A research team from Boston Children's Hospital has created an artificial intelligence model that can successfully predict the likelihood of a patient to miss appointments, according to an April 20 Nature study.
The research team collected data from more than 19,000 patients and over 100,000 hospital appointments from Boston Children’s Hospital’s primary care pediatric clinic. The data included key information such as age, insurance plan and prior no-show rate. They then trained an AI algorithm to predict the likelihood of no-shows to scheduled appointments.
The algorithm successfully predicted the incidence of no-show rates 83 percent of the time and led to a false alert around 17 percent of the time. It was also able to make predictions when other information was missing, but previous no-show rates were the strongest predictor of no-shows.
The model could be used to prompt health systems to send text alerts or reminder prompts to patients likely to miss appointments in an attempt to reduce no-show rates as well as streamline resource allocation for amenities like transportation that can help improve appointment adherence.