Re: How can we produce relevant information for decision makers from small area variation studies?

摘要

The existence and importance of variation between organizational units or between geographical areas has been recognized in numerous studies. The results of such studies are often used as a guide for decision makers in determining appropriate interventions. It is, therefore, of major importance that such studies are carried out within an accurate methodological and conceptual framework; otherwise, the risk of producing misleading information is overwhelming. The article by Morrato et al.recently published in the Journal has, unfortunately, not been performed and interpreted within such a framework. The authors aimed to estimate glucose and lipid testing rates among second-generation antipsychotic-users from three state Medic-aid programs and investigate small area variations, as well as patient and geographic determinants of testing. Using Medicaid claims data from California, Missouri, and Oregon, they found a fourfold difference in glucose testing rates (21%-85%) and a greater than sixfold difference in lipid testing rates (0%-62%) among the 207 counties they investigated.