INTRODUCTION
Although prosthesis use is recommended for persons with major upper-limb amputation (ULA) to improve independence and reduce disability,1 many persons with ULA do not utilize a prosthesis. Although some non-prosthesis users are satisfied with not using a device, others have an unmet prosthetic need, defined in this manuscript as a situation in which an individual not using a prosthesis wishes that they were utilizing one. This definition is largely consistent with that used in a recent scoping review on assistive device supply and demand, which defined unmet need as the proportion of a population “who need and do not use” an assistive device.2
Prosthetic rehabilitation is considered fundamental to rehabilitative care of persons with limb loss.3 Prior studies have found that using an upper-limb prosthesis is associated with greater likelihood of employment,4,5 decreased self-reported disability,6 higher quality of life,5 fewer mental health conditions,7 lower likelihood of needing assistance with activities of daily living,6 and improved performance in hygiene, grooming, and dressing.8 Regular use of a prosthesis may help prevent overuse injuries of the non-amputated side, as well as back and neck pain caused by poor compensatory strategies of individuals with upper-limb amputation.9,10
A systematic review of the literature on upper-limb prosthesis abandonment reported mean rates of 26% for body-powered prostheses and 23% for myoelectric devices in the adult population.11 Prosthesis non-use varies by sex, with women more likely to be non-users then men,12,13 and by amputation level, those with transhumeral (TH)-level and shoulder (SH)-level amputation more likely to be non-users than those with transradial level (TR) amputation.14 Other factors associated with prosthesis non-use include perceived lack of comfort and function of existing devices, perceived impediment to sensory feedback,5,15,16 as well as lower satisfaction with health care services.16
Some persons with ULA choose not to wear a prosthesis because they do not perceive a need for one, or feel that they are more comfortable or functional without a prosthesis.16 Others may lack access to prosthetic expertise, which varies across the US,3,17 or may abandon or reject prostheses because they are unable to afford the prosthesis and/or the associated costs for repair and maintenance.16 ,18 A recent study of US veterans and non-veterans with ULA found that 14% of persons who never used a prosthesis and 9.6–16.5% of prosthesis abandoners reported that affordability was a primary reason for non-use with veterans having 0.20 the odds of paying out-of-pocket costs compared to non-veterans.19
Unmet need for assistive technologies, including mobility devices, has been widely studied,2 and scoping reviews on global needs for rehabilitation services and equipment2 point to unmet need for prostheses, particularly in geographic areas affected by war, where prosthetic expertise and resources are lacking. However, no prior studies have examined the prevalence of unmet prosthesis need amongst persons with ULA who were not using a prosthesis. Thus, in this paper we 1) described the prevalence of unmet need for an upper-limb prosthesis and 2) identified independent correlates associated with unmet need for a prosthesis.
METHODS
RECRUITMENT
Participants in the analyses presented in this manuscript were a subset of those recruited for a larger study. Recruitment for the larger study was conducted from multiple sources including: 1) a Department of Veterans Affairs (VA) study,14 2) a list of all veterans who had received VA care between January 1, 2016 – June 1, 2019, 3) emails sent by the Amputee Coalition of America, and 4) letters from a private prosthetics service company. Eligibility criteria included having a major ULA (at wrist level or higher) and no hearing or cognitive conditions impairing ability to understand study requirements and telephone survey prompts. As shown in Figure 1, the analytic sample for the present analyses excluded those with bilateral amputation (N=43) and individuals who did not use a prosthesis who did not respond to a question regarding whether they wished they used a prosthesis (N=14). Current prosthesis users (N=448) were excluded from comparisons of individuals who did and did not wish that they were using a prosthesis. All participants provided oral informed consent prior to completing the telephone survey. The study and oral consent procedures were approved by the Department of Veterans Affairs Central Institutional Review Board. Data were collected between June 2019 and August 2020.
Figure 1.
Flow diagram
SURVEY
The primary dependent variable was unmet prosthesis need. This variable was derived from the survey item, “Do you wish you were using a prosthesis,” which was asked only of individuals who reported that they did not currently use a prosthesis. Those who indicated that they did not use a prosthesis were asked a follow-up question to determine whether they had ever used a prosthesis. The full survey also included questions about amputation date, level, and etiology. Respondents could indicate more than one amputation etiology: combat, accident, burn, cancer, diabetes, and infection or other health problem. Demographic questions included age, sex, veteran status, race (reclassified as white, black, mixed/other, or unknown), ethnicity, and employment status. The original 6 categories for employment status of employed full-time, employed part-time, student, retired, on medical leave, or other were combined to create four categories: employed/student (employed fulltime or part-time and student), unemployed, retired and not working due to disability (disabled). Free text explanations of the “other” response category were examined and used to recategorize responses into one of the four categories. Years since prosthesis use (for those who abandoned a prosthesis) was categorized based on quartiles in the data as: 1 year or less, >1 to 6 years, >6 to 20 years, and more than 20 years. Participants also indicated reasons (more than one could be chosen) for abandoning or never using a prosthesis. The survey also included the Patient-Reported Outcomes Measurement Information System 13-item short form measure of upper-extremity function scored with amputation-specific Rasch calibration (PROMIS-13 UE AMP).20
STATISTICAL ANALYSIS
Descriptive statistics were examined to characterize the entire sample and the characteristics of respondents in each of three study subgroups (current users, and non-users with and without unmet need for a prosthesis). To characterize similarities and differences between non-users and current users, we compared characteristics using bivariate analyses. To identify correlates of unmet need, we compared characteristics of non-users with and without unmet need for a prosthesis by using bivariate analyses. All bivariate analyses were conducted using t-test and chi-squared tests; Wilcoxon Mann-Whitney and Fisher’s exact tests were also conducted when normality was rejected by the Shapiro-Wilk test or when subgroup sizes were below 5 to confirm results were consistent using both parametric and nonparametric methods. We constructed a multivariate logistic regression model of unmet need by adding variables that were statistically significant at p≤0.2 in bivariate analyses. This model also included amputation level and sex given their theoretical importance and a prior relationship with prosthesis use reported in literature.12 ,13, 14
RESULTS
STUDY PARTICIPANTS
Full Sample
In the full analytic sample, there were 448 (65.4%) current prosthesis users and 237 (34.6%) non-users – including 190 (27.7%) participants who had abandoned a prosthesis and 47 (6.9%) participants who had never used a prosthesis. The mean age was 61.2 (sd 14.7) and mean years living with amputation/limb deficiency was 18.2 (sd 18.5). Most participants were male (80.6%), white (79.9%), retired (61.6%), and veterans (77.4%). TR amputation (56.5%) was most common, followed by TH (32.7%) and SH level (10.8%) amputation. The most common amputation etiology was from accidental injuries (66.7%), followed by combat injuries (24.7%).
Non-users vs Users
As shown in Table 1, on average, years living with amputation/limb deficiency were significantly longer for prosthesis users (mn 30.2, sd 14.8) compared to non-users (mn 26.6, sd 18.7). TR amputation was more common among current users than non-users (70.3% vs 30.4%) while TH (23.4% vs 50.2%) and SH level (6.3% vs 19.4%) amputation were less common among current users compared to non-users. Congenital amputation was more common among current users compared to non-users (9.6% vs 4.2%). Compared to non-users, combat (30.2% vs 14.4%) amputation etiology was more common among current users, while accidental injury (64.4% vs 68.2%) and cancer (4.7% vs 9.7%) were less common among current users. There were no differences in PROMIS-13 UE AMP scores by prosthesis use.
Table 1.
Bivariate analyses comparing users and non-users of upper limb prostheses (n=685) WMW=Wilcoxon Mann-Whitney
| Current nonusers (N=237) | Current prosthesis users (N=448) | p | ||
|---|---|---|---|---|
| Mn (sd) | Mn (sd) | T-test | WMW p | |
| Age | 60.9 (14.6) | 61.4 (14.8) | 0.6661 | 0.3312 |
| Years living with amputation | 26.6 (18.7) | 30.2 (19.9) | 0.0317 | 0.0304 |
| Missing (n) | n=18 | n=55 | ||
| PROMIS UE 13-AMP | 50.2 (9.8) | 50.0 (10.1) | 0.8352 | 0.9450 |
| Missing (n) | n=2 | n=14 | ||
| N (%) | N (%) | Chi-squared p | Fisher’s exact p | |
| Status | 0.2484 | 0.2471 | ||
| Veteran | 176 (74.9) | 342 (78.8) | ||
| Non-Veteran | 59 (25.1) | 92 (21.2) | ||
| Missing (n) | n=1 | n=14 | ||
| Sex | 0.2243 | 0.2250 | ||
| Female | 52 (21.9) | 81 (18.1) | ||
| Male | 185 (78.1) | 367 (81.9) | ||
| Race | 0.9108 | 0.9294 | ||
| White | 190 (80.2) | 357 (76.7) | ||
| Black | 23 (9.7) | 41 (9.2) | ||
| Unknown | 32 (7.1) | 32 (7.1) | ||
| Mixed | 18 (4.0) | 18 (4.0) | ||
| Ethnicity | 0.4663 | 0.5179 | ||
| Hispanic | 18 (7.8) | 27 (6.3) | ||
| Not Hispanic | 213 (92.2) | 402 (83.7) | ||
| Missing (n) | n=6 | n=19 | ||
| Employment | 0.0737 | 0.0648 | ||
| Employed/Student | 47 (19.9) | 123 (28.4) | ||
| Unemployed | 7 (3.0) | 8 (1.9) | ||
| Retired | 152 (64.4) | 260 (80.1) | ||
| Disabled | 30 (12.7) | 42 (9.7) | ||
| Missing (n) | n=1 | n=15 | ||
| Amputation level | <0.0001 | <0.0001 | ||
| Shoulder | 46 (19.4) | 28 (6.3) | ||
| Transhumeral | 119 (50.2) | 105 (23.4) | ||
| Transradial | 72 (30.4) | 315 (70.3) | ||
| Etiology of limb loss/deficiency | ||||
| Congenital | 10 (4.2) | 43 (9.6) | 0.0122 | 0.0153 |
| Combat | 34 (14.4) | 122 (30.2) | 0.0001 | 0.0001 |
| Accident (injury) | 161 (68.2) | 260 (64.4) | 0.0102 | 0.0104 |
| Burn | 14 (5.9) | 40 (9.9) | 0.1558 | 0.1811 |
| Cancer | 23 (9.7) | 19 (4.7) | 0.0047 | 0.0069 |
| Diabetes | 2 (0.8) | 4 (1.0) | 0.9458 | 1.0000 |
| Infection / other health problem | 36 (15.2) | 48 (11.9) | 0.0935 | 0.1112 |
Unmet Need vs No Unmet Need
Among non-users, 119 (50.0%) indicated they wished they were using a prosthesis, while 118 (50.0%) indicated they did not wish they were using one. Characteristics of the two groups are included in Table 2. There were no statistically significant differences in prevalence of unmet need by veteran status. On average, those with an unmet need were significantly younger (58.5 vs 63.2 years old), had been living with limb loss/limb deficiency for less time (21.1 vs 32.2 years) and had lower PROMIS-13 UE AMP scores (47.5 vs 52.9) than those without an unmet need. A greater proportion of those with unmet needs reported their employment status as employed/student (23.5% vs 16.2%) and disabled (19.3% vs 6.0%); whereas a smaller proportion reported their employment status as retired (74.4% vs 54.6%) or unemployed (3.4% vs 2.5%) as compared to those with no unmet need. There was a higher rate of ever using a prosthesis among those with versus without unmet needs (85.7% vs 74.6%). There were no significant differences in reasons for abandoning a prosthesis or never using a prosthesis between those with and without unmet prosthesis needs.
Table 2.
Descriptive statistics and bivariate analyses comparing characteristic by desire for prothesis use (n=237)
| Unmet prosthesis needs (N=119) | No unmet prosthesis needs (N=118) | |||
|---|---|---|---|---|
| Mn (sd) | Mn (sd) | T-test | WMW p | |
| Age | 58.5 (13.7) | 63.2 (15.2) | 0.0141 | 0.0004 |
| Years living with amputation | 21.2 (17.0) | 32.2 (18.7) | <0.0001 | <0.0001 |
| Missing (n) | n=8 | n=10 | ||
| PROMIS UE 13-AMP score | 47.5 (8.3) | 52.9 (10.6) | <0.0001 | <0.0001 |
| Missing (n) | n=1 | n=1 | ||
| N (%) | N (%) | Chi-squared p | Fisher’s exact p | |
| Status | 0.3474 | 0.3699 | ||
| Veteran | 86 (73.3) | 90 (76.9) | ||
| Non-Veteran | 33 (27.7) | 26 (22.2) | ||
| Missing (n) | n=0 | n=1 | ||
| Sex | 0.2220 | 0.2721 | ||
| Female | 30 (25.2) | 22 (18.6) | ||
| Male | 89 (74.5) | 96 (81.4) | ||
| Race | 0.9670 | 0.9726 | ||
| White | 96 (80.7) | 94 (79.7) | ||
| Black | 11 (9.2) | 12 (10.2) | ||
| Unknown | 8 (6.7) | 9 (7.6) | ||
| Mixed | 4 (3.4) | 3 (2.5) | ||
| Ethnicity | 0.3552 | 0.4632 | ||
| Hispanic | 11 (9.4) | 7 (6.1) | ||
| Not Hispanic | 106 (90.6) | 107 (93.9) | ||
| Missing (n) | n=2 | n=4 | ||
| Employment | 0.0036 | 0.0024 | ||
| Employed/Student | 28 (23.5) | 19 (16.2) | ||
| Unemployed | 3 (2.5) | 4 (3.4) | ||
| Retired | 65 (54.6) | 87 (74.4) | ||
| Disabled | 23 (19.3) | 7 (6.0) | ||
| Missing (n) | n=0 | n=1 | ||
| Amputation level | 0.9239 | 0.9239 | ||
| Shoulder | 22 (18.5) | 24 (20.3) | ||
| Transhumeral | 61 (51.3) | 5 (49.2) | ||
| Transradial | 36 (30.3) | 36 (30.6) | ||
| Etiology of limb loss/deficiency | ||||
| Congenital | 6 (5.0) | 4 (3.4) | 0.5270 | 0.5270 |
| Combat | 17 (14.3) | 17 (14.4) | 0.9788 | 1.0000 |
| Accident (injury) | 84 (70.6) | 77 (65.8) | 0.4308 | 0.4852 |
| Burn | 7 (5.9) | 7 (5.9) | 0.9870 | 1.0000 |
| Cancer | 10 (8.4) | 13 (11.0) | 0.4968 | 0.5129 |
| Diabetes | 2 (1.7) | 0 (0.0) | 0.1573 | 0.4979 |
| Infection / other health problem | 23 (19.3) | 13 (11.0) | 0.0747 | 0.1025 |
| Ever used prosthesis | 0.0315 | 0.0315 | ||
| Yes | 102 (85.7) | 88 (74.6) | ||
| No | 17 (14.3) | 30 (25.4) | ||
| Years Since Abandoned Prosthesis Use | <0.0001 | <0.0001 | ||
| 1 or less | 37 (37.0) | 14 (16.3) | ||
| >1 to 6 | 26 (26.0) | 17 (19.8) | ||
| >6 to 20 | 27 (27.0) | 25 (29.1) | ||
| More than 20 | 10 (10.0) | 30 (34.9) | ||
| Missing (n) | n=2 | n=2 | ||
| Reasons stopped using a prosthesis: | ||||
| It was too heavy or fatiguing to wear | 65 (65.0) | 55 (62.5) | 0.7219 | 0.7622 |
| It didn’t fit or was uncomfortable | 71 (70.3) | 57 (64.8) | 0.4178 | 0.4391 |
| It was not functional enough | 66 (64.7) | 68 (77.3) | 0.0582 | 0.0788 |
| It was too much fuss | 59 (58.4) | 62 (71.3) | 0.0666 | 0.0696 |
| It broke or was unreliable | 35 (35.0) | 22 (25.3) | 0.1501 | 0.1564 |
| It was not cosmetic | 16 (15.8) | 14 (16.5) | 0.9075 | 1.0000 |
| It was not intuitive to use | 35 (37.2) | 34 (43.0) | 0.4374 | 0.5332 |
| Could not afford prosthesis | 20 (20.6) | 16 (!8.6) | 0.7323 | 0.8525 |
| Dissatisfaction with devices or care | 7 (6.9) | 6 (6.8) | 0.9903 | 1.0000 |
| Injury or health condition | 13 (12.8) | 7 (8.0) | 0.2833 | 0.3469 |
| Never used a prosthesis reasons: | ||||
| Could not afford a prosthesis | 3 (17.7) | 4 (13.8) | 0.7254 | 1.0000 |
| Did not want a prosthesis | 4 (23.5) | 12 (42.9) | 0.1891 | 0.2187 |
| Could not fit* | 10 (58.8) | 10 (34.5) | 0.1080 | 0.1330 |
| Dissatisfaction with access/process of prosthetic care | 3 (17.7) | 7 (24.1) | 0.6064 | 0.7227 |
| Dissatisfaction with fit, function, or weight | 5 (29.4) | 8 (27.6) | 0.8944 | 1.0000 |
WMW=Wilcoxon Mann-Whitney
’could not fit’ includes individuals who stated that they were never offered a prosthesis
The following variables were included in the multivariate model of unmet need (significant at p≤0.2): age, years since amputation, PROMIS-13 UE AMP score, employment status, limb loss etiology of infection or other health problem (not including cancer, physical injury or diabetes), ever having used a prosthesis, years since last prosthesis use, three reasons for abandonment of prosthesis use (“it was not functional enough” and ‘it was too much fuss,’ and “it was broke or unreliable”), and two reasons for never having used a prosthesis (“did not want a prosthesis,” and “could not fit”). Although they did not meet the p<0.2 threshold for inclusion, we added amputation level and sex to the model due to their theoretical importance. Limb loss etiology of diabetes met the criteria for inclusion but was omitted from the final model because an odds ratio could not be calculated because there were no participants with this etiology in the unmet need group.
Results of the multivariate logistic regression model are shown in Table 3. The odds of having unmet need for a prosthesis were lower for those with better upper-extremity function as measured by the PROMIS score (OR: 0.94). Those who were unemployed had 0.60 times lower odds of unmet prosthesis need compared to those who were employed/students. Those who had used a prosthesis in the past year had 4.30 higher odds of unmet prosthesis need compared to those who had not used a prosthesis for 20 years or more, and those who were unemployed due to disability had 4.01 higher odds of unmet need (p=0.0582). Respondents who abandoned a prosthesis because it was too much fuss had 0.41 lower odds of having unmet prosthesis needs compared those who abandoned for some other reason.
Table 3.
Logistic regression for unmet need for a prosthesis.
| Unmet Prosthesis Need (Yes vs No) | ||
|---|---|---|
| OR (95% CI) | p | |
| Age | 1.00 (0.98, 1.04) | 0.7232 |
| Years living with an amputation | 0.98 (0.96, 1.00) | 0.1017 |
| PROMIS UE 13-AMP score | 0.94 (0.90, 0.98) | 0.0017 |
| Sex | ||
| Female | 1.88 (0.76, 4.64) | 0.1736 |
| Male (ref) | ||
| Employment | ||
| Employed/Student (ref) | ||
| Unemployed | 0.15 (0.02, 0.98) | 0.0471 |
| Retired | 0.60 (0.24, 1.50 | 0.2719 |
| Disabled | 4.01 (0.95, 16.85) | 0.0582 |
| Amputation Level | ||
| Shoulder | 1.17 (0.45, 3.06) | 0.7495 |
| Transhumeral | 1.54 (0.70, 3.38) | 0.2871 |
| Transradial (ref) | ||
| Amputation etiology: | ||
| Infection / other health problem | ||
| Yes | 0.54 (0.18, 1.64) | 0.2752 |
| No (ref) | ||
| Ever used prosthesis | ||
| Yes | 3.58 (0.59, 21.69) | 0.1661 |
| No (ref) | ||
| Years since last prosthesis use | ||
| 1 or less | 4.30 (1.28, 14.51) | 0.0187 |
| >1 to 6 | 2.17 (0.67, 7.04) | 0.1983 |
| >6 to 20 | 1.69 (0.57, 5.02) | 0.3463 |
| More than 20 (ref) | ||
| Reasons stopped using a prosthesis: | ||
| It was not functional enough | ||
| Yes | 0.55 (0.23, 1.29) | 0.1693 |
| No (ref) | ||
| It was too much fuss | ||
| Yes | 0.41 (0.18, 0.92) | 0.0306 |
| No (ref) | ||
| It was broke or unreliable | ||
| Yes | 1.72 (0.78, 3.79) | 0.1770 |
| No (ref) | ||
| Never used a prosthesis reasons | ||
| Did not want a prosthesis | ||
| Yes | 0.21 (0.04, 1.20) | 0.0783 |
| No (ref) | ||
| Could not fit ^ | ||
| Yes | 1.13 (0.22, 5.64) | 0.8968 |
| No (ref) | ||
this model includes variables significant at p≤0.2 in bivariate analyses (N=219) and with theoretical significance
could not fit includes individuals who stated that they were never offered a prosthesis
DISCUSSION
This study aimed to characterize prosthesis use in persons with unilateral ULA, quantify prevalence rates of unmet need for upper-limb prostheses amongst persons who were not using a prosthesis, and identify independent correlates associated with such unmet need. Although factors associated with prosthesis non-use have been previously identified in multiple studies, 5, 12–16 ours is the first study to identify independent correlates of unmet prosthesis need.
The overall rate of prosthesis use in our sample was 65.4%, with rates of use varying substantially by amputation level. Only 38% of persons with SH level amputation were utilizing prostheses compared to 47% of those with TH and 81% of those with TR amputation. These findings are largely consistent with the literature.14 While others have reported that prosthesis non-use varies by sex, and that women were more likely to be non-users then men,12,13 we did not observe this. We also did not see statistically significant differences in rates of prosthesis use in our sample between veterans and non-veterans; this may be attributable to the unequal and smaller sample sized non-veteran group. The rate difference in prosthesis use between veterans and non-veterans in the present study was less than the rate difference we observed between our current veteran sample and veterans in an earlier study (66% vs 60%).14 Our data were collected approximately two years after the earlier study,14 and thus it is possible that the rates of prosthesis use that we observed reflect better adoption rates amongst veterans, perhaps due to coordinated efforts to improve amputation care and access to expertise throughout the VA. Further studies are needed to confirm or refute these findings.
The rates of never using a prosthesis in this study are higher than those previously reported in a nationally representative study of veterans with ULA which found that 6.8% had never utilized a prosthesis. Because our study included about the same proportions of persons with amputation at the TH and SH level as compared to the prior study, we believe that the higher rate of never-users in the current study may be explained by the inclusion of non-veterans and the fact that non-veterans may incur copayments for devices or repairs.
Independent factors associated with higher likelihood of unmet prosthesis need in persons who did not use a prosthesis included worse upper-extremity function scores as measured by the PROMIS-13 UE AMP, and not working due to disability (p=0.0582). We observed a trend with decreasing likelihood of unmet need with greater years since prosthesis use. Individuals who had been recent prosthesis users were more likely to report an unmet need. Whereas respondents who reported that they had abandoned a prosthesis because it was “too much fuss” had a lower likelihood of reporting unmet need, suggesting that this subgroup did not find that the function provided by a prosthesis outweighed the inconvenience of using one.
Differences in upper-limb function that were observed between those with and without unmet need were somewhat surprising, given that there was no difference in upper-limb function score between the larger groups of prosthesis non-users and users. Many persons with ULA have developed successful one-handed or adaptive strategies for performing common tasks, and this may explain similarities in scores between users and non-users. Further, upper-extremity (UE) function varies amongst non-users. The finding that those who were unemployed (compared to employed/student) and those with higher self-reported UE function had lower odds of unmet need suggests that persons who are employed and who have poorer UE function have a greater appreciation for the potential functional benefit of having a prosthesis.
Over 14% of those with an unmet need had never utilized a prosthesis and 58.8% of this subgroup indicated that they had never been offered a device or could not be fit with a prosthesis. Given advances in prosthetic technologies and emerging options for suspension for persons with proximal level amputation, such as osseointegration, these findings suggests that persons with ULA who do not utilize a prosthesis should be re-evaluated regularly to identify whether there is an unmet need and if so, if there are appropriate prosthetic options that would enable prosthesis use.
LIMITATIONS
Several limitations should be considered when interpreting the results. Because our sample was predominantly composed of veterans, there are limits to generalizability beyond the veteran population to all persons with ULA. Our overall sample included 151 non-Veterans, recruited through e-blasts and a prosthetics company. It is possible that overall rates of non-use or of unmet needs were lower than US averages, particularly in the non-veteran sample. However, we cannot say how our non-veteran participants compared to the overall US population of non-veterans with unilateral ULA, in part, because such national data are lacking. Future, national studies are needed to confirm rates of unmet need for prostheses amongst non-users of devices. Our largely veteran sample had likely all received prosthetic assessment and care, and the findings from this sample cannot be generalized to persons who do not receive routine care. Although our sample was relatively large for a study of persons with ULA, some of the subgroups that we examined were small, and we may have been underpowered in detecting statistically significant differences. We noted that there were factors that were statistically significant in bivariate analyses at p≤0.2 that did not rise to significance in multivariable models. These factors included ‘‘could not be fit with a prosthesis,” having a prosthesis that was broken or unreliable, and abandoning a prosthesis because “it was not functional enough.” It is possible that sample sizes in each subgroup were too small to detect statistically significant differences in our regression model. For this reason, we believe that it is important to consider point estimates and confidence intervals around those estimates when making inferences about findings. Future studies with larger samples are needed to confirm or refute our findings.
Our study excluded 14 survey respondents with missing data on unmet need. We cannot say with certainty why these respondents did not answer the item, although 3 ended survey participation early without completing all items. Our definition of unmet needs relied on responses to a single survey question, “Do you wish you were using a prosthesis.” It is possible that some participants may have misunderstood or misinterpreted this question. It is possible that respondents might have answered a differently worded question in a different manner. Future studies that include additional questions on unmet prosthetic needs are needed to confirm our findings.
Our finding that there were no differences in PROMIS-13 UE AMP scores between current prosthesis users and non-users has been previously reported and suggests a limitation of the measure in being able to distinguish persons with ULA by prosthesis use.20 However, our study found that prosthesis non-users with unmet need had lower PROMIS-13 UE AMP scores as compared to non-users without unmet need. This finding supports the utility of the PROMIS-13 UE AMP in prosthesis non-users.
Another limitation of our study is that we did not ask respondents to explain why they were not currently using a prosthesis. However, we did ask prosthesis abandoners the reasons that they had abandoned their last device. There were only 10 persons in each subgroup (those with and without unmet need), and a larger proportion of those with unmet need vs those without (58.8% vs 34.4%) indicated that they were not offered a prosthesis/could not be fit. We also did not ask respondents whether they currently had a prosthesis within their possession. It is very likely that some respondents who had been prior prosthesis users did have devices that they no longer used because the devices no longer met their needs.
Our study focused on non-users of prostheses. Because we did not ask current users if they wished that they were utilizing a different type of device, we could not quantify the extent of under-met needs, defined as, “the proportion of persons using a device that is insufficient to maximize their function.”2 We expect, based on reports from other investigations, that a substantial proportion of persons with ULA have under-met needs for a device. Previous reports found that 23% of Vietnam and 44% of OIF/OEF veterans with unilateral ULA indicated that they wanted to change the type of prosthesis that they used.7 Additional analyses, conducted for the current manuscript, using data from a nationally representative study of veterans with ULA,14 found that 36% of veterans with unilateral ULA expressed a desire to change prosthesis types.14
CONCLUSIONS
Unmet need for prostheses impacts about 50% of persons with upper-limb amputation who are not using a device, and 14% of persons who report never having used a prosthesis. These findings suggest that persons with ULA who do not utilize a prosthesis should be re-evaluated regularly to identify unmet needs. Independent correlates of greater unmet need included worse upper-limb function, and having last used a prosthesis within the prior year. Less unmet need was reported among those who indicated that they stopped using a prosthesis because it was “too much fuss,” and those who were unemployed. Taken together, our findings suggest that recent prosthesis users, those with poorer upper-limb function, and those who are employed should be targeted to ensure that prostheses are repaired, replaced, or provided to best meet their needs. Further studies are needed to understand the barriers to prosthesis use amongst non-users with an unmet need.
Acknowledgments:
Data collection for this study was performed while co-author Melissa Clark was affiliated with the Survey Center of the University of Massachusetts Medical Center
Funding Statement:
Department of Veterans Affairs Rehabilitation Research and Development Service A2936-R and A9264-S. Sponsors had no role in study design, collection, analysis, and interpretation of data.
Footnotes
Manuscript blinded information: The study and oral consent procedures were approved by the VA Central IRB.
Conflict of interests: All authors declare that they have no conflict of interests
REFERENCES
- 1.Crunkhorn A, Andrews E, Fantini C, et al. The Management of Upper Limb Amputation Rehabilitation: Synopsis of the 2022 U.S. Department of Veterans Affairs and U.S. Department of Defense Clinical Practice Guideline for Acquired Amputation. Am J Phys Med Rehabil. Dec 8 2022;doi: 10.1097/PHM.0000000000002164 [DOI] [PubMed]
- 2.Danemayer J, Boggs D, Polack S, et al. Measuring assistive technology supply and demand: A scoping review. Assist Technol. Dec 1 2021;33(sup1):35–49. doi: 10.1080/10400435.2021.1957039 [DOI] [PubMed] [Google Scholar]
- 3.Pasquina CP, Carvalho AJ, Sheehan TP. Ethics in Rehabilitation: Access to Prosthetics and Quality Care Following Amputation. AMA J Ethics. Jun 1 2015;17(6):535–46. doi: 10.1001/journalofethics.2015.17.6.stas1-1506 [DOI] [PubMed] [Google Scholar]
- 4.Raichle KA, Hanley MA, Molton I, et al. Prosthesis use in persons with lower- and upper-limb amputation. J Rehabil Res Dev. 2008;45(7):961–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Yamamoto M, Chung KC, Sterbenz J, et al. Cross-sectional International Multicenter Study on Quality of Life and Reasons for Abandonment of Upper Limb Prostheses. Plast Reconstr Surg Glob Open. May 2019;7(5):e2205. doi: 10.1097/GOX.0000000000002205 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Resnik L, Borgia M, Clark M. Function and Quality of Life of Unilateral Major Upper Limb Amputees: Effect of Prosthesis Use and Type. Arch Phys Med Rehabil. Aug 2020;101(8):1396–1406. doi: 10.1016/j.apmr.2020.04.003 [DOI] [PubMed] [Google Scholar]
- 7.Reiber GE, McFarland LV, Hubbard S, et al. Servicemembers and veterans with major traumatic limb loss from Vietnam war and OIF/OEF conflicts: survey methods, participants, and summary findings. J Rehabil Res Dev. 2010;47(4):275–97. [DOI] [PubMed] [Google Scholar]
- 8.Atkins DJ, Heard DC, Donovan WH. Epidemiologic Overview of lndividuals with Upper-Limb Loss and Their Reported Research Priorities. Journal of Prosthetics & Orthotics. Winter 1996 1996;8(1):0. [Google Scholar]
- 9.Rock Gambrell C. Overuse Syndrome and the Unilateral Upper Limb Amputee: Consequences and Prevention. Journal of Prosthetics & Orthotics (JPO). 2008;20(3):126–132. [Google Scholar]
- 10.Hanley MA, Ehde DM, Jensen M, Czerniecki J, Smith DG, Robinson LR. Chronic pain associated with upper-limb loss. Am J Phys Med Rehabil. Sep 2009;88(9):742–51; quiz 752, 779. doi: 10.1097/PHM.0b013e3181b306ec [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Biddiss EA, Chau TT. Upper limb prosthesis use and abandonment: a survey of the last 25 years. Prosthet Orthot Int. Sep 2007;31(3):236–57. [DOI] [PubMed] [Google Scholar]
- 12.Resnik LJ, Borgia ML, Clark MA. A National Survey of Prosthesis Use in Veterans with Major Upper Limb Amputation: Comparisons by Gender. PM R. Nov 2020;12(11):1086–1098. doi: 10.1002/pmrj.12351 [DOI] [PubMed] [Google Scholar]
- 13.Biddiss EA, Chau TT. Multivariate prediction of upper limb prosthesis acceptance or rejection. Disabil Rehabil Assist Technol. Jul 2008;3(4):181–92. [DOI] [PubMed] [Google Scholar]
- 14.Resnik L, Ekerholm S, Borgia M, Clark MA. A national study of Veterans with major upper limb amputation: Survey methods, participants, and summary findings. PLoS One. 2019;14(3):e0213578. doi: 10.1371/journal.pone.0213578 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Smail LC, Neal C, Wilkins C, Packham TL. Comfort and function remain key factors in upper limb prosthetic abandonment: findings of a scoping review. Disabil Rehabil Assist Technol. Nov 2021;16(8):821–830. doi: 10.1080/17483107.2020.1738567 [DOI] [PubMed] [Google Scholar]
- 16.Biddiss E, Chau T. Upper-limb prosthetics: critical factors in device abandonment. Am J Phys Med Rehabil. Dec 2007;86(12):977–87. [DOI] [PubMed] [Google Scholar]
- 17.Jette AM. The Promise of Assistive Technology to Enhance Work Participation. Phys Ther. Jul 1 2017;97(7):691–692. doi: 10.1093/ptj/pzx054 [DOI] [PubMed] [Google Scholar]
- 18.Etter K, Borgia M, Resnik L. Prescription and repair rates of prosthetic limbs in the VA healthcare system: implications for national prosthetic parity. Disabil Rehabil Assist Technol. Nov 2015;10(6):493–500. doi: 10.3109/17483107.2014.921246 [DOI] [PubMed] [Google Scholar]
- 19.Resnik L, Borgia M, Clark M, Ni P. Out-of-pocket costs and affordability of upper limb prostheses. Under Review. 2023; [DOI] [PubMed]
- 20.Resnik LJ, Stevens PM, Ni P, Borgia ML, Clark MA. Assessment of Patient-Reported Physical Function in Persons with Upper Extremity Amputation: Comparison of Short Form instruments drawn from the PROMIS v2.0 Upper Extremity item bank. Am J Phys Med Rehabil. Jun 15 2022;doi: 10.1097/PHM.0000000000002044 [DOI] [PMC free article] [PubMed]