Multiple Comparisons in Variation of Care Research
Main Article Content
Abstract
Research in hospital variation is important and currently very popular. However, due to the methods employed in such studies—namely, the retrospective mining of large datasets and the use of several alternative variation groupings—some results may be spurious. In this commentary, we perform an empirical analysis of the 50 most highly cited and the 50 most recent papers focusing on variation in medical care. Across these studies, we identify at least 13 unique groupings and could find no single instance where a medical practice was found not to vary. We go on to discuss one example of variation—statin use—in more detail to elucidate the tensions that these studies often create. Together, these results suggest that multiple hypothesis testing is a concern for variation research. Finally, we outline strategies to mitigate this concern.
RÉSUMÉ
La recherche sur la variation hospitalière est importante et actuellement très populaire. Toutefois, en raison des méthodes employées dans de telles études—notamment, l’extraction rétrospective de grands ensembles de données et l’utilisation de plusieurs groupe- ments de variation alternatifs—certains résultats peuvent être fautifs. Dans ce commentaire, nous effectuons une analyse empirique des 50 articles les plus citées et des 50 articles les plus récents se concentrant sur la variation dans les soins médicaux. Dans ces études, nous identifions au moins 13 groupements uniques, et ne pouvions trouver aucun cas où une pratique médicale ne variait pas. Nous discutons ensuite d’un exemple de variation—dans l’utilisation de statines—en plus de détails afin d’élucider les tensions que ces études suscitent souvent. Collectivement, ces résultats suggèrent que la mise à l’essai de multiples hypothèses est une préoccupation lors de la recherche sur la variation. Finalement, nous décrivons des stratégies pour atténuer cette préoccupation.
Article Details
- Authors publishing in the UOJM retain copyright of their articles, including all the drafts and the final published version in the journal.
- While UOJM does not retain any rights to the articles submitted, by agreeing to publish in UOJM, authors are granting the journal right of first publication and distribution rights of their articles.
- Authors are free to submit their works to other publications, including journals, institutional repositories or books, with an acknowledgment of its initial publication in UOJM.
- Copies of UOJM are distributed both in print and online, and all materials will be publicly available online. The journal holds no legal responsibility as to how these materials will be used by the public.
- Please ensure that all authors, co-authors and investigators have read and agree to these terms.
- Works are licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
References
2. Ioannidis JA. The importance of potential studies that have not existed and registration of observational data sets. JAMA. 2012;308(6):575-6.
3. Prasad V, Gall V, Cifu A. The frequency of medical reversal. Arch Intern Med. 2011;171(18):1675-6.
4. Fellström BC, Jardine AG, Schmieder RE, et al. Rosuvastatin and cardiovascular events in patients undergoing hemodialysis. N Engl J Med. 2009;360(14):1395-407.
5. Wanner C, Krane V, März W, et al. Atorvastatin in patients with type 2 dia- betes mellitus undergoing hemodialysis. N Engl J Med. 2005;353(3):238-48.
6. Baigent C, Landray MJ, Reith C, et al. The effects of lowering LDL cholesterol with simvastatin plus ezetimibe in patients with chronic kidney disease (Study of Heart and Renal Protection): a randomised placebo-controlled trial. Lancet. 2011;377(9784):2181-92.
7. Wetmore JB, Mahnken JD, Mukhopadhyay P, et al. Geographic variation in HMG-CoA reductase inhibitor use in dialysis patients. J Gen Intern Med. 2012;27(11):1475-83.
8. Krumholz HM. Variations in health care, patient preferences, and high-quality decision making. JAMA. 2013;310(2):151-2.