Acute myocardial infarctionBMJ 2014; 348 doi: https://doi.org/10.1136/bmj.f7696 (Published 21 January 2014) Cite this as: BMJ 2014;348:f7696
- Lauren Lapointe-Shaw, fellow,
- Chaim M Bell, associate professor
Acute myocardial infarction has high mortality, but early medical and surgical intervention can be lifesaving.1 2 3 4 5 6 Previous studies have shown that the time of day or day of the week when patients seek care can affect outcomes.5 7 8 In most of these studies, patients presenting to hospital with an acute myocardial infarction during off-hours (evenings and weekends) wait longer for interventional treatments than those presenting during regular office hours and have a higher mortality. In a linked paper (doi:10.1136/bmj.f7393), Sorita and colleagues report the first systematic review of the effect of off-hour presentation on outcomes after acute myocardial infarction.9
The authors evaluated the literature on acute myocardial infarction and off-hour care. Outcomes included in-hospital and 30 day mortality, as well as door to balloon time for the subset of patients with ST elevation myocardial infarction. Using a random effects model, they reported pooled odds ratios for each outcome measure. The pooled results confirmed the presence of a 5% relative increase in mortality (both in-hospital and 30 day) as well as a delay of nearly 15 minutes in door to balloon time for patients presenting during off-hours. Meta-regression based on year of data showed an increase in the risk posed by off-hours care over time.
This novel systematic review advances knowledge on quality of care for patients with myocardial infarction, although it is limited by the studies it contains. In the absence of randomization, differences in patient characteristics between compared groups can introduce substantial bias into study results. Because patients cannot be randomized to present during or outside working hours, a common method of adjusting for baseline risk is needed to facilitate meaningful comparison between studies. If the included studies use different methods to control bias, heterogeneity is increased, which limits the conclusions that can be drawn from pooled analyses.10 These are important considerations when undertaking any systematic review of observational studies.
The authors were further challenged by clinical and statistical heterogeneity. The definition of the off-hour time period differed across studies, and varying geographical settings are likely to lead to differences in case mix, time to presentation, physician practices, and hospital characteristics. Such heterogeneity makes it difficult to pool study results and generate a single measure of relative risk. Publication bias, as demonstrated by the absence of small negative studies in the funnel plot, may complicate interpretation still further, although, as the authors point out, there was no significant change in the pooled effect of off-hours presentation after accounting for the missing studies.
Confounding is always a problem in syntheses of observational studies. In Sorita and colleagues analysis it is particularly important to consider whether patients presenting out of hours are systematically different from other patients in such a way that increases their risk of death. They might be sicker, for example, or they may delay calling for medical help for longer. If the last case were true, then delayed presentation would lead to delayed treatment and potentially worse outcomes, which would have little to do with the quality of off-hour care. As the authors point out, the results as to whether time to presentation (delay before reaching hospital) differs significantly between patients presenting during off-hours and working hours are conflicting. If delay in presentation differs between groups, this could bias the measured relative mortality associated with off-hour care.
Although differences in underlying patient characteristics, including time to presentation, can significantly affect mortality, it is less clear how they would affect door to balloon time. Prolongation of door to balloon time is arguably a more robust measure of altered care during off-hours, because it is more likely to be directly controlled by the hospital and care providers. In this case, a process measure (door to balloon time) truly enhances the interpretation of an outcome measure (mortality), albeit for a subgroup of patients. The nearly 15 minute delay in percutaneous coronary intervention experienced by patients presenting with ST elevation myocardial infarction during off-hours provides a potentially causal link between the quality of off-hour care and patient outcomes.
Patients presenting during off-hours experience delays in urgent care and worse outcomes, and the gap seems to be increasing over time. As healthcare managers in many countries move toward performance based remuneration, patient outcomes are increasingly being used to gauge the quality of hospital care. Managers seeking to boost their hospital’s performance for patients with acute myocardial infarction should focus on improving their off-hour care, with the goal of providing consistently high quality care 24 hours a day and seven days a week.
Studies of quality of care and patient outcomes highlight the challenges we face when trying to measure true hospital performance. Administrative data often do not capture all the factors that contribute to baseline patient risk. To properly evaluate the quality of healthcare delivered at all times, we must refine our methods of risk adjustment to include time to presentation and severity of illness. Future studies should try to identify specific deficits in the care pathway during off-hours, allowing differences in outcomes to be linked to differences in processes. We look forward to reading about innovative strategies to deal with this problem. Patients deserve the best possible outcome, at any given time, and on any given day.
Cite this as: BMJ 2014;348:f7696
Competing interests: We have read and understood the BMJ Group policy on declaration of interests and declare the following interests: None.
Provenance and peer review: Commissioned; not externally peer reviewed.