A predictive model for knee joint replacement in older women

Area of Science:

• Gerontology
• Orthopedics
• Epidemiology

Background:

• Knee replacement (KR) is a costly and invasive surgical procedure.
• Currently, no validated predictive algorithms exist to identify individuals at high risk for KR.
• Identifying at-risk populations is crucial for proactive healthcare management and resource allocation.

Purpose of the Study:

• To assess the predictive ability of patient self-reported risk factors for 10-year KR incidence.
• To develop and validate a predictive model for KR risk in women over 70 years old.
• To establish risk categories to guide clinical decision-making and potential interventions.

Main Methods:

• A population-based cohort of 1,462 women aged over 70 years was studied over a 10-year period (1998-2008).
• Hospital records were used to identify prevalent and incident total knee replacements.
• Statistical modeling, including logistic regression and survival analysis, was employed to evaluate risk factors like body mass index, knee pain, previous KR, and analgesia use.

Main Results:

• 129 participants (8.8%) underwent KR during the 10-year follow-up.
• Baseline factors including body mass index, knee pain, prior knee replacement, and analgesic use for joint pain were significant predictors (P < 0.001).
• A 5-variable model demonstrated good predictive discrimination (C-statistic = 0.79 ± 0.02) and calibration, identifying low (<5%), moderate (5-10%), and high (≥10%) risk categories.

Conclusions:

• Five easily obtainable, self-reported patient factors effectively predict 10-year KR risk in older women.
• The developed algorithm shows promise for a patient-based risk calculator to aid in treatment planning.
• This tool could help reduce the necessity for surgery in high-risk elderly populations by enabling targeted interventions.