Date of Award
8-2023
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
Department
Bioengineering
Committee Chair/Advisor
Joseph Singapogu
Committee Member
Zhi Gao
Committee Member
William Richardson
Committee Member
Judy Geissler
Abstract
Dialysis is a vital medical treatment for patients with end-stage renal disease. An arteriovenous fistula, a surgically created connection between an artery and a vein, is the preferred vascular access for dialysis due to superior clinical outcomes. However, cannulation complications caused by needle-related errors are common and dangerous, necessitating the need for effective cannulation training. Despite cannulation being a challenging and complex clinical procedure, little is known about the quantitative aspects of needle insertion dynamics necessary for skilled cannulation. The aim of this study was to develop meaningful and instructive needle-based metrics for cannulation skill assessment and training.
Two iterations of sensor-based dialysis cannulation simulators were created to collect data during cannulation performed on the simulator by participants with varying degrees of clinical experience. Needle motion was tracked using an electromagnetic system along with data collected from other sensor modalities. Fluid-free fistulas of varying geometries were fabricated for the simulator from silicone, and blood “flashback” was simulated using infrared technology. Nursing students and clinical dialysis cannulators participated voluntarily in an Institutional Review Board (IRB) approved study.
Towards examining if objective metrics can quantify aspects of cannulation skill, several needle motion and angle metrics were formulated. Statistical analysis methods (e.g., two-sample t-test and Mann-Whitney U test) and machine learning models (e.g., multiple linear regression, logistic regression, and principal component regression) were used to analyze human subjects’ data. Three types of metrics (sub-task, needle insertion angle, and needle location) were proposed and validated for cannulation outcome and skill assessment. The sub-task metrics demonstrated significant differences between the two insertion phases of cannulation and could predict the risk of infiltration. Needle angle and location metrics were shown to predict the probability of cannulation success and to distinguish between high and low performers, regardless of whether the performance was defined subjectively or objectively. Additionally, these metrics were also related to the flashback quality.
In conclusion, our proposed metrics were validated for skill assessment, and although the findings were based on cannulation simulators rather than in a clinical setting, they could guide competency-based training and teach for needle cannulation.
Recommended Citation
Zhang, Ziyang, "Development & Validation of Needle-based Metrics for Hemodialysis Cannulation Skill Assessment" (2023). All Dissertations. 3402.
https://tigerprints.clemson.edu/all_dissertations/3402