OBJECTIVE: The primary intent for obtaining screening logs in a randomized clinical trial is to assess selection bias in patient recruitment. This is particularly relevant to focused trials in heterogeneous populations such as traumatic brain injury (TBI) patients. We aimed to investigate the benefits of collecting screening logs in two randomized clinical trials conducted in TBI.
METHODS: Screening logs were collected as part of the conduct of two multicenter trials of neuroprotective agents in TBI: the Salzburg Atherosclerosis Prevention Program in Subjects at High Individual Risk study (n = 924) and the dexanabinol study (n = 861). Centers were requested to submit monthly information on all patients with TBI admitted to the intensive care unit, including demographics, time of injury and admission, injury severity, and, if not recruited, the reason(s) for exclusion.
RESULTS: In the Salzburg Atherosclerosis Prevention Program in Subjects at High Individual Risk study, 52 centers submitted admission data on 4166 patients. In the dexanabinol trial, 96 centers submitted data on 7052 patients. On average, only 20% of patients screened for the Salzburg Atherosclerosis Prevention Program in Subjects at High Individual Risk study and 10% for the dexanabinol trial were enrolled. The main reasons for exclusion were neurological status (29 and 26%, respectively), age (24 and 30%, respectively), and admission outside of the time window (17 and 21%, respectively). Differences in patient characteristics between screened and enrolled patients, with substantial country-specific variation, were observed.
CONCLUSION: The collection of screening logs is necessary to report trial results according to the Consolidated Standards of Reporting Trials guidelines and to assess the generalizability of findings. Our experience shows the feasibility of collecting screening logs and illustrates how the potential for selection bias may creep into well-designed randomized clinical trials as a result of factors outside the control of investigators. Consistency and accuracy in screening log completion may further serve as an early indicator of center performance in a trial.
Department of Neurosurgery, Erasmus Medical Center, Rotterdam, The Netherlands (Slieker) (Maas)
Departments of Neurosurgery and Intensive Care, Erasmus Medical Center, Rotterdam, The Netherlands (Kompanje)
Public Health Sciences, University of Edinburgh Medical School, Edinburgh, Scotland (Murray)
Pirkanmaa Hospital District, Tampere, Finland (Öhman)
Ospedale Policlinico Istituto di Ricovero e Cura a Carattere Scientifico, Milan University, Milan, Italy (Stocchetti)
University of Glasgow and Royal College of Physicians and Surgeons, Glasgow, Scotland (Teasdale)
Department of Neurosurgery, University Hospital, Antwerp, Belgium (Maas)
(see Acknowledgments) (SAPHIR and Pharmos TBI Investigators)
Reprint requests: Andrew I.R. Maas, M.D., Ph.D., University Hospital Antwerp, Department of Neurosurgery, Wilrijkstraat 10, 2650 Edegem, Belgium. Email: Andrew.Maas@uza.be
Received, June 15, 2007.
Accepted, March 17, 2007.