Victoria C Weston1, William J Meurer2, Shirley M Frederiksen3, Allison K Fox4, Phillip A Scott5. 1. Department of Emergency Medicine, Northwestern University Feinberg School of Medicine, Chicago, IL; Department of Emergency Medicine, University of Michigan, Ann Arbor, MI. 2. Department of Emergency Medicine, University of Michigan, Ann Arbor, MI; Stroke Program, University of Michigan, Ann Arbor, MI. 3. Department of Emergency Medicine, University of Michigan, Ann Arbor, MI. 4. Department of Emergency Medicine, University of Michigan, Ann Arbor, MI; Oakland University William Beaumont School of Medicine. 5. Department of Emergency Medicine, University of Michigan, Ann Arbor, MI; Stroke Program, University of Michigan, Ann Arbor, MI. Electronic address: phlsctt@med.umich.edu.
Abstract
OBJECTIVES: Cluster randomized trials (CRTs) are increasingly used to evaluate quality improvement interventions aimed at health care providers. In trials testing emergency department (ED) interventions, migration of emergency physicians (EPs) between hospitals is an important concern, as contamination may affect both internal and external validity. We hypothesized that geographically isolating EDs would prevent migratory contamination in a CRT designed to increase ED delivery of tissue plasminogen activator (tPA) in stroke (the INSTINCT trial). METHODS: INSTINCT was a prospective, cluster randomized, controlled trial. Twenty-four Michigan community hospitals were randomly selected in matched pairs for study. Contamination was defined at the cluster level, with substantial contamination defined a priori as greater than 10% of EPs affected. Nonadherence, total crossover (contamination+nonadherence), migration distance, and characteristics were determined. RESULTS: Three hundred seven EPs were identified at all sites. Overall, 7 (2.3%) changed study sites. One moved between control sites, leaving 6 (2.0%) total crossovers. Of these, 2 (0.7%) moved from intervention to control (contamination); and 4 (1.3%) moved from control to intervention (nonadherence). Contamination was observed in 2 of 12 control sites, with 17% and 9% contamination of the total site EP workforce at follow-up, respectively. Average migration distance was 42 miles for all EPs moving in the study and 35 miles for EPs moving from intervention to control sites. CONCLUSION: The mobile nature of EPs should be considered in the design of quality improvement CRTs. Increased reporting of contamination in CRTs is encouraged to clarify thresholds and facilitate CRT design.
RCT Entities:
OBJECTIVES: Cluster randomized trials (CRTs) are increasingly used to evaluate quality improvement interventions aimed at health care providers. In trials testing emergency department (ED) interventions, migration of emergency physicians (EPs) between hospitals is an important concern, as contamination may affect both internal and external validity. We hypothesized that geographically isolating EDs would prevent migratory contamination in a CRT designed to increase ED delivery of tissue plasminogen activator (tPA) in stroke (the INSTINCT trial). METHODS: INSTINCT was a prospective, cluster randomized, controlled trial. Twenty-four Michigan community hospitals were randomly selected in matched pairs for study. Contamination was defined at the cluster level, with substantial contamination defined a priori as greater than 10% of EPs affected. Nonadherence, total crossover (contamination+nonadherence), migration distance, and characteristics were determined. RESULTS: Three hundred seven EPs were identified at all sites. Overall, 7 (2.3%) changed study sites. One moved between control sites, leaving 6 (2.0%) total crossovers. Of these, 2 (0.7%) moved from intervention to control (contamination); and 4 (1.3%) moved from control to intervention (nonadherence). Contamination was observed in 2 of 12 control sites, with 17% and 9% contamination of the total site EP workforce at follow-up, respectively. Average migration distance was 42 miles for all EPs moving in the study and 35 miles for EPs moving from intervention to control sites. CONCLUSION: The mobile nature of EPs should be considered in the design of quality improvement CRTs. Increased reporting of contamination in CRTs is encouraged to clarify thresholds and facilitate CRT design.
Authors: M R Keogh-Brown; M O Bachmann; L Shepstone; C Hewitt; A Howe; C R Ramsay; F Song; J N V Miles; D J Torgerson; S Miles; D Elbourne; I Harvey; M J Campbell Journal: Health Technol Assess Date: 2007-10 Impact factor: 4.014
Authors: Erik von Elm; Douglas G Altman; Matthias Egger; Stuart J Pocock; Peter C Gøtzsche; Jan P Vandenbroucke Journal: J Clin Epidemiol Date: 2008-04 Impact factor: 6.437
Authors: Phillip A Scott; Zhenzhen Xu; William J Meurer; Shirley M Frederiksen; Mary N Haan; Michael W Westfall; Sandip U Kothari; Lewis B Morgenstern; John D Kalbfleisch Journal: Stroke Date: 2010-08-12 Impact factor: 7.914
Authors: Phillip A Scott; William J Meurer; Shirley M Frederiksen; John D Kalbfleisch; Zhenzhen Xu; Mary N Haan; Robert Silbergleit; Lewis B Morgenstern Journal: Lancet Neurol Date: 2012-12-21 Impact factor: 44.182
Authors: William J Meurer; Cemal B Sozener; Zhenzhen Xu; Shirley M Frederiksen; Allison M Kade; Michael Olgren; Sanford J Vieder; John D Kalbfleish; Phillip A Scott Journal: West J Emerg Med Date: 2013-02