Generative Retrieval Does Not Improve Long-Term Retention of Regional Anesthesia Ultrasound Anatomy in Unengaged Learners

Abstract

Ultrasound-guided regional anesthesia is increasingly used in the perioperative period but performance requires a mastery of regional ultrasound anatomy. We aimed to study whether the use of generative retrieval to learn ultrasound anatomy would improve long-term recall. Fourth-year medical students without prior training in ultrasound techniques were randomized into standard practice (SP) and generative retrieval (GR) groups. An initial pre-test consisted of 74 regional anesthesia ultrasound images testing common anatomic structures. During the study/learning session, GR participants were required to verbally identify an unlabeled anatomical structure within 10 seconds of the ultrasound image appearing on the screen. A labeled image of the structure was then shown to the GR participant for 5 seconds. SP participants viewed the same ultrasound images labeled with the correct anatomical structure for 15 seconds. Retention was tested at 1 week and 1 month following the study session. Participants completed a satisfaction survey after each session. Forty-five medical students were enrolled with forty included in the analysis. There was no statistically significant difference in baseline scores (GR = 11.5 ± 4.9; SP = 11.2 ± 6.2; P = 0.84). There was no difference in scores at both the 1-week (SP = 54.5 ± 13.3; GR = 53.9 ± 10.5; P = 0.88) and 1-month (SP = 54.0 ± 14.5; GR = 50.7 ± 11.1; P = 0.42) time points. There was no statistically significant difference in learner satisfaction metrics between the groups. The use of generative retrieval practice to learn regional anesthesia ultrasound anatomy did not yield significant differences in learning and retention compared with standard learning.