Category Archives: Medicine

Technical Note: Interaction and Confounding in Data Analysis

Suppose that we conduct a study to assess if marathon finishing times are associated with diet while controlling for age (this example was inspired by Gretchen Reynolds recent NYT article).  The extraneous variable is age.  Can we ignore age in our analysis and still assess the association between finishing time and diet?

Confounding exists when different interpretations of the relationship between finishing time and diet exist when age is ignored or included in the data analysis.  The assessment requires a comparison between crude estimate of association and an adjusted estimate of association.

Interaction exists when the relationship between finishing time and diet is different for different age groups.  The assessment requires describing the relationship between finishing time and diet for different age groups.

Interaction and confounding can exist in the same data set.  A variable can be a confounding variable and might also have an interaction.  If a strong interaction is found then an adjustment for confounding is inappropriate.

Quality of Toronto Hospitals

Last month (April 2012) The Canadian Institute for Health Information (CIHI) released an online database that compares several patient outcomes for many Canadian hospitals.

A visual examination of eleven indicators quickly reveals how adult hospitals within central Toronto, Scarborough, Mississauga compared to each other.  We grouped Academic and community hospitals together and display the data using side-by-side heat maps.

Toronto Area Hospital QualityThe darker a blue box the worse patient outcomes were for that hospital.

The heat map on the left compares rates per 1000 patients and the heat map on the right compares outcomes per 100 patients.

Academic hospitals report a median of twelve percent more nursing adverse events(*) for surgical patients and seventeen percent more nursing adverse events for medical patients.  The academic hospitals have higher rates compared to the provincial average while the community hospitals are slightly below this benchmark.

(*) A nursing adverse event is either a: urinary tract infection; pressure ulcer; in-hospital fracture; or pneumonia.