Approved Research

Using supervised machine learning to predict the post-operative infection after total knee or hip arthroplasty and phenome-wide association study to examine the causal relationship

Lay summary

Post-operative infection after total knee or hip arthroplasty is rare but causes a significant burden financially for patients and healthcare system. Patients with more comorbidities such as diabetes or anemia would be more likely to have post-operative infection. The aim of this study is to examine the causal relationship between pre-existing comorbidities and post-operative infection. To build up a prediction model of post-operative infection, we use the National Health Insurance Research Database in Taiwan and identify the comprehensive pre-existing comorbidities as predictors. We plan to use the UK Biobank data for two main goals: (1) to replicate our findings in an independent dataset and (2) to identify the genetic effect by using a genetic risk score for the top-associating comorbidities and test its association with all available disease phenotypes. The duration of the project will be a period of 3 years. The completion of this study will help us to understand the causal relationship between pre-existing comorbidities and post-operative infection. We anticipate that this will enable the early prevention of post-operative infection more precisely.