Journal of NeuroEngineering and Rehabilitation

increases the cost by $50,000 or less for every QALY gained, it is considered having good value for the money. MPK has an ICER of $11,606 per QALY, well below the commonly accepted threshold, and thus provides good value for the money. In addition, even when compared to technologies commonly reimbursed by payers in the U.S., MPK fares better. For example, total knee arthro- plasty and prophylactic cardioverter defibrillator im- plantation have an ICER of $14,572 and $76,396 per QALY, respectively [ 52 – 56 ] , both of which are higher than that of MPK. Because the payment system lags behind the advance- ment in technologies and focuses on cost-cutting rather than value, there is a need to shift the dialogue from a cost-driven payment approach to a value-based payment approach. And this is exactly where the U.S. health care system is headed [ 57 ] . However, sophisticated payment approaches such as outcome-based contracts or risk-sharing arrangements require the industry to de- velop sophisticated methodologies and robust evidence for the economic value of new technologies. As reflected in AOPA ’ s Prosthetics 2020 Initiative, the industry is aiming to build the infrastructure needed for evidence generation, such as establishing patient registries and collecting clinical and economic data [ 58 ] . The initiative will help facilitate such a transition, while this analysis and the research gaps identified could serve as a good starting point. Limitations There is a limited number of studies that directly com- pare MPK to NMPK, and some model parameters had to come from studies examining a non-amputee popula- tion. For example, the proportion of medical falls out of all falls came from the non-amputee literature. The model assumes that parameters are generalizable from a non-amputee population to an amputee population. The quality of the studies used to extract model param- eters is suboptimal. For the parameters needed in the model, there are no published randomized clinical trials that compare MPK to NMPK. In addition, studies cited often have small sample sizes that could lead to large un- certainty in the estimates of MPK ’ s impact [ 26 , 59 ] . Also, studies comparing MPK to NMPK often collected data for a limited time period varying from several weeks to a year. For the studies with less than one-year observation period, findings beyond the study period had to be extrapolated for modeling purposes. No studies have examined long-term health outcomes such as obesity, diabetes, and cardiovascular diseases. The lack of studies on these long-term outcomes could potentially lead to an under-estimation of the economic impact of MPK. Existing studies also have a narrow focus in terms of the amputee population, with an average age of between 38 and 62 and a functional level of K3 or K4. As a result, the effects of MPK on various out- comes might not be generalizable to the Medicare population although the Medicare population does contain numerous younger constituents, for example, dual-eligible combat injured Veterans and others. For the same reason, the modeling results for the K1 and K2 population may not be reliable because most model parameters were extracted from studies for K3 and K4 amputees. No prior studies examined directly MPK ’ s impact on incident osteoarthritis and the model relies on expert opinion. While it is generally accepted that differences in gait mechanics manifest in the development of osteo- arthritis, there are no studies that demonstrate the caus- ality. Expert consultation suggested that knee moments may represent a reasonable surrogate for the develop- ment of osteoarthritis; however, in the absence of long-term studies, it is a limitation. Finally, the current Medicare payments were used as the numerator of the ICER. Since payment levels are dif- ferent from the cost of manufacturing MPKs or NMPKs, the estimated numerators of ICERs may not represent the true cost differences between MPKs and NMPKs. If payment levels of MPK and NMPK change in the future, ICER ratios will change accordingly. Conclusions Prior studies have demonstrated that for transfemoral amputees, MPK is superior to NMPK in improving physical function and is associated with sizeable reduc- tions in injurious falls and incident osteoarthritis in the intact limb. Once converted to economic benefits, MPK has an ICER of $11,606 per QALY gained and therefore provides good value for the money compared to NMPK. MPK ’ s economic benefits are comparable to or even greater than widely reimbursed technologies such as total knee replacement and implantable cardioverter defibrillator. Additional file Additional file 1: Table S1. Baseline Characteristics of Medicare Patients with a Unilateral Transfemoral Amputation, 2011 – 2014. (DOCX 50 kb) Abbreviations AOPA: American orthotic and prosthetic association; CMS: Centers for medicare and medicaid services; EQ-5D: EuroQol five dimensions questionnaire; ICER: Incremental cost-effectiveness ratio; K1: Medicare functional classification level 1; K2: Medicare functional classification level 2; K3: Medicare functional classification level 3; K4: Medicare functional classification level 4; MPK: Microprocessor-controlled prosthetic knee; NMPK: Non – microprocessor-controlled prosthetic knee; PEQ: Prosthesis Evaluation Questionnaire; QALY: Quality-adjusted life year; SF-36: 36-Item Short Form Health Survey Chen et al. Journal of NeuroEngineering and Rehabilitation 2018, 15 (Suppl 1):62 Page 56 of 72