Abstract 4143727: AI Integration Decreased Rural Documentation Burden by 40% in Medicare's Chronic Care Management Setting

• DOI: 10.1161/circ.150.suppl_1.4143727
• Publication Date: 2024-11-14T09:02:39Z
• Authors: Jered Miller; Garrett Jimmerson; Callie Miller; Ashley Dey; Caroline Wheeler; Samuel Miller; Ghaith Al Tibi; Nicolas Chronos
• Keywords: Chronic care; Disease management

Efforts to ameliorate rural healthcare burnout stemming from overwork have taken many approaches, including the increased usage of Artificial Intelligence (AI) aid in generating electronic medical records (EMR). This new application AI technology demands characterization possible benefits and efficacy its use. Using a commercially available software (Freed AI) designed for EMR production, we conducted study determine quality generated by AI. Methods: trial analyzed 248 patient-provider interactions within setting Medicare’s Chronic Care Management (CCM program), recorded Freed human scribe. Three blinded readers with clinical background were given 2 notes same patient-encounter. One note was written provider one Each reader performed binary, independent assessment each compared control. Readers trained on definitions categories clarity, accuracy, completeness, relevance. The time required per encounter also analyzed. Discussion: As seen Table 1, better than human-generated fields clarity relevance, suggesting effective formatting well calibrated understanding what information is medically relevant. No significant difference completeness observed, able record as much valuable traditional charting. below standard category accuracy. For example, sometimes misspelled names or misunderstood complex situations being discussed because it not infer certain which human, using outside knowledge past experiences, could. reduction spent while utalizing demonstrates highly decrease workload allows providers freedom allocate more interacting patient hopefully reducing physician burnout.