EPI204 Forum

Dear colleagues,

Although I do not see myself as a defender of the specialty of cardiology, nor as any sort of expert on this data or epidemiology in general, I do think that there may exist alternative interpretations for the results of Jena et al. (Mortality During Dates of National Cardiology Meetings). Specifically, the finding that mortality rates decreased during the meetings may be explained by factors other than those suggested by the authors, even though in their discussion the do present multiple possible explanations. This list is not exhaustive.

1. It is unclear whether there is a decrease in mortality associated with the meetings or an increase in mortality associated with the time before and after the meetings. The meeting days (where a decreased mortality was shown) were compared to non-meeting days in the weeks before and after the conferences. Although we cannot know for certain without comparison to days completely isolated from the meeting, it is possible that the true lesion is an increase in mortality in the weeks before and after a meeting due to the stresses of leaving to go to the meetings (catching up on work, preparing presentations, etc.) rather than a decrease on the actual meeting days. The authors acknowledge this limitation and have agreed they should pursue this analysis.
2. The assertion that decreased utilization of cardiology procedures is the mediator of the decreased mortality does not follow from the data. Although treatment utilization (specifically invasive procedures) among patients admitted to teaching hospitals during meeting dates was significantly less for high risk heart attack patients (for whom there was no difference in adjusted 30-day mortality), there was no decrease in procedures for either high risk heart failure patients or for cardiac arrest patients (for whom there was a decrease in 30 day mortality for those admitted during a cardiology meting). Alternatively, for cardiac arrest patients, those admitted during meeting days had adjusted hospital charges that trended down (mean difference $26,000, p=0.07), not up.
3. Many of the beneficial effects of cardiovascular intervention have been demonstrated with immediate to long-term outcomes (longer than the 30-day mortality looked at in this trial) and they are known to be associated with increased risk of 30 day outcomes (the outcomes looked at in this trial), particularly in high risk patients (they are high risk for mortality and for bad things happening when you try to save them). In the supplemental data for this manuscript they show that the significant difference in 30-day mortality was not apparent when looking at 90-day mortality (the differences were no longer statistically significant).