Canto JG, Rogers WJ, Chandra NC, French WJ, Barron HV, Frederick PD, Maynard C, Every NR, for the National Registry of Myocardial Infarction 2 Investigators. The Association of Sex and Payer Status on Management and Subsequent Survival in Acute Myocardial Infarction. Arch Intern Med. 2002;162(5):587-593. doi:10.1001/archinte.162.5.587
Previous reports have generally shown lower utilization of hospital resources and lower survival in women than men with acute myocardial infarction. However, to our knowledge, no reports have described the influence of payer status on the treatment and outcome of women and men with acute myocardial infarction.
Baseline and clinical presenting characteristics, utilization of hospital resources, and subsequent clinical outcome were ascertained among 327 040 women and men enrolled in a national registry of myocardial infarction from June 1, 1994, to January 31, 1997. Separate Cox regression analyses were performed for Medicare, Medicaid, health maintenance organizations, and commercial payer groups to ascertain variables that were predictive of mortality in the study population.
After adjustment for differences in age and other baseline and presenting characteristics, women were significantly more likely than men to die in the hospital (hazard ratio, 1.13; 95% confidence interval, 1.10-1.16), and this difference was greatest among women with health maintenance organization and commercial insurance (hazard ratios, 1.30 and 1.29, respectively), and least among women with Medicare (hazard ratio, 1.07). However, after adjustment for the additional effect on short-term survival of sex differences in the utilization of both pharmacologic treatments administered within the first 24 hours and invasive cardiac procedures, the mortality difference observed for women and men further diminished (hazard ratio, 1.08; 95% confidence interval, 1.05-1.10).
In this large registry, we did not observe significant variations among payer classes in management and mortality among women and men after acute myocardial infarction.
MILES AND Parker,1 in an article in the New England Journal of Medicine, raised concerns about the lack of data on how private health insurance, Medicare, and Medicaid may serve women and men. Furthermore, these authors suggested that uniform coverage of health services may not necessarily serve both sexes equitably, especially given different life spans and patterns of illnesses, as well as differences in vocational, familial, and political roles and economic status.
Sada et al2 and Canto et al,3 in separate but unique analyses in the National Registry of Myocardial Infarction (NRMI) 2, reported that payer status was strongly associated with the use of pharmacologic therapies and invasive cardiac procedures for acute myocardial infarction (AMI) and observed a higher in-hospital mortality in the Medicaid cohort. Kreindel et al,4 in a population-based study in Worcester, Mass, reported wide variability in hospital case-fatality rates among 5 medical insurance groups after AMI, but after adjustment for differences in clinical characteristics, their study lacked statistical power to find a significant association between insurance status and short-term mortality.
To our knowledge, no large study has described the influence of payer status in the treatment and subsequent outcome of women and men with AMI, especially at a time when many are questioning whether payer status influences the quality of care and outcomes in patients. Therefore, the primary purpose of this investigation was to examine for differences in the contemporary treatment and survival of women and men by primary payer status.
The NRMI 2, a voluntary registry designed to collect, analyze, and report cross-sectional data on patients admitted with confirmed AMI at more than 1470 participating hospitals,5,6 enrolled 446 970 patients from June 1, 1994, to January 31, 1997. Data from each patient were entered onto a 2-page case report form by trained chart abstractors and forwarded to ClinTrials Research Inc (Lexington, Ky). Double key data entry and 87 electronic data checks were routinely performed by the data collection center to help ensure the accuracy, consistency, and completeness of the data. Inaccurate and internally inconsistent case report forms were excluded from analysis and returned to the registry hospital for additional review and correction. Hospitals were strongly encouraged to enroll consecutive patients with AMI. The diagnosis of AMI was based on at least 1 of the following: (1) a level of total creatine kinase or creatine kinase MB fraction greater than or equal to twice the upper limit of normal, (2) electrocardiographic evidence indicative of AMI, (3) alternative enzymatic, scintigraphic, or autopsy evidence indicative of AMI, and (4) International Classification of Diseases, Ninth Revision, Clinical Modification, diagnosis code of 410.X1.
Medicare included all state and federal Medicare-type programs for qualified individuals 65 years or older and for qualified persons with end-stage renal disease who required dialysis or a kidney transplant (regardless of age).7Medicaid included all state and federal Medicaid-type programs for certain low-income individuals. Health maintenance organization (HMO) involved the following models that provided health care services for members on a prepaid basis: staff model, group model, network model, and individual practice associations. Commercial covered all indemnity carriers and preferred provider organizations. At the conception of the NRMI 2 case report form, Medicare and Medicaid HMOs were not considered a separate payer category.
Patients who were transferred out of a registry hospital (21%) were excluded from the analysis, because information regarding their in-hospital course and outcome were usually unknown. However, transferred-in patients were not excluded, because such information was generally captured. Also, patients whose primary payer was Veterans Affairs or the Civilian Health and Medical Program of the Uniformed Services (<1% of total), unknown (3% of total), or uninsured were not included in this report. Although the last group is of increasing importance given the lack of insurance in a substantial segment of the population, they were not included in this analysis given the likely heterogeneous nature of this group, who may represent the poor, middle-class (self-employed), or even those receiving charity care. Thus, inclusion of this payer class may introduce the possibility of confounding or selection bias if differences were or were not observed by sex. A total of 119 930 patients (27%) were excluded in this analysis.
Patients from each payer group were analyzed with respect to (1) baseline and clinical presenting characteristics; (2) utilization of hospital resources such as acute reperfusion strategies, coronary angiography, coronary angioplasty, and coronary artery bypass graft surgery; and (3) mortality. Acute reperfusion strategies, defined as the use of intravenous or intracoronary thrombolytic therapy, primary angioplasty, or immediate coronary artery bypass surgery, were determined among patients with AMI who presented with ST-segment elevation or left bundle-branch block (LBBB) on first electrocardiogram, within 12 hours of symptom onset, and had no contraindication to thrombolytic therapy. Invasive cardiac procedures included any coronary angiography, coronary angioplasty (primary or immediate, rescue, and elective), and coronary artery bypass graft surgery (immediate and elective) performed during the hospitalization period.
Forward stepwise Cox regression analyses8 were used to examine for sex variations in hospital mortality, and whether such differences (if present) were also seen by payer status. Adjusted mortality (hazard ratio ± 95% confidence interval [CI]) was determined among women and men (reference group) in the overall study population and also within each selected payer group. A series of 3 sequential regression models for mortality were developed to ascertain the crude mortality rate, and then to examine the contribution of baseline characteristics and differential use of pharmacologic treatments and invasive cardiac procedures on subsequent survival. Potential prognostic variables were entered into the regression models for mortality as follows: model 1, crude mortality (not adjusted for clinical and treatment differences); model 2, adjusted for age and other demographic features, medical history, presenting characteristics, and hospital characteristics; and model 3, clinical covariates listed in model 2 plus the following treatments and procedures: 10 medications administered within 24 hours (angiotensin-converting enzyme inhibitor, antiplatelet, antithrombin, aspirin, calcium channel blocker, β-blocker, intravenous heparin, intravenous magnesium, intravenous nitroglycerin, and lidocaine), acute reperfusion therapies (thrombolytic therapy, primary angioplasty, and immediate coronary bypass surgery), and any invasive cardiac procedures (coronary angiography, coronary angioplasty, and coronary bypass surgery). Given the large sample size and power in our study, 10 pharmacologic treatments and the invasive cardiac procedures were selected (a priori) to be entered into the multivariable regression model because of their potential to influence mortality. Furthermore, age- and race-adjusted logistic regression analyses were performed to describe differences in the utilization of thrombolytic therapy, primary angioplasty, any acute reperfusion therapy, coronary angiography, coronary angioplasty, and coronary bypass surgery. All statistical analyses were performed with SAS 6.10 statistical package programs (SAS Institute Inc, Cary, NC). This report was based on information processed by the central data collection center as of January 1, 1997.
A total of 327 040 patients fulfilled study criteria. Among all payer groups, women represented 40% and men 60% of the study population; however, the proportion of women was largest among Medicare and Medicaid recipients (48% each), followed by HMO (30%) and commercial (25%).
Women were generally older than men for all payers (mean age, 72.4 vs 64.8 years; median age, 74.2 vs 65.8 years) and within every payer group (Table 1). In the combined sample, the proportion of minorities was similar for women and men (11.9% vs 10.4%); however, the payer-stratified data showed that, within the Medicaid and HMO subsets, women were significantly more likely to come from a minority group, whereas in those with commercial insurance, men were more likely to be part of a minority group.
Women were generally less likely to have a history of AMI or previous coronary revascularization procedures, although they were more likely to have previously documented heart failure. Diabetes and hypertension were significantly more prevalent among women than men, but women were less likely to smoke or have a family history of coronary artery disease. Generally speaking, these differences were seen between women and men within each payer group.
Women were less likely than men to call for an ambulance before hospital arrival, present to the hospital with chest pain, or have ST-segment elevation or LBBB on initial electrocardiogram (Table 2). However, women were more likely than men to delay before presenting to the hospital and more likely to have a Killip class greater than I on initial evaluation. There were similar proportions of Q-wave and anterior infarctions. With the exception of Medicare-insured women, who were slightly more likely to present with ST-segment elevation or LBBB, the variations observed among all women and men were similar to those observed within each payer group.
After adjustment for age and race among eligible candidates who presented within 12 hours of symptom onset with ST-segment elevation or LBBB on the initial electrocardiogram, and no contraindication to reperfusion therapies, women were still 24% less likely than men to receive intravenous thrombolytic therapy (odds ratio [OR], 0.76; 95% CI, 0.74-0.79) and 17% less likely to undergo primary angioplasty (OR, 0.83; 95% CI, 0.79-0.86) (Table 3). Taken together, women were 31% less likely than men to receive any appropriate therapy, after accounting for age and race (OR, 0.69; 95% CI, 0.67-0.71). Similar degrees of lower utilization were evident in every payer category (the ORs varied little among the payer groups and ranged from 0.68 to 0.76).
After adjustment for differences in age and race, women were significantly less likely than men to undergo coronary angiography (OR, 0.74; 95% CI, 0.73-0.75). Furthermore, after angiography, women were significantly less likely to undergo either coronary angioplasty (OR, 0.81; 95% CI, 0.80-0.83) or coronary bypass surgery (OR, 0.64; 95% CI, 0.63-0.66) while in the hospital.
However, the degree of sex difference in the utilization of invasive cardiac procedures among women and men was generally greatest among Medicare recipients (coronary angiography: OR, 0.68; coronary angioplasty: OR, 0.79; and coronary bypass surgery: OR, 0.58) and least among those with commercial and Medicaid insurance (coronary angiography: OR, 0.88 and 0.86, respectively; coronary angioplasty: OR, 0.88 and 0.85, respectively; and coronary bypass surgery: OR, 0.76 and 0.90, respectively).
Women were generally less likely than men to receive aspirin, β-blocker, intravenous heparin, or nitrate therapies within the first 24 hours, a finding present regardless of payer status (Table 4).
For every payer group, women were significantly more likely than men to die in the hospital (model 1), and this difference was greatest among women with HMO and commercial insurance (hazard ratios, 1.61 and 1.64, respectively), and least among women with Medicare (hazard ratio, 1.18) (Table 5). After adjustment for age, baseline, and clinical presentation (model 2), women were still more likely than men to die in the hospital, and this was highest among women with HMO and commercial insurance, who were still 1.3 times more likely than men to die in the hospital. Women with Medicaid may be 1.2 times more likely than men to die in the hospital, but the CI was very wide. Finally, after adjustment for the combination of clinical characteristics, hospital characteristics, the pharmacologic treatments administered within the first 24 hours, and invasive cardiac procedures performed in the hospital (model 3), most of the differences in mortality observed among the payer group for women and men disappeared.
The higher mortality observed in women as compared with men in large part was attributed to sex differences in the baseline characteristics (70%), and in small part was attributed to the differential use of pharmacologic treatments and invasive cardiac procedures (10%). However, the degree of these sex differences in mortality did not appear to vary significantly by payer status, as the adjusted mortality stratified by payer group did not significantly differ from that of the overall study population. All the adjusted models achieved reasonable fit, as indicated by the c statistic, which ranged from 0.65 to 0.86. In all cases, the model χ2 performed significantly better in the higher-order model (model 3) than either of the lower-order models (model 1 or model 2).
Our data, which reflect contemporary, real-world practice patterns among patients enrolled from all 50 states and the District of Columbia, depict sex-specific differences in the utilization of hospital resource and hospital mortality, a finding that was present irrespective of payer status. Among patients hospitalized with AMI in this national registry, overall, women clearly received fewer cardiac treatments and procedures and had a worse outcome than men, although payer status did not appear to explain these differences (with the exception of some minor differences in the utilization rates of invasive cardiac procedures by payer groups). Although payer status was not responsible for the worse survival observed in women, the degree of mortality variation appeared to be highest among women enrolled in HMO and commercial insurance plans and lowest among women with Medicare, a finding that may warrant further investigation.
Women were significantly less likely to receive important pharmacologic treatments such as thrombolytic therapy, aspirin, β-blocker, and intravenous heparin therapies within 24 hours of hospital admission and less likely to receive invasive cardiac procedures such as coronary angiography and subsequent revascularization in the NRMI 2, but was the lack of these treatments necessarily associated with worse outcome? In the present analysis of NRMI 2, the Cox proportional hazard models for hospital mortality among all payer groups suggested that clinical factors such as older age and higher proportion of other comorbidities were largely responsible for the higher hospital mortality observed among women in the overall study population and also among each payer group as compared with men (models 1 and 2). However, sex differences in the utilization of pharmacologic treatments instituted within 24 hours and invasive cardiac procedures (generally lower usage in women) contributed additional information in explaining the higher hospital mortality observed in women (model 3 statistically performed better than model 2 irrespective of payer status, and there was no overlap of the CIs between these 2 models in the overall population).
Payer status may represent more than just who pays for one's health care; some may argue that it is also a surrogate marker of social class or socioeconomic status (data that were not collected in the NRMI). It is important to emphasize that this analysis was not designed to directly compare each respective payer group, which comparison has been reported in previous analyses of the NRMI,2,3 but rather aimed at addressing the magnitude of differences in the management and outcome of men and women within each payer group. Our results by payer status are consistent with a small population-based study of the Worcester, Mass, metropolitan area4 that also did not find a significant overall association between insurance status and short-term mortality in the general population. However, to our knowledge, our study is the first to report on the influence of payer status on the treatment and outcome of women and men with AMI.
Given the higher short-term risk of mortality from AMI, our data suggest that women, regardless of payer status, who often have other coexisting factors associated with a worse prognosis, may warrant at least equal (not less) aggressive in-hospital assessments and treatments known to improve short-term survival. Chandra et al9 reported that women enrolled in the NRMI 1 from 1990 to 1994 received less aggressive pharmacologic treatments and invasive cardiac procedures and had a lower hospital survival than men. However, whether the lower utilization of hospital resources predicted worse survival in this cohort of women in Chandra and coworkers' earlier analysis was not rigorously examined. In our study, sex differences in the receipt of important pharmacologic treatments and invasive procedures contributed in small part to the increase in mortality.
The NRMI 2 represents an observational study, and therefore we could not fully adjust for unmeasured confounders such as socioeconomic status, which may be associated with sex, payer status, and outcome. However, the NRMI 2 has collected more than 225 data entry variables related to clinical covariates, acute treatments, cardiac procedures, and subsequent hospital outcome, which may allow for more extensive adjustments of important baseline and presenting characteristics. Also, primary payer status was determined by chart review from trained abstractors, but proper coding of primary payer status was not directly verified or validated. Therefore, payer misclassification is possible, especially in situations where more than 1 payer group existed. Nonetheless, if this were the case, a similar degree of bias (ie, a nondifferential misclassification in bias) would likely be seen among women and men within each payer group, and therefore this (theoretical) limitation should not compromise the validity of our findings. Also, at the conception of the NRMI 2 case report form, Medicare and Medicaid HMOs were not prevalent and were not considered a separate payer category.
The primary purpose of this study was to examine whether there were differences in treatment and outcome of women and men with MI when analyzed by payer status, and exclusion of the uninsured would not likely have changed our conclusions.
This report including 327 040 women and men enrolled in the NRMI 2 from June 1994 to January 1997 suggests significant sex differences in the acute management and outcome after hospitalization for AMI, although payer status was not likely responsible for these disparities. Despite the acknowledgment of limitations in the NRMI 2, this analysis, which describes the clinical characteristics and in-hospital experience of women and men among 4 major payer groups, may provide a unique opportunity to better understand the utilization of hospital resources in the management of AMI and its effect on hospital outcome among all women and men.
Accepted for publication July 2, 2001.
This study was supported in part by grant HS08843 from the Agency for Healthcare Research and Quality, Rockville, Md. The NRMI 2 is supported by Genentech, Inc, South San Francisco, Calif.
Corresponding author: John G. Canto, MD, MSPH, University of Alabama at Birmingham, 363 BDB, 1808 Seventh Ave S, Birmingham, AL 35294-0012.