Author + information
- Received November 21, 2001
- Revision received March 11, 2002
- Accepted March 29, 2002
- Published online June 19, 2002.
- D.A Alter, MD, PhD*,‡,* (, )
- C.D Naylor, MD, Dphil*,∥,¶,
- P.C Austin, PC, PhD*,∥ and
- J.V Tu, MD, PhD*,†,§∥
- ↵*Reprint requests and correspondence:
Dr. David A. Alter, Institute for Clinical Evaluative Sciences, G106-2075 Bayview Avenue, Toronto, Ontario, Canada M4N 3M5.
Objectives The goal of our study was to examine how age and gender affect the use of coronary angiography and the intensity of cardiac follow-up care within the first year after acute myocardial infarction (AMI). Another objective was to evaluate the association of age, gender and treatment intensity with five-year survival after AMI.
Background Utilization rates of specialized cardiac services inversely correlate with age. Gender-specific practice patterns may also vary with age in a manner similar to known age–gender survival differences after AMI.
Methods Using linked population-based administrative data, we examined the association of age and gender with treatment intensity and long-term survival among 25,697 patients hospitalized with AMI in Ontario between April 1, 1992, and December 31, 1993. A Cox proportional hazards model was used to adjust for socioeconomic status, illness severity, attending physician specialty and admitting hospital characteristics.
Results After adjusting for baseline differences, the relative rates of angiography and follow-up specialist care for women relative to men, respectively, fell 17.5% (95% confidence interval [CI], 13.6 to 21.3, p < 0.001) and 10.2% (95% CI, 7.1 to 13.2, p < 0.001) for every 10-year increase in age. Conversely, long-term AMI survival rates in women relative to men improved with increasing age, such that the relative survival in women rose 14.2% (95% CI, 10.1 to 17.5, p < 0.001) for every 10-year age increase.
Conclusions Gender differences in the intensity of invasive testing and follow-up care are strongly age-specific. While care becomes progressively less aggressive among older women relative to men, survival advantages track in the opposite direction, with older women clearly favored. These findings suggest that biology is likely to remain the main determinant of long-term survival after AMI for women.
Gender differences in cardiovascular care after acute myocardial infarction (AMI) have been studied in a number of settings (1–6). Variations in the use of evidence-based therapies between men and women are generally smaller and less consistent than those associated with the use of more discretionary interventions where the benefits are less clear (2,3,7). Studies of gender-treatment differences are also confounded by age effects. Treatment selection and referral behaviors are influenced by age (2,8–10); therefore, the delayed onset of coronary disease in women probably contributes to the observed variations in clinical decision-making. Indeed, the relationship between gender and disease severity may also vary by age. For example, in order for coronary disease to occur in younger women, the disease must be particularly aggressive or driven by the presence of multiple risk factors (11,12). Physicians may accordingly adopt more aggressive referral patterns in younger women. While investigators have routinely studied the relationship between gender and treatments within age-specific strata, few have examined how gender-specific treatments differ across all ages.
A parallel concern is the fact that many studies have demonstrated an age–gender mortality interaction for AMI such that younger, but not older, women have higher mortality rates when compared with men of similar ages (13–15). Given age–gender differences in treatment, one might reasonably ask: to what extent do gender-specific differences in outcomes after AMI reflect biology as opposed to age/gender bias?
The objectives of this study were to examine whether gender-related differences in the use of specialized cardiac services vary according to age and, if so, to what extent such differences in service intensity are associated with variations in longer-term mortality between men and women after AMI.
We collected information from the Ontario Myocardial Infarction Database (OMID), which links together a variety of administrative databases in Ontario. Complete details regarding the construction, eligibility and accuracy of OMID have been previously described (16–18). Using the Canadian Institute for Health Information (CIHI) hospital discharge database, we examined all Ontarians admitted with a most responsible diagnosis of AMI (International Classification of Diseases, 9th Revision, Clinical Modification [ICD-9-CM] code 410) between April 1, 1992 to December 31, 1993.
Baseline sociodemographic, clinical and hospital characteristics
Baseline patient characteristics including age, gender, socioeconomic status and disease severity were obtained from CIHI discharge abstracts of the index AMI admission. Socioeconomic data were obtained using neighborhood income level obtained from 1991 official Canadian census data. To control for variations in patient severity of illness on admission, we used the Ontario AMI mortality prediction rule for 30-day mortality (18). The variables in this model include age, gender, cardiac severity (e.g., congestive heart failure, cardiogenic shock, arrhythmias) and comorbid status (e.g., diabetes mellitus, stroke, acute and chronic renal disease and malignancy) as derived from the ICD-9 codes present in the 15 secondary diagnostic fields of the CIHI database. The predictive models showed good predictive power, with areas under the receiver operating characteristics cure of 0.78 and 0.79 for 30-day and for one-year mortality, respectively. The derivation and validation of the Ontario AMI prediction rule is described elsewhere (18,19). Excluding age and gender from the mortality prediction rule did not significantly alter our results.
Available evidence suggests that AMI processes of care are also influenced by other physician and hospital characteristics. These include attending physician specialty (20), hospital volumes (based on the annual number of patients admitted to the myocardial infarction facility) (18,21), the presence or absence of on-site procedural characteristics at the admitting hospital (22), hospital teaching status (17)and geographical proximity to the nearest tertiary facility (from latitude and longitude coordinates, “as the crow flies”) (23). Attending physician specialty and physician specialty visits were identified using hospital discharge abstracts and Ontario Health Insurance Plan claims data, respectively. During the study period, there were 197 acute care hospitals in Ontario. Three rural isolated hospitals were excluded from the analysis given that census data suppress information needed to obtain socioeconomic status in isolated rural communities. Four institutions had on-site angiography-only facilities comprising 3.5% of the sample population and were also excluded from the analysis because of small sample size.
Specialized cardiac care and outcome
Coronary angiography and out-patient follow-up care served as the two process indicators of intensity of service, while mortality was the sole outcome examined in this study. Coronary angiography was examined within the first six months of the index AMI, in order to allow for post-MI risk stratification and lengthy waiting list delays for coronary angiography in Ontario at that time (23). Mortality was assessed at five years after MI to allow for a sufficient opportunity to observe long-term prognostic benefits attributable to procedures if such were to exist (24).
We began by subdividing our cohort into four pre-specified age groups (ages 20 to 49, 50 to 64, 65 to 74 and 75+) in order to examine baseline characteristics and crude outcomes between men and women. The age-specific breakdown classification used in this study has been previously incorporated by our group in other studies (23). Unadjusted gender differences were examined using chi-square tests of proportions.
We then examined age as a continuous variable, evaluating age-specific gender effects applying an age–gender interaction term. Multivariate analyses using Cox proportional hazard models were developed to examine the association between age–gender, process (coronary angiography, post-discharge follow-up care) and outcome (mortality) after adjusting for age, socioeconomic status, illness severity (probability of 30-day mortality), attending physician specialty, and the four hospital characteristics described above. Formal diagnostic testing for collinearity across hospital- and/or physician-level variables did not reveal any variance inflation factor to be >5.0. Therefore, collinearity was not a significant issue for this analysis (i.e., the maximum Variance Inflation Factor across explanatory variables for the 190 hospitals was 2.6) (25).
All models were derived in a similar fashion. Patient characteristics (age, gender, socioeconomic status, illness severity) were forced into the statistical model, while physician and hospital characteristics whose variables were significant at a p value of 0.2 in univariate fashion were selected using backward stepwise regression techniques by comparing the −2 log likelihoods ratio test of the Cox proportional hazards. Because early mortality results in less opportunity for patients to receive treatments or specialty follow-up care, death was the main censoring variable. Adjusted survival rates were examined among patients who received coronary angiography or post-discharge follow-up care and among those who received neither. To examine the robustness of our analyses, a parametric survival model that did not assume proportional hazards was fit to our data; similar results were obtained.
Statistical significance was defined as p < 0.05 for all analyses. SAS version 8.2 statistical software was used for all statistical analyses.
The cohort consisted of 25,697 patients (16,756 men, 8,941 women) and had a mean age of 66.6 years (64.1 vs. 71.2 for men and women, respectively, p < 0.001). While the distribution of age was more homogenous within the four pre-specified age groups, women were still, on average, significantly older than men by 0.4 years, 0.6 years and 1.1 years in the 50 to 64 year, 65 to 74 year and 75+ year age groups, respectively (p < 0.001 for each). The utilization rates of six-month coronary angiography and one-year cardiology follow-up were 23.6% and 38.7%, respectively; 40.7% of the cohort died within five years after AMI.
Table 1demonstrates the unadjusted odds ratios in women relative to men for each baseline characteristic across the four pre-specified patient subgroups. Among the younger age categories, women were significantly poorer and sicker than were men. However, gender differences in socioeconomic status and illness severity markedly narrowed among elderly age groups. While the distribution of hospital services was similar among women and men across all age groups, elderly women were significantly more likely to be seen by general practitioners than were elderly men.
Unadjusted rates of specialty care and outcomes
Table 2illustrates that the unadjusted odds ratio of receiving coronary angiography and cardiac follow-up specialty care after discharge in women relative to men decreased with advancing age. The likelihood that women, compared with men, were not followed up by any specialist (either a cardiologist or an internist) grew with increasing age. Despite the inverse relationship between age and treatment aggressiveness in women relative to men, crude five-year mortality rates in women relative to men decreased with advancing age.
Adjusted rates of specialty care and outcomes
Figure 1illustrates the age/gender-specific rates of coronary angiography and post-discharge follow-up care after adjusting for baseline case-mix, attending physician and admitting hospital characteristics. For each of the three variables examined, younger women were treated more aggressively than younger men, while older women were treated less aggressively than older men. For every 10 years of increasing age, the relative decreases in the risk ratios for coronary angiography and cardiology specialty follow-up in women relative to men were 17.5% (95% confidence interval, 13.6 to 21.3, p < 0.001) and 10.2% (95% confidence interval, 7.1 to 13.2, p < 0.001), respectively. Absolute differences in treatments across age groups were greater than the absolute differences in treatments between men and women (Fig. 1). Accordingly, while a significant age–gender interaction existed, age was a much stronger predictor of utilization than was gender.
Figure 2illustrates age/gender-specific survival rates after adjusting for baseline case-mix, attending physician and admitting hospital characteristics. Among younger patients, women had lower five-year adjusted survival rates than men; among older patients, women had higher five-year survival rates than men. The interaction between age and gender was significant. Specifically, for every 10 years of increasing age, the relative hazard change of death in women as compared with men decreased by 14.2% (95% confidence interval, 10.1 to 17.5, p < 0.001). Moreover, the relationship between age, gender and mortality was similar between those who received coronary angiography or cardiac specialty care and those who did not. Absolute differences in survival across age groups were greater than the absolute differences in survival between men and women (Fig. 2). Accordingly, while the age–gender mortality interaction was consistent across subgroups, age again was a stronger predictor of mortality than was gender. In sum, the relationship between age–gender and process of care factors was discordant to that between age–gender and outcomes. Survival differences in women improved with increasing age despite receiving less aggressive interventions.
Our study demonstrates a paradoxical age–gender interaction for long-term treatments and outcomes for patients hospitalized with AMI in Ontario. The aggressiveness of physician referrals for coronary angiography and follow-up specialty care diminished with increasing age, but did so faster among women than among men after AMI. In contrast, survival rates in women relative to men improved with increasing age at presentation.
The age–gender treatment interaction
Many cardiovascular studies have demonstrated less aggressive diagnostic and management strategies in women relative to men. Researchers have referred to such differences as a reflection of a “gender bias” (4,6). Although our results do support the presence of gender-specific treatment differences for two key markers of service intensity, the association was complex and age-dependent. Most studies have focused on gender as an isolated entity when determining its relationship to treatments, ignoring its interaction with other biological factors. The age-dependent effects of gender may account for inconsistencies in findings across studies examining for gender variations in treatments and outcomes after AMI (1–6,10,13–15,26–29). Although these studies have routinely adjusted for differences in age, few have examined whether gender effects differ according to the age at which patients initially present with AMI (13–15). Our results would suggest that researchers and clinicians who draw inferences on “overall” gender-treatment effects must do so with caution. Moreover, given that age was a more important determinant of process and outcomes of care than gender, treatment differences in men and women may largely reflect an issue of “age bias” rather than “gender bias,” especially when considering that women present with AMI at considerably older ages than do men.
The higher rate of interventions among younger women observed in this study may be appropriate given their higher relative rates of congestive heart failure, shock and diabetes as compared with similarly aged men. While illness severity appears to be the driving determinant of physician aggressiveness among patients in younger age groups, advancing age leads to reduced intensity of care, particularly among women. A conservative approach to treatment among older women may reflect the perception that elderly women incur greater risk but fewer benefits when subjected to cardiac interventions compared with elderly men (1,2,7,10,30–32).
Perceptions of risk-benefit tradeoffs may also reflect service supply or system capacity. When confronted with limited capacity for specialized cardiac services, physicians become more stringent when selecting patients for discretionary interventions (33). Thus, any age or gender effects on physician decision-making might be more evident in a jurisdiction such as Ontario given the associated limited availability of coronary angiography and specialty cardiac physicians (23).
The age–gender mortality interaction
The age–gender mortality interaction noted in our study confirms the findings of at least one recent study that examined 8,277 patients hospitalized in Worcester, Massachusetts, between 1975 and 1995 (14). Despite differences in the U.S. and Canadian health care systems and variation in time periods examined between the two cohorts, the relative changes in magnitude for mortality with increasing age in women relative to men were strikingly similar between the two cohorts (decreases of 15.4% and 14.2% in the relative hazards of death for every 10-year increase in age among U.S. and Canadian cohorts, respectively).
In our study, the relationship between age–gender and mortality bore no relation to that between age–gender and treatments after AMI. Other studies examining gender variations in the use of in-hospital services have also failed to account for outcome differences between men and women after AMI (3,15). One recent population-based study has suggested that the age–gender mortality interaction among patients with AMI may be attributable to selection bias or a survivor effect arising from gender differences in the rates of out-of-hospital deaths before AMI presentation (34). Whether or not that finding can be confirmed, the interactions demonstrated here between age, gender, service patterns and outcomes support the hypothesis that intrinsic biological and/or psychosocial factors are more likely to explain gender-related outcome differences after MI than gender bias in the care of post-AMI patients (35–39).
Our study focused on only two process-of-care measures. However, several studies have demonstrated a positive relationship between specialized cardiac service intensity and outcomes after AMI (17,40,41). Accordingly, we believe these measures serve as a reasonable test case for a relationship between process and outcomes of care. Second, our data lacked clinical detail. Accordingly, we had no information directly indicating the initial infarct location, hemodynamic status and degree of impairment of left ventricular function. While we did adjust for a number of clinical factors, it is possible that the inclusion of additional clinical variables might have altered our results. Third, we examined patients admitted to Ontario hospitals between 1992 and 1993. We did so to ensure a minimum of five-year follow-up on all patients, but it is uncertain whether the observed practice patterns would be the same today. Finally, without studying the perspectives of physicians, it is not possible to explain the observed interaction between age and gender in determining treatments after AMI. Nonetheless, these limitations must be traded off against the comprehensiveness of a very large population-based AMI cohort that is highly representative of the Canadian population.
Gender differences in treatments and outcomes after AMI are highly complex phenomena. The relationships between gender, patterns of practice and long-term outcomes vary according to age. Gender differences in treatments strongly correlated with age differences of the relevant patient subgroups. Long-term outcome differences between men and women are more likely explained by intrinsic biological factors rather than by variations in processes of care after AMI.
☆ Supported by an operating grant from the CIHR. The Institute for Clinical Evaluative Sciences is supported, in part, by a grant from the Ontario Ministry of Health. The results, conclusions, and opinions are those of the authors, and no endorsement by the Ministry, the Institute, or the CIHR is intended or should be inferred. Dr. Tu is supported by a Canada Research Chair in Health Services Research. Dr. Alter is a New Investigator at the Canadian Institutes of Health Research.
- acute myocardial infarction
- Canadian Institute for Health Information database
- Ontario Myocardial Infarction Database
- Received November 21, 2001.
- Revision received March 11, 2002.
- Accepted March 29, 2002.
- American College of Cardiology Foundation
- Bell M.R.,
- Berger P.B.,
- Holmes D.R.J.,
- Mullany C.J.,
- Bailey K.R.,
- Gersh B.J.
- Pettersen K.I.
- Tu J.V.,
- Naylor C.D.,
- Austin P.
- Tu J.V.,
- Austin P.,
- Naylor C.D.,
- Iron K.,
- Zhang H.
- Tu J.V.,
- Austin P.,
- Walld R.,
- Roos L.L.,
- Aguglia F.,
- McDonough P.
- Alter D.A.,
- Austin P.,
- Tu J.V.
- Fox J.
- Roeters vL.J.,
- Zwinderman A.H.,
- Roeters vL.H.,
- et al.
- Kostis J.B.,
- Wilson A.C.,
- O’Dowd K.,
- et al.
- MacIntyre K.,
- Stewart S.,
- Capewell S.,
- et al.
- Mennen L.I.,
- Balkau B.,
- Charles M.A.,
- D’Hour A.,
- le Mauff J.M.
- Billing E.,
- Hjemdahl P.,
- Rehnqvist N.