The identification of gaze behaviour associated with decision-making in surgery: an eye- tracking study
File(s)
Author(s)
Ashraf, Hajra
Type
Thesis
Abstract
Adverse surgical events remain at an unacceptably high level despite multiple global safety initiatives being introduced to practise over the last 15 years. As yet, however there is no conclusive evidence to identify whether physiological markers can be used to predict whether a surgeon will make an error. The aim of this thesis is to identify whether physiological metrics, specifically gaze related metrics can be used to predict whether a surgeon is going to make an error. Thus, my research hypothesis was gaze metrics can be analysed to provide a greater understanding of perceptual contributions to decision- making in healthcare and aid in the formulation of innovative solutions.
The research question was initially approached by the undertaking of a literature review, outlining the current knowledge surrounding the uses of eye- tracking technology in healthcare, the findings of which are outlined in chapter one. Next, I undertook two studies recruiting surgeons to perform a series of laparoscopic cholecystectomy tasks on the LapMentor Simbionix trainer. Fatigue severity score (FSS) was calculated prior to the task. Physiological metrics, including gaze tracking, heart rate and skin conductance were measured while they completed the tasks. LightGBM and CatBoost were used to predict the physiological metric most useful in predicting whether a surgeon was about to make an error. My studies indicated that the most predictive feature was the correlation of the skin conductance measure against itself shifted by two time steps (4ms, or 250hz). Skin conductance was twice as likely to successfully predict impending error compared to heart rate and fatigue severity scale.
Gaze metrics were found to be unimportant thus rendering our hypothesis incorrect. However, when I added gaze
features, overall model performance improved by 6.4%. This leads us to surmise that gaze metrics are relevant, however future work would need to establish the importance of this improvement in performance.
This work advances our understanding of physiological changes that occur during decision making and has important implications in the surgical field, as skin conductance could potentially be used as a warning system to indicate when surgeons are at high risk of making an error. The potential for reduction in surgical error rate and improvement in patient safety are important factors to consider.
The research question was initially approached by the undertaking of a literature review, outlining the current knowledge surrounding the uses of eye- tracking technology in healthcare, the findings of which are outlined in chapter one. Next, I undertook two studies recruiting surgeons to perform a series of laparoscopic cholecystectomy tasks on the LapMentor Simbionix trainer. Fatigue severity score (FSS) was calculated prior to the task. Physiological metrics, including gaze tracking, heart rate and skin conductance were measured while they completed the tasks. LightGBM and CatBoost were used to predict the physiological metric most useful in predicting whether a surgeon was about to make an error. My studies indicated that the most predictive feature was the correlation of the skin conductance measure against itself shifted by two time steps (4ms, or 250hz). Skin conductance was twice as likely to successfully predict impending error compared to heart rate and fatigue severity scale.
Gaze metrics were found to be unimportant thus rendering our hypothesis incorrect. However, when I added gaze
features, overall model performance improved by 6.4%. This leads us to surmise that gaze metrics are relevant, however future work would need to establish the importance of this improvement in performance.
This work advances our understanding of physiological changes that occur during decision making and has important implications in the surgical field, as skin conductance could potentially be used as a warning system to indicate when surgeons are at high risk of making an error. The potential for reduction in surgical error rate and improvement in patient safety are important factors to consider.
Version
Open Access
Date Issued
2019-02
Date Awarded
2020-12
Copyright Statement
Creative Commons Attribution NonCommercial Licence
License URL
Advisor
Sodergren, Mikael
Mylonas, George
Darzi, Ara
Publisher Department
Department of Surgery & Cancer
Publisher Institution
Imperial College London
Qualification Level
Masters
Qualification Name
Master of Philosophy (MPhil)