Disparities in functional recovery after dysvascular lower limb amputation are associated with employment status and self-efficacy

Abstract

Purpose:

Employment status is considered a determinant of health, yet returning to work is frequently a challenge after lower limb amputation. No studies have documented if working after lower limb amputation is associated with functional recovery. The study purpose was to examine the influence of full-time employment on functioning after lower limb amputation.

Methods:

Multisite, cross-sectional study of 49 people with dysvascular lower limb amputation. Outcomes of interest included performance-based measures, the Component Timed-Up-and-Go test and the 2-minute walk test, and self-reported measures of prosthetic mobility and activity participation.

Results:

Average participant age was 62.1 ± 9.7 years, 39% were female and 45% were persons of color. Results indicated 80% of participants were not employed full-time. Accounting for age, people lacking full-time employment exhibited significantly poorer outcomes of mobility and activity participation. Per regression analyses, primary contributors to better prosthetic mobility were working full-time (R² ranging 0.06–0.24) and greater self-efficacy (R² ranging 0.32–0.75).

Conclusions:

This study offers novel evidence of associations between employment and performance-based mobility outcomes after dysvascular lower limb amputation. Further research is required to determine cause-effect directionalities. These results provide the foundation for future patient-centered research into how work affects outcomes after lower limb amputation.

Introduction

Initiatives by the World Health Organization (WHO) and National Institutes of Health (NIH) recognize employment as a social determinant of health [1, 2], for the contribution work has toward economic and social resources, and health disparities. Therefore, is imperative that work-related research becomes a priority to address quality of life for all persons across the life span. Lower limb amputation (LLA) is one of the costliest [3, 4], most debilitating, and potentially preventable chronic conditions affecting global society [5], and long-term disablement after LLA is a significant concern [6]. In the U.S., it is estimated within the next 30 years there will be 3.6 million people living with LLA [7], with the majority of LLAs due to dysvascular conditions like diabetes mellitus and peripheral arterial disease. Problematically, the incidence of dysvascular LLA is increasing in an ever-younger population, with people 18–64 years of age exhibiting increased rates of LLA [8]. Causes for the observed increase are likely multifactorial, potentially involving issues of early intervention, barriers to care, etc. This is concerning, as younger individuals are still of viable working age and often have personal and family obligations that necessitate employment. Yet, gaps in evidence remain regarding how employment may be beneficial to functional recovery, and no work-related research has focused exclusively on people with dysvascular LLA.

As full-time employment in the U.S. is often associated with higher income and health insurance benefits, not working full-time threatens to relegate people with LLA to poorer outcomes, greater reliance on federal assistance, and decreased quality of life. Detrimental changes to quality of life frequently accompany LLA resulting from impaired physical abilities [9], interruption of employment [6], and barriers to community reintegration [10]. Working-aged individuals with LLA experience extended periods of unemployment or underemployment post-surgery [6, 11], often not returning to full-time work for years [12], if ever. Return to work rates after LLA vary widely, with rates between 48–89% having been reported [6]. White/Caucasian men, those with higher annual income, more years of education, and a stronger social support network [13, 14] are more likely to be employed after LLA. Additionally, evidence suggests that factors like perceived physical functioning and walking ability [12], and higher self-efficacy contribute to employment after LLA [6, 14]. Self-efficacy, in particular, could also result from the sense of pride someone experiences with working, and is a component of overall well-being [15]. However, there is a dearth of research on employment and functioning after LLA and results are largely variable [6, 13].

Prosthetic mobility is a construct encompassing daily functioning with a prosthetic limb, including, but not limited to, tasks of walking, sit-to-stand transitions, and negotiation of obstacles [16]. One of the limitations of previous work-related research in the LLA population was the exclusive use of self-reported (perceived ability to perform a task) measures in lieu of performance-based (physical capacity to perform a task) measures to examine associations between functioning and employment. This approach offers no evidence on how a person’s ability to physically execute prosthetic mobility tasks could be associated with their work status. This discrepancy could be of particular importance in the LLA population, where self-reported and performance-based prosthetic mobility outcomes are only moderately correlated [17] and excluding one type of measure risks losing information about an individual’s ability to utilize the prosthetic limb. Therefore, the primary purpose of this pilot study was to explore prosthetic mobility, using both performance-based and self-report outcome measures, to assess how full-time employment might serve as a determinant of health in its association to functioning after LLA. We hypothesized that lack of full-time employment would be associated with poorer prosthetic mobility and participation in social activities. Examining prosthetic mobility as the dependent variable offers a unique approach to our study that differs from typical employment-focused research where return to work serves as the main outcome of interest.

Materials and Methods

Study Design and Participants

This was a cross-sectional study of English and Spanish speaking, community-dwelling (not living in any type of assisted living environment) people with dysvascular LLA. Data was collected, during an 11-month period, at prosthetic clinics in four distinct geographical locations in the U.S., including metropolitan areas of Miami and Tampa, Florida, Las Vegas, Nevada and the more suburban area of St. George, Utah. Inclusion criteria were age 18–80 years, having LLA above ankle disarticulation on at least one leg due to dysvascular causes (e.g., diabetes, peripheral vascular disease), and ability to walk at least 30 feet without assistance of another person. The broad age range allowed for inclusion of adults with dysvascular LLA across the lifespan and provides a unique opportunity to explore the findings in people who were working beyond the typical “retirement age” of 65 years. The Institutional Review Board, located at Florida International University approved all study protocols including the necessary mitigation procedures to safeguard research personnel and study participants at each study site due to the COVID-19 pandemic. Informed consent was provided by all participants prior to initiation of study procedures. All screening and data collection was performed or supervised by research personnel who were licensed physical therapists. Order of study related procedures is presented in figure 1.

Figure 1:

Flow diagram of study procedures.

Measures

Variables of interest were classified by the domains of the WHO’s International Classification of Functioning, Disability, and Health (ICF) model [18] to include physical impairment (amputation level), activity limitations (performance-based and perceived prosthetic mobility tests), and participation restrictions (perceived ability to participate in work and social activities, current employment status), as well as personal (race, sex, self-efficacy) and environmental (neighborhood, community accessibility) factors. Outcome measures of interest are indicated by bold text (figure 2). Previous amputation-related research has utilized the ICF model as a framework to examine interactions amongst variables that affect prosthetic mobility [19, 20], providing an opportunity to promote comprehensive data collection, interpretation, and dissemination across professions.

Figure 2.

Proposed International Classification of Functioning, Disability, and Health (ICF) theoretical model for examining the interaction of variables associated with functioning after LLA.

All study data were collected electronically using REDCap [21] to ensure standardization of study procedures and secure data storage across collection sites. Potential participants could complete study consents and surveys at home to minimize participant/research personnel physical interaction due to the COVID-19 pandemic. Participants were personally contacted by study personnel to clarify any missing data. All performance-based measures were administered in-person at designated study sites.

Activity limitations

To measure participants ability to perform basic prosthetic mobility tasks within the activity limitation domain, two performance-based and one self-report outcome measure were administered. The Component Timed-Up-and-Go test (cTUG) is based on the traditional Timed-Up-and-Go test (TUG) [22] but provides additional information regarding important motor task performance like turning and sit to stand transfers [16]. For study purposes, the total time was used to assess basic prosthetic mobility, with a lower total time equating to better functioning. The second performance-based measure was the 2-minute walk test (2-MWT). The 2-MWT is frequently measured in patients with LLA to determine prosthetic mobility and endurance [23, 24]. The test was administered using rectangular paths and instructing the participant to, “Cover as much distance as possible in two minutes.” Upon “Go,” the participant walked continuously for two minutes, resting as needed, and upon completion the distance walked was measured in meters. Better mobility and endurance resulted in greater total distance ambulated.

To measure self-reported prosthetic mobility the Prosthetic Limb Users Survey of Mobility (PLUS-M^™) 12-item short form was administered. The PLUS-M asks individuals to report on their perceived ability to perform walking tasks while wearing their prosthesis. The PLUS-M has been found valid and reliable in the LLA population and uses t-scores to allow for comparisons to a development sample and higher scores indicate better prosthetic mobility [25, 26]. Data is continuous in nature and the maximum achievable score is 71.4 points.

Participation restrictions

Questionnaires were used to assess social and work-related constructs within the participation restriction domain of the ICF model. The Patient-Reported Outcomes Measurement Information System (PROMIS) [27] Ability to Participate in Social Roles and Activities survey was administered to capture participants’ perceptions of their capability to engage in leisure and work activities. The eight-item short form includes two items specific to work, including work at home, and six items related to leisure activities with family and friends. Maximum achievable score is 65.4 points, and a lower score indicates more perceived difficulty engaging in activities. The PROMIS family of measures were developed by the NIH to provide psychometrically sound tools for clinical and research purposes, with the intent of providing a universal, standardized means of collecting patient-reported information. All PROMIS measures provide continuous data, with a raw score conversion to a t-score offering the opportunity for comparisons to the general population. Additionally, participants were asked to self-report their current employment status (choices included working full-time or part-time, unemployed, retired, or disabled).

Personal and environmental factors

To capture variables associated with the personal factors domain, demographic information including age, sex, race/ethnicity, and yearly family income were collected for descriptive purposes. Additionally, the Functional Comorbidity Index [28] was administered to assess the number and type of comorbidities reported by participants. This measure is an 18-item list of diagnoses (e.g., arthritis, stroke, diabetes) that are associated with physical functioning ability. A higher score indicates greater presence of comorbidities. Lastly, self-efficacy was assessed using the 8-item short-form PROMIS Self-Efficacy for Managing Chronic Conditions-Managing Daily Activities. This outcome measures someone’s confidence in performing common daily activities (e.g., shopping, chores, usual work activities) without assistance. Maximum achievable score is 60.74 points and poorer self-efficacy is indicated by a lower score.

To determine the effects of environmental factors on prosthetic functioning, four discrete items were chosen from the Environmental Factors Item Bank (EFIB) [29] to measure the participant’s natural and built environment, and accessibility within their community. The EFIB was developed based on the ICF’s taxonomy for classifying environmental factors as barriers or facilitators to health and well-being. Participants are asked to rate their perceived ability to move around within their home and community (e.g., stores, neighborhoods) and determine barriers to their mobility (furniture, accessible sidewalks, safety, etc.) Items are scored on a 5-point scale (1=no difficulty to 5=extreme difficulty) resulting in a t-score for comparison to the sample mean, and a percentage is provided. The higher the percentage, the more environmental barriers the participant perceives. Additionally, the PROMIS Instrumental Support eight-item short form was administered to evaluate participants’ social support and perceived availability of assistance with their needs in their immediate environment. Lastly, adapted from the EFIB, participants were asked to answer yes or no to two statements regarding the availability of necessary rehabilitation services and availability of adequate medical insurance [30].

Statistical analyses

For the purposes of this study, participants were categorized by presence or absence of full-time (FT) employment, with those having reported part-time, unemployment, retired, or disabled being categorized as no full-time (NFT) employment. Because the study purpose was to determine the impact of FT employment on functioning, participants who did not self-categorize as working FT were examined separately, regardless of age. Having participants self-categorize provided them the ability to determine for themselves their perceived contribution to the work force, potentially taking into consideration elective retirement or non-traditional employment situations. Descriptive statistics were used to analyze demographics and outcome measure results stratified by the two categories of work status. Normality of the score distributions was determined using the Shapiro–Wilk test [31] and visual inspection of histograms and Quantile–Quantile (QQ) plots. Simple, pairwise comparisons of categorical variables used Fisher’s exact test and chi-square test for binomial distributions depending on the assumptions being met, respectively. Parametric, or non-parametric with non-normally distributed data, group comparison analyses were used to examine differences between employment groups (FT vs. NFT). Effect sizes were calculated using Hedges’ g to provide estimates of the magnitude of any statistically significant differences. Indices indicating small (0.2), medium (0.5), large (0.8), and very large (1.3) effect sizes were applied [32].

To identify how employment status and other clinically relevant variables contributed to prosthetic mobility outcomes (cTUG, 2-MWT, PLUS-M), simple linear regression was applied to determine the unadjusted contribution of each variable, followed by stepwise backward multiple linear regression to control for each variable. A two-stage process was used in all regression analyses to reduce the number of independent variables in the final models, using care not to overfit the models due to our sample size limitations. Independent variables were considered for the multivariate analysis if they were a statistically significant contributor to the dependent variable at an α < 0.1, or were reasonable to include in the model based on clinical importance. In addition to employment status, available data for other independent variables of interest included age [6, 33], amputation level [6, 25], and self-efficacy [14, 34] as all have been shown to influence both prosthetic mobility and return to work. Similarly, multivariate logistic regression analyses were used to address the logical argument that someone was working full-time as a result of better prosthetic mobility. Age, amputation level, and prosthetic mobility performance were factors modeled for their association with employment status. For all regression analyses, independent variables were assessed for multicollinearity using Pearson product moment correlations, serving as point biserial correlation [35] for categorical variables. Correlation coefficients above r = 0.75 were considered evidence of multicollinearity [36]. All analyses were conducted in SAS version 9.4.

Results

Descriptive Statistics

Forty-nine people with dysvascular LLA completed all study outcome measures. The mean age of participants was 62.1 ± 9.7 years, with ages ranging from 38–78 years old. Sixty-one percent of participants were under the age of 65 years. Mean time since amputation was 6.9 ± 6.6 years. Per self-report, 80% (n=39) of participants were not working FT. Participant characteristics are exhibited in table 1. No significant differences in race/ethnicity or sex were found between employment status groups or income levels. Additionally, there were no differences in age, number of comorbidities, or level of amputation based on employment status. Furthermore, there were no differences in reported availability of insurance or rehabilitation services, or in perceived environmental barriers between the groups. Lastly, 69.4% (n=34) individuals reported some level of employment prior to LLA.

Table 1.

Participant characteristics.

Variable	n (%)
Sex
Male	30 (61)
Female	19 (39)
Amputation level
Transtibial	32 (65)
Transfemoral	8 (16)
Bilateral	9 (18)
Employment status
Full-time	10 (20)
Part-time	1 (2)
Unemployed	5 (10)
Retired	15 (31)
Disabled	18 (37)
Race
Non-Hispanic White	27 (55)
Person of color (Black, Hispanic, Native American, Asian)	22 (45)
Yearly income
Less than $20K	16 (33)
$20K–$50K	14 (29)
Greater than $50K	15 (31)
Refused to report	4 (8)

Group Differences Based on Presence or Absence of Full-Time Employment

Measures of prosthetic mobility were significantly lower in participants who were categorized as NFT employment. This group exhibited poorer scores on both performance-based outcomes of the cTUG (p = 0.02) and 2-MWT (p = 0.03), and the self-reported PLUS-M (p < 0.01). Additionally, the NFT employment cohort reported significantly lower PROMIS scores on ability to participate in activities (p < 0.01) and self-efficacy (p < 0.01) than the FT employment group. Effect sizes were moderate to very large. Table 2 presents group comparisons details.

Table 2.

Group comparisons by employment status.

Variable	FT employment (n=10) Mean (SD) [range]	NFT employment (n=39) Mean (SD) [range]	p value	Hedges’ Effect size g
Age (years)	61.7 (8.5) [47.9–74.0]	62.2 (10.1) [38.9–78.5]	0.87	-----
cTUG (sec)	14.6 (8.0) [9.2–36.7]	20.3 (9.5) [8.5–48.3]	0.02*	0.62
2-MWT distance (m)	92.6 (29.2) [31.2–131.3]	68.7 (31.4) [12.0–156.7]	0.03*	0.77
PLUS-M	58.0 (9.4) [40.9–71.4]	44.7 (9.7) [21.8–64.5]	< 0.01*	1.38
PROMIS Ability to participate	53.7 (6.8) [45.0–65.4]	43.5 (10.8) [25.9–65.4]	< 0.01*	1.00
PROMIS Self-efficacy	54.3 (6.0) [45.6–60.7]	42.6 (8.6) [31.2–60.7]	< 0.01*	1.43

^*Statistical significance at p = 0.05.

FT=full-time employment; NFT=no full-time employment; 2-MWT=2-minute walk test; cTUG=Component Timed-Up-and-Go; PLUS-M^™=Prosthetic Limb Users Survey of Mobility. Effect size: small = 0.2, medium = 0.5, large = 0.8, and very large =1.3.

Univariate and Multivariate Regression Analyses

When simple linear regression analyses were executed to examine contributions of the variables of interest to the three outcomes of prosthetic mobility, the variables of employment status and self-efficacy demonstrated associations at an acceptable level of α less than 0.1. Neither age nor amputation level exhibited this degree of association but were included in the final model due to their clinically relevant relationships to prosthetic mobility. Hence, age, amputation level, employment status, and self-efficacy were the final variables for the full multiple linear regression analysis. With differences in reported self-efficacy between employment groups (p < 0.01), it was found that self-efficacy drove the prosthetic mobility scores within the multiple regression model and masked the contribution of FT employment status. Therefore, to simplify the model and create a more reasonable story, separate multiple linear regression models were created to examine the unique contributions of employment status (table 3) and self-efficacy (table 4) to 2-MWT, cTUG, and PLUS-M scores. Variables satisfied the assumption of absence of multicollinearity.

Table 3.

Multiple linear regression modeling of prosthetic mobility measures with amputation level, age, and FT employment.

2-MWT Model: (F (1, 47) = 4.72, p = 0.03)
Independent Variable	Parameter estimate	SE	R2	F	p value
Amputation level	0.65	5.76	0.09	0.01	0.91
Age	0.01	0.03	0.09	0.14	0.71
FT employment status	23.89	11.0	0.09	4.72	0.03*
cTUG Model: (F (1, 47) = 3.11, p = 0.08)
Amputation level	0.63	1.70	0.08	0.14	0.71
Age	−0.01	0.01	0.08	0.70	0.40
FT employment status	−5.78	3.30	0.06	3.11	0.08*
PLUS-M Model: (F (1, 47) = 15.21, p < 0.01)
Amputation level	−2.12	1.71	0.31	1.55	0.21
Age	0.02	0.01	0.29	2.81	0.10*
FT employment status	13.3	3.41	0.24	15.21	< 0.01*

^*Statistical significance at p = 0.10.

FT=full-time; 2-MWT=2-minute walk test; cTUG=Component Timed-Up-and-Go; PLUS-M^™=Prosthetic Limb Users Survey of Mobility.

Table 4.

Multiple linear regression modeling of prosthetic mobility measures with amputation level, age, and self-efficacy.

2-MWT Model: (F (1, 47) = 22.43, p < 0.01)
Independent Variable	Parameter estimate	SE	R2	F	p value
Amputation level	3.27	4.94	0.33	0.44	0.51
Age	< 0.01	0.03	0.33	< 0.01	0.97
Self-efficacy	1.95	0.41	0.32	22.43	< 0.01*
cTUG Model: (F (1, 47) = 25.22, p < 0.01)
Amputation level	−0.16	1.44	0.35	0.01	0.91
Age	−0.01	0.01	0.35	0.39	0.54
Self-efficacy	−0.59	0.12	0.35	25.22	<0.01*
PLUS-M Model: (F (2, 46) = 67.93, p < 0.01)
Amputation level	−0.86	1.04	0.75	0.69	0.41
Age	0.01	< 0.01	0.75	3.59	0.06*
Self-efficacy	0.98	0.09	0.75	128.69	< 0.01*

^*Statistical significance at p = 0.10.

2-MWT=2-minute walk test; cTUG=Component Timed-Up-and-Go; PLUS-M^™=Prosthetic Limb Users Survey of Mobility.

In the final models, employment status uniquely explained 9.11%, 6.2%, and 24.2% of the variance in 2-MWT, cTUG, and PLUS-M scores, respectively, and reported self-efficacy explained 32.9%, 34.6%, and 70.6%, respectively. The evidence suggests that both FT employment status and self-efficacy explained the majority of the variance in their respective models. Data indicates that FT employment is associated with walking approximately 24 meters further during the 2-MWT, and six seconds faster during the cTUG.

Multivariate logistic regressions were run to estimate the likelihood that better prosthetic mobility (per cTUG and 2-MWT scores), age, amputation level, and working before LLA contributed to FT employment status. Due to strong correlations between cTUG and 2-MWT scores (r = −0.81, p < 0.01), two separate models were examined. Results indicated that age, amputation level, and cTUG were not associated with FT employment. The data suggests that better 2-MWT performance (OR = 1.03, 95% CI 0.99–1.05, p = 0.06), and working prior to LLA (OR = 5.73, 95% CI 0.55–59.6, p = 0.14) were associated with working FT in our study participants. Indicating that for every additional meter walked during the 2-MWT, odds of working FT increase by 3% and that working prior to LLA results in five times greater odds of working FT after LLA. Hosmer and Lemeshow goodness of fit test was consistent with satisfactory model fit to the data (p=0.09).

Discussion

The objective of this pilot study was to explore mechanisms by which employment may function as a social determinant of health as measured by prosthetic mobility outcomes in people with dysvascular LLA. To our knowledge, this is the first study in the LLA population to examine prosthetic mobility disparities relative to employment status. Furthermore, it is the only work-related study to incorporate performance-based measures to assess prosthetic mobility in a diverse cohort of people with LLA. Additionally, the administration of the cTUG and 2-MWT, both clinically useful and recommended measures of prosthetic mobility [37], enhance the study rigor and clinical utility of the results. Participants who were not employed FT exhibited significantly worse scores on all pertinent outcome measures, and associated effect sizes were moderate to large. These are promising findings in lieu of our limited sample size, and support the study hypothesis.

To begin to examine work as a social determinant of health, it is vital to recognize the complex interactions amongst the panoply of factors related to living with LLA. Use of the ICF model provided a framework to categorize the variables of interest, and affords a standardized means of disseminating the results across disciplines. Our study sample reflected a high number of people (80%) with dysvascular LLA are lacking FT employment and these individuals are reporting and exhibiting significantly poorer functional recovery. Physical deficits threaten quality of life in people with LLA [38]. In working-aged individuals this may promote a vicious cycle, as poorer perceived mobility [6, 39], and mental and behavioral disorders [40] have been associated with lack of employment [41].

Per final regression models, working full-time was significantly associated with better scores on all outcomes of prosthetic mobility in our study. This novel approach exposed the positive influence of FT employment in the recovery of people with LLA. However, return to work and reintegration into the community is difficult after LLA [6, 42] and often vocational interests are not a priority during prosthetic rehabilitation. Therefore, it is imperative to understand by what mechanisms FT employment contributes to better mobility. One possible mechanism may be the mediating effect of self-efficacy. Self-efficacy has been shown to influence both prosthetic mobility [43] and return to work [14] in people with LLA and offers a common, potentially modifiable factor [44] for intervention during rehabilitation. Our study results indicated that reported self-efficacy was the strongest predictor of prosthetic mobility and masked the contribution of FT employment. Clinically, this suggests the importance of considering frequently overlooked socioenvironmental variables, like self-efficacy and employment, to assess their possible influences on successful functional recovery.

Identifying employment status and self-efficacy as the major contributors to prosthetic mobility in this study bears significance in that they outweighed typically influential variables of age and level of amputation [19, 24]. This is despite our sample including older adults with varying levels of amputation. The lack of significance of amputation level in our regression models may indicate that certain often-ignored sociodemographic variables provide more information about functional recovery than has been traditionally considered. Yet, the final regression models for the cTUG and 2-MWT do not explain the majority of the variance in prosthetic mobility (FT employment models R²=0.06, 0.09; self-efficacy models R²=0.32, 0.35, respectively). Employment status and self-efficacy contributed more heavily to the PLUS-M models, which was the only self-reported outcome measure of mobility. This could be due to employment status and self-efficacy also being self-reported, increasing the associations between the variables. The remaining unexplained variance in the mobility regression models could potentially be explained by other factors like sex and/or race/ethnicity which were not included in the models for sake of parsimony. Further quantitative and qualitative prospective research on a larger diverse sample population, with focused recruitment of FT employed individuals with LLA, is necessary. This will offer the opportunity to increase the understanding of how mobility is affected by employment, and other non-traditional factors. Ultimately, our study results suggest that socioenvironmental factors are associated with functional recovery after LLA. Considering a patient’s work status can help guide expectations and clinical decision making, and may indicate the need for referral to employment-related resources or vocational rehabilitation.

Strengths and Limitations

A major strength of this study is the novel way of examining prosthetic mobility outcomes with respect to employment. Typically, return to work is the outcome of interest, but in our study, work was examined as a determinant of health for its contribution toward prosthetic mobility. Furthermore, unique to our study, prosthetic mobility was assessed using performance-based outcome measures in lieu of only self-reported mobility.

One primary study limitation is our sample size. Additionally, detailed queries regarding work history and conditions were not administered, and specific definitions of “full-time,” “part-time,” etc. were not provided. Furthermore, we made no distinction between paid or unpaid employment or details of retirement which may have affected employment categorization. Therefore, conclusions drawn from this research, though generalizable to the dysvascular LLA population due to our diverse sampling, must be viewed as bases on which future research will be founded. Because only 10 of our participants were employed FT, we are underpowered to confidently conclude whether better mobility contributes to FT employment or vice versa. However, our study provides novel evidence that lack of FT employment is associated with poorer prosthetic outcomes, but definitive causes and effects remain hypothetical. Furthermore, the cross-sectional design limits any conclusions about causation or temporal relationships, making it unclear as to which factors are the driving force. Additional prospective, qualitative research is needed to address a patient-centered approach to examining work as a social determinant of health in people with LLA.

Conclusion

The study results indicate that a potentially high number of people may not attain FT employment after dysvascular LLA, and it is evident that a lack of FT employment is associated with poorer functional outcomes. Those who are not working full-time exhibit significantly worse scores on measures of prosthetic mobility, activity participation, and self-efficacy. However, it appears that self-efficacy may be the mechanism by which both prosthetic mobility and employment status can be addressed, as it is a modifiable variable influencing both constructs. Our results suggest that practitioners should include assessment of personal and socioenvironmental factors related to employment when considering potential barriers and facilitators to optimal prosthetic rehabilitation.

Implications for rehabilitation.

• Lower limb amputation can pose barriers to employment and activity participation, potentially affecting quality of life.

• This study found that the majority of people living with lower limb amputation due to dysvascular causes were not employed full-time and were exhibiting poorer prosthetic outcomes.

• Healthcare practitioners should consider the modifiable variable of employment when evaluating factors that may affect prosthetic mobility.

• The modifiable variable of self-efficacy should be assessed by healthcare professionals when evaluating factors that may affect prosthetic mobility.

Data availability:

The datasets generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request.