Invited Commentary
Tragedy of conducting a clinical trial; generic alert system needed

https://doi.org/10.1016/j.jclinepi.2007.06.004Get rights and content

Abstract

Stopping a clinical trial without reaching the final objective is not the ideal outcome any researcher wants; sometimes ceasing is inevitable. Due to marginal inclusion of patients we were forced to cease our randomized clinical trial on the effectiveness of proprioceptive training on the development of chronic whiplash complaints a year after the start. Although incidence figures demonstrate that recruitment of the planned number of whiplash patients would be easily feasible, we were unable to enroll the amount of subjects. Several motives can be proposed that would have prevented this obliged halting from happening. Other studies also report impracticability of the planned number of whiplash injury patients.

Introduction

In spite of preceding investigation of social relevance and feasibility, sometimes ending a clinical trial without achieving the aims is inevitable. Research about stopping clinical trials is often related to the likelihood of achieving statistical significant results at all [1] or stopping a trial early for benefit [2], [3]. Sometimes other factors make ceasing inevitable. Although successful research endings generally result in publication, failures are usually sentenced to disappearance. Unfortunately, in this way it is not educational. Publication of such failures and their circumstances may increase our knowledge of the associated determinants, which may lead to better prevention of future failures.

Example

Last year, we started a randomized clinical trial in patients suffering from acute Whiplash Associated Disorders [4] (WAD) evaluating additional proprioceptive training (done by physiotherapists).

We planned to include 120 WAD patients between 4 days and 12 weeks after the car collision. Three measurements were scheduled. Each measurement included a questionnaire and measurement of the optokinetic reflex, cervico ocular reflex, and vestibular ocular reflex. Randomization took place after baseline measurement. We planned to measure again all patients of both groups at 9 (T2) and 26 (T3) weeks after randomization. Patients were recruited by participating General Practitioners (GPs) or at Emergency Departments (EDs) within Rotterdam and surroundings (6,61,887 persons [2005] [5]). A total of 260 GPs were contacted by phone, and 105 were willing to participate. Three out of five contacted EDs contributed in the trial and received patient information forms and posters to attract the physicians' and patients' attention.

The medical ethical committee of the Erasmus MC approved the study.

After 12 months, 79 probably eligible whiplash patients were referred to us (21 by their GP, 25 by the EDs, the rest of the subjects entered after radio or newspaper publicity).

Finally, only 11 whiplash patients met the selection criteria. Major reasons for exclusion were collision appeared more than 3 months ago (n = 8), subjects already received physiotherapy (n = 29), and no rear-end collision (n = 20). Reasons for withdrawal were recovery (n = 2), unknown (n = 1). Two patients did not show up for the baseline measurement even after three subsequent appointments were made.

Section snippets

What can go wrong in conducting a trial?

Although randomized controlled trials are a widely used design in medical research, often recruitment problems give rise to bringing the trial to a standstill. According to the literature most studies suffer from Lasagna's Law, meaning that as soon as a study starts to recruit participants, 90% of the eligible participants disappear to return only after the end of the recruitment period [6], [7]. Failure to recruit sufficient numbers of patients frequently threatens a successful completion of

What can be done?

There are many interventions that could potentially improve clinician and patient participation in trials. Finding which ones are effective would be of benefit to the research community and society. A Cochrane review was published on strategies to improve recruitment [15]. The authors found only strategies aimed at patients, none was aimed at research collaborators (e.g., doctors), whereas factors mentioned in literature mainly focus on factors associated to the clinician [8]. Besides using a

General learning points and recommendations

Recently, an international trial register has been created, in which clinical trials are registered prospectively to prevent double studies and positive publication bias. In the future, such a register can also determine the number of ceased trials [17]. Reasons for cessation will be recorded and a better overview of these reasons can be provided. Also, strategies to improve patient recruitment can be retrieved.

Randomized controlled trials are a valid method to determine the effectiveness of

Acknowledgments

During this research M.F. and C.Z. were sponsored by the NWO-VIDI-program. I.M. was sponsored by Revolving Fund of the Erasmus MC.

References (18)

There are more references available in the full text version of this article.

Cited by (4)

  • The Effect of a Personalized Newsletter to Physical Therapists on Patient Recruitment: A Cluster Randomized Trial in Primary Physiotherapy Care

    2020, Journal of Manipulative and Physiological Therapeutics
    Citation Excerpt :

    Nevertheless, decision makers are increasingly looking at the results of RCTs to guide practice.11,12 Lasagna's law is an important phenomenon in research, where researchers and clinicians overestimate the number of suitable patients they will be able to include in a study.6,12-17 Recruitment rates depend on the willingness of patients to participate.

  • Discontinuation of a randomised controlled trial in general practice due to unsuccessful patient recruitment

    2017, BJGP Open

  • Enrollment into a time sensitive clinical study in the critical care setting: Results from computerized septic shock sniffer implementation

    2011, Journal of the American Medical Informatics Association

  • The term 'whiplash' is insufficient: Commentary on multidisciplinary guideline for uncomplicated whiplash

    2009, Nederlands Tijdschrift voor Geneeskunde
View full text
Copyright © 2008 Elsevier Inc. All rights reserved.