Journal of NeuroEngineering and Rehabilitation

quality of life, this leads to 91.4 more QALYs in MPK users compared to NMPK users (453.3 vs. 361.9). The probabilistic sensitivity analysis supports the same con- clusions: on average, the number of life years increases by14 years, ranging from 5 to 25 years per 100 MPK users, and the discounted QALYs gained average 102 years, ranging from 82 to 125 years per 100 MPK users. Economic benefits Direct healthcare cost and indirect cost The simulation for a 10-year period shows reductions in falls and incident osteoarthritis of the intact knees correspond to savings in direct healthcare cost of $3496 and $180 per person per year, respectively. Overall, on a per person per year basis, MPK users have a lower direct healthcare cost than NMPK users, $2890 vs. $6566 (Fig. 3 ) . Moreover, MPK is associated with a reduction of $909 ($4268 vs. $5177) in indirect cost, which includes lost wages, caregiving expenses, and transportation expenses. Cost of device acquisition Over a 10-year time period, MPK acquisition and repair cost amounts to $7925 per person per year taking into account the effect of survival. The estimate varies from $6054 to $8379 in the probabilistic sensitivity analysis. Similarly, on a per person per year basis, the acquisition and repair of NMPK cost $1638, varying from $785 to $2183 according to the probabilistic sensitivity analysis. Total cost The resulting total cost in the simulation, defined as the sum of direct ($2890 vs. $6566), indirect ($4268 vs. $5177), and device acquisition and repair cost ($7925 vs. $1638), is $15,083 and $13,382 per person per year for MPK and NMPK users, respectively. The total cost esti- mates for both MPK and NMPK users are sensitive to the proportion of medical falls among all falls, the aver- age number of falls per faller per year, medical cost per major or minor injury fall, osteoarthritis-related medical cost, and discount rate, as indicated in the one-way sen- sitivity analyses. In the best scenario, the total cost per person per year for MPK users is $5042 lower than that of NMPK users; in the worst scenario, MPK users cost Fig. 2 Quality of Life Among MPK and NMPK Users, Note: MPK: microprocessor-controlled knees; NMPK: non-microprocessor controlled knees. Source: [ 15 , 28 , 30 , 35 – 40 ] Fig. 3 Savings Derived From the Use of MPK in Direct Healthcare Cost and Indirect Cost, Note: MPK: microprocessor-controlled knees; NMPK: non-microprocessor controlled knees. Results are reported on a per person per year basis. All costs are in 2016 U.S. dollars Chen et al. Journal of NeuroEngineering and Rehabilitation 2018, 15 (Suppl 1):62 Page 54 of 72