Real-time Visualisation and Analysis of Clinicians' Performance during Palpation in Physical Examinations.

Abstract

Motivated by the fact that palpation skills are challenging to learn and teach, particularly during Digital Rectal Examinations (DRE), and the lack of understanding of what constitutes adequate performance, we present a visualisation and analysis system that uses small position and pressure sensors located on the examining finger, allowing the quantitative analysis of duration, steps and pressure applied. The system is first described, followed by an experimental study of twenty experts from four clinical specialties performing ten DREs each on a benchtop model using the proposed system. Analysis of the constitutive steps was conducted to improve understanding of the examination. A Markov model representing executed tasks and analysis of pressure applied is also introduced. The proposed system successfully allowed the visualisation and analysis during the experimental study. General Practitioners and Nurses were found to execute more tasks compared to Urologists and Colorectal Surgeons. Urologists executed the least number of tasks and were the most consistent group compared to others. The ability to "see through" allowed us to better characterise the performance of experts when conducting a DRE on a benchtop model, comparing the performance of relevant specialties, and studying executed tasks and the pressure applied. The Markov model presented summarises task execution of experts and could be used to compare performance of novices against that of experts. This approach allows for the analysis of performance based on continuous sensor data recording that can be easily extended to real subjects and other types of physical examinations.