Understanding repair and replacement of prosthetic limbs using routinely-collected data: a retrospective study over three decades in Cambodia

Prosthetic limbs deliver major quality of life and socioeconomic benefits for people with amputation, particularly in low-resource settings. The value of administrative data analysis to enable sustainable health care improvement is established, but there has been limited research into the failure, repair, and replacement of prosthetic limbs. Survivorship data is sparse and highly variable, and rarely addresses differences between demographic groups. Therefore, we investigated the distribution of time between device delivery, repair and replacement for a Cambodian cohort, considering the influence of a range of service delivery, user demographic and health characteristics. We conducted Kaplan-Meier survival analysis, and a Cox model to compare repair and replacement likelihood between groups. The study explored 14,822 device deliveries to 6,986 clients, with median 3 devices/person (interdecile range IDR 1-9), and 22,205 repairs, median 1 repair/device (IDR 0-4). The median device survival before repair was 237 days (IDR 38-854), and replacement was 727 days (IDR 208-2154). Devices used by people in more active occupations were repaired earlier, and devices were replaced earlier when used by children, replaced later for upper- than lower-limb devices, and replaced earlier for volume change than for wear and tear. Several less intuitive trends were revealed such as different preferences or capacities between clinics for device repair vs. replacement, and earlier device repair and replacement for women than men. Prosthetic limb repair and replacement is influenced both by the devices durability and the users access to well-resourced physical rehabilitation services. A worn-out device may indicate poor quality, or the opposite: that it fitted well and enabled great mobility. However, such analysis may enable us to identify groups who are less well-served by current devices or rehabilitation models, and contribute to cost effectiveness analysis of current services. Furthermore, the findings represent benchmark data against which engineers could measure new technologies, to ensure that innovation justifies its inherent risk by offering a genuine improvement which balances functionality, cost, and durability.