Re: BC physicians reduce unnecessary antibiotic use—and costs

The BC Centre for Disease Control article by Dr David Patrick and colleagues [BCMJ 2016;58:501,503] proposes an association between antibiotic cost reduction in British Columbia and Ministry of Health–funded education programs. Whereas the potential for such an association raises hopes for future gains in the same endeavors, it is questionable as to whether the science is supported by the data. Essentially, the authors examine the costs over a 10-year period and assume that a reduction in costs over the last 3 years of that term is indicative of an association with the educational program.

Such unidimensional analysis is fraught with difficulty and engenders skepticism. If one had conducted such an analysis in 2011, an abstraction from their formative processes could have suggested then that there was a net increase in costs due to the educational program. As epidemiologists, scientists, and physicians, the authors would know that any such analysis of the educational program and the cost of antibiotics should be multidimensional, and a complex multivariate analysis would be more logical to determine putative cause and effect. For example, other major changes in health care and epidemiology would need to be factored into any such analysis. The plague of community-associated methicillin-resistant Staphylococcus aureus infections in the general populace has dramatically declined over the last 3 or more years. The pandemic of H1N1 influenza of 2009–2010 is likely to have affected antibiotic use considerably and then to have subsided in the last 3 or more years. Declines in antibiotic prescription costs due to generic pricing do not necessarily have a direct inverse impact against a rising provincial population number. In such a casual analysis, one can only observe, at best, that the educational program’s continuation was associated with variable changes in antibiotic costs over a 10-year period. Analyses of costs and benefits are worthy but in a more complex multivariate model. Indeed, it would be worth knowing what costs the educational program had for the province to reach the provisional analysis thus far.
—Nevio Cimolai, MD, FRCPC
Department of Pathology and Laboratory Medicine
UBC Faculty of Medicine