Non-technical skills
(NTS) are cognitive and interpersonal skills that are relevant to task
completion such as situation awareness, decision-making, teamwork, and
leadership. NTS in clinical environments, such as surgery, have been identified
to contribute to patient safety and team performance, which in turn affects
clinical outcomes. Assessment tools of these skills in surgery exist; however,
current evaluations are limited in that they require trained raters, are
subjective, are time-intensive, and are checklist-based. Therefore, there is a
need for objective measurement of NTS that addresses the limitations of the
rating-based techniques. The purpose of this Ph.D. dissertation work is to
identify physiological and behavioral metrics that measure NTS objectively and
investigate the application of objective metrics to measure intraoperative NTS
of surgeons. Through a scoping review of engineering, behavioral science, and
medical literature, behavioral and physiological metrics that quantified NTS
constructs of surgeons were identified. The synthesized literature was used to
build a framework integrating objective metrics to NTS constructs. To develop
an objective model of surgeons’ NTS, subjective and objective behavioral data
of surgeons were collected in the operating room and prediction models were created.
Results found that objective metrics such as communication, speech, and
proximity features can be used to predict subjective NTS. Furthermore,
objective task features (e.g., time and number of incidents during an operation)
has the potential to also model subjective NTS, and these task features can be
predicted by the behavioral metrics; thus, triangulation is obtained with the
three NTS metrics: subjective score, objective behavioral metrics, and task
performance metrics. The relationship between the two objective metrics shows the
possibility of achieving a fully objective model of surgeons’ NTS. The
consolidation of current objective measurement techniques can provide a
foundation in further understanding NTS beyond assessments based on observed
behaviors, and the developed models can be expanded and implemented for real-time
NTS assessment of clinical teams to improve patient care.