Does the Surgical Innovation Evidence Apply to You? How Surgeon Volume and Experience Determine Whether Adoption Helps or Harms

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Background Studies of new surgical techniques usually report the result without separating outcomes by factors such as surgeon volume, experience, or practice setting. But those same factors shape the result, so a technique that benefits patients in the hands of the surgeons who reported it might actually harm patients in the hands of surgeons who meaningfully differ. Published outcomes describe the surgeons and conditions that produced them, but there is little to guide surgeons considering a new approach beyond a nonspecific sense of risk versus reward. Questions/purposes (1) Does the decision to adopt a surgical technique reverse depending on surgeon-level variables, even when the underlying evidence is the same? (2) Can a formal model identify the conditions under which adoption produces net harm to a surgeon’s patients? Methods A volume-stratified decision-analytic model was developed that calculates a surgeon-specific break-even horizon, the point at which cumulative benefit first exceeds cumulative learning-curve harm. The model was applied to direct anterior approach versus posterior approach hemiarthroplasty, with parameter values drawn from the published direct anterior approach and posterior approach evidence and stratified by annual case volume. The model integrates three parameters: the number needed to treat (NNT) for a single claimed benefit of the new technique; a learning-curve decay constant (λ) derived from cumulative sum (CUSUM) analyses, where a higher value means faster learning (λ = 0.07 implies proficiency in roughly 40 cases, λ = 0.03 in roughly 100); and a complication severity weight of harm (W h ) anchored to the risk of reoperation. The benefit is granted favorably and weighed against learning-curve harm assigned to the new approach alone, ignoring the established approach’s advantages and the cost of switching; every simplification therefore favors adoption. An interactive tool lets readers enter their own values. Results The decision to adopt a new surgical technique depended heavily on surgeon-specific factors, chiefly annual case volume. The same published evidence pointed toward adoption for a higher-volume surgeon and against it for a lower-volume one. Three parameter scenarios were tested across three volume tiers (8, 20, and 50 patients per year). Under innovation-favorable parameters (NNT = 50, λ = 0.07, W h = 1), all three tiers reached breakeven within 4 years, and 20-year cumulative utility was positive throughout (2.5 to 19.3). Under default parameters (NNT = 100, λ = 0.05, W h = 3), breakeven ranged from 5 years 10 months at 50 patients per year to beyond 20 years at the lowest tier (20-year utility 7.1 to -1.3), and the 8-patient-per-year surgeon never reached breakeven. Under innovation-unfavorable parameters (NNT = 140, λ = 0.04, W h = 4), only the 50-patient surgeon reached breakeven, at 13 years 9 months (2.2); the lower tiers ended in net harm (-2.0 and -3.7). In every scenario, the higher-volume surgeon did better. What shifted with the parameters was only whether adoption was favorable at all: uniformly so under optimistic assumptions, volume-dependent at default values, and favorable only at the highest volume under unfavorable ones. Conclusion Whether a newly reported technique helps or harms depends heavily on how often the operating surgeon performs the procedure and how much relevant experience they bring, factors that the published studies rarely report. This holds for any innovation that offers a modest benefit and carries a learning curve, not only the approach examined here. Clinical Relevance Published innovation studies report outcomes from one set of surgeons, at particular volumes and levels of experience, but are often read as if they applied to everyone. In deciding whether to adopt a new technique, surgeons should weigh their own case volume and relevant experience alongside the published result, recognizing that a beneficial-sounding result may not reach breakeven at lower volume within a clinically meaningful period. Future studies could extend this work by reporting outcomes stratified by surgeon volume and experience so that readers can locate themselves rather than assume that a published result transfers. An interactive version of the model, which lets readers enter their own volume and experience, is available at https://rcpmodel.netlify.app.