Review Of: Shen YC, Hsia RY. Association between ambulance diversion and survival among patients with acute myocardial infarction. JAMA. 2011;305(23): 2440–2447.
This study comes out of four densely populated California counties primarily urban in nature, and it attempts to study the effect of hospital diversion on patients suffering from an acute myocardial infarction (AMI). The population of interest were Medicare recepients within these counties, who experienced an AMI between 2000 and 2006. The investigators broke down the amount of diversion into the following categories: no diversion, <6 hours, 6 to <12, and 12 hours. Mortality was then measured at seven days, 30 days, 90 days, nine months, and one year. The study found that Medicare patients experiencing an AMI with exposure to greater than 12 hours of diversion were associated with increased mortality at 30 days, 90 days, nine months and one year.
Medic Marshall: I’m a little torn on this study because it’s not specifically aimed at EMS. The authors lumped in all patients that presented to the emergency department (ED), so that includes all the patients having a myocardial infarction (MI) that drove themselves in. This may negatively impact mortality, for example a patient who walks into a very busy triage with atypical signs and symptoms of an MI may not receive immediate attention if they are poorly triaged. So of course you would expect those patients not to fare as well as those with immediate evaluation.
Though I don’t often run into hospital diversion on the streets, it can be annoying when I do. Letting the patient know they can’t go to the hospital of their choice and potentially turning completely around is not ideal. Plus if the patient is “sick,” it makes your job that much more difficult—especially if you have to decide to take the patient to a specialty resource center or the nearest appropriate facility. But what else can you do?
Personally, I don’t believe in hospital diversion. I find it to be a symptom of a greater problem with our healthcare system. Overcrowding is a problem in EDs; I understand that, but there needs to be more of an effort to fix this. Alternative destinations and patient follow-up with clinics and their primary physicians are means to reduce the number of patients presenting to EDs for primary care. But until the bigger issues are addressed, we’re just going to have to deal with “diversion.”
Doc Wesley: I agree and disagree with Marshall on this one. First, this paper confirms what other papers and systems have seen. An “association” between diversion and the mortality of our patients exists. But the “cause” is not explained by this or other studies. To understand this, we have to look closely at this study. It suffers from some significant “selection bias” because it doesn’t differentiate patients who arrived by ambulance and those who walked into the ED.
Additionally, the study excluded 25% of the Medicare patients who were admitted directly to the hospital bypassing the ED. Who were these patients? Were they transfers from over-crowded hospitals? If so, it would be interesting to see if their outcomes were better or worse than those admitted through the ED. Not knowing how these four counties perform EMS ST-elevation myocardial infarction (STEMI) alerts further compounds the issue. In many systems the medics activate the catheterization lab and bypass the ED and take the patient directly to the catheterization lab where they’re then admitted to the hospital. It’s reasonable to assume those patients have better outcomes. Does this study represent part of that 25% that were excluded? The authors don’t tell us.
Secondly, this study examined only Medicare patients, who based on age, represent only 50–60% of AMI patients. Do younger patient receive the same effect of diversion as older patients?
Third, if you look closely at the data, the group of patients exposed to >12 of diversion that had the greatest effect on mortality were black. Perhaps this is an incrimination of the lower quality of healthcare available to this socioeconomic population.
Without knowing specifically the impact on EMS delivered patients, I can’t make any policy decisions as it relates to diversion. Perhaps a prohibition of diversion would simply exacerbate overcrowding and further worsen the outcomes of the patients who drive themselves to the ED.
Finally, this study didn’t differentiate diversion status of hospitals by those that were percutaneous coronary intervention (PCI) capable and those that were not. It clearly demonstrated that patients not receiving PCI had a higher mortality, but this simply confirms what we already know about the superiority of PCI. It’s a thought-provoking article and one that I’m sure will lead to a closer examination of this significant issue.
Shen YC, Hsia RY. JAMA. 2011;305(23):2440–2447. Epub 2011 Jun 12.
Association between ambulance diversion and survival among patients with acute myocardial infarction.
Ambulance diversion, a practice in which emergency departments (EDs) are temporarily closed to ambulance traffic, might be problematic for patients experiencing time-sensitive conditions, such as acute myocardial infarction (AMI). However, there is little empirical evidence to show whether diversion is associated with worse patient outcomes.
To analyze whether temporary ED closure on the day a patient experiences AMI, as measured by ambulance diversion hours of the nearest ED, is associated with increased mortality rates among patients with AMI. DESIGN, STUDY, AND PARTICIPANTS: A case-crossover design of 13,860 Medicare patients with AMI from 508 zip codes within four California counties (Los Angeles, San Francisco, San Mateo, and Santa Clara) whose admission date was between 2000 and 2005. Data included 100% Medicare claims data that covered admissions between 2000 and 2005, linked with date of death until 2006, and daily ambulance diversion logs from the same 4 counties. Among the hospital universe, 149 EDs were identified as the nearest ED to these patients.
Main Outcome Measures:
The percentage of patients with AMI who died within 7 days, 30 days, 90 days, 9 months, and 1 year from admission (when their nearest ED was not on diversion and when that same ED was exposed to <6, 6 to <12, and ≥12 hours of diversion out of 24 hours on the day of admission).
Between 2000 and 2006, the mean (SD) daily diversion duration was 7.9 (6.1) hours. Based on analysis of 11,625 patients admitted to the ED between 2000 and 2005, and whose nearest ED had at least 3 diversion exposure levels (3541, 3357, 2667, and 2060 patients for no exposure, exposure to <6, 6 to <12, and ≥12 hours of diversion, respectively), there were no statistically significant differences in mortality rates between no diversion and exposure to less than 12 hours of diversion. Exposure to 12 or more hours of diversion was associated with higher 30-day mortality vs. no diversion status (unadjusted mortality rate, 392 patients [19%] vs. 545 patients [15%]; regression adjusted difference, 3.24 percentage points; 95% confidence interval [CI], 0.60-5.88); higher 90-day mortality (537 patients [26%] vs. 762 patients [22%]; 2.89 percentage points; 95% CI, 0.13-5.64); higher 9-month mortality (680 patients [33%] vs. 980 patients [28%]; 2.93 percentage points; 95% CI, 0.15-5.71); and higher 1-year mortality (731 patients [35%] vs. 1034 patients [29%]; 3.04 percentage points; 95% CI, 0.33-5.75).
Among Medicare patients with AMI in four populous California counties, exposure to at least 12 hours of diversion by the nearest ED was associated with increased 30-day, 90-day, 9-month, and 1-year mortality.