The Hidden Cost of Mid-Week Surgical Backlog: A Pre-Registered Value-Based Analysis of Day-of-Week Scheduling, Length of Stay, and Cost After Primary Knee Arthroplasty

Objectives. To quantify the health-economic burden of day-of-week scheduling patterns after primary knee arthroplasty and to test whether scheduling information adds value to machine-learning (ML) prediction of extended length of stay (eLOS), within a pre-registered regression + causal-mediation + ML framework using a value-based-care lens. Methods. Pre-registered (OSF 10.17605/OSF.IO/8EM7Y), locked-dataset analysis of 1,167 consecutive primary total or unicompartmental knee arthroplasties at a Chinese tertiary center (2019–2025). Primary outcome: eLOS (≥12 days). We estimated adjusted day-of-week effects (logistic and Gamma-generalized-linear-model regression), decomposed the admission-day→time-to-surgery→LOS pathway (mediation, 1,000 bootstraps), fitted nine ML algorithms across four feature sets (±day-of-week), and translated effects into attributable inpatient cost and a scheduling-reform cost offset (CHEERS 2022). Results. Operative volume peaked mid-week from weekend-closure backlog. Wednesday operations carried the highest adjusted eLOS odds of any operation day (aOR 1.87; 95% CI 1.23–2.85; 1.82 under reference-free deviation coding) and 8% longer LOS. Each additional pre-operative day increased LOS by 8.5%, implying ≈¥3,210 (US$447) avoidable bed-day cost per patient per waiting day; a one-day reduction in mean time-to-surgery projected ≈¥0.93 million (US$130,000) in annual cost offset at this center. Adding day-of-week features did not improve ML discrimination (mean ΔAUC −0.007) or net benefit; time-to-surgery dominated all scheduling features (mean|SHAP| 1.16 vs 0.13). Conclusions. The economic value of scheduling information is fully captured by time-to-surgery. Investment should target throughput, not day-of-week redesign; the pre-registered null constrains researcher degrees of freedom and yields actionable, economically grounded policy guidance.