BACKGROUND: Training patients are the first individuals in whom a physician uses an investigational device. There is great variability in the use of data from training patients in the absence of guidelines. The prevalence and extent of data reporting from training patients in cardiovascular device studies submitted for US Food and Drug Administration (FDA) approval has not been characterized. METHODS: Information on training patients was abstracted from the Summary of Safety and Effectiveness Data summarizing cardiovascular device premarket applications approved by the FDA from 2000 through 2007. We examined the numbers and characteristics of training patients and the inclusion of their results in end-point analyses. RESULTS: There were 78 cardiovascular device summaries in this 8-year period, of which 17 (22%) involved training patients. Of the 123 studies in the summaries, 20 (16%) used training patients. All studies excluded training patients from efficacy analyses and 19 of 20 (95%) excluded them from safety analyses. Sixteen of 20 (80%) did not provide any outcome data, and 15 of 20 (75%) did not check for outcome differences between training and nontraining treatment patients. Eighteen of 20 (90%) did not provide demographic information on training patients, and 14 of 20 (70%) did not prespecify guidelines for their enrollment. CONCLUSIONS: Training patients comprise a considerable proportion of patients receiving investigational cardiovascular devices, but their results are excluded from FDA submissions. Their exclusion from analyses means that safety and efficacy outcomes may look better than actual results. Guidelines on the use and inclusion of results for training patients would improve accuracy on results reporting.
BACKGROUND: Training patients are the first individuals in whom a physician uses an investigational device. There is great variability in the use of data from training patients in the absence of guidelines. The prevalence and extent of data reporting from training patients in cardiovascular device studies submitted for US Food and Drug Administration (FDA) approval has not been characterized. METHODS: Information on training patients was abstracted from the Summary of Safety and Effectiveness Data summarizing cardiovascular device premarket applications approved by the FDA from 2000 through 2007. We examined the numbers and characteristics of training patients and the inclusion of their results in end-point analyses. RESULTS: There were 78 cardiovascular device summaries in this 8-year period, of which 17 (22%) involved training patients. Of the 123 studies in the summaries, 20 (16%) used training patients. All studies excluded training patients from efficacy analyses and 19 of 20 (95%) excluded them from safety analyses. Sixteen of 20 (80%) did not provide any outcome data, and 15 of 20 (75%) did not check for outcome differences between training and nontraining treatment patients. Eighteen of 20 (90%) did not provide demographic information on training patients, and 14 of 20 (70%) did not prespecify guidelines for their enrollment. CONCLUSIONS: Training patients comprise a considerable proportion of patients receiving investigational cardiovascular devices, but their results are excluded from FDA submissions. Their exclusion from analyses means that safety and efficacy outcomes may look better than actual results. Guidelines on the use and inclusion of results for training patients would improve accuracy on results reporting.