Back

Effect of hospital assignment on mortality for AMI patients

Groesland, M.; Telle, K. E.; Oeien, H.

2022-07-10 health economics
10.1101/2022.07.08.22277206 medRxiv
Show abstract

International guidelines recommend percutaneous coronary intervention (PCI) to treat acute myocardial infarction (AMI) if PCI can be performed within two hours. PCI is a centralized treatment, and therefore a common trade-off is whether to send AMI patients directly to a hospital that performs PCI, or postponing a potential PCI-treatment by first receiving acute treatment at a non-PCI-hospital. In this paper, we estimate the effect of sending patients directly to a PCI-hospital on AMI mortality. Since the underlying health of patients may affect both hospital assignment and mortality, estimates from traditional multivariate risk adjustment models are likely biased. We therefore apply an instrumental variable (IV) model using the historical municipal share that is sent directly to a PCI-hospital as an instrument for being sent directly to a PCI-hospital. Our results disclose that patients sent directly to a PCI-hospital are younger and have fewer comorbidities than patients who are first sent to a non-PCI-hospital. Our IV results suggest that those initially sent to PCI-hospitals have 4.8 percentage points decrease (95% CI (-18.1)-8.5) in mortality after one month compared to those initially sent to non-PCI-hospitals. However, the estimates are too imprecise to conclude that health personnel should change their practice.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.