- Split View
-
Views
-
Cite
Cite
Maria Letícia L. Lana, Andrea Z. Beaton, Luisa C. C. Brant, Isadora C. R. S. Bozzi, Osias de Magalhães, Luiz Ricardo de A. Castro, Francisco César T. da Silva Júnior, José Luiz P. da Silva, Antonio Luiz P. Ribeiro, Bruno R. Nascimento, Factors associated with compliance to AHA/ACC performance measures in a myocardial infarction system of care in Brazil, International Journal for Quality in Health Care, Volume 29, Issue 4, August 2017, Pages 499–506, https://doi.org/10.1093/intqhc/mzx059
- Share Icon Share
Abstract
To evaluate compliance with American Heart Association/American College of Cardiology (AHA/ACC) performance measures for adults with acute myocardial infarction (AMI) and to investigate the factors associated with compliance, in an AMI System of Care in Brazil.
Observational longitudinal study.
A high-complexity University Hospital, part of the AMI System of Care implemented in Belo Horizonte, Brazil, in 2010.
Of note, 1129 patients with ST-elevation myocardial infarction (STEMI) and non-ST-elevation myocardial infarction (NSTEMI) admitted to a single center over 36 months (between 2011 and 2014).
Compliance with 13 pre-specified AHA/ACC AMI performance measures was evaluated for patients with AMI, observing exclusion criteria and appropriate numerators and denominators. Median compliance was calculated and variables independently associated with compliance rates were evaluated.
Median age was 60 (51/68) years, 67.7% male, 69.8% presented with STEMI and hospital mortality was 8.7%. Median compliance with performance measures was 83% (75/88). Among patients with STEMI, 56% received reperfusion therapy. Overall, 67.3% of patients complied with ≥80% of quality measures. Factors independently associated with better compliance were later date of presentation (semester), likely reflecting ongoing training (OR = 1.19, 95% CI: 1.10–1.28, P < 0.001), male gender (OR = 1.33, 95% CI: 1.00–1.76, P < 0.046), Killip I/II on admission (OR = 1.95, 95% CI: 1.36–2.80, P < 0.001) and diagnosis of NSTEMI (OR = 5.0, 95% CI: 3.51–7.11, P < 0.001).
Compliance with AHA/ACC AMI performance measures remains below target in Brazil, but the time trends observed suggest improvement. Continuing education, reduction of system delays and prioritizing high-risk groups are needed to optimize AMI systems of care and improve patient outcomes.
Introduction
Quality metrics, derived from evidence-based practice guidelines, are emerging as an integral component of contemporary healthcare. Patients, providers, administrators and payers are all increasingly seeking out quality metrics to inform their healthcare decisions. In the USA, adherence to these metrics is being tied to institutional rankings and reimbursement structures. Globally, adoption of quality metrics has been slower, with resultant lag in improvement of patient outcomes.
Acute myocardial infarction (AMI) is an exemplary target for global implementation of quality metrics, as ideal care for patients with AMI is well defined by a large number of randomized control trials. The American Heart Association (AHA) and the American College of Cardiology (ACC) have complied these data into a summary statement outlining best practices, aiming to develop detailed performance measures than can be used to improve care [1], focusing on areas with the most potential for impact, along with the strongest consensus about best practice. Prospective observational studies have demonstrated a strong association between institutional adherence to these recommendations and improved patient outcomes in the USA and on a small scale in Brazil [2–4].
Improving outcomes for AMI in middle-income countries, such as Brazil, could contribute substantially toward the World Health Organization goal of, ‘A 25% relative reduction in the overall mortality from cardiovascular diseases… by the year 2025 [5]’. In 2015, Brazilians suffered 15.1% mortality from AMI [6], far exceeding AMI death percentages in the USA (10.1% in 2006) [7]. Inconsistent adherence to AMI standards of care likely contributes to this discrepancy [3, 4].
In 2010, the Federal University of Minas Gerais (UFMG) implemented a standardized AMI system of care [2] in-line with the AHA/ACC consensus statement [1], with short-term improvement in mortality [2]. The objective of this investigation was to evaluate compliance at UFMG with AHA/ACC AMI performance measures [1] and to investigate factors associated with compliance.
Methods
The AMI system of care was implemented in 2010 in the metropolitan area of Belo Horizonte, a large city in southeast Brazil, with 2.5 million inhabitants [8]. The system of care, overall characteristics and initial results are described elsewhere [2]. The intervention consisted of continuing education in pre-hospital and hospital facilities, improvement of infrastructure, public emergency transportation (SAMU) and availability of drug therapy, along with definition and tracking of quality metrics. Initially, two high-complexity hospitals were equipped for 24 h availability of interventional cardiology facilities and coronary intensive care unit (CICU) beds for immediate referral of urgent cases, and a third one was incorporated 2 years after. The flowchart of the AMI system of care is depicted in Fig. 1.
Inclusion criteria
This is an observational longitudinal study that enrolled consecutive patients admitted to the CICU of the Hospital das Clínicas, Universidade Federal de Minas Gerais (HC-UFMG)—one of the three percutaneous coronary intervention (PCI)-capable units of the system of care—with the diagnosis of AMI: ST-elevation myocardial infarction (STEMI) and non-ST-elevation myocardial infarction (NSTEMI) from December 2011 to December 2014. Patients aged ≥18 years who were diagnosed with out-of-hospital AMI were consecutively included, regardless of the therapy or reperfusion strategy applied. Patients who presented more than once to HC-UFMG with AMI were only included in their index admission. Counseling measures were reinforced by written educational material and referrals (e.g. rehabilitation) were considered according to attendance.
Exclusion criteria
Patients with final diagnoses other than AMI at the initial presentation or after hospital complimentary investigations.
Ethics
Approval was obtained from the institutional review board of UFMG. This study complied with the REporting of Studies Conducted using Observational Routinely-collected health Data (RECORD) Statement [9].
Data collection
All clinical and demographic variables and data from complimentary tests (ECG, functional tests, echocardiography and coronary angiography), interventions and treatments were systematically collected in an online dedicated database (Core Ware®, São Paulo—SP, Brazil; www.coreware.com.br). The attending physician of the CICU collected baseline data and pre-hospital information from the referral facility or emergency department (ED) on admission. Medical students, under direct supervision of three investigators, were then responsible for collecting the remainder of data after patient's discharge or death, including in-hospital clinical outcomes. Two independent investigators routinely checked data quality.
We prospectively recorded 13 Class I AHA/ACC performance measures [1] (Appendix 1). Data about patient eligibility for each measure were systematically collected according to defined AHA/ACC guideline indications and reported contraindications for treatments and procedures. Patients who died before arrival at HC-UFMG were not considered for analysis and those dying anytime during their hospital stay were excluded from discharge care assessment. Patient composite adherence scores (compliance rates) were calculated as the proportion (%) of quality metrics achieved, provided the patient's total number of eligible measures. When the AHA/ACC statement did not pre-specify ideal cut-offs for system delays, we considered the recommendations of current guidelines [10]. The nine test measures included in the 2008 guidelines were not included in this analysis for being relatively specific for certain situations (e.g. excess dosing of antiplatelets and anticoagulants) and less representative of the global improvement of healthcare processes.
Demographic variables (age, sex, race and location of initial presentation) clinical variables (family history of coronary artery disease (CAD), hypertension, diabetes mellitus, current/recent smoker, hypercholesterolemia, prior MI, prior PCI, prior coronary artery bypass graft (CABG), prior congestive heart failure, prior stroke) and pre-hospital and in-hospital variables (date of presentation, diagnosis, Killip class, signs of congestive heart failure, blood pressure and heart rate) were considered for this analysis.
Statistical analysis
Statistical analysis was performed using SPSS software version 22.0 for Mac OSX (SPSS Inc., IBM©, Chicago, IL, USA). Descriptive analysis of demographic and clinical variables, in addition to the AHA/ACC performance measures, was performed for groups (STEMI and NSTEMI). The distribution pattern of variables was assessed with the Shapiro–Wilk test. Continuous variables were expressed as mean ± standard deviation or median and quartiles 25%/75% (Q1/Q3) when appropriate. Categorical variables were expressed as absolute values and percentages.
The between-group comparison was performed using the Student t-test for continuous variables with normal distribution and the Mann–Whitney U-test for those with non-normal distribution. The comparison of categorical variables was performed using the Fisher's exact test and Chi-square Pearson's test. The independent variable for the evaluation of compliance with AHA/ACC performance measures was the percent compliance (≥80% of the applicable measures) for each patient, observing exclusion criteria [1]. Univariable and multivariable logistic regression models were adjusted including plausible variables: demographics, clinical presentation, risk factors, past medical history (diagnoses/interventions), Killip class on admission (I/II or III/IV), location of initial presentation [(i) HC-UFMG, (ii) Pre-hospital emergency units of Belo Horizonte, (iii) other hospitals or health centers in Belo Horizonte, (iv) cities in the metropolitan region of Belo Horizonte and (v) other cities], diagnosis (STEMI/NSTEMI) and date of presentation (semester) to the AMI system of care. When necessary, transformations were done for analysis of variance. A two-tailed significance level of 0.05 was considered statistically significant.
Results
During the 36-month enrollment period, 1258 patients with acute coronary syndromes were admitted to HC-UFMG; 129 were excluded due to the diagnosis of unstable angina. As such, 1129 patients were included, with a median age of 60 (51/68) years, 67.7% male. Of these, 69.8% had a diagnosis of STEMI and 30.2% of NSTEMI; overall hospital mortality was 8.7% (9.1% STEMI vs. 7.6% NSTEMI, P = 0.49). Respectively, 15.6% and 10.2% of the patients were admitted with Killip Classes III and IV. Detailed demographic and clinical characteristics of the sample are depicted in Table 1.
Variable . | OVERALL (N = 1129) . | STEMI . | NSTEMI . | P-value . |
---|---|---|---|---|
N = 788 (69.8%) . | N = 341 (30.2%) . | |||
Demographic characteristics | ||||
Age (years, median, Q1/Q3) | 60 (51/68) | 59 (51/67) | 62 (54/72) | <0.001 |
Sex (N, %) | ||||
Male | 765 (32.2) | 544 (69) | 221(64.8) | 0.17 |
Female | 354 (67.8) | 244 (31) | 120 (35.2) | |
Race (N, %) | ||||
White | 444 (39.3) | 300 (38.1) | 144 (42.2) | 0.56 |
Asiatic | 10 (0.9) | 6 (0.8) | 4 (1.2) | |
Black | 149 (13.2) | 109 (13.8) | 40 (11.7) | |
Mixed race | 410 (36.3) | 285 (36.2) | 125 (36.7) | |
Not reported | 116 (10.3) | 88 (11.2) | 28 (8.2) | |
Location of initial presentation (N, %) | ||||
HC-UFMG ED | 51 (4.5) | 21 (2.7) | 30 (8.8) | <0.001 |
Pre-hospital emergency units | 267 (23.6) | 193 (24.5) | 74 (21.7) | |
Hospitals in Belo Horizonte | 278 (24.6) | 169 (21.4) | 109 (32) | |
EDs in the metropolitan area of Belo Horizonte | 407 (36.0) | 319 (40.5) | 88 (25.8) | |
EDs in other cities | 104 (9.2) | 74 (9.4) | 30 (8.8) | |
Not reported | 22 (1.9) | 12 (1.5) | 10 (2.9) | |
Clinical characteristics | ||||
Killip class on admission (N, %) | ||||
Killip I | 729 (64.6) | 490 (62.2) | 239 (70.1) | <0.001 |
Killip II | 242 (21.4) | 175 (22.2) | 67 (19.6) | |
Killip III | 64 (5.7) | 39 (4.9) | 25 (7.3) | |
Killip IV | 94 (8.3) | 84 (10.7) | 10 (2.9) | |
Past medical history (N, %) | ||||
Hypertension | 796 (70.6) | 523 (66.4) | 273 (80.1) | <0.001 |
Diabetes mellitus | 297 (26.3) | 190 (24.1) | 107 (31.4) | 0.01 |
Dyslipidemia | 541 (48.0) | 347 (44) | 194 (56.9) | <0.001 |
Smoker | 438 (38.8) | 344 (43.7) | 94 (27.6) | <0.001 |
Family history of CAD | 272 (24.1) | 186 (23.6) | 86 (25.2) | 0.60 |
MI | 145 (12.8) | 59 (7.5) | 86 (25.2) | <0.001 |
PCI | 37 (3.3) | 11 (1.4) | 26 (7.6) | <0.001 |
CABG | 18 (1.6) | 8 (1.0) | 10 (2.9) | 0.03 |
CVD | 57 (5.0) | 42 (5.3) | 15 (4.4) | 0.56 |
Variable . | OVERALL (N = 1129) . | STEMI . | NSTEMI . | P-value . |
---|---|---|---|---|
N = 788 (69.8%) . | N = 341 (30.2%) . | |||
Demographic characteristics | ||||
Age (years, median, Q1/Q3) | 60 (51/68) | 59 (51/67) | 62 (54/72) | <0.001 |
Sex (N, %) | ||||
Male | 765 (32.2) | 544 (69) | 221(64.8) | 0.17 |
Female | 354 (67.8) | 244 (31) | 120 (35.2) | |
Race (N, %) | ||||
White | 444 (39.3) | 300 (38.1) | 144 (42.2) | 0.56 |
Asiatic | 10 (0.9) | 6 (0.8) | 4 (1.2) | |
Black | 149 (13.2) | 109 (13.8) | 40 (11.7) | |
Mixed race | 410 (36.3) | 285 (36.2) | 125 (36.7) | |
Not reported | 116 (10.3) | 88 (11.2) | 28 (8.2) | |
Location of initial presentation (N, %) | ||||
HC-UFMG ED | 51 (4.5) | 21 (2.7) | 30 (8.8) | <0.001 |
Pre-hospital emergency units | 267 (23.6) | 193 (24.5) | 74 (21.7) | |
Hospitals in Belo Horizonte | 278 (24.6) | 169 (21.4) | 109 (32) | |
EDs in the metropolitan area of Belo Horizonte | 407 (36.0) | 319 (40.5) | 88 (25.8) | |
EDs in other cities | 104 (9.2) | 74 (9.4) | 30 (8.8) | |
Not reported | 22 (1.9) | 12 (1.5) | 10 (2.9) | |
Clinical characteristics | ||||
Killip class on admission (N, %) | ||||
Killip I | 729 (64.6) | 490 (62.2) | 239 (70.1) | <0.001 |
Killip II | 242 (21.4) | 175 (22.2) | 67 (19.6) | |
Killip III | 64 (5.7) | 39 (4.9) | 25 (7.3) | |
Killip IV | 94 (8.3) | 84 (10.7) | 10 (2.9) | |
Past medical history (N, %) | ||||
Hypertension | 796 (70.6) | 523 (66.4) | 273 (80.1) | <0.001 |
Diabetes mellitus | 297 (26.3) | 190 (24.1) | 107 (31.4) | 0.01 |
Dyslipidemia | 541 (48.0) | 347 (44) | 194 (56.9) | <0.001 |
Smoker | 438 (38.8) | 344 (43.7) | 94 (27.6) | <0.001 |
Family history of CAD | 272 (24.1) | 186 (23.6) | 86 (25.2) | 0.60 |
MI | 145 (12.8) | 59 (7.5) | 86 (25.2) | <0.001 |
PCI | 37 (3.3) | 11 (1.4) | 26 (7.6) | <0.001 |
CABG | 18 (1.6) | 8 (1.0) | 10 (2.9) | 0.03 |
CVD | 57 (5.0) | 42 (5.3) | 15 (4.4) | 0.56 |
CVD, cerebrovascular disease; Q1/Q3, quartiles 25/75.
Variable . | OVERALL (N = 1129) . | STEMI . | NSTEMI . | P-value . |
---|---|---|---|---|
N = 788 (69.8%) . | N = 341 (30.2%) . | |||
Demographic characteristics | ||||
Age (years, median, Q1/Q3) | 60 (51/68) | 59 (51/67) | 62 (54/72) | <0.001 |
Sex (N, %) | ||||
Male | 765 (32.2) | 544 (69) | 221(64.8) | 0.17 |
Female | 354 (67.8) | 244 (31) | 120 (35.2) | |
Race (N, %) | ||||
White | 444 (39.3) | 300 (38.1) | 144 (42.2) | 0.56 |
Asiatic | 10 (0.9) | 6 (0.8) | 4 (1.2) | |
Black | 149 (13.2) | 109 (13.8) | 40 (11.7) | |
Mixed race | 410 (36.3) | 285 (36.2) | 125 (36.7) | |
Not reported | 116 (10.3) | 88 (11.2) | 28 (8.2) | |
Location of initial presentation (N, %) | ||||
HC-UFMG ED | 51 (4.5) | 21 (2.7) | 30 (8.8) | <0.001 |
Pre-hospital emergency units | 267 (23.6) | 193 (24.5) | 74 (21.7) | |
Hospitals in Belo Horizonte | 278 (24.6) | 169 (21.4) | 109 (32) | |
EDs in the metropolitan area of Belo Horizonte | 407 (36.0) | 319 (40.5) | 88 (25.8) | |
EDs in other cities | 104 (9.2) | 74 (9.4) | 30 (8.8) | |
Not reported | 22 (1.9) | 12 (1.5) | 10 (2.9) | |
Clinical characteristics | ||||
Killip class on admission (N, %) | ||||
Killip I | 729 (64.6) | 490 (62.2) | 239 (70.1) | <0.001 |
Killip II | 242 (21.4) | 175 (22.2) | 67 (19.6) | |
Killip III | 64 (5.7) | 39 (4.9) | 25 (7.3) | |
Killip IV | 94 (8.3) | 84 (10.7) | 10 (2.9) | |
Past medical history (N, %) | ||||
Hypertension | 796 (70.6) | 523 (66.4) | 273 (80.1) | <0.001 |
Diabetes mellitus | 297 (26.3) | 190 (24.1) | 107 (31.4) | 0.01 |
Dyslipidemia | 541 (48.0) | 347 (44) | 194 (56.9) | <0.001 |
Smoker | 438 (38.8) | 344 (43.7) | 94 (27.6) | <0.001 |
Family history of CAD | 272 (24.1) | 186 (23.6) | 86 (25.2) | 0.60 |
MI | 145 (12.8) | 59 (7.5) | 86 (25.2) | <0.001 |
PCI | 37 (3.3) | 11 (1.4) | 26 (7.6) | <0.001 |
CABG | 18 (1.6) | 8 (1.0) | 10 (2.9) | 0.03 |
CVD | 57 (5.0) | 42 (5.3) | 15 (4.4) | 0.56 |
Variable . | OVERALL (N = 1129) . | STEMI . | NSTEMI . | P-value . |
---|---|---|---|---|
N = 788 (69.8%) . | N = 341 (30.2%) . | |||
Demographic characteristics | ||||
Age (years, median, Q1/Q3) | 60 (51/68) | 59 (51/67) | 62 (54/72) | <0.001 |
Sex (N, %) | ||||
Male | 765 (32.2) | 544 (69) | 221(64.8) | 0.17 |
Female | 354 (67.8) | 244 (31) | 120 (35.2) | |
Race (N, %) | ||||
White | 444 (39.3) | 300 (38.1) | 144 (42.2) | 0.56 |
Asiatic | 10 (0.9) | 6 (0.8) | 4 (1.2) | |
Black | 149 (13.2) | 109 (13.8) | 40 (11.7) | |
Mixed race | 410 (36.3) | 285 (36.2) | 125 (36.7) | |
Not reported | 116 (10.3) | 88 (11.2) | 28 (8.2) | |
Location of initial presentation (N, %) | ||||
HC-UFMG ED | 51 (4.5) | 21 (2.7) | 30 (8.8) | <0.001 |
Pre-hospital emergency units | 267 (23.6) | 193 (24.5) | 74 (21.7) | |
Hospitals in Belo Horizonte | 278 (24.6) | 169 (21.4) | 109 (32) | |
EDs in the metropolitan area of Belo Horizonte | 407 (36.0) | 319 (40.5) | 88 (25.8) | |
EDs in other cities | 104 (9.2) | 74 (9.4) | 30 (8.8) | |
Not reported | 22 (1.9) | 12 (1.5) | 10 (2.9) | |
Clinical characteristics | ||||
Killip class on admission (N, %) | ||||
Killip I | 729 (64.6) | 490 (62.2) | 239 (70.1) | <0.001 |
Killip II | 242 (21.4) | 175 (22.2) | 67 (19.6) | |
Killip III | 64 (5.7) | 39 (4.9) | 25 (7.3) | |
Killip IV | 94 (8.3) | 84 (10.7) | 10 (2.9) | |
Past medical history (N, %) | ||||
Hypertension | 796 (70.6) | 523 (66.4) | 273 (80.1) | <0.001 |
Diabetes mellitus | 297 (26.3) | 190 (24.1) | 107 (31.4) | 0.01 |
Dyslipidemia | 541 (48.0) | 347 (44) | 194 (56.9) | <0.001 |
Smoker | 438 (38.8) | 344 (43.7) | 94 (27.6) | <0.001 |
Family history of CAD | 272 (24.1) | 186 (23.6) | 86 (25.2) | 0.60 |
MI | 145 (12.8) | 59 (7.5) | 86 (25.2) | <0.001 |
PCI | 37 (3.3) | 11 (1.4) | 26 (7.6) | <0.001 |
CABG | 18 (1.6) | 8 (1.0) | 10 (2.9) | 0.03 |
CVD | 57 (5.0) | 42 (5.3) | 15 (4.4) | 0.56 |
CVD, cerebrovascular disease; Q1/Q3, quartiles 25/75.
Overall, there was a median 83% (75/88) compliance with the applicable AHA/ACC performance measures, being 82% (72/88) for STEMI and 86% (83/100) for NSTEMI (P < 0.001). Composite compliance of ≥80% was achieved in only 67.3% of patients (58.8% STEMI vs. 87.1% NSTEMI, P < 0.001). Table 2 shows the compliance with each of the 13 performance measures: the best rates were observed for those related to drug therapy and counseling and the worst for metrics of system delays and access to reperfusion therapy. In the last semester, median overall compliance was 85% (80/88), with the highest rates of the composite compliance (≥80%): 76.3% (69.1% STEMI and 96.1% NSTEMI).
Performance measure (N, %) . | STEMI . | NSTEMI . | |
---|---|---|---|
N = 788 (69.8%) . | N = 341 (30.2%) . | ||
1. | Aspirin at arrival | 788 (100) | 341 (100) |
2. | Aspirin prescribed at discharge | ||
Yes | 675 (99.1) | 298 (98.3) | |
No | 6 (0.9) | 5 (1.7) | |
N/A | 107 | 38 | |
3. | Beta-blocker prescribed at discharge | ||
Yes | 642 (95.4) | 277 (90.2) | |
No | 31 (4.6) | 30 (9.8) | |
N/A | 115 | 34 | |
4. | Statin prescribed at discharge | ||
Yes | 702 (99.0) | 313 (99.4) | |
No | 7 (1.0) | 2 (0.6) | |
N/A | 79 | 26 | |
5. | Evaluation of LVSF (%) | ||
Yes | 660 (83.8) | 258 (75.7) | |
No | 128 (16.2) | 83 (24.3) | |
6. | ACEi or ARB for LVSD (LVEF < 40%) | ||
Yes | 91 (90.1) | 23 (82.1) | |
No | 10 (9.9) | 5 (17.9) | |
N/A | 687 | 313 | |
7. | Time from hospital arrival to fibrinolysis ≤30 min | ||
Yes | 36 (13.6) | ||
No | 228 (86.4) | ||
N/A | 524 | 341 | |
8. | Time from hospital arrival to primary PCI ≤ 90 min | ||
Yes | 128 (75.3) | ||
No | 42 (24.7) | ||
N/A | 618 | 341 | |
9. | Reperfusion therapy | ||
Yes | 400 (56.0) | ||
No | 314 (43.1) | ||
N/A | 74 | 341 | |
10. | Time from ED arrival at STEMI referral facility to arrival at a STEMI-receiving facility ≤60 min | ||
Yes | 10 (5.2) | ||
No | 181 (94.8) | ||
N/A | 597 | 341 | |
11. | Time from patient arrival at a STEMI referral facility's ED to time of primary PCI ≤120 min | ||
Yes | 7 (4.2) | ||
No | 159 (95.8) | ||
N/A | 622 | 341 | |
12. | Cardiac rehabilitation patient referral | ||
Yes | 551 (77.2) | 213 (67.6) | |
No | 163 (22.8) | 102 (32.4) | |
N/A | |||
13. | Adult smoking cessation advice/counseling | ||
Yes | 311 (98.1) | 88 (98.9) | |
No | 6 (1.9) | 1 (1.1) | |
N/A | 471 | 252 |
Performance measure (N, %) . | STEMI . | NSTEMI . | |
---|---|---|---|
N = 788 (69.8%) . | N = 341 (30.2%) . | ||
1. | Aspirin at arrival | 788 (100) | 341 (100) |
2. | Aspirin prescribed at discharge | ||
Yes | 675 (99.1) | 298 (98.3) | |
No | 6 (0.9) | 5 (1.7) | |
N/A | 107 | 38 | |
3. | Beta-blocker prescribed at discharge | ||
Yes | 642 (95.4) | 277 (90.2) | |
No | 31 (4.6) | 30 (9.8) | |
N/A | 115 | 34 | |
4. | Statin prescribed at discharge | ||
Yes | 702 (99.0) | 313 (99.4) | |
No | 7 (1.0) | 2 (0.6) | |
N/A | 79 | 26 | |
5. | Evaluation of LVSF (%) | ||
Yes | 660 (83.8) | 258 (75.7) | |
No | 128 (16.2) | 83 (24.3) | |
6. | ACEi or ARB for LVSD (LVEF < 40%) | ||
Yes | 91 (90.1) | 23 (82.1) | |
No | 10 (9.9) | 5 (17.9) | |
N/A | 687 | 313 | |
7. | Time from hospital arrival to fibrinolysis ≤30 min | ||
Yes | 36 (13.6) | ||
No | 228 (86.4) | ||
N/A | 524 | 341 | |
8. | Time from hospital arrival to primary PCI ≤ 90 min | ||
Yes | 128 (75.3) | ||
No | 42 (24.7) | ||
N/A | 618 | 341 | |
9. | Reperfusion therapy | ||
Yes | 400 (56.0) | ||
No | 314 (43.1) | ||
N/A | 74 | 341 | |
10. | Time from ED arrival at STEMI referral facility to arrival at a STEMI-receiving facility ≤60 min | ||
Yes | 10 (5.2) | ||
No | 181 (94.8) | ||
N/A | 597 | 341 | |
11. | Time from patient arrival at a STEMI referral facility's ED to time of primary PCI ≤120 min | ||
Yes | 7 (4.2) | ||
No | 159 (95.8) | ||
N/A | 622 | 341 | |
12. | Cardiac rehabilitation patient referral | ||
Yes | 551 (77.2) | 213 (67.6) | |
No | 163 (22.8) | 102 (32.4) | |
N/A | |||
13. | Adult smoking cessation advice/counseling | ||
Yes | 311 (98.1) | 88 (98.9) | |
No | 6 (1.9) | 1 (1.1) | |
N/A | 471 | 252 |
ACEi, angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker; LVEF, left ventricular ejection fraction; LVSD, left ventricular systolic dysfunction; LVSF, left ventricular systolic function; N/A, not applicable.
Performance measure (N, %) . | STEMI . | NSTEMI . | |
---|---|---|---|
N = 788 (69.8%) . | N = 341 (30.2%) . | ||
1. | Aspirin at arrival | 788 (100) | 341 (100) |
2. | Aspirin prescribed at discharge | ||
Yes | 675 (99.1) | 298 (98.3) | |
No | 6 (0.9) | 5 (1.7) | |
N/A | 107 | 38 | |
3. | Beta-blocker prescribed at discharge | ||
Yes | 642 (95.4) | 277 (90.2) | |
No | 31 (4.6) | 30 (9.8) | |
N/A | 115 | 34 | |
4. | Statin prescribed at discharge | ||
Yes | 702 (99.0) | 313 (99.4) | |
No | 7 (1.0) | 2 (0.6) | |
N/A | 79 | 26 | |
5. | Evaluation of LVSF (%) | ||
Yes | 660 (83.8) | 258 (75.7) | |
No | 128 (16.2) | 83 (24.3) | |
6. | ACEi or ARB for LVSD (LVEF < 40%) | ||
Yes | 91 (90.1) | 23 (82.1) | |
No | 10 (9.9) | 5 (17.9) | |
N/A | 687 | 313 | |
7. | Time from hospital arrival to fibrinolysis ≤30 min | ||
Yes | 36 (13.6) | ||
No | 228 (86.4) | ||
N/A | 524 | 341 | |
8. | Time from hospital arrival to primary PCI ≤ 90 min | ||
Yes | 128 (75.3) | ||
No | 42 (24.7) | ||
N/A | 618 | 341 | |
9. | Reperfusion therapy | ||
Yes | 400 (56.0) | ||
No | 314 (43.1) | ||
N/A | 74 | 341 | |
10. | Time from ED arrival at STEMI referral facility to arrival at a STEMI-receiving facility ≤60 min | ||
Yes | 10 (5.2) | ||
No | 181 (94.8) | ||
N/A | 597 | 341 | |
11. | Time from patient arrival at a STEMI referral facility's ED to time of primary PCI ≤120 min | ||
Yes | 7 (4.2) | ||
No | 159 (95.8) | ||
N/A | 622 | 341 | |
12. | Cardiac rehabilitation patient referral | ||
Yes | 551 (77.2) | 213 (67.6) | |
No | 163 (22.8) | 102 (32.4) | |
N/A | |||
13. | Adult smoking cessation advice/counseling | ||
Yes | 311 (98.1) | 88 (98.9) | |
No | 6 (1.9) | 1 (1.1) | |
N/A | 471 | 252 |
Performance measure (N, %) . | STEMI . | NSTEMI . | |
---|---|---|---|
N = 788 (69.8%) . | N = 341 (30.2%) . | ||
1. | Aspirin at arrival | 788 (100) | 341 (100) |
2. | Aspirin prescribed at discharge | ||
Yes | 675 (99.1) | 298 (98.3) | |
No | 6 (0.9) | 5 (1.7) | |
N/A | 107 | 38 | |
3. | Beta-blocker prescribed at discharge | ||
Yes | 642 (95.4) | 277 (90.2) | |
No | 31 (4.6) | 30 (9.8) | |
N/A | 115 | 34 | |
4. | Statin prescribed at discharge | ||
Yes | 702 (99.0) | 313 (99.4) | |
No | 7 (1.0) | 2 (0.6) | |
N/A | 79 | 26 | |
5. | Evaluation of LVSF (%) | ||
Yes | 660 (83.8) | 258 (75.7) | |
No | 128 (16.2) | 83 (24.3) | |
6. | ACEi or ARB for LVSD (LVEF < 40%) | ||
Yes | 91 (90.1) | 23 (82.1) | |
No | 10 (9.9) | 5 (17.9) | |
N/A | 687 | 313 | |
7. | Time from hospital arrival to fibrinolysis ≤30 min | ||
Yes | 36 (13.6) | ||
No | 228 (86.4) | ||
N/A | 524 | 341 | |
8. | Time from hospital arrival to primary PCI ≤ 90 min | ||
Yes | 128 (75.3) | ||
No | 42 (24.7) | ||
N/A | 618 | 341 | |
9. | Reperfusion therapy | ||
Yes | 400 (56.0) | ||
No | 314 (43.1) | ||
N/A | 74 | 341 | |
10. | Time from ED arrival at STEMI referral facility to arrival at a STEMI-receiving facility ≤60 min | ||
Yes | 10 (5.2) | ||
No | 181 (94.8) | ||
N/A | 597 | 341 | |
11. | Time from patient arrival at a STEMI referral facility's ED to time of primary PCI ≤120 min | ||
Yes | 7 (4.2) | ||
No | 159 (95.8) | ||
N/A | 622 | 341 | |
12. | Cardiac rehabilitation patient referral | ||
Yes | 551 (77.2) | 213 (67.6) | |
No | 163 (22.8) | 102 (32.4) | |
N/A | |||
13. | Adult smoking cessation advice/counseling | ||
Yes | 311 (98.1) | 88 (98.9) | |
No | 6 (1.9) | 1 (1.1) | |
N/A | 471 | 252 |
ACEi, angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker; LVEF, left ventricular ejection fraction; LVSD, left ventricular systolic dysfunction; LVSF, left ventricular systolic function; N/A, not applicable.
Among patients with STEMI, only 56% received any reperfusion therapy; only 13.6% of those receiving thrombolytics had a door-to-needle time ≤30 min. Of patients undergoing primary PCI, 75.3% had a door-to-balloon time ≤90 min. The most frequent reasons for deferral of thrombolysis were delayed presentation (20.5%), lack of medication (13.6%) and non-diagnostic ECG (5.3%), as reported by the ED staff. In 35.6% of the cases, the reason was not adequately reported. We observed considerable delays in ED arrival to thrombolysis (median time 100 min [60/176]), from arrival at STEMI referral facility to arrival at a STEMI-receiving facility (median time 270 min [180/390]), and from arrival at a STEMI referral facility ED to primary PCI (352 min [234/501]) (Table 3).
Time (min) . | Minimum . | Maximum . | Median . | Quartile 1 . | Quartile 3 . |
---|---|---|---|---|---|
From hospital arrival to thrombolysis | 0 | 2424 | 100 | 60 | 176 |
From ED arrival at STEMI referral facility to arrival at a STEMI-receiving facility | 50 | 1020 | 270 | 180 | 390 |
From patient arrival at a STEMI referral facility's ED to time of primary PCI | 80 | 1072 | 352 | 234 | 501 |
From hospital arrival to primary PCI | 10 | 715 | 66 | 40 | 90 |
Time (min) . | Minimum . | Maximum . | Median . | Quartile 1 . | Quartile 3 . |
---|---|---|---|---|---|
From hospital arrival to thrombolysis | 0 | 2424 | 100 | 60 | 176 |
From ED arrival at STEMI referral facility to arrival at a STEMI-receiving facility | 50 | 1020 | 270 | 180 | 390 |
From patient arrival at a STEMI referral facility's ED to time of primary PCI | 80 | 1072 | 352 | 234 | 501 |
From hospital arrival to primary PCI | 10 | 715 | 66 | 40 | 90 |
Time (min) . | Minimum . | Maximum . | Median . | Quartile 1 . | Quartile 3 . |
---|---|---|---|---|---|
From hospital arrival to thrombolysis | 0 | 2424 | 100 | 60 | 176 |
From ED arrival at STEMI referral facility to arrival at a STEMI-receiving facility | 50 | 1020 | 270 | 180 | 390 |
From patient arrival at a STEMI referral facility's ED to time of primary PCI | 80 | 1072 | 352 | 234 | 501 |
From hospital arrival to primary PCI | 10 | 715 | 66 | 40 | 90 |
Time (min) . | Minimum . | Maximum . | Median . | Quartile 1 . | Quartile 3 . |
---|---|---|---|---|---|
From hospital arrival to thrombolysis | 0 | 2424 | 100 | 60 | 176 |
From ED arrival at STEMI referral facility to arrival at a STEMI-receiving facility | 50 | 1020 | 270 | 180 | 390 |
From patient arrival at a STEMI referral facility's ED to time of primary PCI | 80 | 1072 | 352 | 234 | 501 |
From hospital arrival to primary PCI | 10 | 715 | 66 | 40 | 90 |
In univariable analysis, better compliance with AHA/ACC performance measures (≥80% of applicable measures) was associated with male gender (OR: 1.31, 95% CI: 1.01–1.31, P = 0.042), past history of AMI (OR: 1.76, 95% CI: 1.17–2.65, P < 0.001), higher number of angina episodes in the past 24 h (OR: 1.22, 95% CI: 1.03–1.43, P = 0.018), diagnosis of NSTEMI (OR: 4.74, 95% CI: 3.35–6.70, P < 0.001) (Fig. 2), Killip Class I/II on admission (OR: 2.15, 95% CI: 1.52–3.02, P < 0.001) (Fig. 3) and later date (semester) of presentation to the AMI system of care (OR: 1.16, 95% CI: 1.08–1.24, P < 0.001). In the multivariable model, adjusted for covariates, the variables independently associated with better compliance (≥80%) were later date of presentation (semester), male gender, Killip I/II on admission and diagnosis of NSTEMI (Table 4).
Variable . | OR . | 95% CI . | P-value . |
---|---|---|---|
Sex (male) | 1.33 | 1.01–1.76 | 0.046* |
Killip class I/II on admission | 1.95 | 1.36–2.80 | <0.001* |
Diagnosis of NSTEMI | 5.00 | 3.51–7.12 | <0.001* |
Date of presentation to the AMI system of care (semester) | 1.19 | 1.10–1.28 | <0.001* |
Variable . | OR . | 95% CI . | P-value . |
---|---|---|---|
Sex (male) | 1.33 | 1.01–1.76 | 0.046* |
Killip class I/II on admission | 1.95 | 1.36–2.80 | <0.001* |
Diagnosis of NSTEMI | 5.00 | 3.51–7.12 | <0.001* |
Date of presentation to the AMI system of care (semester) | 1.19 | 1.10–1.28 | <0.001* |
CI, confidence interval; OR, odds ratio. *P-value < 0.05.
Variable . | OR . | 95% CI . | P-value . |
---|---|---|---|
Sex (male) | 1.33 | 1.01–1.76 | 0.046* |
Killip class I/II on admission | 1.95 | 1.36–2.80 | <0.001* |
Diagnosis of NSTEMI | 5.00 | 3.51–7.12 | <0.001* |
Date of presentation to the AMI system of care (semester) | 1.19 | 1.10–1.28 | <0.001* |
Variable . | OR . | 95% CI . | P-value . |
---|---|---|---|
Sex (male) | 1.33 | 1.01–1.76 | 0.046* |
Killip class I/II on admission | 1.95 | 1.36–2.80 | <0.001* |
Diagnosis of NSTEMI | 5.00 | 3.51–7.12 | <0.001* |
Date of presentation to the AMI system of care (semester) | 1.19 | 1.10–1.28 | <0.001* |
CI, confidence interval; OR, odds ratio. *P-value < 0.05.
Discussion
In our study sample, overall compliance with AHA/ACC performance measures for adults with AMI [1] was considerably below target. Of particular concern, compliance measures were less applied in high-risk groups (advanced Killip class, diagnosis of STEMI). System delays and access to reperfusion were identified as drivers of poor compliance and remain significant challenges within the AMI system of care of Belo Horizonte, a large city in Brazil. Our data show that later date of presentation to the AMI system of care was associated with the delivery of better quality of care, suggesting that ongoing provider education and improvement of infrastructure may have led to improvement in care.
Our results demonstrate that in-hospital care for AMI was comparable to several national and international studies [1, 3, 11–18]. Low overall performance on metrics was mainly driven by low rates of reperfusion therapy in patients with STEMI (56.0%)—which indeed was very low compared to other centers [12, 13, 16, 19, 20]—and unacceptable system delays [18, 21–25], both of which have been shown to negatively affect outcome [26]. This partially explains the significantly higher compliance rates observed for patients with NSTEMI, for whom reperfusion and system delays metrics do not apply. These results emphasize the complexity of an AMI system of care, in which all components must be integrated to achieve best outcomes. The AMI system of care in Belo Horizonte is a service fully provided and managed by SUS (Sistema Único de Saúde), the Brazilian public health system, from basic pre-hospital EDs—where the majority of patients enter the system—to high-complexity and STEMI-receiving units, as well as post-discharge care. The difficulty in achieving acceptable rates for reperfusion therapy and system delays can be explained by the characteristic of the health facilities and providers of this service.
In high-complexity hospitals, adequate training and availability of resources, including medications and dedicated beds in CICU or coronary inpatient units, enabled the incorporation of evidence-based protocols in line with current guidelines, even though some structural problems still precluded prompt access to imaging modalities (especially nuclear imaging and echocardiography), and systematic referral for rehabilitation. Despite the considerable effort by the local health authorities to increase the number of CICU beds in two high-complexity hospitals, including HC-UFMG, paying a fixed reimbursement for beds to be kept available for patients with STEMI or acute coronary syndromes (ACSs) with unfavorable initial presentation, the number of dedicated beds is still limited, making access to these centers not timely available to all AMI patients.
In the pre-hospital phase, our results may reflect some particular drawbacks of the public health system in Belo Horizonte region, such as the lack of thrombolytics, and other critical AMI medications—especially out of the capitol. Referring facilities are often inadequately educated (largely due to high turnover of medical staff, frequently composed by early career physicians). Therefore, expertise is limited, and suboptimal diagnosis of AMI and uncertainty about the initiation of drug therapy is a continuous challenge. And, when hospital beds are available, there is often delay secondary to inadequate staffing and resources for transfer, a problem that has been previously reported in Brazil [27, 28].
Over the duration of our study, plans were in place to overcome these challenges. The implementation of these actions may have been responsible for the better compliance observed for patients admitted in the later semesters. In the AMI referral units, additional CICU beds were added (in 2014, another high-complexity hospital joined the AMI system of care) and wider availability of fibrin-specific thrombolytics, and adequate dual antiplatelet and anticoagulation regimens were targeted. The triage protocols of the public emergency transportation system (SAMU) within Belo Horizonte were reviewed and STEMI and ACS with clinical instability prioritized. There was also an investment in healthcare worker education and improved access to cardiology care in remote areas. However, municipalities outside of the capitol were not targeted, leading to considerable continued delays.
To deal with the healthcare provider education and training, there has been a substantial investment in a wide-reaching telemedicine system. The Cardiology department and the Telehealth Center at UFMG created a formal cardiovascular educational curriculum and implemented remote ECG interpretation of ECGs transmitted from pre-hospital units to the CICUs [2]. Again, education was not delivered to referral units outside the city of Belo Horizonte, which may have impacted the results.
Besides the diagnosis of NSTEMI and later date presentation to the AMI system of care, lower Killip class and male sex also predicted better compliance to quality metrics. It is possible that patients with higher Killip class required longer stabilization prior to transfer and did not tolerate drug therapy. The converse is also plausible that system delays caused instability, causing a higher Killip class on admission to UFMG. Poorer adherence to quality metrics for women is consistent with previous literature showing longer decision times in the ED, delayed thrombolysis, and less optimal in-hospital drug therapy among women [29–32].
Interestingly, observing the demographics of our sample evidences other particular characteristics of Belo Horizonte's AMI system of care. The proportion of STEMI (69.8%) observed was higher than expected [12, 23], and was probably increased due to the commitment by HC-UFMG and other PCI-capable hospitals to permanently keep available CICU beds primarily for STEMI patients referred for primary, rescue or emergency PCI. Furthermore, the median age of patients is below that of international studies [19, 33], but in accordance with some studies in Brazil [15, 16, 27, 34], illustrating what has been described about the earlier onset of AMI in low- and middle-income countries and its higher impacts on the economically productive ages [35, 36]. The mortality rates for both STEMI and NSTEMI were higher than some international and national studies of public and private facilities [10, 12, 15, 23, 34], but lower than rates in the Public Health System database (DATASUS, Brazilian Health Ministry Data Processing System) [6, 27]. Compared to remote and extremely low resourced areas of the country, Belo Horizonte—a big metropolitan area located in the highest income Brazilian region (Southeast)—is relatively better served with health resources even at baseline, what may account for the lower mortality compared to the whole country.
Despite the indubitable accomplishments of the AMI system of care, the compliance with performance measures remain below target. After 5 years of operation of the AMI system of care in Belo Horizonte, our observations suggest that the improvements related to health infrastructure and administration are more sustainable in the long run, while the effects of education depend on continuing training. With the improvement of performance measures, it is expected that the positive effects on mortality observed in a preliminary analysis of the SUS administrative database [2] will be sustainable and improve in medium and long terms. The implementation of AMI system of care in low- and middle-income countries are challenging, but must be confronted as a way to deliver good quality of care to the increasing numbers of patients presenting with AMI in these countries [5].
Limitations
The generalizability of our data is limited as it comes from a single center experience, limiting the extrapolation to other regions and countries. Similarly, the above-mentioned high turnover of medical staff is difficult to be distinguished from insufficient training to explain some drawbacks of the educational process—a key feature of the system of care. Moreover, considering the complexity of this system, systematic and reliable data collection was only feasible for 13 AHA/ACC performance measures, some of them adapted for the study. In some situations, only the initiation of a measure (e.g. drug therapy, counseling)—and not adherence to it—was recorded. As mid- and long-term follow-up data were not yet available, it was not possible to evaluate the impact of compliance on patient's clinical outcomes. These data will be explored in future publications.
Finally, performance measures and their impacts on clinical endpoints are multifactorial and associated with a number of conditions not addressed by our analytic model (socioeconomic, cultural and geographic issues, population education and awareness, working system of healthcare providers, organization of the health system, etc.).
Conclusion
Compliance with AHA/ACC AMI performance measures remains below target in Brazil—what may significantly impact the outcomes—but the time trends observed suggest improvement. The lowest compliance rates were observed for patients with STEMI—mainly due to considerable system delays/prolonged transport times—and for female patients and those with advanced Killip class at presentation. The better compliance observed in patients with later dates of presentation to the system of care highlights the importance of developing expertise. Continuing education, reduction of system delays, improvement of health infrastructure and prioritization of groups at higher risk are needed to optimize AMI systems of care and improve patient outcomes.
Supplementary material
Supplementary material is available at International Journal for Quality in Health Care online.
Conflicts of interest statement
None declared.
References
Instituto Brasileiro de Geografia e Estatística, 2016. http://cidades.ibge.gov.br/ (27 July 2016, date last accessed).
- myocardial infarction, acute
- myocardial infarction
- non-st elevated myocardial infarction
- st segment elevation myocardial infarction
- reperfusion therapy
- adult
- american heart association
- brazil
- education, continuing
- hospital mortality
- male
- diagnosis
- system of care
- american college of cardiology
- patient-focused outcomes
- performance measures