed a meta-analysis to examine the efficacy of multicomponent integrated care on mortality and hospitalization outcomes among patients with ACS in outpatient settings.

Data sources and searches
We performed literature search on PubMed, EMBASE, Ovid, and The Cochrane Library databases for RCTs on multicomponent integrated care and ACS, which were published in English language between January 1980 (inception) and November 2020. The Medical Subject Headings (MeSH) search terms are listed in the supplementary material online, Table S1. We reported the results in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guideline. The protocol is prospectively registered on International Prospective Register of Systematic Reviews (CRD42020220107).

Study selection
Based on the PICO framework, we included RCTs which recruited at least 100 adults aged 18 years or above with either pre-existing or new-onset ACS managed in outpatient settings, with the implementation of multicomponent integrated care for a minimum of 1 month. Similar to that in our previous work, 16 we defined multicomponent integrated care as the implementation of two or more QI strategies from two out of three domains namely the healthcare system, healthcare providers, and patients (Supplementary material online, Table S2).
The primary outcome was mortality (all-cause and cardiovascular-related). The secondary outcomes were hospitalization (all-cause and cardiovascular-related), health care utilisation (emergency department visits and unscheduled outpatient visits), medication prescription, and the control of cardiometabolic risk factors.

Data extraction and quality assessment
Two authors (J.X.H and Y.F.Y) independently reviewed all publication titles and abstracts to exclude articles not meeting the eligibility criteria. Using a standardised case record form, two authors extracted data including sociodemographic, healthcare settings, sample size, characteristic of study population (age, gender, ethnicity, and cardiometabolic risk factors), duration of intervention, type of QI strategies, outcomes of interest, and medication prescription. Using

Data synthesis and analysis
To explore the effect of multicomponent integrated care, we determined the net intervention with consideration of the number and type of QI strategies, as well as the outcome measures, between the intervention and control arms. In the present metaanalysis, there were seven RCTs with >2-arm study design, in which the intervention arm with the highest number of QI strategies was selected.
We calculated the risk ratios (RR) and mean post-interventional differences with 95% confidence intervals (CI) for binary and continuous outcomes, respectively. We pooled the available trial-level data to estimate measurements of central tendency and dispersion. Continuous variables that were reported as median [interquartile range (IQR)] were converted to mean ± standard deviation (SD) following the established method. 18,19 We pooled the risk ratio (RR) with 95% confidence interval (CI) for the association between multicomponent integrated care and clinical outcomes using the Mantel-Haenszel test. We performed I 2 statistic and funnel plots to assess study heterogeneity and publication bias, respectively. 18 An I 2 statistic of >50% was considered significant study heterogeneity and therefore, random-effect model was used. On the other hand, fixed-effect model was performed for I 2 statistic of <50%. We performed two sensitivity analyses: (1) stratified by median duration of follow-up (≤12 months vs. >12 months) and (2) excluded RCTs that were conducted before 2010 as usual care might have transformed over time. To explore the significance of the differences in RR and the possible influence of confounding factors, we performed meta-regression and subgroup analyses for possible sources of heterogeneity including age, gender, and use of organ-protective drugs (antiplatelet/antithrombotic therapy, RAAS inhibitors, and lipid-lowering therapy). 20 All statistical analyses were performed using RevMan software version 5.4.1 and R version 4.1.3 ('metareg' package). A two-sided P <0.05 denoted statistical significance. Figure 1 shows the PRISMA flow diagram. The database search generated 1759 citations. After removal of duplicates, 1263 articles underwent abstract and title screening. A total of 74 RCTs (n = 93 278 patients) fulfilled the eligibility criteria and were analysed.

Study characteristics
Baseline characteristics of the RCTs included in the meta-analysis are listed in the supplementary material online, Table S3. The mean ± SD age of the study population was 62.4 ± 11.9 years and 67 253 (72.1%) were males. A total of 15 (20.3%) RCTs were conducted in Asia, 29 (39.2%) in Europe, 17 (22.9%) in North America, and 13 (17.6%) in other countries/regions. The median duration of intervention was 6 months (IQR 3-10). The median number of QI strategies implemented was 4 (IQR 2-7). The most common QI  Table S3).
The assessment of study quality is presented in the supplementary material online, Figure S1. The majority of included RCTs were of low risk of bias. There were 31 (41.9%) RCTs with performance bias due to a lack of blinding of participants and study personnel, followed by attrition bias in 18
In addition, multicomponent integrated care was associated with improvements in the diastolic blood pressure (mean difference  Figure S9). There were no significant improvements with glycaemia and systolic blood pressure (Supplementary material online, Figure S9).

Heterogeneity analysis
We conducted meta-regression analysis based on the results of meta-analysis to determine potential sources of heterogeneity. As shown in the supplementary material online Table S6, use of antiplatelet/antithrombotic therapy might influence the effect of multicomponent integrated care on cardiovascular-related mortality. When the proportion of antiplatelet/antithrombotic therapy was >50%, multicomponent integrated care was associated with a pooled RR of 0.79 (95% CI 0.70-0.89) for cardiovascular-related mortality (Supplementary material online, Figure S10). Regarding the risk for all-cause hospitalization, mean age ≥60 years (RR 0.81, 95% CI 0.69-0.94), >90% proportional use of antiplatelet/antithrombotic therapy (RR 0.81, 95% CI 0.66-1.00), and <90% proportional use of lipid-lowering drugs at baseline (RR 0.71, 95% CI 0.55-0.92) were associated with reduced risk for all-cause hospitalization (Supplementary material online, Figure S11). Mean age ≥60 years was also associated with reduced risk for cardiovascular-related hospitalization (RR 0.85, 95% CI 0.77-0.93) (Supplementary material online, Figure S12).

Discussion
To the best of our knowledge, this is the largest meta-analysis investigating the effects of multicomponent integrated care on mortality and hospitalization among patients with ACS managed in the Figure 4 Meta-analysis results of the effects of individual quality improvement strategy on (A) all-cause mortality and (B) cardiovascular-related mortality among patients with acute coronary syndrome. Forest plots were derived from fixed-effects meta-analysis models. SE, standard error; 95% CI, 95% confidence interval; Continuous QI, continuous quality improvement. outpatient settings. We reported that the implementation of multicomponent integrated care for at least one month was associated with a 11-19% reduced risk for mortality and hospitalization among patients with ACS, particularly in those aged ≥60 years, with >50% use of antiplatelet/antithrombotic therapy and <90% use of lipidlowering therapy at baseline. We also reported greater risk reductions with a longer duration of follow-up, highlighting that ongoing multicomponent integrated care program is needed to sustain clinical benefits.
Inter-professional collaboration has become an essential skill to meet complex care needs of patients with chronic health conditions including ACS. Compared to published reviews, the present meta-analysis was the largest by including 74 RCTs involving 93 278 patients with ACS. [21][22][23][24] Furthermore, the present meta-analysis utilized a stricter definition of multicomponent integrated care (defined as two or more QI strategies from different health domains), whereas previous reviews included those with single domain. Compared with usual care, the present meta-analysis reported that patients receiving multicomponent integrated care had higher medication prescription rate and improved levels of diastolic blood pressure, lipid profile, and BMI. A meta-analysis of 52 RCTs involving adults receiving cholesterol treatment reported that every 1 mmol/L reduction in LDL-cholesterol was associated with a 19% relative risk reduction for major adverse cardiovascular events including cardiovascular mortality, non-fatal myocardial infarction, non-fatal ischaemic stroke, and coronary revascularization. 25 Another metaanalysis of 54 RCTs involving adults with obesity also reported that weight loss interventions were associated with an 18% relative risk reduction for all-cause mortality, but not for cardiovascular mortality. 26 In the present meta-analysis, fasting plasma glucose and systolic blood pressure were not significantly improved likely due to lower levels at baseline (median 6.1 mmol/L and 129 mmHg, respectively).
After an initial presentation of ACS, there is an increased risk for recurrence with shortened survival, especially in the presence patients with acute coronary syndrome. Forest plots were derived from random-effects meta-analysis models. SE, standard error; 95% CI, 95% confidence interval; Continuous QI, continuous quality improvement.
of sub-optimally controlled cardiometabolic risk factors. 11,27 Moreover, the complexity of treatment regimen in patients with chronic diseases is a barrier to adherence to guideline-directed medical therapy. 28 As shown in the present meta-analysis, multicomponent integrated care can enhance patient-provider communication and care continuity, as well as promote patient education and empowerment for better self-management, resulting in improved adherence to guideline-directed medical therapy and control of multiple cardiometabolic risk factors. 29 In our subgroup analysis, multicomponent integrated care was particularly effective in those with >50% use of antiplatelet/antithrombotic therapy and <90% use of lipidlowering therapy at baseline, likely due to its impact in overcoming therapeutic inertia among healthcare providers, as well as in promoting self-management among these very high-risk patients.
Patients with ACS tend to have recurrent hospitalizations especially in the first 12 months post-discharge. This is likely due to infections, comorbidities, and medication-related adverse events such as bleeding. 30 Hence, post-discharge treatment plan is crucial to improve clinical outcomes, prevent hospitalization, and mortality among patients with ACS. 31,32 Among 12 quality strategies, case management, team-based care (involving allied health personnel such as nurses, pharmacists, and dietitians), and facilitated patientprovider communication (using either the manual/electronic tools, allied health personnel, or peer supporters) are most effective in preventing mortality and hospitalization due to any causes among patients with ACS. Our results are consistent with a meta-analysis involving patients with type 2 diabetes. 16 In contrast to patients with type 2 diabetes, 16 continuous QI, audit and feedback, and clinician education were the additional QI strategies that could reduce risk for mortality and hospitalization in patients with ACS. Several studies and disease registries had shown that up to 40% of patients with ACS had coexisting diabetes. 4,33,34 As such, compared with those with type 2 diabetes, patients with ACS are likely to have a more complex clinical course needing multicomponent integrated care with continuous refinement of the approach, along with audit and feedback at the system-and provider-level. In addition, clinician education is important as the management of ACS has progressed thus far. Decision-making is dependent on the patient's history, haemodynamic stability and timing of initiation, and risk-benefit ratio of available therapies (pharmacological approach vs. coronary revascularization). 35 Our main strength is by performing an updated meta-analysis of RCTs examining the efficacy of multicomponent integrated care on mortality and hospitalization among patients with ACS in outpatient settings. We also used stricter criteria for defining multicomponent integrated care to emphasize the need for interdisciplinary collaboration in managing patients with chronic diseases including ACS. Given that usual care might transform over time, we reported consistent benefits on mortality and hospitalization with multicomponent integrated care after excluding RCTs conducted before 2010.
We acknowledge several study limitations. First, classification of QI strategies was challenging as they were not clearly defined in lowquality RCTs. Second, as in other RCTs, it was not possible to eliminate selection bias and observation bias in the present meta-analysis. Although the funnel plots showed minor asymmetry corresponding to study heterogeneity (Supplementary material online, Figure S13), our results were consistent in several sensitivity and subgroup analyses. Last, this was a trial-level meta-analysis and the lack of access to individual-level data might have limited the robustness of our results. Hence, we were not able to perform subgroup analysis by the status of smoking, diabetes, and obesity.

Conclusions
Multicomponent integrated care was associated with a reduced risk for mortality and hospitalization among patients after ACS. This was likely due to better treatment adherence and control of multiple cardiometabolic risk factors. Further cost-effectiveness analysis is needed to inform, change, and influence.

Supplementary material
Supplementary material is available at European Heart Journal-Quality of Care and Clinical Outcomes online.