Article Text

other Versions

PDF
Prevention of false positive findings in observational studies: registration will not work but replication might
  1. P de Jonge1,2,
  2. H J Conradi1,3,
  3. B D Thombs4,
  4. J G M Rosmalen1,
  5. H Burger1,5,
  6. J Ormel1
  1. 1ICPE (Interdisciplinary Center for Psychiatric Epidemiology), Department of Psychiatry and Internal Medicine, University Medical Center Groningen, University of Groningen, Groningen, the Netherlands
  2. 2CORPS (Centre of Research on Psychology and Somatic Disease), Department of Medical Psychology, Tilburg University, Tilburg, the Netherlands
  3. 3Department of Clinical Psychology, University of Amsterdam, Amsterdam, the Netherlands
  4. 4Departments of Psychiatry; Epidemiology, Biostatistics, and Occupational Health; and Medicine, McGill University, Montreal, Quebec, Canada; Department of Psychiatry, Center for Clinical Epidemiology and Community Studies, and Division of Rheumatology, Jewish General Hospital, Montreal, Quebec, Canada
  5. 5Department of Epidemiology, University Medical Center Groningen, University of Groningen, Groningen, the Netherlands
  1. Correspondence to Prof dr Peter de Jonge, Interdisciplinary Center of Psychiatric Epidemiology, Department of Psychiatry, UMCG University of Groningen, CC72, PO Box 30 001, Groningen 9700 RB, the Netherlands; peter.de.jonge{at}med.umcg.nl

Statistics from Altmetric.com

Introduction

Progress in science is built on a balance between curiosity and scepticism, and between creativity and rigour.1 Although progress in science inevitably needs the generating of findings that may not be verified in subsequent studies, there is a considerable risk that exploratory studies introduce bias in the body of scientific knowledge. This risk may be substantial in observational studies, including cohort studies, case–control studies and cross-sectional studies.2 Observational studies play an essential role in medical research as they are often conducted to evaluate research questions that cannot be addressed by clinical trials.3 As shown by Ioannidis, however, the likelihood that any given finding from a published observational study is true in reality is limited.2 Aside from problems of uncontrolled confounding and other biases, observational studies are often generated from databases comprising many variables. Researchers often explore an unknown quantity of potential relationships capitalising on the chance of obtaining positive findings. In this editorial, we will explore two potential solutions to reduce the risk of false positive findings from observational studies: a registration requirement; and a replication requirement.

Registration requirement for observational research

With a registration requirement, scientific journals would require researchers to register their observational studies in a manner similar to what has become policy for clinical trials.4 5 Study registration would include the recording of well grounded hypotheses that will be tested and data to be collected in a given study, recently …

View Full Text

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.