Jump to comment:
- Page navigation anchor for RE: association versus causationRE: association versus causation
The authors are to be congratulated on an important study. They conclude that "a province-wide program to promote team-based primary care is a feasible and effective way to reduce use of emergency departments" which implies a causal association between Primary Care Networks and reductions in emergency department use. Their study was cross-sectional and although it controlled for a number of relevant factors it did not examine pre-existing patterns of emergency department use.
In a previous study (http://www.cmaj.ca/content/180/11/E72) we found substantial differences in emergency department use between physician payment models in Ontario but on examination of an earlier time period before the models were implemented we found the patterns to be pre-existing. Self-selection of physicians into particular models of care appears to involve important differences in physician and patient characteristics, not all of which might be measurable using existing data (https://www.ncbi.nlm.nih.gov/pubmed/26190516).
Without a longitudinal design that accounts for baseline patterns of care, conclusions about causation may not be warranted.Competing Interests: None declared.