Dr. Usha Govindarajulu, biostatistician at SUNY Downstate Medical Center School of Public Health, and her colleagues developed a methodology to model the probability of successful implementation of surgical devices and procedures. To do this, she and her team studied learning curve effects of practitioners, which are a critical component of medical device surveillance. Learning curve effects were estimated by evaluating multiple methods for modeling success rates within a complex, simulated dataset representing patients treated by physicians clustered within institutions. The team employed unique modeling of learning curves that incorporated the learning hierarchy between institution and physicians, and then further evaluated the learning curves using established statistical methods for hierarchical data. Findings of best ways to model the hierarchical learning curves with various shapes and modeling will serve to improve device/procedure safety and surveillance.
The article, “Learning curve estimation in medical devices and procedures: hierarchical modeling” was published in Statistics in Medicine July 30, 2017, 36(17): 2764-2785.