Abstract
STUDY OBJECTIVE
The "illusion of validity" is a cognitive bias in which the ability to interpret and predict surgical performance accurately is overestimated. To address this bias, we assessed participants comparing fundamentals of laparoscopic surgery (FLS) and non-FLS tasks with cadaveric vaginal cuff suturing to determine the most representative simulation task for laparoscopic vaginal cuff suturing.
DESIGN
Validity (Messick framework) study comparing FLS and non-FLS tasks with cadaveric vaginal cuff suturing.
SETTING
Simulation center cadaver laboratory.
PARTICIPANTS
Obstetrics and gynecology residents (n = 21), minimally invasive gynecologic surgery fellows (n = 3), gynecologic surgical subspecialists (n = 4), general obstetrician/gynecologists (n = 10).
INTERVENTIONS
Tasks included a simulated vaginal cuff (ipsilateral port placement), needle passage through a metal eyelet loop (contralateral and ipsilateral), and intracorporeal knot tying (contralateral and ipsilateral). Simulation task times were compared with the placement of the first cadaveric vaginal cuff suture time, as well as the in-person and blinded Global Operative Assessment of Laparoscopic Skills (GOALS) score ("relations to other variables" validity evidence). Statistical analyses included Spearman's test of correlation (continuous and ordinal variables) or Wilcoxon rank sum test (categoric variables).
MEASUREMENTS AND MAIN RESULTS
There was a stronger association with cadaver cuff suturing time for simulated vaginal cuff suturing time (r = 0.73, p <.001) compared with FLS intracorporeal contralateral suturing time (r = 0.54, p <.001). Additional measures associated with cadaveric performance included subspecialty training (median: 82 vs 185 seconds, p = .002), number of total laparoscopic hysterectomies (r = -0.53, p <.001), number of laparoscopic cuff closures (r = -0.61, p <.001), number of simulated laparoscopic suturing experiences (r = -0.51, p <.001), and eyelet contralateral time (r = 0.52, p <.001). Strong agreement between the in-person and blinded GOALS (intraclass correlation coefficient = 0.80) supports response process evidence. Correlations of cadaver cuff time with in-person (Spearman's r = -0.84, p <.001) and blinded GOALS (r = -0.76, p <.001) supports relations to other variables evidence CONCLUSION: The weaker correlation between FLS suturing and cadaver cuff suturing compared with a simulated vaginal cuff model may lead to an "illusion of validity" for assessment in gynecology. Since gynecology specific validity evidence has not been well established for FLS, we recommend prioritizing the use of a simulated vaginal cuff suturing assessment in addition to FLS.
Collapse