We are aiming to help doctors predict a patient’s future risk of needing a knee or hip replacement for osteoarthritis, an often progressive painful condition involving the joint. This could help focus efforts on earlier and more intensive non-surgical treatment that might delay or prevent need for joint replacement. It may also help with more timely discussion on referral for, joint replacement.
Our study will use information routinely recorded in patients’ medical records. This includes information like patient age at time of diagnosis, whether they are overweight and previous injury to the joint. All of these are known to be risk factors for osteoarthritis getting worse but there may be others, such as the strength of painkillers prescribed, the presence of other illnesses, and which part of the country the patient lives. Our study will look across a wide range of factors and work out which ones, used in combination, are best at predicting joint replacement.
Similar types of risk prediction tools have been developed and are regularly used by doctors and their patients for predicting risk of heart attacks, stroke, and fractures. We think this is the first study to apply these methods to the prediction of joint replacement.
To advance knowledge of individual patient prognosis for osteoarthritis, we propose to derive and internally validate multivariable models to predict the risk of future primary hip and knee replacement in an open cohort of patients aged ≥40 years presenting to general practice with osteoarthritis or hip pain or knee pain in the period 1993-2015. Our study has 5 main phases: (i) evaluating in the primary care record the feasibility of using prognostic factors identified from a systematic literature review; (ii) a case-control study to discover other potential prognostic factors available in the primary care record. (iii) determining a final candidate list of prognostic factors through consultation with our expert collaborators, panel of GPs, and Research User Group members; (iv) derivation, internal validation, evaluation of discrimination/calibration of multivariable risk prediction models for primary hip arthroplasty and primary knee arthroplasty using Cox proportional hazards models; (v) secondary analyses to evaluate sources of bias, model assumptions, and model performance in specified subgroups (e.g. at point of first OA diagnosis). A future ambition, beyond the scope of this proposal, is to explore with our collaborators opportunities to externally validate our multivariable risk prediction models in primary care databases in Spain.
Dahai Yu - Chief Investigator - Keele University
George Peat - Corresponding Applicant - Keele University
Avril Poyner - Collaborator - Not from an Organisation
Christine Walker - Collaborator - Keele University
Daniel Prieto-Alhambra - Collaborator - University of Oxford
John Edwards - Collaborator - Keele University
Kelvin Jordan - Collaborator - Keele University
Vincent Ukachukwu - Collaborator - Keele University
Patient Level Index of Multiple Deprivation;Practice Level Index of Multiple Deprivation