Predictive factors and outcomes in primary hip arthroplasty
Author(s)
Alagha, M. Abdulhadi
Type
Thesis or dissertation
Abstract
Cognisant prognostication involves informed predictions and is of a particular importance in surgical practice where significant variations prevail for patients with end stage hip pathologies necessitating arthroplasty. A comprehensive evaluation of Machine Learning (ML) models was undertaken using two of the largest hip registries, from Sweden and the United Kingdom. Failure rates (re-operation, revision, periprosthetic femoral fracture, and mortality), and patient-reported outcome measures (PROMs) were forecasted for primary elective total hip replacement. ML algorithms had a superior binary discriminative power in predicting all outcomes and achieved “good” to “excellent” accuracy in forecasting significant and substantial clinical improvement for PROMs.
Multivariate analysis of the LASSO*-penalised CoxNet** survival model revealed a statistically significant trade-off encompassing a heightened mortality risk in patients with low body mass index and an augmented revision risk in obese individuals. Likewise, whilst cemented fixation implants were associated with a diminished likelihood of re-do procedures, they were observed to have an increased mortality risk. SHAR data was further evaluated to a) predict the risk of femoral fractures requiring re-do surgery, and b) identify the effect of fixation methods on outcomes in a matched cohort. Cementless fixation, appeared to increase this likelihood with statistically significant differences between the matched cohorts.
Two laboratory-based experiments investigated the effect of impaction energy levels on bone strain behaviour, peri-prosthetic femoral fracture risk, and implant stability using two types of specimens (synthetic and cadaveric hips) and two biomechanical approaches (digital image correlation and rosette strain gauges). The findings of these seem to suggest that a high number of low-energy strikes during femoral broaching and implant seating reduces the risk of femoral fractures without affecting implant fixation.
Using personalised risk stratification models in orthopaedic practice may help balance safety and functional outcomes within the context of variability in patient characteristics, surgical practices, and implant designs.
Multivariate analysis of the LASSO*-penalised CoxNet** survival model revealed a statistically significant trade-off encompassing a heightened mortality risk in patients with low body mass index and an augmented revision risk in obese individuals. Likewise, whilst cemented fixation implants were associated with a diminished likelihood of re-do procedures, they were observed to have an increased mortality risk. SHAR data was further evaluated to a) predict the risk of femoral fractures requiring re-do surgery, and b) identify the effect of fixation methods on outcomes in a matched cohort. Cementless fixation, appeared to increase this likelihood with statistically significant differences between the matched cohorts.
Two laboratory-based experiments investigated the effect of impaction energy levels on bone strain behaviour, peri-prosthetic femoral fracture risk, and implant stability using two types of specimens (synthetic and cadaveric hips) and two biomechanical approaches (digital image correlation and rosette strain gauges). The findings of these seem to suggest that a high number of low-energy strikes during femoral broaching and implant seating reduces the risk of femoral fractures without affecting implant fixation.
Using personalised risk stratification models in orthopaedic practice may help balance safety and functional outcomes within the context of variability in patient characteristics, surgical practices, and implant designs.
Version
Open Access
Date Issued
2023-09-29
Date Awarded
2024-07-01
Copyright Statement
Creative Commons Attribution NonCommercial Licence
License URL
Advisor
Cobb, Justin
Liddle, Alexander
Mohaddes, Maziar
Sponsor
Imperial College London
Publisher Department
Department of Surgery & Cancer
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)