1. Evidence Base and Model Rationale
- Primary baseline: the revision hazard follows a piecewise, time varying schedule anchored to Katz et al. (2012), a US Medicare cohort study that measured 12-year revision risk after primary total hip replacement with competing risk methods. The hazard is highest in the first year (infection, dislocation, periprosthetic fracture), reaches its low point in years 2 to 5, then climbs slowly as polyethylene wear and aseptic loosening accumulate.
- Evans et al. (2019) context: that multinational registry meta-analysis (non-US, with late follow-up driven mostly by the Finnish Arthroplasty Register) reported 25-year revision free survivorship of 57.9 percent. It serves here only as a long term plausibility reference, not as the model baseline.
- Endpoint: revision surgery, defined as component replacement or removal. Short term complication endpoints such as irrigation and debridement alone are excluded from the baseline schedule.
- Medicare Advantage ascertainment correction: a 1.15 hazard multiplier is applied from year 13 onward only. It corrects for revisions that Traditional Medicare claims miss as patients move to Medicare Advantage plans, calibrated to the 92 percent versus 90 percent survivorship divergence at 15 years reported by Lacny et al. (2023).
2. Model Coefficients
| Variable | Source | Functional Form | Effect Size |
|---|
| Age at surgery (continuous) | Bayliss et al., 2017 | exp(−0.025 × [age − 70]) | HR ≈ 2.0 at age 42; ≈ 0.61 at age 90 (reference age 70) |
| Male sex (year 0 to 1) | Katz et al., 2012 | Constant HR, time varying | 1.35 |
| Male sex (year 1 and later) | Katz et al., 2012 | Constant HR, time varying | 1.15 |
| BMI (continuous, above 25) | Meta-analytic synthesis | exp(0.030 × max(0, BMI − 25)) | HR ≈ 1.16 at BMI 30; ≈ 1.35 at BMI 35 |
| Inflammatory arthritis | Bozic et al., 2014 | Constant HR | 1.32 |
| Prior orthopedic surgery | Wright et al., 2012 | Constant HR | 1.35 (converted from OR 1.45) |
| Complicated diabetes | SooHoo et al., 2010 | Year 0 to 1 only | 1.50 (down weighted; short term endpoint) |
Peripheral vascular disease was dropped from the model because SooHoo (2010) supports only short term complication risk, not long term revision risk. Height and weight toggles were replaced with a continuous BMI computed from the entered measurements.
3. Statistical Method
The model runs a cause specific hazard competing risk recursion in 1-year steps (the Beyersmann et al. formulation) to produce the Cumulative Incidence Function (CIF) for revision:
- Revision hazard (hR): time varying baseline λ(t) multiplied by the composite HR. The baseline schedule is 1.20 percent in year 0 to 1, 0.45 percent in year 1 to 2, 0.35 percent in years 2 to 5, 0.45 percent in years 5 to 10, 0.60 percent in years 10 to 15, 0.75 percent in years 15 to 20, 0.90 percent in years 20 to 25, and 1.05 percent from year 25 on. The 1.15 Medicare Advantage correction applies from year 13 onward.
- Death hazard (μD): computed from the US Life Tables 2023 as μ = −ln(1 − qx), which converts the annual death probability into a proper continuous time, cause specific mortality hazard.
- CIF increment: CIFR(t+1) = CIFR(t) + S(t) × [1 − exp(−(hR + μD))] × hR / (hR + μD).
- Overall survival: S(t+1) = S(t) × exp(−(hR + μD)).
- Terminal age: the recursion runs to age 100; ages beyond the embedded life table use a default high mortality rate.
This structure is why an older patient sees a lower lifetime number: competing mortality absorbs at-risk years, so an 85 year old cannot accumulate 30 years of revision hazard.
4. Plausibility Checks
- Reference patient (female, age 70, BMI 25, no comorbidities): 12-year CIF near 5 percent, consistent with the Katz lower bound reference, and a remaining lifetime CIF near 10 percent.
- Younger patient (female, age 50, BMI 25): remaining lifetime CIF near 35 percent, reflecting roughly 50 years at risk.
- Elderly patient (female, age 80, BMI 25): remaining lifetime CIF near 4 percent, bounded by competing mortality.
- High risk combination (male, age 50, BMI 40, inflammatory arthritis plus prior surgery plus diabetes): remaining lifetime CIF near 65 percent.
Each figure is a remaining lifetime probability, not a fixed horizon rate.
5. Bibliography
- Katz JN, et al. Twelve-year risk of revision after primary total hip replacement in the US Medicare population. J Bone Joint Surg Am. 2012;94(20):e149.
- Bayliss LE, et al. The effect of patient age at intervention on risk of implant revision after total replacement of the hip or knee. Lancet. 2017;389:1424-1430.
- Bozic KJ, et al. Risk of complication and revision after primary total hip arthroplasty among Medicare patients. Clin Orthop Relat Res. 2010.
- Wright EA, et al. Predictors of 10-year survivorship for total hip replacement. J Bone Joint Surg Am. 2012.
- SooHoo NF, et al. Factors that predict short-term complication rates after total hip arthroplasty. Clin Orthop Relat Res. 2010;468:2363-2371.
- Evans JT, et al. How long does a hip replacement last? A systematic review and meta-analysis of international registries (non-US; late follow-up driven by Finnish Arthroplasty Register). Lancet. 2019;393:647-654.
- National Center for Health Statistics. United States Life Tables, 2023. National Vital Statistics Reports.
- Lacny S, et al. Is the rise of Medicare Advantage impacting the fidelity of implant registries? J Bone Joint Surg Am. 2023.
- Beyersmann J, et al. Competing Risks and Multistate Models with R. Springer, 2012.
