Author + information
- Gagan Kumar, MD,
- Abhishek Deshmukh, MD,
- Ankit Sakhuja, MD,
- Amit Taneja, MD,
- Nilay Kumar, MD,
- Elizabeth Jacobs, MD,
- Rahul Nanchal, MD∗ (, )
- Milwaukee Initiative in Critical Care Outcomes Research (MICCOR) Group of Investigators
- ↵∗Division of Pulmonary and Critical Care, Medical College of Wisconsin, 9200 West Wisconsin Avenue, Milwaukee, Wisconsin 53226
Previous work has shown that persons admitted over the weekend for certain time-sensitive acute conditions, including acute myocardial infarction (AMI), have increased mortality risk compared with similar counterparts admitted on weekdays (1). This excess mortality risk for AMI has been partly attributed to lower rates of revascularization procedures performed over the weekend, although these studies were performed when adherence to the benchmarks for timeliness of diagnoses and therapy were not widespread (1).
We used the Healthcare Cost and Utilization Project Nationwide Inpatient Sample (NIS), a large, publicly-available database regarding inpatient care in the United States, to examine whether the weekend effect for AMI had diminished over time (2). We examined 3 distinct epochs (2000 to 2002, 2003 to 2005, and 2006 to 2008) and sought to describe changes in the utilization of revascularization procedures and their impact on weekend mortality.
We included patients age ≥18 years, discharged with a primary diagnosis of AMI (International Classification of Diseases-Ninth Revision-Clinical Modification code 410.x1) (3). We excluded patients with subendocardial myocardial infarction (International Classification of Diseases-Ninth Revision-Clinical Modification code 410.71). We defined weekend admissions as those occurring between 12:01 am Saturday through 11:59 pm on Sunday and considered all other admissions to be weekday admissions (2). Our primary outcome was in-hospital mortality. We also investigated the frequency and timing of revascularization procedures.
We constructed multivariable models to determine the independent association of patient characteristics with in-hospital mortality. We used the propensity score, which included factors that would affect the likelihood of being admitted over the weekend. Finally, we forced the revascularization procedure performed into our previously-developed model to determine whether the weekend effect reflected the differential use of this procedure.
There were an estimated 2,288,392 discharges with a primary diagnosis of AMI from 2000 to 2008. Of these, 1,691,583 (73.9%) were admitted on weekdays and 596,809 (26.1%) were admitted on weekends. The overall rates of cardiac catheterization rose from 60.7% in 2000 to 82.6% in 2008, and a larger proportion of these procedures were performed on the day of admission during 2006 to 2008 than the other 2 earlier time periods.
During the time periods of 2000 to 2002 and 2003 to 2005, a significantly larger proportion of persons admitted on weekdays received cardiac catheterization on the day of hospital admission as compared with weekends. This weekend-weekday difference declined over each time period, and by 2006 to 2008, this difference disappeared (Figure 1).
Overall in-hospital mortality for persons admitted with AMI declined significantly (9.4% in 2000 vs. 7.1% in 2008). We observed this decline in persons admitted over weekends as well as weekdays (Figure 1). Despite this decline, persons admitted over weekends continued to have significantly higher mortality as compared with those admitted over weekdays until 2005. From 2006 onwards, this difference was not detectable. Adjusted in-hospital mortality was significantly higher for weekend admissions for the time periods of 2000 to 2002 (odds ratio [OR]: 1.10; 95% confidence interval [CI]: 1.05 to 1.15) and 2003 to 2005 (OR: 1.11; 95% CI: 1.05 to 1.17), whereas during 2006 to 2008, it was not significant (OR: 1.02; 95% CI: 0.97 to 1.09).
For the time periods 2000 to 2002 and 2003 to 2005, when we forced the cardiac catheterization into the model, the mortality risk on weekends declined and became similar to that on weekdays. This suggests that the disparity in mortality observed on weekends was secondary to the disproportionately lower use of cardiac catheterization on weekends. Our findings are remarkably similar to those of Kostis et al. (1), as well as Magid et al. (4), who also demonstrated weekend-weekday mortality differences for AMI, which disappeared after we controlled for revascularization procedures.
The adjusted odds of receiving cardiac catheterization on weekends were significantly lower for the time periods 2000 to 2002 (adjusted OR: 0.70; 95% CI: 0.68 to 0.72; p < 0.001) and 2003 to 2005 (adjusted OR: 0.79; 95% CI: 0.77 to 0.81; p < 0.001) when compared with weekends during 2006 to 2008 (adjusted OR: 0.88; 95% CI: 0.85 to 0.91; p < 0.001) (interaction p value for both time periods <0.01). Changes over time in rates of cardiac catheterization likely reflect the more liberal use of primary PCI and better adherence to guideline recommendations due to public reporting and government oversight of compliance with core measures.
Limitations of our study include varying coding practices across U.S. hospitals, although these are unlikely to differ between weekdays and weekends in a given hospital. We do not have data for time of onset of chest pain, medications used, electrocardiographic findings, cardiac biomarkers, and coronary angiography findings due to the administrative nature of the NIS database. If available, these details would have allowed us to adjust for confounders more robustly. Inability to determine the cause of death also limits our interpretation of results. Therefore, despite multivariable adjustment, we cannot exclude unmeasured confounders as a cause for our results. Finally, the time to procedure is coded in days and not in hours or minutes. This has significance, as door-to-balloon times are typically quoted in minutes.
Despite these limitations, we demonstrate the disappearance of the weekend effect for AMI. This has important implications for other time-sensitive disease processes, such as pulmonary embolism, where the weekend effect persists (5). Health care systems should strive to emulate processes for AMI that have led to persons receiving equivalent care, regardless of the day of admission.
- American College of Cardiology Foundation