Trends in Acute Myocardial Infarction in 4 US States Between 1992 and 2001
Clinical Characteristics, Quality of Care, and Outcomes
Background— Because of the health impact of acute myocardial infarction (AMI), substantial resources have been dedicated to improving AMI care and outcomes. Long-term trends in the clinical characteristics, quality of care, and outcomes for AMI over time from the health system perspective in geographically diverse populations are not well known.
Methods and Results— The present study included 20 550 Medicare patients aged ≥65 years hospitalized in 4 US states (Alabama, Connecticut, Iowa, Wisconsin) with the confirmed primary discharge diagnosis of AMI in 4 periods: 1992–1993 (n=10 292), 1995 (n=5566), 1998–1999 (n=2413), and 2000–2001 (n=2279). With the use of standard quality indicator definitions, treatment of ideal candidates with aspirin and β-blockers within 24 hours after presentation, β-blockers, and angiotensin-converting enzyme inhibitors at discharge was assessed. Multivariable models were constructed to calculate adjusted 1-year mortality. The hospitalized Medicare population with AMI changed substantially during 1992–2001, with increasing age, more comorbidity, and fewer meeting ideal treatment criteria. Although treatment rates increased significantly for all medications, aspirin, β-blockers, and angiotensin-converting enzyme inhibitors were not provided at discharge to 12.6%, 19.7%, and 25.2% of ideal candidates, respectively, in 2000–2001. Crude 1-year mortality increased (27.6%, 28.3%, 30.6%, and 31.0%; P=0.003 for trend, but adjusted mortality declined (compared with 1992–1993, relative risk in 1995=0.94 [95% CI, 0.88 to 1.01]; relative risk in 1998–1999=0.91 [95% CI, 0.85 to 0.98]; relative risk in 2000–2001=0.87 [95% CI, 0.81 to 0.94]).
Conclusions— The quality of care and adjusted 1-year mortality improved significantly for Medicare beneficiaries with AMI during 1992–2001. Nevertheless, fewer were ideal for guideline-based therapy, and absolute mortality remains high, suggesting the need for treatment strategies applicable to a broader range of older patients.
Received January 1, 2006; revision received September 29, 2006; accepted October 10, 2006.
Despite the substantial resources dedicated to efforts to translating clinical trials of acute myocardial infarction (AMI) treatment into practice and to improving AMI care, comprehensive assessments of recent temporal trends in clinical characteristics, treatment, and outcomes of patients with the condition are lacking. Existing epidemiological studies of this topic provide population-based assessments of AMI but are not able to provide insights into temporal changes in the burden of AMI from the health system perspective.1–4 Available studies of hospitalized populations have generally been constrained by a limited geographic scope,5–8 lack of detailed clinical data,9 or the unavailability of long-term outcomes.10 Although 2 recent studies have reported on contemporary patterns of AMI treatment in national samples, neither assessed underlying population clinical characteristics, subsequent patient outcomes, or long-term trends.11,12 Information about secular trends in representative, community-based populations hospitalized with AMI would complement traditional epidemiological studies and is critically important in assessing past progress in improving quality of care and outcomes, evaluating past policy initiatives, and planning future quality improvement initiatives and clinical trials.
Clinical Perspective p 2814
In 1992, the Health Care Financing Administration (HCFA) (now the Centers for Medicare and Medicaid Services [CMS]) initiated the Cooperative Cardiovascular Project (CCP) pilot project, a study in 4 states (Connecticut, Iowa, Alabama, and Wisconsin) to assess the clinical characteristics and quality of care for Medicare beneficiaries with AMI.13 The CCP represented a transition in Medicare peer review from a paradigm of identifying clear outliers in quality to the goal of elevating care for all patients.14 As a result of the success of the CCP pilot, HCFA/CMS extended the initiative nationally in subsequent projects in 1995, 1998–1999, and 2000–2001.
Although some of the data from these projects have been published,15–18 a longitudinal assessment including all 4 time periods has not been available. The simultaneous study of all of these projects thus provides a unique opportunity to assess long-term trends in large community-based populations of older patients hospitalized with AMI from the health system perspective. Our objectives were to assess trends across a decade in the clinical history and presentation; eligibility for treatment with aspirin, β-blockers, and angiotensin-converting enzyme (ACE) inhibitors; rates of treatment with these medications; and mortality during and after AMI hospitalization.
Primary Clinical Data Sources
Analyses were based on data collected for the HCFA/CMS-sponsored AMI projects in 4 time periods: 1992–1993 (the CCP pilot); 1995 (the CCP); 1998–1999 (the National Heart Care [NHC] Project baseline), and 2000–2001 (the NHC remeasurement). Because the CCP pilot study was conducted in 4 US states (Alabama, Connecticut, Iowa, and Wisconsin), we restricted our analyses in all time frames to these 4 states to eliminate the introduction of geographic variability over time among samples.
Within each state, all discharges with the principal discharge diagnosis of AMI were identified from Medicare fee-for-service claims data with the use of the International Classification of Diseases, 9th Revision, Clinical Modification codes 410.xx. Because the projects were intended to assess the care of patients with an acute presentation, those patients with code 410.x2 were not included in the sampling because this code pertains to patients with subsequent episodes of care after a recent AMI.
In the CCP pilot sample, all claims for AMI submitted by any hospital in the 4 states between June 1992 and February 1993 were included. In the second CCP sample, all AMI claims in these states between August 1995 and November 1995 were included. The NHC baseline sample in the 4 states included hospital discharges between July 1998 and March 1999, and the NHC remeasurement included discharges between October 2000 and June 2001. In both of the NHC samples, all claims within each state in a 6-month time frame were sorted by age, race, and hospital provider number, and a systematic random sample of up to 850 discharges per state within the time frame was included. The numbers of patients for each of the 4 states in each of the sampling periods are included in Table I in the online-only Data Supplement.
Copies of the complete medical records for patients in the samples were provided for data abstraction. In the CCP pilot, records were abstracted at the individual state Peer Review Organizations (now known as Quality Improvement Organizations). For the CCP and both NHC samples, records were abstracted in 2 Clinical Data Abstraction Centers operating in contract with HFCA/CMS. In all cases, trained abstractors reviewed the medical records for the hospitalization using detailed standardized definitions for collected variables and entered data into electronic databases. The reliability of the abstracted variables and the quality indicators was assessed with random record reabstraction and subsequent calculation of rates of interrater agreement and κ statistics. Overall variable agreement averaged >90% for all variables in all samples, and κ exceeded 0.41, indicating at least good agreement.15,16 Agreement on the quality indicators exceeded 95% for all indicators in all samples, with κ exceeding 0.88, indicating excellent agreement for all indicators.19 For all variables used in the present study, the definitions used during the abstraction were consistent across time periods.
Only those patients with confirmed AMI were included in the study samples. In addition to the primary discharge diagnosis of AMI, cases were confirmed with the use of 1 of 2 criteria. The first was an elevation of cardiac markers, defined as (a) a lactate dehydrogenase (LDH) level >1.5 times the upper limit of normal or an LDH-1>LDH-2 on the highest LDH measurement, (b) a creatinine kinase muscle-brain isoenzyme fraction of total creatinine kinase >0.05, or (c) a troponin-T or troponin-I level exceeding the upper limits of normal. The second was any 2 of the following 3 characteristics: chest pain on presentation; at least a 2-fold elevation of total creatinine kinase; or evidence of new myocardial infarction (MI) on the ECG. Thus, all patients met the criteria for either definite or probable AMI according to the case definitions proposed by the American Heart Association.20 Of note, LDH elevations were noted in only 5.8% of records from 1992–1993, 2.6% in 1995, and <1% in 1998–1999 and 2000–2001. Troponin levels were not used clinically in the first 2 time frames, and the proportions of patients meeting the creatinine kinase muscle-brain isoenzyme criterion in each of the sample time frames declined over time (66.2%, 68.6%, 62.8%, and 58.0%; P for trend <0.001).
For the purposes of the present study, patients younger than 65 years were excluded because persons in this younger age group qualify for Medicare only if they have long-term disability or end-stage renal disease and thus are not representative of the general younger population. Patients with terminal illnesses documented in the medical record; those with unknown vital status; and patients who died, were discharged, or were transferred on the first day of the hospitalization were also excluded. The proportions of patients excluded did not differ significantly among sample time frames except for terminal illnesses, which were more prevalent in the 1992–1993 and 1995 samples. For patients who had >1 hospitalization during any of the sampling periods (n=1750, or 5.8% of the final sample), the clinical data from only the first hospitalization were included.
Quality Indicators and Ideal Treatment Candidates
The quality indicators assessed were the standard core measures used for the HCFA/CMS national efforts and were defined consistently throughout the 4 sample time frames: (1) the administration of aspirin within 24 hours after presentation; (2) the administration of a β-blocker within 24 hours after presentation; (3) the prescription of aspirin at hospital discharge; (4) the prescription of a β-blocker at hospital discharge; and (5) the prescription of an ACE inhibitor at hospital discharge among patients with documented left ventricular systolic dysfunction (left ventricular ejection fraction <0.40 or at least moderate left ventricular systolic dysfunction).15,16,18
All indicator rates were assessed only in those candidates considered ideal for treatment. First, those patients meeting minimal treatment requirements (“eligible candidates”) were identified. For the early aspirin and β-blocker measures, patients who were transferred from another acute care facility were not considered eligible. For the discharge treatment indicators, those patients who died during hospitalization, who were transferred to another acute care facility, or whose discharge status was unknown were not considered eligible. Among the cohorts of eligible candidates, ideal candidates were identified by excluding any patients with a documented contraindication to treatment. In general, the allowed contraindications were permissive, creating cohorts with a low likelihood of meaningful contraindications to treatment. The criteria for the quality measures are detailed in Table II in the online-only Data Supplement.
The dependent variables included the following: (1) meeting the ideal treatment criteria for the 5 quality indicators; (2) treatment with guideline-based medications among ideal candidates for the 5 quality indicators; and (3) outcomes of the hospitalization and subsequent clinical course, including (a) death during the index hospitalization and within 30 days, 6 months, or 1 year of admission and (b) hospital length of stay in days. In-hospital mortality was ascertained from the medical record and discharge abstracts, and dates of death within a year after hospital discharge were obtained from the Medicare Enrollment Database.21
Survey analysis methods were implemented to account for the clustering of patients by hospital and the sampling fraction at the state level.22 Survey mean methods were used to report and test the bivariate differences in continuous variables, and survey frequency methods were used to report and test the bivariate differences in categorical variables. Mortality differences over the time periods were assessed with the use of multivariable survey logistic regression models with adjustment for confounding factors. To assess the significance of differences over the time periods, 3 indicator variables were constructed representing the later sampling time frame in which the patient was treated, with 1992–1993 as the referent. Subsequently, models including these dummy variables for time were constructed including groups of potential confounders in the following order: no confounders, demographics, medical history, and clinical characteristics at presentation. Because of the possibility that increases in comorbidity prevalence could reflect trends in the coding of diagnoses, the risks for mortality for all comorbidities were calculated in models stratified by time period, and the significance of the interaction terms between time and each comorbidity was calculated in the full models. Because the sample size permitted, the models did not exclude any candidate variables to control maximally for confounding. The variables included in the models are shown in Table 1⇓.
Odds ratios were converted to estimated risk ratios.23 For each time period, weights proportional to the inverse sampling fraction within each state were applied to account for differences in the sampling fraction by state. Statistical analyses were conducted with the use of Stata/SE 8.0 (Stata Corporation, College Station, Tex) and SAS 9.0 (SAS, Inc, Cary, NC).
The authors had full access to the data and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.
Trends in Clinical Characteristics
The mean age of the study populations increased progressively from 75.7 years in 1992–1993 to 78.3 years in 2000–2001 (Table 1). By the last time period, nearly one quarter of the study population was at least 85 years old. There were significant increases over time in the proportion of the population who were female and nonwhite. By 2000–2001, 8% of patients were admitted from a skilled nursing facility. With respect to medical history, there were progressively greater proportions with additional cardiovascular conditions including prior MI, hypertension, heart failure, revascularization, and stroke, as well as other conditions including chronic lung disease, diabetes, and dementia. There were consistently higher proportions of patients admitted who were taking aspirin, β-blockers, ACE inhibitors, and statins.
There were also significant trends in the presentation of AMI over the time periods studied. Patients presenting in the later time periods had lower systolic blood pressures and lower heart rates. The proportion of patients with ST-segment elevation at presentation decreased from 37.4% to 23.5%, whereas that with left bundle-branch block increased from 3.1% to 10.8%. There were significant increases in the number of patients presenting with radiographic evidence of heart failure and in the number experiencing shock or heart failure during their hospitalization. Rates of percutaneous coronary interventions during the hospitalization increased from 12.7% to 24.6%.
Trends in Proportions of Ideal Candidates
The proportions of patients considered ideal candidates declined significantly over time for most of the quality indicators (Table 2). The absolute proportions ideal for aspirin or β-blockers at admission declined by nearly 20% for both indicators; the declines in the proportions of ideal candidates for aspirin and β-blockers at discharge were smaller but were also significant. Because these proportions were lower than for the admission indicators at all time frames, however, declines in eligibility resulted in the applicability of these indicators to relatively small minorities of patients. By 2000–2001, only 36.4% of patients surviving hospitalization were ideal candidates for aspirin, and only 11.1% were ideal candidates for β-blockers. In contrast, there was no significant change in the proportion of the population considered ideal for ACE inhibitor treatment at discharge. The stability of the proportion ideal for ACE inhibitors reflected an increase in the number of patients with a determination of left ventricular systolic function counterbalanced by an increasing number of patients with exclusionary conditions.
Trends in Treatment
In the entire study population regardless of treatment contraindications, discharge therapy with aspirin, β-blockers, and ACE inhibitors increased significantly over the time periods (Table 1). Among ideal candidates, proportions treated increased significantly (P<0.001) for all quality indicators. Of the indicators of treatment during the hospitalization, the absolute rates of treatment with aspirin increased by 8.0%, and the absolute rates of treatment with β-blockers increased by 33.5% (Figure 1). In absolute terms, rates of treatment at hospital discharge among ideal candidates increased by 10.8% for aspirin, 34.3% for β-blockers, and 25.7% for ACE inhibitors (Figure 2).
Trends in Hospital Length of Stay and Mortality
Between the first and last time period, the mean length of hospital stay declined from 10.1 days (SD=7.6) in 1992–1993 to 6.9 days (SD 12.6) in 2000–2001, representing a decrease of 30.7% (Table 3). In-hospital mortality and mortality within 1 month of admission did not change significantly over time, but death rates within 6 months and 1 year increased progressively during the time of the study (probability values for difference of 0.007 and 0.003, respectively; Table 3).
Compared with 1992–1993, unadjusted mortality rates within 1 year of admission were significantly higher in both 1998–1999 and 2000–2001 (Figure 3). After adjustment for differences in patient demographic and clinical characteristics, however, mortality risk declined progressively and was significantly lower in 1998–1999 and 2000–2001 compared with the initial time period, with a relative risk for mortality of 0.91 (95% CI, 0.85 to 0.98) in 1998–1999 and 0.87 (95% CI, 0.81 to 0.94) in 2000–2001.
In general, the relationship between individual patient factors and mortality was not significantly different across the time periods studied. Interactions with time were significant only for diabetes (P for interaction=0.05) and prior heart failure (P=0.04), neither of which were components of the criteria defining ideal candidacy for treatment for any of the indicators.
Our results provide several important insights into the population trends associated with AMI from the health system perspective. First, the clinical profile of Medicare patients hospitalized with AMI changed dramatically between 1992 and 2001. Second, with the use of consistent definitions of eligibility for therapy, the proportion of patients eligible for therapy with aspirin, β-blockers, and ACE inhibitors has declined over time. Third, although the proportions of this diminishing population of ideal candidates who were treated with evidence-based medical treatment improved significantly, important gaps in care persist. Fourth, in parallel with these improvements in care, the average length of hospital stay after AMI has decreased >30%. Finally, crude mortality rates within a year after AMI in the elderly population have increased since 1992. This trend, however, was associated with a worsening clinical risk profile of the population, and after adjustment for differences in patient characteristics, the risk of death declined.
The present study identifies marked changes in older patient population hospitalized with AMI in a geographically diverse sample, including lower proportions with ST-segment elevation, more frequent in-hospital adverse events, a greater burden of comorbidity, and more frequent contraindications to treatment. A changing nature of the MI population has been suggested previously in population- and hospital-based studies, including trends toward more non–ST-segment elevation MI2,6 and greater MI case severity.2,4 These findings have not been consistent, however.8,24
Although progressively fewer older patients with AMI were considered ideal for the application of standard guidelines, our data demonstrate consistent and significant improvements in care among those patients who were likely to benefit from treatment. The improvements in treatment rates, especially for β-blockers and ACE inhibitors, corroborate and expand on data from other studies, which indicate that care for hospitalized AMI patients has been improving.7,10,11,25 Despite these improvements, however, discharge prescriptions for aspirin, β-blockers, and ACE inhibitors were not provided to 12.6%, 19.7%, and 25.2% of ideal candidates, respectively, in 2000–2001, emphasizing the importance of ongoing efforts to translate clinical evidence into practice.
Concurrent with improvements in AMI care, adjusted mortality at 1 year after hospital discharge declined by 13% in relative terms between 1992 and 2001. This decline in mortality occurred while hospital lengths of stay decreased by >30%. These encouraging findings suggest the remarkable potential of advances in the clinical science supporting AMI care and the translation of these findings into practice. Persistently high mortality rates, however, also highlight the need to expand the evidence base for AMI care and for effective strategies to ensure that relevant advances in care are effectively disseminated and incorporated on a widespread basis into clinical practice.
Our results, which focus on trends in the clinical characteristics, care, and outcomes of older patients hospitalized with the principal discharge diagnosis of AMI, complement existing epidemiological studies designed to identify trends in cardiovascular risk factors, incident MI, and case fatality rates. Typically, these studies have assessed the contributions of a wide range of risk factors on events and outcomes.1–4 The present study also provides additional information that has not been available in previous studies of hospitalized patients, which have been limited by relatively narrow geographic scope,5–8 lack of detailed clinical data,9 or the inability to ascertain long-term outcomes.10 In contrast to existing published data, our present study focuses on AMI from the health system perspective among older persons, who compose the majority of AMI hospitalizations in the United States; studies clinical characteristics of >30 000 patients hospitalized in 4 geographically diverse US states; assesses the impact of changes in population characteristics on eligibility for therapy; and quantifies trends in the quality of care using explicit and consistent standards across a 10-year time frame.
The striking trends identified in the present study have important implications for clinical trials, guidelines, and performance measurement, all of which must adapt to the changing population with AMI to remain relevant to contemporary clinical care.26 Although past clinical trials in cardiovascular disease have excluded many patients treated in practice,27 the present study suggests that the differences between patients enrolled in clinical trials, who are typically relatively young and without substantial comorbidity, are becoming more pronounced. Because guidelines must rely on the best available clinical evidence, they are thus also limited in their capacity to address effectively and explicitly the best practices for caring for the very old patients with multiple comorbidities who make up an increasing proportion of the Medicare population. In turn, because the most robust performance measures are informed by guideline recommendations,28 the measures currently used to assess the quality of AMI care apply to smaller proportions of patients. With this trend, the precision of the estimates of quality based on these measures will decline progressively. Measures lacking the ability to discriminate because of small sample sizes are less useful for the purposes of public reporting or remuneration based on performance. Thus, the present study signals the urgent need for the generation of evidence for new strategies to improve outcomes for older patients with AMI.
The trends identified by the present study and the implications of these changes also confirm the importance of a national surveillance system for AMI. Although the CMS samples used for this study provided important insights on population trends, there are no plans to perpetuate the collection of clinical and outcomes data on a national scale. Existing longitudinal registries of adequate scope are currently supported by the industry.10,29 The future of these registries is subject to the interest of their sponsors, the access to registry data may be limited, the populations enrolled may include only those eligible for specific treatments, and the studies performed with the data may be influenced by the funding sources. An ongoing national registry with accessible data on the clinical presentation, treatment, and outcomes of AMI would have important public health benefits by providing timely insights into the changing population of patients with a condition that remains a leading cause of death in the United States.
Certain factors should be considered in the interpretation of these results. First, because of changes in diagnostic criteria for AMI, particularly those due to changes in cardiac biomarkers implemented in practice over time, some of the population trends in clinical characteristics and mortality may reflect differences in case identification due to differences in case identification, including the greater accuracy of the troponin assay. A prior study in Europe found that definitions using troponin identified an older patient cohort with greater comorbidity and higher crude mortality rates.30 However, most of the trends in the present study existed during 1992–1995, when only creatinine kinase was available, and continued during 1998–2001, when troponin was primarily used, suggesting that the population trends do not solely reflect differences in surveillance methods. It is also known that the 410 International Classification of Diseases, 9th Revision, Clinical Modification codes do not identify all cases of AMI, and the present study could not identify patients in the population who did not present to the hospital or whose diagnosis was missed. However, the validity of these codes has been shown to be stable over time.31 Regardless, case identification would not have an important impact on conclusions about the applicability of treatment because all patients in the present study had a primary discharge diagnosis of AMI, and those with contraindications were excluded from the ideal candidate groups. Finally, comprehensive risk adjustment was performed to provide an assessment of mortality trends after we accounted for differences in age, case severity, and comorbidity.
It is also possible that the documentation of comorbidities in medical records changed over time, which in turn could influence the documented prevalence of these associated conditions. For most of the measured comorbidities, however, the relationship with mortality over time was not significantly different, suggesting relatively consistent ascertainment of these factors during the time periods included in the present study. Furthermore, the prevalence of many comorbidities increased in parallel with increasing age, supporting true increases in the burden of comorbidities as opposed to changes in ascertainment. As a study of older Medicare beneficiaries, most of whom were white, these results may not apply to younger populations or nonwhite patients. Because the majority of acute coronary events in the United States occur in the older population, however, the present study is applicable to a substantial segment of the population with AMI. Finally, because the quality measures implemented in these CMS projects did not include assessments of the provision of immediate reperfusion or statin therapy to appropriate patients, trends in the use of these therapies in ideal subsets could not be studied.
In conclusion, the clinical characteristics, treatment, and outcomes of older patients presenting with AMI in 4 states during 1992–2001 has changed substantially. These findings illustrate the urgent need for additional evidence-based strategies applicable to this population to provide guidance to practitioners who care for increasingly complex elderly AMI patients. The rapid changes in this population also highlight the substantial potential for a system of national surveillance of patients with AMI to identify important trends promptly, with the goal of informing enrollment in future clinical trials, measures of performance, and health policy.
Sources of Funding
Dr Foody is supported by a National Institutes of Health/National Institute on Aging Research Career Award K08-AG20623 and National Institute on Aging/Hartford Foundation Fellowship in Geriatrics.
Dr Masoudi has served on speaker’s bureaus for AstraZeneca and Pfizer. The analyses on which this publication is based were performed under contract 500-02-CO 01 entitled “Utilization and Quality Improvement Organization for the State of Colorado” sponsored by the Centers for Medicare & Medicaid Services, Department of Health and Human Services. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US government. The authors assume full responsibility for the accuracy and completeness of the ideas presented. This article is a direct result of the Health Care Quality Improvement Program initiated by the Centers for Medicare & Medicaid Services, which has encouraged identification of quality improvement projects derived from analysis of patterns of care, and therefore required no special funding on the part of this contractor. Ideas and contributions to the author concerning experiences in engaging with issues presented are welcomed.
Hellermann JP, Reeder GS, Jacobsen SJ, Weston SA, Killian JM, Roger VL. Longitudinal trends in the severity of acute myocardial infarction: a population study in Olmsted County, Minnesota. Am J Epidemiol. 2002; 156: 246–253.
Tunstall-Pedoe H, Vanuzzo D, Hobbs M, Mahonen M, Cepaitis Z, Kuulasmaa K, Keil U. Estimation of contribution of changes in coronary care to improving survival, event rates, and coronary heart disease mortality across the WHO MONICA Project populations. Lancet. 2000; 355: 688–700.
Furman MI, Dauerman HL, Goldberg RJ, Yarzebski J, Lessard D, Gore JM. Twenty-two year (1975 to 1997) trends in the incidence, in-hospital and long-term case fatality rates from initial Q-wave and non-Q-wave myocardial infarction: a multi-hospital, community-wide perspective. J Am Coll Cardiol. 2001; 37: 1571–1580.
Edlavitch SA, Crow R, Burke GL, Baxter J. Secular trends in Q wave and non-Q wave acute myocardial infarction: the Minnesota Heart Survey. Circulation. 1991; 83: 492–503.
Rogers WJ, Canto JG, Lambrew CT, Tiefenbrunn AJ, Kinkaid B, Shoultz DA, Frederick PD, Every N. Temporal trends in the treatment of over 1.5 million patients with myocardial infarction in the US from 1990 through 1999: the National Registry of Myocardial Infarction 1, 2 and 3. J Am Coll Cardiol. 2000; 36: 2056–2063.
Marciniak TA, Ellerbeck EF, Radford MJ, Kresowik TF, Gold JA, Krumholz HM, Kiefe CI, Allman RM, Vogel RA, Jencks SF. Improving the quality of care for Medicare patients with acute myocardial infarction: results from the Cooperative Cardiovascular Project. JAMA. 1998; 279: 1351–1357.
Burwen DR, Galusha DH, Lewis JM, Bedinger MR, Radford MJ, Krumholz HM, Foody JM. National and state trends in quality of care for acute myocardial infarction between 1994–1995 and 1998–1999: the Medicare health care quality improvement program. Arch Intern Med. 2003; 163: 1430–1439.
Luepker RV, Apple FS, Christenson RH, Crow RS, Fortmann SP, Goff D, Goldberg RJ, Hand MM, Jaffe AS, Julian DG, Levy D, Manolio T, Mendis S, Mensah G, Pajak A, Prineas RJ, Reddy KS, Roger VL, Rosamond WD, Shahar E, Sharrett AR, Sorlie P, Tunstall-Pedoe H. Case definitions for acute coronary heart disease in epidemiology and clinical research studies. Circulation. 2003; 108: 2543–2549.
STATA Survey Data Reference Manual. College Station, Tex: STATA Press; 2005.
McNutt LA, Wu C, Xue X, Hafner JP. Estimating the relative risk in cohort studies and clinical trials of common outcomes. Am J Epidemiol. 2003; 157: 940–943.
Salomaa V, Koukkunen H, Ketonen M, Immonen-Raiha P, Karja-Koskenkari P, Mustonen J, Lehto S, Torppa J, Lehtonen A, Tuomilehto J, Kesaniemi YA, Pyorala K. A new definition for myocardial infarction: what difference does it make? Eur Heart J. 2005; 26: 1719–1725.
Rosamond WD, Chambless LE, Sorlie PD, Bell EM, Weitzman S, Smith JC, Folsom AR. Trends in the sensitivity, positive predictive value, false-positive rate, and comparability ratio of hospital discharge diagnosis codes for acute myocardial infarction in four US communities, 1987–2000. Am J Epidemiol. 2004; 160: 1137–1146.
Although substantial resources have been dedicated to improving the quality of care and outcomes of patients with acute myocardial infarction, recent trends in the clinical characteristics, medical care, and mortality in this patient population have not been well characterized. The present study included 20 550 Medicare beneficiaries at least 65 years old hospitalized with acute myocardial infarction in 4 US states during 1992–2001. The clinical characteristics of the population changed markedly over time, with increasing age, a greater burden of comorbid conditions, and a larger proportion presenting with non–ST-segment elevation myocardial infarction. With the use of standard quality indicator definitions, the proportions of patients ideal for treatment with aspirin and β-blockers at hospital admission and at discharge and with angiotensin-converting enzyme inhibitors at discharge among those with left ventricular systolic dysfunction declined significantly over time. Treatment rates increased significantly for all therapies. In 2000–2001, however, aspirin, β-blockers, and angiotensin-converting enzyme inhibitors were not provided at discharge to 12.6%, 19.7%, and 25.2% of ideal candidates, respectively. Although crude 1-year mortality increased (27.6% to 31.0%), adjusted mortality in 2000–2001 was significantly lower than in 1992–1993 (relative risk=0.87; 95% CI, 0.81 to 0.94). These findings have important implications for clinical investigation, performance measurement, and health policy. Practitioners and health systems must adapt to the rapidly changing population with acute myocardial infarction and develop strategies to provide care to increasingly complex patients. Although treatment is improving, gaps in care persist, supporting ongoing efforts to enhance the use of existing therapies. Finally, declining proportions ideal for established treatments and persistently high mortality rates demonstrate the need to develop new therapies applicable to larger numbers of patients.
Guest Editor for this article was Gregory L. Burke, MD, MSc.
The online-only Data Supplement, consisting of tables, is available with this article at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.106.611707/DC1.