Validating the Use of Smartphone-Based Accelerometers for Performance Assessment in a Simulated Neurosurgical Task

Abstract

Abstract BACKGROUND: Reductions in working hours affect training opportunities for surgeons. Surgical simulation is increasingly proposed to help bridge the resultant training gap. For simulation training to translate effectively into the operating theater, acquisition of technical proficiency must be objectively assessed. Evaluating “economy of movement” is one way to achieve this. OBJECTIVE: We sought to validate a practical and economical method of assessing economy of movement during a simulated task. We hypothesized that accelerometers, found in smartphones, provide quantitative, objective feedback when attached to a neurosurgeon's wrists. METHODS: Subjects (n = 25) included consultants, senior registrars, junior registrars, junior doctors, and medical students. Total resultant acceleration (TRA), average resultant acceleration, and movements with acceleration >0.6g (suprathreshold acceleration events) were recorded while subjects performed a simulated dural closure task. RESULTS: Students recorded an average TRA 97.0 ± 31.2 ms−2 higher than senior registrars (P = .03) and 103 ± 31.2 ms−2 higher than consultants (P = .02). Similarly, junior doctors accrued an average TRA 181 ± 31.2 ms−2 higher than senior registrars (P < .001) and 187 ± 31.2 ms−2 higher than consultants (P < .001). Significant correlations were observed between surgical outcome (as measured by quality of dural closure) and both TRA (r = .44, P < .001) and number of suprathreshold acceleration events (r = .33, P < .001). TRA (219 ± 66.6 ms−2; P = .01) and number of suprathreshold acceleration events (127 ± 42.5; P = .02) dropped between the first and fourth trials for junior doctors, suggesting procedural learning. TRA was 45.4 ± 17.1 ms−2 higher in the dominant hand for students (P = .04) and 57.2 ± 17.1 ms−2 for junior doctors (P = .005), contrasting with even TRA distribution between hands (acquired ambidexterity) in senior groups. CONCLUSION: Data from smartphone-based accelerometers show construct validity as an adjunct for assessing technical performance during simulation training.