Changes in treatment and mortality of acute myocardial infarction in Estonian tertiary and secondary care hospitals in 2001 and 2007

In the last decade, Estonia has reported one of the highest rates of mortality due to ischemic heart diseases in Europe [1]. At the same time, as in other East European countries, the health care system in Estonia has undergone considerable changes. Lead by the Estonian Society of Cardiology, much effort has been made to improve the quality of care for acute myocardial infarction (AMI) patients through better application of the diagnosis and treatment guidelines [2–7].

One of the main priorities has been to increase access to percutaneous coronary interventions (PCI) and to enable more ST-segment elevation AMI (STEMI) patients receive reperfusion, including primary PCI. According to a recent study, the rates of primary PCI in Estonia are now comparable to those in such Nordic countries as Norway and Denmark [8].

Previous studies have mainly focused on the overall changes in the treatment and mortality of AMI patients [9–14]. Changes in different types of hospitals with unequal availability of coronary intervention facilities have received little attention. Still, such information is crucial in a country with limited health care resources aiming to provide equal care for all AMI patients.

This study aimed to determine the changes in in-hospital treatment and 30-day mortality and 3-year mortality of AMI patients hospitalized into tertiary and secondary care hospitals in Estonia in 2001 and 2007.

We conducted a retrospective cross-sectional study based on patient records. The formation of the study samples is presented in Figure 1. The list of AMI cases (main diagnosis code I21-I22 according to the International Statistical Classification of Diseases and Related Health Problems 10^th revision [15]) hospitalized from January 1 to December 31, 2001 and 2007, was obtained from the database of the Estonian Health Insurance Fund (EHIF). The Estonian health insurance system is a social insurance relying on the principle of solidarity and of the 1.3 million inhabitants about 95% are insured. Consistency in reporting to the EHIF database and the validity of the data has been established [16].

The EHIF applied the following exclusion criteria for case selection: (1) patients who were not first hospitalized into one of the study hospitals; (2) patients who were re-admitted with AMI within 28 days of the first admission; (3) patients whose length of hospital stay was less than 3 days if they were discharged alive and were not transferred to another hospital, which made the diagnosis of AMI unlikely.

In 2001, according to the EHIF database, 2365 AMI cases were hospitalized during the study period in Estonia. Management of AMI patients was shared among 27 Estonian hospitals with a different number of beds. As we aimed to evaluate the management of AMI patients in hospitals that treat the major proportion of annual AMI cases, the study included 9 hospitals: 2 tertiary PCI-capable (only during working hours) and 7 secondary care hospitals. In the secondary care hospitals the number of annual AMI cases ranged from 7 to 165 cases and the tending physicians were mainly anesthesiologists or internists and in some hospitals also cardiologists. After the application of the exclusion criteria by the EHIF, 1955 cases remained, out of which a random sample of 520 cases (27%) was formed. The sampling was performed in clusters according to the hospitals involved.

In 2007, 3251 AMI cases were hospitalized in Estonia. Due to the changes in the hospital network only 20 hospitals admitted AMI patients in 2007. The study included 16 hospitals that treated the major proportion of annual AMI cases: 2 tertiary 24/7 PCI-capable and 14 secondary care hospitals (one hospital had PCI availability during working hours). The tertiary care hospitals were the same in both time periods. In the secondary care hospitals the number of annual AMI cases ranged from 16 to 267 cases and the tending physicians were mainly anesthesiologists or internists and in some hospitals also cardiologists. After the application of the exclusion criteria by the EHIF, 2862 cases remained, out of which a random sample of 800 cases (28%) was formed. The sampling was performed in clusters according to the hospitals involved.

Out of the random study samples (n = 520 in 2001, n = 800 in 2007), 40 (7.8%) cases in 2001 and 62 (7.8%) cases in 2007 were excluded from the analysis because of selection errors that arose in the forming of the study population by the EHIF (the study sample included cases that were not first hospitalized into one of the study hospitals).

Among the remaining cases, the diagnosis of AMI was confirmed for the following patients: 93.3% in 2001 and 95.3% in 2007 (P = 0.304) in the tertiary care hospitals; 83.5% in 2001 and 91.1% in 2007 (P = 0.003) in the secondary care hospitals. The diagnosis of AMI in both study periods was based on the redefinition document of myocardial infarction published in 2000 [2].

The final study samples included 423 cases in 2001 (210 from the tertiary and 213 from the secondary care hospitals) and 687 cases in 2007 (327 from the tertiary and 360 from the secondary care hospitals).

Mortality

In crude and baseline-adjusted analysis there were no significant differences in the 30-day mortality and 3-year mortality within the two hospital types in 2007 compared to 2001 (Table 4 and Figure 2).

	Year 2001	Year 2007
Mortality	%	%	P	HR (95% CI)	Adjusted HR (95% CI)a
30-day
Tertiary care hospitals	17.6	13.2	0.156	0.71 (0.44-1.15)	0.69 (0.42-1.13)
Secondary care hospitals	20.2	22.5	0.516	1.17 (0.78-1.77)	1.01 (0.66-1.55)
3-year
Tertiary care	35.7	32.7	0.475	0.90 (0.66-1.22)	0.90 (0.65-1.23)
Secondary care hospitals	38.5	45.3	0.113	1.24 (0.94-1.64)	1.03 (0.77-1.37)

CI, confidence interval; HR, hazard ratio

^a Adjusted for age, sex, AMI subtype (STEMI or non-STEMI), diabetes mellitus, arterial hypertension, previous acute myocardial infarction and previous chronic heart failure.

Comparison of mortality between the two hospital types in 2001 showed no significant differences (results not presented). However, in 2007, the 30-day mortality was significantly lower in the tertiary care hospitals than in the secondary care hospitals in crude (HR 0.54; 95% CI 0.36-0.81) and baseline-adjusted analysis (HR 0.61; 95% CI 0.41-0.93). Also after 3 years mortality in the tertiary care hospitals remained lower in crude (HR 0.66; 95% CI 0.51-0.85) and baseline-adjusted analysis (HR 0.73; 95% CI 0.57-0.95).

This country-wide study shows that between 2001 and 2007, the treatment of AMI patients improved in the tertiary and secondary care hospitals in Estonia, although the rates of reperfusion and revascularization increased mainly in the tertiary care setting. At the same time, improvements in the treatment did not translate into significantly lower mortality within the hospital types even after accounting of the differences in the baseline characteristics. Instead, in 2007, a marked mortality gap can be observed between the tertiary and the secondary care hospitals.

The positive trend in AMI treatment during hospitalization has also been reported by previous studies [9–14, 17–19]. The observed improvements in Estonia are probably connected to better access to coronary intervention facilities, the release of new European and Estonian guidelines during the study period, and the launching of several training programs by the Estonian Society of Cardiology [3–7].

Still, in 2007, the mean use of beta-blockers and ACEI/ARB during hospitalization and for outpatient use remained below 83%, while the rates in the secondary care hospitals were even lower. The use of statins in the secondary care hospitals causes even more concern as although the rates increased markedly during the study period, their use is still below 38%. The fact that the quality of care of AMI patients in the secondary care hospitals lags behind that in their tertiary care counterparts may be due to the known slower implementation of guideline-recommended medications in secondary care hospitals [20]. As the study demonstrated, patients in the secondary care hospitals tend to be older and have more co-morbidities, which may influence management decisions.

Better access to invasive coronary care facilities and recent research findings contributed to the wider use of PCI in the tertiary care hospitals. At the same time, in the secondary care hospitals transfer for further cardiac testing and revascularisation into tertiary care hospitals increased. Still, the transfer rates were low compared to previous studies [21, 22]. Although the mean age of transferred patients increased, transferred patients still tended to be younger than the non-transferred patients (data not presented).

While in the tertiary care hospitals reperfusion rates increased and primary PCI became the preferred method of reperfusion for STEMI patients, then in the secondary care hospitals reperfusion rates showed little change. It is possible that STEMI patients are usually transferred before receiving medical reperfusion. Still, a previous study revealed that the reasons for not receiving reperfusion may be unknown in up to 45% of cases [23].

Several studies have demonstrated that closer adherence to published guidelines for AMI management results in improved short-and long-term outcomes and this even despite the growing prevalence of risk factors (older age, history of hyperlipidaemia and diabetes mellitus) at presentation [10–14, 16, 24–26]. Although our study demonstrates a marked improvement in the treatment quality for AMI patients in 2007 compared to 2001, especially in tertiary care hospitals, it fails to show a significant decrease in 30-day and 3-year mortality. As the study sample of the secondary care hospitals demonstrated, this may largely be due to the higher age and cardiovascular risk among the study samples in 2007. For instance, the rates of diabetes had almost doubled in both types of hospitals. In order to further clarify this issue, we performed a sub-analysis to compare the baseline characteristics, quality of care provided, and mortality separately among patients < 75 and ≥ 75 years in 2001 and 2007 in tertiary care hospitals (data not presented). On the basis of this analysis we may hypothesize that in tertiary care hospitals the reason why short-and long-term mortality has not improved despite better overall quality of care is that firstly the improvement is more targeting those younger and healthier and secondly the rate of patients over 75 years with more co-morbidities has increased.

Interesting findings of the study were the differences in short and long-term mortality between the two hospital types in 2007. This can probably be explained by a combined effect of improved management possibilities in the tertiary care hospitals and different patient baseline characteristics in the two hospital types. The patients admitted to the secondary care hospitals are more likely to be older and have a higher cardiovascular risk. Moreover, as the elderly often present with atypical symptoms and have a greater burden of cardiac and non-cardiac co-morbidities, physicians are more reluctant to treat them aggressively, the more so when the outcomes of interventions and surgery may be poorer [27, 28].

Firstly, this study includes only random study samples which may not represent the true patient populations. As a consequence, the practice patterns observed in the present study may not be representative of the general AMI patient population. Secondly, the sample of 2007 included patients from seven secondary care hospitals that were not included in 2001 (accounting for 25% of the cases from the secondary care hospitals). This corresponds to the study principle of including hospitals that treat the major proportion of annual AMI cases. Limiting the analysis to only those secondary care hospitals that participated both in 2001 and 2007 produced similar treatment and mortality results (data not presented). Thirdly, the study did not capture information on co-morbid conditions such as chronic kidney disease and cancer, which may have precluded the use of certain pharmacologic and interventional treatments. Fourthly, owing to the study design, we were not fully able to collect information on smoking and lipid profiles and hence we could not account for these variables in mortality analysis.

In Estonia, both types of hospitals underwent considerable improvement in AMI treatment between 2001 and 2007. Still, changes were more pronounced in the tertiary care setting, especially with respect to the use of reperfusion and revascularisation. Low rates of reperfusion and transfer for PCI clearly need to be addressed in the secondary care hospitals. The improvement in treatment did not result in a significant change in short-and long-term mortality in either hospital type. Higher age and cardiovascular risk are posing a challenge for AMI management.

In addition to addressing the gaps in the quality of AMI treatment across Estonian hospitals, there is a need for high-quality data from a prospective nationwide acute myocardial infarction register for continuous follow-up of AMI treatment and outcomes.

MB, TA, TM, and JE are the members of the working group of Acute Coronary Syndromes of the Estonian Society of Cardiology as well as the members of the Scientific Board of the Estonian Myocardial Infarction Registry.