Original article
Fitting a routine health-care activity into a randomized trial: An experiment possible without informed consent?

https://doi.org/10.1016/S0895-4356(96)00422-2Get rights and content

Abstract

Due to possible methodological and practical problems, many researchers refrain from using a randomized controlled trial design to evaluate procedures already embedded in routine health care. We performed a randomized controlled trial on the effects of routine individual feedback on test ordering behavior of family physicians. The trial started after 4 years of feedback and lasted for 2.5 years.

With some adaptations a randomized trial proved to be possible. In evaluating health-care procedures that cannot be blinded in a traditional way, asking full and study-specific informed consent may conflict with the validity of the design. In such studies, an alternative procedure is to be considered. Our trial, with doctors as study subjects, was held on an already accepted routine procedure (feedback). This made it possible to refrain from obtaining study-specific informed consent. Consequently, a Hawthorne effect and contamination of the trial arms through information leakage could be avoided. Justification and general criteria for not obtaining full and study-specific informed consent are worked out. In health-care research on the performance of doctors or on interventions into the quality of care, obtaining a general informed consent in advance is an acceptable alternative approach.

References (10)

  • SJ Pocock
  • KE Grady
  • TD Cook et al.
  • JA Knottnerus et al.

    Effectiviteitsonderzoek in de huisartspraktijk I

    Huisarts Wet

    (1987)
    JA Knottnerus et al.

    Effectiviteitsonderzoek in de huisartspraktijk II

    Huisarts Wet

    (1987)
    JA Knottnerus et al.

    Effectiviteitsonderzoek in de huisartspraktijk III

    Huisarts Wet

    (1987)
  • FJ Roethlisberger et al.
There are more references available in the full text version of this article.

Cited by (19)

  • Conducting Clinical Research in Post-acute and Long-term Nursing Home Care Settings: Regulatory Challenges

    2019, Journal of the American Medical Directors Association
    Citation Excerpt :

    If the NH agrees to adopt 1 of the 2 approaches for research, individual residents would not be able to decline participation if the approaches were considered standard of care. Several studies in non-NH settings have argued that research processes embedded in routine health care activities and of minimal risk to the residents do not require informed consent.20,21,38 Depending on the interventions studied, investigators may consider informing residents and families either during or after the study.

  • Cluster randomised trials: An important method in primary care research

    2009, Zeitschrift fur Evidenz, Fortbildung und Qualitat im Gesundheitswesen
  • Impact of 2 different levels of performance feedback on compliance with infection control process measures in 2 intensive care units

    2008, American Journal of Infection Control
    Citation Excerpt :

    The majority of respondents was also aware that their infection control practices were being observed during the final study phase. The Hawthorne effect, a phenomenon of altered behavior resulting from awareness of being observed, typically occurs during monitoring and feedback.15 To reduce this effect, we randomly observed the process measures by ICPs who routinely conducted a surveillance of HAIs in that unit, and the observers were as unobtrusive as possible.

  • Block design allowed for control of the Hawthorne effect in a randomized controlled trial of test ordering

    2004, Journal of Clinical Epidemiology
    Citation Excerpt :

    Such a design might provide more accurate answers to the problem of the magnitude of the nonspecific effects. Moreover, we did not handle other nonspecific effects, such as the fact that contact between intervention and control physicians can influence outcome, as physicians talk about the quality strategy under study, an effect commonly known as across subject contamination effect or leaking effect [8,25]. We assume that this effect is not particularly great, because GPs normally do not discuss test-ordering performance among themselves and the teams were located in different regions.

View all citing articles on Scopus
View full text