Abstract WP257: Accuracy of Pre-enrollment Last Known Well Time in a Large Prehospital Stroke Clinical Trial

Abstract

Background: Clinical trials of stroke therapies require accurate documentation of last known well time (LKWT) to account for injury accumulation prior to treatment start. For prehospital studies, this requirement is particularly challenging, as paramedic-determined and final-determined LKWTs in routine practice are concordant (within 15 mins) in only half of cases. We sought to determine the accuracy of LKWT obtained in the field by a two-tier enrollment system of study-trained paramedics and cellphone-connected physician-investigators. Methods: Paramedics screened consecutive transports for participation in the NIH Field Administration of Stroke Therapy-Magnesium (FAST-MAG) clinical trial. Paramedic screening criteria included LKWT <2 hours. Physician-investigators by cellphone confirmed or refined the LKWT after conversation with paramedics and patients or legally authorized representative. Prehospital LKWT was compared with post-arrival LKWT determined by trained study nurses after ED arrival by speaking with patients, family and other sources. We describe the number of enrollment calls with inaccurate LKWT at the paramedic-screening level and at the post paramedic plus physician-investigator telephone screening level. Results: A total of 4458 post-screening enrollment calls were made by paramedics from January 2005 to December 2012 of which 539 (12%) were determined by physician-instigators to have inaccurate LKWT leading to non-enrollment. Of the remaining 3919 calls, 1700 led to enrollments in the study and 2219 were not enrolled for a reason other than inaccurate LKWT. Among enrolled cases, exact congruence between prehospital and post-arrival LKWTs occurred 72% (n=1220), concordance within 15 minutes in 87%, within 30 minutes in 93%, and within 1 hour in 97%. Among enrolled cases, final-determined LKWT was within the study entry window of 2h in 96.3%. Conclusions: A 2-tiered system of paramedic screening followed by physician-investigator cellphone assessment led to high congruence between prehospital-determined and post-arrival-determined of LKWT. This system can be used in future trials of prehospital, paramedic-in initiated stroke therapy when accuracy of LKWT is important for intervention evaluation.