Key Points
Question
Can a multifaceted quality improvement intervention increase health care professionals’ adherence to prescribing evidence-based therapies for patients at high cardiovascular risk?
Findings
In this cluster randomized clinical trial that included 1619 patients from 40 outpatient clinics or primary care units in Brazil, at 12 months, those in intervention clusters were more likely to receive a prescription of all eligible evidence-based therapies than those in control clusters (73.5% [515 of 701] vs 58.7% [493 of 840]).
Meaning
Compared with usual care, a multifaceted intervention resulted in improved prescription rates of evidence-based therapies when assessed as an all-or-none measure.
This cluster randomized clinical trial evaluates the effect of a multifaceted quality improvement intervention on the prescription of statins, antiplatelet therapy, and angiotensin-converting enzyme inhibitors or angiotensin receptor blockers in patients at high cardiovascular risk in Brazil.
Abstract
Importance
Studies have found that patients at high cardiovascular risk often fail to receive evidence-based therapies in community practice.
Objective
To evaluate whether a multifaceted quality improvement intervention can improve the prescription of evidence-based therapies.
Design, Setting, and Participants
In this 2-arm cluster randomized clinical trial, patients with established atherothrombotic disease from 40 public and private outpatient clinics (clusters) in Brazil were studied. Patients were recruited from August 2016 to August 2017, with follow-up to August 2018. Data were analyzed in September 2018.
Interventions
Case management, audit and feedback reports, and distribution of educational materials (to health care professionals and patients) vs routine practice.
Main Outcomes and Measures
The primary end point was prescription of evidence-based therapies (ie, statins, antiplatelet therapy, and angiotensin-converting enzyme inhibitors or angiotensin receptor blockers) using the all-or-none approach at 12 months after the intervention period in patients without contraindications.
Results
Of the 1619 included patients, 1029 (63.6%) were male, 1327 (82.0%) had coronary artery disease (843 [52.1%] with prior acute myocardial infarction), 355 (21.9%) had prior ischemic stroke or transient ischemic attack, and 197 (12.2%) had peripheral vascular disease, and the mean (SD) age was 65.6 (10.5) years. Among randomized clusters, 30 (75%) were cardiology sites, 6 (15%) were primary care units, and 26 (65%) were teaching institutions. Among eligible patients, those in intervention clusters were more likely to receive a prescription of evidence-based therapies than those in control clusters (73.5% [515 of 701] vs 58.7% [493 of 840]; odds ratio, 2.30; 95% CI, 1.14-4.65). There were no differences between the intervention and control groups with regards to risk factor control (ie, hyperlipidemia, hypertension, or diabetes). Rates of education for smoking cessation were higher among current smokers in the intervention group than in the control group (51.9% [364 of 701] vs 18.2% [153 of 840]; odds ratio, 11.24; 95% CI, 2.20-57.43). The rate of cardiovascular mortality, acute myocardial infarction, and stroke was 2.6% for patients from intervention clusters and 3.4% for those in the control group (hazard ratio, 0.76; 95% CI, 0.43-1.34).
Conclusions and Relevance
Among Brazilian patients at high cardiovascular risk, a quality improvement intervention resulted in improved prescription of evidence-based therapies.
Trial Registration
ClinicalTrials.gov identifier: NCT02851732
Introduction
Cardiovascular diseases remain the leading cause of death globally, particularly in low- and middle-income countries, where 80% of the burden resides.1,2 It is well established that statins,3,4 antiplatelet therapy,5 and angiotensin-converting enzyme (ACE) inhibitors6,7 or angiotensin receptor blockers (ARBs)8,9 reduce the risk of clinical events in patients with cardiovascular disease. Nevertheless, registries have consistently demonstrated that the implementation of these therapies in practice is suboptimal, especially in low- and middle-income countries.10,11
Prior systematic reviews have suggested that certain quality improvement (QI) tools are associated with better quality of care.12,13,14 These include reminders, educational outreach visits, audit and feedback, case management, and distribution of educational materials to health care professionals. Combined strategies targeting different barriers are more likely to be effective than single interventions.15 Nevertheless, trials testing the effects of QI interventions have rarely been conducted in lower-resource settings. To assess the effectiveness of a multifaceted QI intervention in patients at high cardiovascular risk in Brazil, we conducted a cluster randomized clinical trial, the Brazilian Intervention to Increase Evidence Usage (BRIDGE) Cardiovascular Prevention study.
Methods
Study Design
The trial protocol and the statistical analysis plan are available in Supplement 1, and the methods were previously published.16 Briefly, the BRIDGE Cardiovascular Prevention study was a pragmatic 2-arm, cluster randomized clinical trial. The main objective was to evaluate whether a multifaceted QI intervention could improve the prescription of evidence-based therapies for patients at high cardiovascular risk at 12 months. All clusters submitted the protocol for approval by their research ethics boards; written informed consent was obtained at the cluster and at the patient level. The trial is registered at ClinicalTrials.gov (NCT02851732). The enrollment period was from August 2016 to August 2017. Follow-up was completed in August 2018.
Clusters
Clusters were outpatient clinics from public or private hospitals or primary care units from all regions in Brazil. Details of the Brazilian health care system are provided in eAppendix 1 in Supplement 2. Both teaching and nonteaching units were eligible. We identified clusters through previous registries, our research network, and medical societies.
Patients
We enrolled consecutive male and female patients older than 40 years who had established coronary artery disease, a prior ischemic stroke, or peripheral artery disease. Detailed eligibility criteria are shown in Supplement 1 and in eAppendix 2 in Supplement 2.
Randomization and Allocation Concealment
Clusters were randomized 1:1 to a multifaceted QI improvement strategy or to routine practice. All clusters were randomized at once by a statistician using a central web-based randomization system before enrollment of the first patient. Randomization was stratified according to practice type (specialty vs primary care).
Blinding
Patients and investigators were not blinded to the allocation of treatment. Adjudication of clinical events was performed in a blinded fashion. Statisticians were blinded to the nature of the intervention.
Intervention
The QI intervention included case management, audit and feedback, decision support tools based on current guidelines, and distribution of educational materials (eApppendix 3 in Supplement 2). A nurse trained in the intervention acted as the case manager. The case manager (usually 1 per cluster) was responsible for patient evaluations (lasting 15 to 20 minutes) conducted immediately before and after the physician visit. The prephysician visit evaluations were conducted with a checklist that included information on cardiovascular comorbidities, risk factor control, and current medications. This checklist was organized in 4 colored sections comprehending lipid profile control (red), blood pressure control and use of ACE inhibitors or ARBs (green), glycemic control (blue), and antiplatelet therapy (yellow). The case manager provided the filled checklist to the physician together with the patient medical records and prompted the physician in case any abnormality was detected. After the physician visit, the case manager conducted another patient evaluation to guarantee that each section of the checklist had been addressed. In case actions were not taken, the physician was prompted again to check if further management decisions were needed.
We provided physicians from intervention clusters with a single-page decision support algorithm. The algorithm was also organized in different colored sections comprehending lipid profile control (red), blood pressure control and ACE inhibitor or ARBs usage (green), glycemic control (blue), and antiplatelet therapy (yellow). The decision support system summarized the key recommendations adapted from different guidelines in 1 page.
Monthly audit and feedback reports were provided to both research staff and health care professionals from clusters allocated to the intervention group. These reports contained information on adherence to performance measures, such as prescription of evidence-based therapies and risk factor control, and were periodically discussed during web or telephone conferences with the coordinating center.
Finally, we distributed educational materials to patients (ie, a folder including recommendations on lifestyle modification) and to physicians (ie, a folder with information on how to prescribe evidence-based therapies). Health care professionals from intervention clusters (at least 1 physician who acted as the local leader as well as case managers) were invited to attend a 1-day workshop on cardiovascular prevention and received training on how to implement the quality improvement intervention. In addition, all clusters randomized to the intervention received a 1-day on-site training visit complemented by web-based and telephone training. During visits, we discussed the trial design and how to apply the tools. Visits were attended by leadership, physicians, case managers, and other health care professionals.
End Points
The primary end point was prescription of evidence-based therapies (ie, statins, antiplatelet therapy, and ACE inhibitors or ARBs) using the all-or-none approach at 12 months after the intervention period in patients without contraindications. According to these criteria, to be classified as yes, a patient must have received all therapies; otherwise, the patient was classified as no in the database. We chose to define the primary end point using the all-or-none approach because the all-or-none approach is more conservative than using composite scores, and there is robust evidence from large-scale trials to suggest that these medications reduce major cardiovascular events. In addition, trial data suggest that their effects are additive.
Secondary outcomes were prescription of individual components of the primary end point at 12 months, a combined therapy of statins, antiplatelet therapy, and ACE inhibitors or ARBs at 6 months after the intervention period in patients without contraindications, percentage of eligible patients with low-density lipoprotein cholesterol levels less than 70 mg/dL (to convert to micromoles per liter, multiply by 0.0259) 12 months after the intervention period, and prescription of high-dose statins in patients without contraindications. Clinical events were considered as secondary end points, including a combined end point of cardiovascular mortality, nonfatal myocardial infarction, and nonfatal stroke at 12 months. Outcomes related to prescription patterns and laboratory tests were ascertained by an independent data collector. Clinical events were ascertained during in-person visits by health care professionals who also provided documents for adjudication. In addition, the independent data collector spoke with patients by telephone and reviewed medical records to identify other potential events. Clinical events were centrally adjudicated based on standardized definitions (Supplement 1). In specific prespecified subgroups, we also measured the following end points at 6 and 12 months after the intervention period: prescription of β-blockers in patients with previous myocardial infarction without contraindications; smoking cessation education rates (ascertained by patient interview and medical record review) in current smokers; percentage of patients with hypertension with systolic blood pressure less than 140 mm Hg and diastolic blood pressure less than 90 mm Hg; percentage of patients with hypertension with systolic blood pressure less than 120 mm Hg and diastolic blood pressure less than 80 mm Hg; percentage of patients with diabetes with normal fasting plasma glucose levels (less than 126 mg/dL [to convert to micromoles per liter, multiply by 0.0555]) and less than 110 mg/dL; and percentage of patients with diabetes with hemoglobin A1c less than 7%.
Data Collection
In all clusters, data were collected prospectively by trained data collectors not involved in patient care. To avoid ascertainment bias, only these independent data collectors acquired data. The independent data collectors also assessed adherence to the QI tools by reviewing whether the checklists and decision support tools were filled and whether the visits were registered in medical records. Data quality control was assured by automated data entry checks, central statistical monitoring checks, and on-site monitoring.
Baseline Survey
Baseline performance (prescription rates of evidence-based therapies) was obtained from the Registry of Clinical Practice in High Risk Patients (REACT) Registry,11 previously conducted in several clusters that have joined the current trial. In clusters that were not part of the REACT Registry, we conducted a prerandomization survey using the same eligibility criteria for patient inclusion. Details are presented in eAppendix 4 and eTable 2 in Supplement 2.
Sample Size
To detect a 20% absolute improvement rate in the primary end point considering a 50% rate in the control group, an intracluster correlation coefficient (ICC) of 0.2011 with 90% power, and a median of 40 patients per cluster, we needed to randomize at least 21 clusters per group (a total of 42 clusters). A 2-tailed α of 5% was set for statistical significance.
Statistical Analysis
All analyses followed the intention-to-treat principle. Prespecified comparisons between groups were conducted using logistic regression with random effects corrected for the baseline performance. Missing data were not imputed. The effects were presented as odds ratios (ORs) with 95% CIs. A description of how ORs were obtained from logistic regression with random effects is provided in eAppendix 5 in the Supplement 2. Clinical events were compared using Cox regression with random effects, and results are presented as hazard ratios with 95% CIs. Interim analyses were not performed.
We conducted 2 post hoc sensitivity analyses for the primary end point (1) considering multiple imputation for missing data and (2) adjusted for patient age, unit profile (hospital outpatient clinic vs primary care center), unit status (teaching unit vs nonteaching unit), and general vs specialized unit. We also conducted post hoc analyses using the Benjamini-Hochberg correction for multiple comparisons for secondary outcomes at 12 months.17 The effect of the intervention was compared in the following prespecified subgroups: hospital outpatient clinics vs primary care units, teaching vs nonteaching institutions, general vs specialized unit, and presence or not of polyvascular disease.
All analyses were performed using R version 3.5.1 (The R Foundation). A 2-sided P value less than .05 was established as the level of significance for all tests. P values for ORs were calculated using the Wald test and for hazard ratios using normal approximation to the distribution of the restricted maximum likelihood estimators. The Benjamini-Hochberg procedure was also applied to the P values regardless of the model to control for false discovery rate.
Results
From 123 potentially eligible clusters that were invited, 83 were excluded (63 did not answer the invitation despite at least 2 attempts, 15 declined participation, and 2 did not meet inclusion criteria). From the remaining 43 clusters that confirmed interest, 3 withdrew postrandomization but prior to intervention because they were unable to send any research staff to attend training sessions. The characteristics of the 3 clusters excluded postrandomization are presented in eTables 1 and 2 in Supplement 2. From the 40 randomized clusters that completed the study, a total of 1619 patients were enrolled prospectively (Figure 1).
Figure 1. Study Flow Diagram.
Baseline Characteristics
Baseline cluster and patient characteristics were generally similar in each group (Table 1). From the included clusters, 30 (75%) were cardiology sites, 6 (15%) were primary care units, 26 (65%) were teaching hospitals, and the median volume of patients seen was approximately 240 patients per month. The mean (SD) age of the patients enrolled was 65.6 (10.5) years, and 1029 (63.6%) were male, 1327 (82.0%) had coronary artery disease (843 [52.1%] with prior myocardial infarction), 355 (21.9%) had prior ischemic stroke or transient ischemic attack, 197 (12.2%) had peripheral vascular disease, 1431 (84.4%) had a history of hypertension, and 624 (38.5%) had diabetes. The mean (range) number of patients in each cluster was 40 (7-51).
Table 1. Baseline Characteristics of Participating Clusters and Patients.
| Baseline Characteristic | No. (%) | |
|---|---|---|
| Intervention | Control | |
| Cluster | ||
| Total, No. | 18 | 22 |
| Predominant clinical specialty | ||
| Cardiology | 14 (78) | 16 (73) |
| Neurology | 0 | 1 (5) |
| Vascular surgery | 1 (6) | 1 (5) |
| Internal medicine | 1 (6) | 0 |
| Primary care unit | 2 (11) | 4 (18) |
| Previous participation in clinical trials | 16 (89) | 16 (73) |
| Teaching unit | 13 (72) | 13 (59) |
| Volume of patients seen in ambulatory per mo, median (IQR) | 285 (150- 550) | 200 (100-489) |
| Public units | 9 (50) | 11 (50) |
| Structured protocol for care and management of patients at high cardiovascular risk | 7 (39) | 11 (50) |
| Patient | ||
| Total, No. | 726 | 893 |
| Male | 462 (63.6) | 567 (63.5) |
| Age, mean (SD), y | 65.9 (10.4) | 65.4 (10.7) |
| Previous CAD | 612 (84.3) | 715 (80.1) |
| Previous acute myocardial infarction | 370 (51.0) | 473 (53.0) |
| Previous stroke or TIA | 146 (20.1) | 209 (23.4) |
| Peripheral arterial disease | 83 (11.4) | 114 (12.8) |
| Systemic arterial hypertension | 662 (91.2) | 769 (86.1) |
| Diabetes | 308 (42.4) | 316 (35.4) |
| Dyslipidemia | 535 (73.7) | 563 (63.0) |
| Diabetic nephropathy | 25 (3.4) | 23 (2.6) |
| Asymptomatic carotid disease | 40 (5.5) | 86 (9.6) |
| Family history of CAD | 272 (37.5) | 333 (37.3) |
| Heart failure | 110 (15.2) | 161 (18.0) |
| Renal insufficiency | 26 (3.6) | 24 (2.7) |
| Valvular heart disease | 10 (1.4) | 45 (5.0) |
| COPD | 30 (4.1) | 35 (3.9) |
| Bronchial asthma | 18 (2.5) | 24 (2.7) |
| Body mass index >30, No./total No. (%)a | 231/725 (31.9) | 274/892 (30.7) |
| Current smoker | 74 (10.2) | 108 (12.1) |
| Blood pressure, mean (SD), mm Hg | ||
| Systolic | 131 (20.2) | 131.7 (20.3) |
| Diastolic | 76.7 (11.2) | 79.7 (11.1) |
Abbreviations, CAD, coronary artery disease; COPD, chronic obstructive pulmonary disease; IQR, interquartile range; TIA, transient ischemic attack.
Calculated as weight in kilograms divided by height in meters squared.
Adherence to the QI Intervention
In the intervention group, nurses were able to act as case managers for 654 of 713 included patients (91.7%; prephysician and postphysician evaluation visits were completed in 652 patients [91.4%]), adherence to the decision support tools by physicians occurred in 658 cases (92.3%), and lifestyle recommendations brochures were distributed to 637 patients (89.3%) (eTable 3 in the Supplement). All clusters from the intervention group received monthly audit and feedback reports.
Effects on Prescription of Evidence-Based Therapies
The effects of the QI intervention on prescription rates of evidence-based therapies at 12 months are shown in Table 2. Effects of the intervention on prescription rates at admission and at 6 months are shown in eTables 4 and 5 in Supplement 2.
Table 2. Effects of a Quality Improvement Intervention on Prescription of Evidence-Based Therapies and Risk Factor Control at 12-Month Follow-up.
| Outcome | No./Total No. (%) | OR (95% CI) | ICC | P Value | P Valuea | ||
|---|---|---|---|---|---|---|---|
| Intervention | Control | Unadjusted | Adjusted | ||||
| Adherence to evidence-based therapiesb | |||||||
| Admission | 492/726 (67.8) | 547/893 (61.3) | 1.33 (1.06-1.67) | 1.49 (0.86-2.60) | 0.11 | .10 | NA |
| 12 mo | 515/701 (73.5) | 493/840 (58.7) | 1.95 (1.53-2.48) | 2.30 (1.14-4.65) | 0.18 | .01 | NA |
| Statins | 656/701 (93.6) | 686/840 (81.7) | 3.27 (2.21-4.85) | 4.04 (1.50-10.89) | 0.28 | <.001 | .01 |
| High-dose statinsc | 74/701 (10.6) | 59/840 (7.0) | 1.56 (1.05-2.32) | 2.64 (0.80-8.72) | 0.34 | .06 | .34 |
| Antiplatelet therapy | 659/701 (94.0) | 725/840 (86.3) | 2.49 (1.64-3.77) | 3.13 (1.29-7.60) | 0.21 | <.001 | .02 |
| Aspirin | 637/692 (92.1) | 682/836 (81.6) | 2.62 (1.81-3.78) | 3.15 (1.37-7.26) | 0.21 | <.001 | .01 |
| Clopidogrel | 118/701 (16.8) | 159/838 (19.0) | 0.86 (0.64-1.16) | 0.95 (0.48-1.87) | 0.15 | .86 | .97 |
| Ticagrelor | 7/701 (1.0) | 9/840 (1.1) | 0.93 (0.31-2.82) | 1.12 (0.16-8.00) | 0.43 | .89 | .97 |
| Prasugrel | 0 | 2/840 (0.2) | NA | NA | NA | NA | NA |
| ACEi or ARB | 561/699 (80.3) | 625/840 (74.4) | 1.40 (1.07-1.83) | 1.44 (0.88-2.36) | 0.08 | .09 | .35 |
| ACEi | 273/699 (39.1) | 347/840 (41.3) | 0.91 (0.72-1.14) | 1.02 (0.57-1.82) | 0.12 | .94 | .97 |
| ARB | 303/700 (43.3) | 290/840 (34.5) | 1.45 (1.15-1.82) | 1.35 (0.77-2.37) | 0.11 | .21 | .58 |
| β-Blockers in patients with MI | 281/358 (78.5) | 332/456 (72.8) | 1.36 (0.95-1.96) | 1.37 (0.58-3.22) | 0.21 | .40 | .77 |
| Smoking cessation education | 364/701 (51.9) | 153/840 (18.2) | 4.85 (3.76-6.25) | 11.24 (2.20-57.43) | 0.56 | <.001 | .01 |
| Risk factor control | |||||||
| LDL-C level, mg/DL | |||||||
| Mean (SD) | 81.5 (32.7) | 86.4 (36.5) | −4.88 (−13.78 to 4.02)d | −5.19 (−14.28 to 3.91)e | 0.09 | .25 | .58 |
| <100 | 330/432 (76.4) | 272/380 (71.6) | 1.28 (0.91-1.82) | 1.23 (0.67-2.26) | 0.09 | .43 | .77 |
| <70 | 177/432 (41.0) | 146/380 (38.4) | 1.11 (0.81-1.52) | 1.12 (0.60-2.09) | 0.11 | .67 | .90 |
| BP in patients with hypertension, mm Hg | |||||||
| Systolic BP, mean (SD) | 124.3 (32.8) | 126.7 (34.9) | −2.37 (−14.78 to 10.04)d | −2.05 (−14.92 to 10.80)e | 0.27 | .75 | .95 |
| Diastolic BP, mean (SD) | 74.1 (19.1) | 75.1 (20.4) | −1.00 (−8.23 to 6.23)d | −0.46 (−7.91 to 7.00)e | 0.27 | .90 | .97 |
| <140/90 | 433/638 (67.9) | 432/722 (59.8) | 1.42 (1.11-1.82) | 1.37 (0.87-2.18) | 0.07 | .11 | .35 |
| <120/80 | 132/638 (20.7) | 134/722 (18.6) | 1.14 (0.85-1.54) | 1.22 (0.58-2.57) | 0.19 | .54 | .87 |
| Systolic BP <120 | 158/638 (24.8) | 166/722 (23.0) | 1.10 (0.83-1.46) | 1.00 (0.51-1.94) | 0.14 | .99 | .99 |
| Fasting glucose levels in patients with diabetes, mg/dL | |||||||
| ≤110 | 68/202 (33.7) | 53/155 (34.2) | 0.98 (0.60-1.59) | 1.12 (0.60-2.11) | 0 | .67 | .90 |
| ≤126 | 110/202 (54.5) | 82/155 (52.9) | 1.06 (0.67-1.69) | 1.13 (0.60-2.16) | 0 | .65 | .90 |
| Hemoglobin A1c level ≤7.0% in patients with diabetics | 83/165 (50.3) | 67/139 (48.2) | 1.09 (0.66-1.79) | 1.37 (0.52-3.60) | 0.03 | .44 | .77 |
Abbreviations: ACEi, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; BP, blood pressure; ICC, intracluster correlation coefficient; LDL-C, low-density lipoprotein cholesterol; MI, myocardial infarction; NA, not applicable; OR, odds ratio.
SI conversion factors: To convert LDL-C to micromoles per liter, multiply by 0.0259; to convert glucose to micromoles per liter, multiply by 0.0555.
P value adjustment for multiple testing (Benjamini–Hochberg procedure).
Complete adherence to statins, antiplatelet therapy (in patients with MI up to 12 months, dual antiplatelet therapy is considered), and ACEi or ARBs.
High doses defined as 80 mg of simvastatin, 80 mg of pravastatin, 80 mg of atorvastatin, 40 mg of rosuvastatin, 80 mg of fluvastatin, or 4 mg of pitavastatin.
Mean difference and 95% CI estimated by mixed-effect regression model using the center as the random intercept.
Mean difference and 95% CI estimated by mixed-effect regression model using the center as the random intercept and corrected for the baseline survey values as a fixed effect.
Among eligible patients, those in intervention clusters were more likely to receive a prescription of all eligible evidence-based therapies (ie, statins, antiplatelet therapy and ACE inhibitors or ARBs) than those in control clusters (73.5% [515 of 701] vs 58.7% [493 of 840]; OR, 2.30; 95% CI, 1.14-4.65; ICC, 0.17). Post hoc sensitivity analyses for the primary end point yielded similar results (eTable 6 in Supplement 2).
Prescription of statins among 801 of 1150 eligible patients (69.7%) at 12 months was higher in intervention clusters vs control clusters (93.6% [656 of 701] vs 81.7% [686 of 840]; OR, 4.04; 95% CI, 1.50-10.89; ICC, 0.28). Overall prescription of antiplatelet therapy was higher in patients from clusters in the intervention group than in patients from the control group (94.0% [659 of 701] vs 86.3% [725 of 840]; OR, 3.13; 95% CI, 1.29-7.60; ICC, 0.21), mainly driven by increased prescription of aspirin in patients from intervention clusters vs control clusters (92.1% [637 of 692] vs 81.6% [682 of 836]; OR, 3.15; 95% CI, 1.37-7.26; ICC, 0.20). Prescription of ACE inhibitors or ARBs were also higher in QI intervention clusters than in control clusters (80.3% [561 of 699] vs 74.4% [625 of 840]; OR, 1.44; 95% CI, 0.88-2.36; ICC, 0.07), but the difference was not statistically significant.
Effects on Risk Factors Control
Table 2 shows the effects of the intervention on risk factor control at 12 months. There were no differences between the intervention and control groups with regards to the proportion of patients who achieved low-density lipoprotein cholesterol levels less than 70 mg/dL.
Among patients with hypertension, the intervention had no effects on the rate of patients who achieved systolic blood pressure levels less than 140 mm Hg or diastolic blood pressure levels less than 90 mm Hg (67.9% [433 of 638] vs 59.8% [432 vs 722]; OR, 1.37; 95% CI, 0.87-2.18; ICC, 0.07). Similarly, there was no difference between groups in the proportion of patients who achieved systolic blood pressure less than 120 mm Hg.
Education for smoking cessation rates were higher in the intervention than in the control group (51.9% [364 of 701] vs 18.2% [153 of 840]; OR, 11.24; 95% CI, 2.20-57.43; ICC, 0.56). There were no differences between groups regarding diabetes control.
Effects on Clinical Events
Table 3 shows the effects of our intervention on clinical events at 12 months. A total of 19 of 705 patients (2.7%) in the intervention group and 30 of 844 (3.6%) in the control group experienced major cardiovascular events, without a statistically significant difference (hazard ratio, 0.76; 95% CI, 0.43-1.34; ICC, 0).
Table 3. Effects of a Quality Improvement Intervention on Major Cardiovascular Events at 12-Month Follow-up.
| Outcome | No. of Events/Person-Years (Incidence Rate per 100 Person-Years) | Hazard Ratio (95% CI)a | ICC | P Value | P Valueb | |
|---|---|---|---|---|---|---|
| Intervention | Control | |||||
| MACEc | 19/705 (2.69) | 30/844 (3.56) | 0.76 (0.43-1.34) | NA | .34 | .73 |
| Cardiovascular mortality | 10/710 (1.41) | 13/851 (1.53) | 0.92 (0.41-2.11) | 0 | .85 | .97 |
| Stroke | 7/707 (0.99) | 18/844 (2.13) | 0.46 (0.18-1.18) | 0.05 | .11 | .35 |
| Myocardial infarction | 7/708 (0.99) | 4/851 (0.47) | 2.10 (0.61-7.17) | 0 | .24 | .58 |
| Total mortality | 19/710 (2.68) | 19/851 (2.23) | 1.21 (0.64-2.28) | 0 | .56 | .87 |
Abbreviations: ICC, intracluster correlation coefficient; MACE, major cardiovascular events.
Presented as hazard ratio estimates from unadjusted frailty Cox proportional hazard models with random effect by cluster.
P value adjustment for multiple testing (Benjamini–Hochberg procedure).
Combined occurrence of a first cardiovascular event (ie, cardiovascular mortality, acute myocardial infarction, and stroke).
Subgroup Analysis
The subgroup analysis is shown in Figure 2. The effect of the intervention on the primary end point was greater in teaching vs nonteaching clusters. Despite the fact that testing for interaction did not suggest evidence of other subgroup effects, these findings should be interpreted with caution.
Figure 2. Primary End Point According to Specified Subgroups.
ACEi indicates angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker.
Discussion
In this cluster randomized clinical trial, a QI intervention including case management, audit and feedback reports, decision support tools, and distribution of educational materials was effective in improving prescription rates of evidence-based therapies at 12 months. Results were consistent for the components of the primary end point, particularly statins and antiplatelet therapies, among different subgroups but with greater effect in clusters with teaching units. In addition, our intervention also improved education for smoking cessation rates. The intervention had no effect on risk factor control or clinical events.
Previous trials aimed at improving the care of patients with or at risk for cardiovascular diseases conducted in developed countries have had mixed results. The Thrombus Obliteration by Rapid Percutaneous Endovenous Intervention in Deep Venous Occlusion (TORPEDO) trial18 involved more than 38 000 people and 60 clusters in Australia and tested a decision support system combined with audit and feedback. The intervention resulted in a 10% absolute improvement in screening for cardiovascular disease risk. However, there were no significant improvements in prescribing recommended medicines to people at high cardiovascular risk.18 As opposed to this trial, in the TORPEDO trial, there were no adjustments for baseline care.
The Improved Delivery of Cardiovascular Care (IDOCC) project19 was a cluster randomized clinical trial designed to improve the delivery of evidence-based cardiovascular care in 48 primary care practices in Canada. Participants were randomly assigned by region to one of 3 steps. Practice facilitators were intended to visit practices every 3 to 4 weeks (year 1—intensive phase) or 6 to 12 weeks (year 2—sustainability phase) to support changes in practice behavior. The primary outcome was adherence to indicators of evidence-based care measured at the patient level. After adjustment for patient and health care professional characteristics, the intervention did not improve adherence to best practices.19 Similar efforts are currently being conducted in the United States, such as the HealthyHearts NYC trial.20
The main differences between our trial and these studies are (1) inclusion of a broader range of clusters, including outpatient clinics of different specialties from teaching and nonteaching hospitals as well as primary care units; (2) focus on patients with established atherothrombotic disease; and (3) use of a multifaceted intervention, including case management, audit and feedback reports, decision support tools, and distribution of educational materials for health care professionals and patients. To our knowledge, the BRIDGE Cardiovascular Prevention study represents one of the first trials testing a QI intervention in patients with established atherothrombotic disease conducted in a middle-income country. Our results suggest that QI interventions may be feasible in these settings, especially using interventions such as the one shown in this study, which is simple and does not rely on expensive information technology or on complex human interventions. Further trials are needed in lower resource settings.13 An ongoing trial from China (31 708 high-risk patients from 67 clusters)21 is testing a package of interventions to increase adherence to lifestyle modifications and prescription of evidence-based therapies. The primary outcome is the 24-month incidence of cardiovascular events. Finally, the Cardiovascular Risk Reduction in South Asia (CARRS) trial22 is testing a multifactorial diabetes care delivery strategy aimed at improving risk factor control and treatment plans in South Asia.
Limitations
Our trial had several limitations. Most participating sites were cardiovascular clinics, which may limit the external validity of our results. Additionally, our trial did not assess the cost-effectiveness of the intervention, which may limit its widespread utilization owing to costs associated with case management and audit and feedback. Although centers were requested to enroll consecutive patients, we did not implement a system of registration of potentially eligible patients to confirm whether that actually happened. Our intervention was delivered over 12 months, and this may be too short to detect changes in practice and in clinical end points. Our trial did not detect differences in risk factor control, as reflected by the neutral results in blood pressure and lipid control, suggesting that the increase in drugs prescribed was not followed by an increase in drugs actually taken. There is evidence that patient-directed interventions combined with physician-focused strategies may be more effective than the latter alone.23 Thus, the full implementation of evidence-based therapies and translation to improved health outcomes depends on combined strategies. In addition, our study was underpowered to detect differences in clinical outcomes. However, we consider that improving the prescription of statins, antiplatelet therapy, and ACE inhibitors or ARBs represents a valid end point, since it is well established that these therapies have additive effects in reducing major cardiovascular events.24 Our trial is prone to recruitment bias since there were baseline differences in the primary end point between the groups. Additionally, we had more teaching units in the intervention group. We attempted to correct for this limitation by performing our main analyses adjusted for baseline performance.
Conclusions
In conclusion, in patients at high cardiovascular risk, a multifaceted QI intervention resulted in significant improvement in the prescription of evidence-based therapies. The tools tested in our trial can become the starting point for developing future programs to maximize the use of evidence-based interventions for the management of patients with established cardiovascular disease, especially in settings with limited resources.
Trial protocol.
eAppendix 1. The Brazilian health care system.
eAppendix 2. Eligibility criteria.
eAppendix 3. BRIDGE CV quality improvement intervention.
eAppendix 4. Details of the baseline survey.
eAppendix 5. Statistical methods details.
eTable 1. Prescription of evidence-based therapies in eligible patients from BRIGDE Cardiovascular Prevention baseline survey and REACT Registry.
eTable 2. Overall comparison between baseline characteristics from all included and excluded clusters.
eTable 3. Adherence to the multifaceted quality improvement intervention.
eTable 4. Effects of a quality improvement intervention on prescription of evidence-based therapies in eligible patients and risk factors control on admission.
eTable 5. Effects of a quality improvement intervention on prescription of evidence-based therapies in eligible patients, major cardiovascular events, and risk factors control at 6 months follow-up.
eTable 6. Post hoc sensitivity analysis for the primary outcome considering multiple imputations.
Data sharing statement.
References
- 1.GBD 2016 Causes of Death Collaborators Global, regional, and national age-sex specific mortality for 264 causes of death, 1980-2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet. 2017;390(10100):1151-1210. doi: 10.1016/S0140-6736(17)32152-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.França EB, Passos VMA, Malta DC, et al. Cause-specific mortality for 249 causes in Brazil and states during 1990-2015: a systematic analysis for the Global Burden of Disease Study 2015. Popul Health Metr. 2017;15(1):39. doi: 10.1186/s12963-017-0156-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Baigent C, Keech A, Kearney PM, et al. ; Cholesterol Treatment Trialists’ (CTT) Collaborators . Efficacy and safety of cholesterol-lowering treatment: prospective meta-analysis of data from 90,056 participants in 14 randomised trials of statins. Lancet. 2005;366(9493):1267-1278. doi: 10.1016/S0140-6736(05)67394-1 [DOI] [PubMed] [Google Scholar]
- 4.Baigent C, Blackwell L, Emberson J, et al. ; Cholesterol Treatment Trialists’ (CTT) Collaboration . Efficacy and safety of more intensive lowering of LDL cholesterol: a meta-analysis of data from 170,000 participants in 26 randomised trials. Lancet. 2010;376(9753):1670-1681. doi: 10.1016/S0140-6736(10)61350-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Antithrombotic Trialists’ Collaboration Collaborative meta-analysis of randomised trials of antiplatelet therapy for prevention of death, myocardial infarction, and stroke in high risk patients. BMJ. 2002;324(7329):71-86. doi: 10.1136/bmj.324.7329.71 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Yusuf S, Sleight P, Pogue J, Bosch J, Davies R, Dagenais G; Heart Outcomes Prevention Evaluation Study Investigators . Effects of an angiotensin-converting-enzyme inhibitor, ramipril, on cardiovascular events in high-risk patients. N Engl J Med. 2000;342(3):145-153. [DOI] [PubMed] [Google Scholar]
- 7.PROGRESS Collaborative Group Randomised trial of a perindopril-based blood-pressure-lowering regimen among 6,105 individuals with previous stroke or transient ischaemic attack. Lancet. 2001;358(9287):1033-1041. doi: 10.1016/S0140-6736(01)06178-5 [DOI] [PubMed] [Google Scholar]
- 8.Yusuf S, Diener HC, Sacco RL, et al. ; PRoFESS Study Group . Telmisartan to prevent recurrent stroke and cardiovascular events. N Engl J Med. 2008;359(12):1225-1237. doi: 10.1056/NEJMoa0804593 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Yusuf S, Teo KK, Pogue J, et al. ; ONTARGET Investigators . Telmisartan, ramipril, or both in patients at high risk for vascular events. N Engl J Med. 2008;358(15):1547-1559. doi: 10.1056/NEJMoa0801317 [DOI] [PubMed] [Google Scholar]
- 10.Yusuf S, Islam S, Chow CK, et al. ; Prospective Urban Rural Epidemiology (PURE) Study Investigators . Use of secondary prevention drugs for cardiovascular disease in the community in high-income, middle-income, and low-income countries (the PURE Study): a prospective epidemiological survey. Lancet. 2011;378(9798):1231-1243. doi: 10.1016/S0140-6736(11)61215-4 [DOI] [PubMed] [Google Scholar]
- 11.Berwanger O, Piva e Mattos LA, Martin JF, et al. Evidence-based therapy prescription in high-cardiovascular risk patients: the REACT study. Arq Bras Cardiol. 2013;100(3):212-220. doi: 10.5935/abc.20130062 [DOI] [PubMed] [Google Scholar]
- 12.Grimshaw JM, Shirran L, Thomas R, et al. Changing provider behavior: an overview of systematic reviews of interventions. Med Care. 2001;39(8, suppl 2):II2-II45. [PubMed] [Google Scholar]
- 13.Rowe AK, Rowe SY, Peters DH, Holloway KA, Chalker J, Ross-Degnan D. Effectiveness of strategies to improve health-care provider practices in low-income and middle-income countries: a systematic review. Lancet Glob Health. 2018;6(11):e1163-e1175. doi: 10.1016/S2214-109X(18)30398-X [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Viswanathan M, Golin CE, Jones CD, et al. Interventions to improve adherence to self-administered medications for chronic diseases in the United States: a systematic review. Ann Intern Med. 2012;157(11):785-795. doi: 10.7326/0003-4819-157-11-201212040-00538 [DOI] [PubMed] [Google Scholar]
- 15.Grimshaw J, Eccles M, Thomas R, et al. Toward evidence-based quality improvement: evidence (and its limitations) of the effectiveness of guideline dissemination and implementation strategies 1966-1998. J Gen Intern Med. 2006;21(suppl 2):S14-S20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Machline-Carrion MJ, Soares RM, Damiani LP, et al. Rationale and design for a cluster randomized quality-improvement trial to increase the uptake of evidence-based therapies for patients at high cardiovascular risk: the BRIDGE-Cardiovascular Prevention trial. Am Heart J. 2019;207:40-48. doi: 10.1016/j.ahj.2018.10.001 [DOI] [PubMed] [Google Scholar]
- 17.Benjamini Y, Hochbeg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Series B Stat Methodol. 1995;57(1):289-300. [Google Scholar]
- 18.Peiris D, Usherwood T, Panaretto K, et al. Effect of a computer-guided, quality improvement program for cardiovascular disease risk management in primary health care: the treatment of cardiovascular risk using electronic decision support cluster-randomized trial. Circ Cardiovasc Qual Outcomes. 2015;8(1):87-95. doi: 10.1161/CIRCOUTCOMES.114.001235 [DOI] [PubMed] [Google Scholar]
- 19.Liddy C, Hogg W, Singh J, et al. A real-world stepped wedge cluster randomized trial of practice facilitation to improve cardiovascular care. Implement Sci. 2015;10:150. doi: 10.1186/s13012-015-0341-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Shelley DR, Ogedegbe G, Anane S, et al. Testing the use of practice facilitation in a cluster randomized stepped-wedge design trial to improve adherence to cardiovascular disease prevention guidelines: HealthyHearts NYC. Implement Sci. 2016;11(1):88. doi: 10.1186/s13012-016-0450-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Wei X, Zou G, Gong W, et al. Cardiovascular disease risk reduction in rural China: a clustered randomized controlled trial in Zhejiang. Trials. 2013;14:354. doi: 10.1186/1745-6215-14-354 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Shah S, Singh K, Ali MK, et al. ; CARRS Trial Writing Group . Improving diabetes care: multi-component cardiovascular disease risk reduction strategies for people with diabetes in South Asia—the CARRS multi-center translation trial. Diabetes Res Clin Pract. 2012;98(2):285-294. doi: 10.1016/j.diabres.2012.09.023 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Tricco AC, Ivers NM, Grimshaw JM, et al. Effectiveness of quality improvement strategies on the management of diabetes: a systematic review and meta-analysis. Lancet. 2012;379(9833):2252-2261. doi: 10.1016/S0140-6736(12)60480-2 [DOI] [PubMed] [Google Scholar]
- 24.Yusuf S. Two decades of progress in preventing vascular disease. Lancet. 2002;360(9326):2-3. doi: 10.1016/S0140-6736(02)09358-3 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Trial protocol.
eAppendix 1. The Brazilian health care system.
eAppendix 2. Eligibility criteria.
eAppendix 3. BRIDGE CV quality improvement intervention.
eAppendix 4. Details of the baseline survey.
eAppendix 5. Statistical methods details.
eTable 1. Prescription of evidence-based therapies in eligible patients from BRIGDE Cardiovascular Prevention baseline survey and REACT Registry.
eTable 2. Overall comparison between baseline characteristics from all included and excluded clusters.
eTable 3. Adherence to the multifaceted quality improvement intervention.
eTable 4. Effects of a quality improvement intervention on prescription of evidence-based therapies in eligible patients and risk factors control on admission.
eTable 5. Effects of a quality improvement intervention on prescription of evidence-based therapies in eligible patients, major cardiovascular events, and risk factors control at 6 months follow-up.
eTable 6. Post hoc sensitivity analysis for the primary outcome considering multiple imputations.
Data sharing statement.