This Article Full Text (PDF) References Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Saved Citations Download to citation manager Request Permissions Request Reprints Add to My Marked Citations Citing Articles Citing Articles via Google Scholar Citing Articles via Scopus Google Scholar Articles by To, T. Articles by Goel, V. Search for Related Content PubMed PubMed Citation Articles by To, T. Articles by Goel, V. Social Bookmarking                         What's this?

Comparison of methods to identify outliers observed in health services small area variation studies

Population Health Sciences, Research Institute, Hospital for Sick Children, Toronto, Canada; Institute for Clinical Evaluative Sciences in Ontario and Clinical Epidemiology Unit, Sunnybrook Health Science Centre, North York, Ontario, Canada; Department of Public Health Sciences, University of Toronto, Ontario, Canada, teresa.to{at}sickkids.ca

J Ivan Williams

Institute for Clinical Evaluative Sciences in Ontario and Clinical Epidemiology Unit, Sunnybrook Health Science Centre, North York, Ontario, Canada; Department of Public Health Sciences, University of Toronto, Ontario, Canada

Keyi Wu

Institute for Clinical Evaluative Sciences in Ontario and Clinical Epidemiology Unit, Sunnybrook Health Science Centre, North York, Ontario, Canada

Marc-Erick Theriault

Institute for Clinical Evaluative Sciences in Ontario and Clinical Epidemiology Unit, Sunnybrook Health Science Centre, North York, Ontario, Canada

Vivek Goel

Institute for Clinical Evaluative Sciences in Ontario and Clinical Epidemiology Unit, Sunnybrook Health Science Centre, North York, Ontario, Canada; Department of Health Administration, University of Toronto, Ontario, Canada

Small area variation analysis (SAV) is an established methodology in health services and epidemiological research. The goal is to demonstrate that rates differ across areas, and to explain these differences by differences in physician practice styles or patient characteristics. While the SAV statistics provide an overall variation estimate, they do not provide a statistical means to identify significant outliers. We compared the chi-square (²) test with three approaches in determining significant outliers in SAV. We used data from the Canadian Institute for Health Information (CIHI) for Ontario residents discharged between 1989 and 1991. Coronary artery bypass surgery, hysterectomy and hip replacement data were used to compare four statistics in determining outliers: the ² test, Swift’s approximate bootstrap confidence interval (ABC), Carriere’s T²(T²) with simultaneous confidence intervals (SCI), and Gentleman’s normalized scores (GNS). Both the ABC and SCI correct the skewness of the distribution of the adjusted rates. With large data, confidence intervals calculated by the normal or the ABC methods are indistinguishable. The T² can be applied to also nonbinary events. For binary events, it is asymptotically the same as the ². The GNS ranks the rates, but the distribution of these ranks does not differ significantly from that of the adjusted rates. We concluded that when using large data with binary events, there is little advantage in using the ABC, SCI or GNS over the commonly known ². The ² remains a useful tool in small area variation analysis to ‘screen’ or ‘ag potential differences beyond chance alone.

Statistical Methods in Medical Research, Vol. 12, No. 6, 531-546 (2003)
DOI: 10.1191/0962280203sm350oa


CiteULike    Complore    Connotea    Del.icio.us    Digg    Reddit    Technorati    Twitter    What's this?