JBJS - 2026-09-01 - Journal Article
Heterogeneity of Patients' Preferences for Bone-Anchored Prostheses After Lower-Extremity Amputation (PREFER-BAP-2): A Latent Class Analysis of a Discrete Choice Experiment in The Netherlands.
Saleib SM, Jonker MF, Verhofstad MHJ, Paping MA, Van Vledder MG, Leijendekkers RA, Van Waes OJF
Topics
Key Takeaway
Latent class analysis of 247 Dutch BAP patients identified 3 distinct preference subgroups, with QoL driving 26–53% of decision weight across classes and long-term complications dominating Class 3 (24% of patients, predominantly male, lower education, pre-surgical).
Summary Depth
Choose how much analysis to show on this article page.
Summary
This discrete choice experiment analyzed preference heterogeneity for bone-anchored prosthesis (BAP) attributes among 247 Dutch lower-extremity amputees using a latent class logit model. Three subgroups emerged: Class 1 (49%) weighted QoL and cost; Class 2 (27%) prioritized QoL and was insensitive to short-term complications; Class 3 (24%) weighted long-term complications most and was cost-sensitive only at extreme out-of-pocket levels (€20,000–€25,000). Background characteristics—including amputation level, education, income, pre-treatment functional status, and surgical timing—differed significantly across classes.
Key Limitation
Cross-sectional preference elicitation at a single time point cannot capture how BAP experience alters attribute weighting, meaning pre-surgical patients (concentrated in Class 3) may shift preference class after implantation.
Original Abstract
BACKGROUND
Patients with a bone-anchored prosthesis (BAP) previously expressed preferences regarding (i.e., indicated the importance of) 5 treatment attributes: change in quality of life (QoL), short- and long-term complications, implant survival, and out-of-pocket contributions (costs). However, preference heterogeneity is unclear. We aimed to identify latent preference subgroups, quantify attribute importance per subgroup, and describe background characteristics associated with each subgroup.
METHODS
A discrete choice experiment to reveal preferences for BAP characteristics among 247 patients from The Netherlands was analyzed using a latent class logit model. Subgroups (classes) were interpreted using relative attribute weights and patient characteristics.
RESULTS
A model with 3 classes (with 49% [n = 121], 27% [n = 67], and 24% [n = 59] of the patients in Classes 1, 2, and 3, respectively) fit best. Across classes, short-term complications were least important, and patients considered opting out of BAP treatment when outcomes were unfavorable. Within the classes, when BAP was selected, QoL remained a key driver (31% of decision weight in Class 1, 53% in Class 2, and 26% in Class 3). Class 1 weighted QoL and out-of-pocket contributions most; Class 2 prioritized QoL and disregarded short-term complications; Class 3 placed the most weight on long-term complications and was sensitive to out-of-pocket contributions mainly at high cost levels (€20,000 to €25,000) and to osseointegrated implant survival only when survival dropped from 20 to 5 years. Background profiles differed: compared with the other classes, Class 1 (66% male) had a greater proportion of patients with transtibial amputations and longer BAP experience; Class 2 (48% male) included a greater proportion with higher education and income levels and worse pre-treatment mobility, pain, and anxiety; and Class 3 (80% male) had a greater proportion with lower education and income levels, and the highest proportion awaiting surgery.
CONCLUSIONS
Patient preferences for key aspects of BAP treatment are heterogeneous. We identified 3 distinct subgroups, and a patient's background characteristics can help estimate membership in a particular subgroup. Patient-reported outcomes should be interpreted in light of these preference differences. Future work should integrate preference assessment into pre-treatment triage to better align counseling and selection with what patients value, ideally using a brief choice-task screening tool; the optimal screening tasks remain to be established.
LEVEL OF EVIDENCE
Therapeutic Level III. See Instructions for Authors for a complete description of levels of evidence.