Institutional Variability in a Minimal Risk, Population-Based Study: Recognizing Policy Barriers to Health Services Research

Authors

  • Craig D. Newgard,

    Search for more papers by this author
    • Address correspondence to Craig D. Newgard, M.D., M.P.H., Center for Policy and Research in Emergency Medicine, Department of Emergency Medicine, Oregon Health & Science University, Portland, OR. Sai-Hung Joshua Hui, B.S., and Roger J. Lewis, M.D., Ph.D., are with the David Geffen School of Medicine at UCLA, University of California, Los Angeles, CA. Patrick Stamps-White, M.D., was previously with the University of Iowa School of Medicine, Iowa City, IA. Roger J. Lewis, M.D., Ph.D., is with the Department of Emergency Medicine Harbor-UCLA Medical Center, Torrance, CA. Craig D. Newgard, M.D., M.P.H., and Roger J. Lewis, M.D., Ph.D., are with the Los Angeles Biomedical Research Institute at Harbor-UCLA Medical Center, Torrance, CA. Dr. Stamps-White is now at William Beaumont Hospital in Royal Oak, MI.

  • Sai-Hung Joshua Hui,

  • Patrick Stamps-White,

  • Roger J. Lewis


Abstract

Objective. To describe (1) the institutional variability in study approval and willingness to obtain federal assurance documents for a federally supported, minimal risk health services research project conducted during the implementation of the Privacy Rule and federalwide assurances, and (2) the potential impact of such policy on selection of research subjects and generalizability of study results.

Data Sources. Primary data collection from 2001 to 2004.

Study Design. We provide a descriptive analysis of a prospective, observational, out-of-hospital study.

Study Setting. Twenty-seven pediatric receiving hospitals were approached for participation in a study validating a decision rule to identify seriously injured children involved in motor vehicle crashes in Los Angeles County. Two federal research policies, the Privacy Rule and the requirement for federalwide assurances, were implemented during the project.

Data Collection. All 27 hospitals were sent an identical research protocol requesting approval to review charts of children transported to their facility. The research protocol included strict confidentiality protections, was noninterventional, did not alter the standard of care at the scene or at the hospital, and met requirements for waivers of both informed consent and the Health Insurance Portability and Accountability Act. Because the project was federally supported, all participating hospitals were required to have a federalwide assurance. The hospitals receiving the research protocol were the unit of analysis and outcomes included: approval of the research protocol, total number of days to study approval, and successfully obtaining a federalwide assurance.

Principle Findings. Overall, 6 of 27 (22 percent) hospitals refused to participate in the study, all of which were community hospitals. The median time from submitting an application to study approval was 118 days (interquartile range 34–254, range 12–960 days) and time to study approval differed when hospitals were categorized by type and the presence of an institutional review board (p=.053). No institutional review resulted in a change in the basic study protocol, although one hospital required paramedic consent. Following intensive efforts to secure federalwide assurances, 12 of 27 hospitals (44 percent) possessed the necessary assurance to conduct the study. If all patients transported to hospitals that failed to obtain such an assurance were omitted, the sample size would have been reduced by 62 percent and would have excluded all children transported to community hospitals.

Conclusions. There is substantial institutional variability in approval of a minimal risk observational study and in willingness to obtain a federalwide assurance, particularly among community hospitals. Federal research policy involving patient privacy and institutional assurances may be contributing to this variability, which can adversely affect selection of research subjects, disrupt population-based study design, and threaten the generalizability of study results.

Ancillary