Barriers and Facilitators to Lifestyle Intervention Engagement and Weight Loss in People Living With Spinal Cord Injury

Abstract

Background: Individuals living with spinal cord injury (SCI) have a high prevalence of obesity and unique barriers to healthy lifestyle. Objective: To examine barriers and facilitators to engagement and weight loss among SCI participants enrolled in the Group Lifestyle Balance Adapted for individuals with Impaired Mobility (GLB-AIM), a 12-month intensive lifestyle intervention. Methods: SCI participants (N = 31) enrolled in a wait-list, randomized controlled trial where all participants received intervention between August 2015 and February 2017. Analyses of pooled data occurred in 2020 to examine cross-sectional and prospective associations of hypothesized barriers and facilitators with (1) intervention engagement, comprised of attendance and self-monitoring, and (2) percent weight change from baseline to 12 months. We performed multivariable linear regression on variables associated with outcomes at p < .05 in bivariate analyses and controlled for intervention group. Results: Participants were middle-aged (mean age, 48.26 ± 11.01 years), equally male (50%) and female, White (80.7%), and unemployed (65.6%). In participants who completed baseline surveys (n = 30), dietary self-efficacy explained 26% of variance in engagement (p < .01); among the 12-month study completers (n = 22, 71.0%), relationship issues explained 23% of variance in engagement (p < .01). Money problems, health issues unrelated to SCI, lack of motivation, and experimental group explained 57% of variance in weight loss (p for model < .01), with lack of motivation uniquely explaining 24% of variance (p < .01). Conclusion: Improving engagement and weight loss for persons with SCI in the GLBAIM program may be achieved by addressing lack of motivation, relationship issues, and nutrition self-efficacy.

People with spinal cord injury (SCI) have higher prevalence of obesity, estimated at 40% to 70%,^1–5 and obesity-related chronic conditions such as diabetes, dyslipidemia, and cardiovascular disease^6–12 than the general population. Additionally, at the same weight, people with SCI have higher percent body fat,^13,14 which has led to the recommendation to use a lower body mass index (BMI) value to indicate obesity in people with SCI.^15–17 Greater fat mass accumulates in the lower extremities and abdomen due to muscle paralysis and resultant immobility.¹⁰ Due to the negative health impacts of this constellation of factors, clinical practice guidelines published by the Consortium for Spinal Cord Medicine¹⁷ advise that providers identify and manage cardiometabolic disease (CMD), which occurs when someone experiences three or more of the following risk factors: abdominal adiposity, dyslipidemia, hypertension, insulin resistance or glucose intolerance, proinflammatory state, and prothrombotic state.¹⁸ CMD increases risk for diabetes, cardiovascular disease, and mortality.¹⁹ People with SCI experience greater risk for each of these risk factors than people without SCI.¹⁷ Notably, the consortium recommends lifestyle intervention targeting nutrition and physical activity (PA) as the first line of prevention and treatment, based on available evidence and expert consensus.

Despite the importance of weight management to health and function for people with SCI,^17,20 providers have noted a lack of evidence-based interventions as a barrier to promoting weight loss in their patients with SCI.²¹ Although multiple studies have targeted PA,^22–24 to our knowledge, only three published studies have investigated weight loss approaches combining PA and nutrition specifically in people with SCI.^25–27 One was a case series with three individuals²⁷ and the other two reported on outcomes of 13²⁵ and 16²⁶ participants in single-group, pre-post studies. All interventions included weekly in-person educational sessions addressing nutrition, physical activity, and behavioral change strategies over 12^25,26 or 24²⁷ weeks. The studies reported retaining 82% of participants and an average (weighted by sample size) weight loss of 5.01 kg or 5.0% of starting weight. These pilot results offer encouraging evidence that lifestyle interventions can promote behavior change and weight loss in individuals with SCI.

Maximizing uptake and effectiveness of lifestyle interventions for people with SCI requires understanding what helps or hinders their efforts, yet few studies have assessed participant-reported barriers and facilitators to program engagement and subsequent weight loss. Even though Bailey and colleagues²⁸ conducted a qualitative study of barriers and facilitators to an anti-inflammatory diet in people with SCI, the sample consisted of only six participants. Other studies have examined frequently cited barriers to exercise among people with SCI, which include internal factors (e.g., lack of time or motivation, low self-efficacy for PA), resource-related factors (e.g., not having funds to join an exercise program), and structural factors (e.g., lack of accessible gyms or equipment in the community).^29,30 However, a recent survey of 180 people with SCI found that the barriers most highly correlated with exercise levels were internal, rather than environmental, and less frequently cited barriers were more strongly associated with exercise status. ³¹

We conducted a 12-month randomized controlled trial (RCT) of the intensive lifestyle intervention (ILI), Group Lifestyle Balance Adapted for Impaired Mobility (GLB-AIM), an adaptation of the evidence-based Diabetes Prevention Program Group Lifestyle Balance,^32,33 with a sample that included 31 participants with SCI. To identify opportunities to enhance intervention effectiveness for this population uniquely at risk for adverse effects of obesity, this secondary analysis explores potential barriers and facilitators of engagement and weight loss in the subset of SCI participants who enrolled in the GLB-AIM. Specifically, we sought to identify barriers and facilitators associated with program engagement, defined as session attendance and self-monitoring, and percent weight loss achieved by 12 months. We focused on factors that can be addressed in the context of an accessible intervention that targets individual behavior change.

Methods

This secondary analysis pools data from the subset of SCI participants who enrolled in a 12-month RCT of the GLB-AIM. Participants were randomized to receive the intervention immediately (Group 1) or to a wait-list control that received it after a 6-month delay (Group 2). Primary RCT outcomes have been published elsewhere.³⁴ This analysis combines groups and uses data collected from August 2015 to February 2017 in lab-based visits at study baseline (survey), each group’s baseline to initiate the intervention (weight), and after the 12-month intervention period for each group (survey and weight).

GLB-AIM Intervention

This group-based ILI targets 5% to 7% weight loss through limiting daily caloric intake and increasing time spent in weekly physical activity to achieve an energy deficit. The GLB-AIM encourages participants to adopt a moderate calorie and moderate fat diet and to progressively increase exercise to meet American College of Sports Medicine and American Heart Association recommendations of 150 minutes per week.³⁵ Based on Social Cognitive Theory principles,³⁶ the program targets behavior change primarily through daily goal-setting and self-monitoring of caloric intake and physical activity. Trained lifestyle coaches led 13 weekly core sessions, followed by three support sessions, occurring every other week, and seven monthly support sessions. Sessions teach skills to support these core behaviors, including adaptive cooking and exercise, problem solving, assertiveness, managing environmental cues, enlisting social support, and managing negative thinking and stress. To address transportation barriers, sessions were delivered as a combination of group-based conference calls (twelve 60-minute sessions) and monthly in-person meetings (eleven 90-minute sessions), where lifestyle coaches weighed participants using a wheelchair accessible scale to provide feedback on their weight loss efforts. Participants were asked to self-monitor fat and caloric intake over the first 13 weeks and share their logs with interventionists for feedback. A detailed description of all program sessions is available elsewhere.³⁷

Participants

Recruitment occurred via physical medicine and rehabilitation outpatient clinics of two large medical centers in the Dallas-Fort Worth metropolitan area. Recruitment also included disseminating fliers to organizations serving individuals with mobility impairment, including independent living centers, durable medical equipment centers, and community organizations.

Participants were required to be ≥18 years old, have had impaired mobility for ≥1 year, be overweight based on BMI >22 for those with SCI (recommendation given injury-related changes in fat/muscle mass),¹⁵ and have sufficient upper arm mobility to exercise, access to a telephone, and physician clearance. Exclusion criteria included cognitive impairment that limited autonomy, medical issues where exercise is contraindicated, >75 years old, pregnant, or not fluent in English. The larger RCT included 66 participants from among 92 eligible individuals with mobility impairment from various etiologies (e.g., multiple sclerosis, arthritis). The local institutional review board approved the study, and staff obtained informed consent in person at the baseline assessment. Participants received gift cards for attending testing sessions (maximum $100–$150 depending on group assignment) and tools for measuring foods and exercise aids (arm ergometer, TheraBand, DVD).

Independent variables

We assessed a variety of potential barriers and facilitators consistent with Social Cognitive Theory’s central tenets of reciprocal determinism (i.e., that personal factors, behavior, and environment are mutually influential) and primacy of self-efficacy (confidence in one’s ability to perform a behavior).³⁶

Barriers assessed at study baseline included potentially relevant domains from International Classification of Functioning³⁸ components of body function, activities and participation, and personal factors. Measures of body function included self-rated physical function limitations, secondary conditions, and bodily pain. Activities and participation included limitations in daily activities and preparing/eating meals. Table 1 presents construct operationalizations, measures, scoring, and reliability and validity information. In brief, physical function limitation was measured by the SF-36 Enabled Version Physical Functioning subscale,³⁹ pain was measured by the SF-36 Bodily Pain subscale, conditions secondary to SCI were self-reported and summed, limitation preparing and eating meals was measured by the LIFE-H^40,41 Diet and Meal Preparation subscale, and daily activity limitation was measured by LIFE-H Daily Activity subscale.

Table 1.

Barriers and facilitators measured at study baseline

Domain/level	Construct	Operationalization	Measure	Items and scoring	Reliability and validity
Barriers
Body function	Physical function limitations	Degree of limitation while performing physical tasks with one’s usual assistive device (e.g., wheelchair)	Medical Outcomes Study Short Form (SF-36) Enabled Version, Physical Functioning subscale	10 items averaged to obtain a score ranging from 0 to 100, with higher scores corresponding with better function	The SF-36 was found to have high internal consistency and construct validity among 215 people with SCI, with reliability across the 8 health scales ranging from 0.76 to 0.90; no significant correlation was found between the Mental and Physical Component Summaries.63The enabled version used here was found to achieve similar results as the original SF-36 (ICCs were above 0.60 for all scales and summary scales across the 2 versions), while avoiding terms like “walking” and “climbing.”39
	Bodily pain	Frequency of bodily pain and perceived level of interference with normal work	SF-36 Enabled Version, Bodily Pain subscale	2 items averaged to obtain a score ranging from 0 to 100, with higher scores corresponding with better function/less pain
	Secondary conditions	Occurrence of health conditions related to SCI in the year prior to the study: pressure injuries, urinary tract infections, bowel incontinence, kidney stones, bone fracture, autonomic dysreflexia, and edema	Sum of health conditions related to SCI that an individual reported experiencing in the year before starting the study	Sample ranged from 0 to 12 instances occurring over the previous year	N/A
Activities and participation	Limitations in daily activity	Ratings of difficulty and need for assistance with performing activities of daily living such as dressing, preparing and eating meals, maintaining physical fitness, shopping, working, etc.	Assessment of Life Habits (LIFE-H) General Short Form, Daily Activities subscale	Sum of 10 items comprising the LIFE-H Daily Activities life habits scale. Score ranges from 0 to 90. Higher scores correspond with higher levels of difficulty and need for assistance.	The LIFE-H was developed as a measure of participation among people with physical disabilities.40 The LIFE-H has shown acceptable internal consistency in adults and children with ICCs > 0.5 for each life habit domain, and good test–retest reliability among adults with spinal cord injury (ICC = 0.74).40
	Limitations preparing/eating meals	Ratings of difficulty and need for assistance with tasks related to preparing and eating meals	LIFE-H General Long Form Version 3.0, Diet and Meal Preparation subscale	Sum of 7 items comprising the LIFE-H Diet and Meal Preparation life habits scale. Score ranges from 0 to 63. Higher scores correspond with higher levels of difficulty and need for assistance.
Facilitators
Personal factors	Self-efficacy for health behaviors	Confidence in one’s ability to perform healthy behaviors such as maintain an exercise routine, manage stress levels, find a doctor, etc.	Self-Rated Abilities for Health Practices Scale (SRAHP)	We used 27 of 28 items and summed scores for a total score range of 0 to 108, omitting an item on toothbrushing that did not load onto the nutrition subscale in factor analysis.4 Higher scores indicate higher levels of self-efficacy.	The SRAHP has demonstrated high internal consistency across scales (Cronbach’s alpha ranged from .77 to .91) for a sample with disabilities and good content, discriminant, and convergent validity.42
	Nutrition self-efficacy	Confidence in one’s ability to maintain a healthy diet, including selecting, purchasing, and eating healthy foods	Self-Rated Abilities for Health Practices Scale (SRAHP), Nutrition subscale	Sum of 6 out of the 7 items from the Nutrition Subscale (omitting the toothbrushing item) with a score range of 0 to 24. Higher scores indicate higher levels of self-efficacy.
Activities and participation	Social participation	Ability to use transportation, participate in groups and activities, carry out responsibilities and activities related to maintaining social relationships and volunteering, etc.	LIFE-H Social Roles subscale	Sum of 5 items comprising the LIFE-H Social Roles subscale. Score ranges from 0 to 40. Higher scores correspond with higher levels of difficulty and need for assistance.	See LIFE-H description above.

Note: ICC = intraclass correlation coefficient.

Table 1.

Barriers and facilitators measured at study baseline

Domain/level	Construct	Operationalization	Measure	Items and scoring	Reliability and validity
Barriers
Body function	Physical function limitations	Degree of limitation while performing physical tasks with one’s usual assistive device (e.g., wheelchair)	Medical Outcomes Study Short Form (SF-36) Enabled Version, Physical Functioning subscale	10 items averaged to obtain a score ranging from 0 to 100, with higher scores corresponding with better function	The SF-36 was found to have high internal consistency and construct validity among 215 people with SCI, with reliability across the 8 health scales ranging from 0.76 to 0.90; no significant correlation was found between the Mental and Physical Component Summaries.63The enabled version used here was found to achieve similar results as the original SF-36 (ICCs were above 0.60 for all scales and summary scales across the 2 versions), while avoiding terms like “walking” and “climbing.”39
	Bodily pain	Frequency of bodily pain and perceived level of interference with normal work	SF-36 Enabled Version, Bodily Pain subscale	2 items averaged to obtain a score ranging from 0 to 100, with higher scores corresponding with better function/less pain
	Secondary conditions	Occurrence of health conditions related to SCI in the year prior to the study: pressure injuries, urinary tract infections, bowel incontinence, kidney stones, bone fracture, autonomic dysreflexia, and edema	Sum of health conditions related to SCI that an individual reported experiencing in the year before starting the study	Sample ranged from 0 to 12 instances occurring over the previous year	N/A
Activities and participation	Limitations in daily activity	Ratings of difficulty and need for assistance with performing activities of daily living such as dressing, preparing and eating meals, maintaining physical fitness, shopping, working, etc.	Assessment of Life Habits (LIFE-H) General Short Form, Daily Activities subscale	Sum of 10 items comprising the LIFE-H Daily Activities life habits scale. Score ranges from 0 to 90. Higher scores correspond with higher levels of difficulty and need for assistance.	The LIFE-H was developed as a measure of participation among people with physical disabilities.40 The LIFE-H has shown acceptable internal consistency in adults and children with ICCs > 0.5 for each life habit domain, and good test–retest reliability among adults with spinal cord injury (ICC = 0.74).40
	Limitations preparing/eating meals	Ratings of difficulty and need for assistance with tasks related to preparing and eating meals	LIFE-H General Long Form Version 3.0, Diet and Meal Preparation subscale	Sum of 7 items comprising the LIFE-H Diet and Meal Preparation life habits scale. Score ranges from 0 to 63. Higher scores correspond with higher levels of difficulty and need for assistance.
Facilitators
Personal factors	Self-efficacy for health behaviors	Confidence in one’s ability to perform healthy behaviors such as maintain an exercise routine, manage stress levels, find a doctor, etc.	Self-Rated Abilities for Health Practices Scale (SRAHP)	We used 27 of 28 items and summed scores for a total score range of 0 to 108, omitting an item on toothbrushing that did not load onto the nutrition subscale in factor analysis.4 Higher scores indicate higher levels of self-efficacy.	The SRAHP has demonstrated high internal consistency across scales (Cronbach’s alpha ranged from .77 to .91) for a sample with disabilities and good content, discriminant, and convergent validity.42
	Nutrition self-efficacy	Confidence in one’s ability to maintain a healthy diet, including selecting, purchasing, and eating healthy foods	Self-Rated Abilities for Health Practices Scale (SRAHP), Nutrition subscale	Sum of 6 out of the 7 items from the Nutrition Subscale (omitting the toothbrushing item) with a score range of 0 to 24. Higher scores indicate higher levels of self-efficacy.
Activities and participation	Social participation	Ability to use transportation, participate in groups and activities, carry out responsibilities and activities related to maintaining social relationships and volunteering, etc.	LIFE-H Social Roles subscale	Sum of 5 items comprising the LIFE-H Social Roles subscale. Score ranges from 0 to 40. Higher scores correspond with higher levels of difficulty and need for assistance.	See LIFE-H description above.

Note: ICC = intraclass correlation coefficient.

Throughout the 12-month intensive intervention, participants discussed barriers and facilitators with lifestyle coaches during formal sessions as part of the curriculum and in other study contacts (e.g., scheduling phone calls). To assess frequency of these factors across the study sample, the study team developed a battery of 18 specific barriers (Table 3; organized by ICF components). These barriers were assessed at 12 months in a postintervention survey that asked participants to assess each potential issue as follows: “Please rate how much the barriers listed below kept you from participating in GLBAIM lifestyle change activities over the past year.” Participants rated each barrier on a scale of 1 (not at all) to 5 (totally). Because some participants responded “not applicable” (N/A), we dichotomized responses such that N/A, 1 (not at all) or 2 (a little) were coded as no barrier (0), and responses of ≥3 (some, a lot, totally) were coded as a barrier (1).

Table 3.

Participant responses to barrier and facilitator items at baseline and 12-month exit surveys and engagement and percent weight change outcomes

Baseline measures (score range)	n	M	SD
Barriers
Body function
Secondary conditions	31	0.61	0.38
Physical function limitations (0–100)	31	32.36	10.70
Bodily pain (0–100)	30	44.39	11.83
Activities and participation
Limitations in daily activity (0–90)	30	47.70	16.64
Limitations preparing/eating meals (0–72)	28	45.04	14.37
Facilitators
Activities and participation
Social participation (0–45)	30	28.60	10.38
Personal factors
Self-efficacy for health behaviors (0–112)	30	74.47	20.18
Nutrition self-efficacy (0–24)	30	14.73	6.53
12-month measures (n = 22)	n	%
Barriers
Body function
SCI-related health issues (e.g., UTI)	11	50.00
Pain	7	31.82
Mental health issues (e.g., depression)	5	22.73
Lack of hunger sensation	3	13.64
Other health issues (e.g., cold, flu)	2	9.09
Activities and participation
Difficulty preparing healthy foods	11	50.00
Relationship issues	8	36.36
Employment issues	4	18.18
Personal factors
Lack of motivation	9	40.91
Not reaching my goals	9	40.91
Relationship issues	8	36.36
Adjusting to disability	6	27.27
Negative thoughts	5	22.73
Difficulty getting enough calories	3	13.64
Environmental factors
Transportation	6	27.27
Access to healthy food	6	27.27
Money problems	4	18.18
Other issues	4	18.18
Technology issues	1	4.55
Facilitators: Sources of social support
Spouse/significant other	12	54.55
Family	11	50.00
Friends	11	50.00
Healthcare provider	10	45.45
Caregiver	7	31.82
Religious/spiritual community	7	31.82
People at your gym	7	31.82
	n	M	SD
Total barriers reported (0–18)	22	4.73	3.35
Social support sum scale (0–7)		2.95	2.15
Outcome measures (n = 22)	n	M	SD
Engagement, full sample (0–36)	31	20.23	10.74
Engagement, completers (0–36)	22	24.60	9.09
Weight change, completers, kg	20	−5.03	8.58
Weight change, completers, %	20	−4.14	6.53

Note: UTI = urinary tract infection.

Facilitators assessed at study baseline included self-efficacy for healthy behaviors, measured by the Self-Rated Abilities for Health Practices (SRAHP) scale,⁴² nutrition self-efficacy measured by the SRAHP nutrition subscale,⁴² and social participation measured by the LIFE-H^43–45 Social Roles subscale.

Facilitators identified during the course of intervention and assessed in the 12-month postintervention survey included social support for GLB-AIM lifestyle change activities from seven sources (full list in Table 3) that include spouse/significant other, family, friends, etc. Participants rated each source of social support on a helpfulness scale from 1 (very unhelpful) to 5 (very helpful). We dichotomized responses such that N/A or scores ≤3 (very unhelpful, mostly unhelpful, no opinion) were coded as no facilitator (0), and responses >3 (mostly helpful, very helpful) were coded as facilitator (1).

Dependent variables

These outcomes reflect participant engagement and weight change over the 12-month program. Program engagement reflects participants’ program involvement over the year, based on their (a) attendance to the 23 sessions and (b) dietary self-monitoring over the 13 core sessions. Participants received one point for each session attended, and one point for each food log returned >50% complete, for a total score range of 0 to 36. Percent weight change was calculated by subtracting weight at intervention initiation from weight after 12 months of intervention, divided by weight at initiation, and multiplied by 100. Participants were weighed on a Seca platform scale (model #672), with their wheelchair weight and/or other assistive devices (e.g., ankle-knee orthosis) subtracted from their total weight. Weight was measured twice and the average recorded unless they differed by more than 1 kg, then weight was measured again and all three averaged.

Data analysis

Descriptive statistics examined socio-demographic and injury characteristics, including age, sex, level of injury, years living with disability, age at injury, race/ethnicity, education level, marital status, employment status, and self-reported health issues, and Fisher’s exact tests examined differences between study completers and those who did not complete the study.

Least squares linear regression examined the barriers and facilitators significantly associated with intervention engagement and weight change. For engagement, two models examined associations in (1) the full sample, including those lost to follow-up, using independent variables measured at baseline, and (2) those who completed postintervention surveys (completers), using independent variables measured at baseline or after 12 months of intervention. We excluded observations where the independent or dependent variables were missing. Barriers and facilitators significantly associated with each outcome at p < .05 in bivariate analyses were included in the multivariable model. Where variables were highly correlated (p ≥ .5), we included only the variable most strongly associated with each outcome. Models also controlled for intervention group, given a protocol change that more strongly facilitated electronic self-monitoring in Group 2. Statistical significance was defined as p < .05. All analyses were conducted in SAS version 9.4 (SAS Institute, Cary, NC).

Results

Study sample

Participants with SCI (n = 31) were middle aged (48.26 ± 11.01 years old) (Table 2), nearly equally split between men and women (48.39% vs 51.61%), and predominantly non-Hispanic White (80.65%); three-quarters (77.42%) attended at least some college, and two-thirds were unemployed (64.52%). There were no significant differences between study completers and noncompleters.

Table 2.

Participant characteristics at baseline

	All participants (N = 31)		Completers (n = 22)		Non-completers (n = 9)
Variables	M	SD	M	SD	M	SD	p
Age, years	48.26	11.01	48.72	11.19	47.13	11.14	0.593
Age at injury, years	32.97	13.51	34.13	13.73	30.13	13.29	0.480
Time with injury, years	15.29	12.41	14.59	12.55	17	12.63	0.708
BMI	33.2	6.46	33.83	6.82	31.68	5.54	0.273
Body weight, pounds	224.55	54.44	230.34	59.21	210.38	39.88	0.480
Waist circumference, inches	45.23	6.77	46.46	7.34	42.24	4	0.111
	n	%	n	%	n	%	p
Group assignment							0.456
1 (intervention)	14	0.00	11	50.00	3	0.00
2 (wait-list control)	17	0.00	11	50.00	6	0.00
Injury type							0.693
Paraplegia	17	54.84	13	59.09	4	44.44
Tetraplegia	14	45.16	9	40.91	5	55.56
Sex							0.704
Male	15	48.39	10	45.45	5	55.56
Female	16	51.61	12	54.55	4	44.44
Hispanic ethnicity	2	6.45	2	9.09	0	0.00	0.089
Race							0.344
White	25	80.65	19	86.36	6	66.67
Black/AA	4	12.90	2	9.09	2	22.22
Other	2	6.45	1	4.55	1	11.11
Marital status							0.693
Single/separated/divorced	14	45.16	9	40.91	5	55.56
Married/living with partner	17	54.84	13	59.09	4	44.44
Education							0.161
High school or below	7	22.58	4	18.18	3	33.33
Some college/associates	9	29.03	5	22.73	4	44.44
Bachelors	8	25.81	8	36.36	0	0.00
Some post-grad/master's	7	22.58	5	22.73	2	22.22
Employment status							1.000
Employed	11	35.48	8	36.36	3	33.33
Unemployed	14	45.16	10	45.45	4	44.44
Student/homemaker/retired	6	19.35	4	18.18	2	22.22
Diabetes diagnosis	3	9.68	3	13.64	0	0.00	0.538

Note: AA = African American; BMI = body mass index.

The 12-month retention rate was 71% (n = 22/31). Reasons for participants not completing the study are provided in the CONSORT diagram (Figure 1). Of the 22 who completed postintervention surveys, pressure injuries prevented two from returning for weight measurement. There were no significant differences between study completers and noncompleters in sociodemographic or injury characteristics.

Figure 1.

CONSORT flow diagram.

Barriers and facilitators of intervention engagement

Table 3 presents mean scores for scales and percent endorsed for dichotomous variables. Participants’ mean engagement (n = 31) was 20.23 ± 10.74 on a 36-point scale. In the final adjusted model (Table 4) for engagement in the full sample (n = 30), higher nutrition self-efficacy significantly predicted higher intervention engagement, accounting for 26% of the variance (p < .01, sr² = 0.26). In the final model for intervention completers (n = 22), those who reported relationship issues at 12 months had significantly lower engagement, with this factor explaining 23% of variance (p = .02, sr² = 0.23). No other barriers or facilitators were significant for engagement in either the full sample or completers.

Table 4.

Results of linear regression models predicting engagement and percent weight change

Model 1: Predictors of intervention engagement in full sample (n= 30)							Model 2: Correlates of intervention engagement in completers (n= 22)
Variables	B	SE(B)	95% CI		sr2	p	B	SE(B)	95% CI		sr2	p
Intercept	0.82	7.04	−13.63	15.27		.908	20.73	5.57	9.08	32.38		.076
Facilitator: SRAHP nutrition	0.86	0.27	0.30	1.41	0.04	.004**
Barrier: Relationship issues							−8.75	3.56	−16.20	−1.30	0.23	.024*
Group assignment	4.40	3.48	−2.75	11.55	0.26	.217	4.18	3.42	−2.99	11.35	0.06	.237
Model 3: Correlates of percent weight change in completers (n= 20)							Model statistics
Variables	B	SE(B)	95% CI		sr2	p		R2	Adj. R2	F	p for model
Intercept	−8.07	3.47	−15.48	−0.67		.035	Model 1	0.29	0.24	5.62	.009**
Barrier: Other health issues	4.63	3.56	−2.97	12.22	0.04	.214	Model 2	0.28	0.21	3.77	.042*
Barrier: Money problems	3.88	2.65	−1.77	9.53	0.05	.164	Model 3	0.66	0.57	7.19	.002**
Barrier: Lack of motivation	7.34	2.27	2.49	12.19	0.24	.006**
Group assignment	−0.16	2.02	−4.46	4.13	0.00	.936

Note: SRAHP = Self-Rated Abilities for Health Practices scale.

^*Significance at p< .05.

^**Significance at p< .01

Barriers and facilitators of weight loss

Completers with a final weight recorded (n = 20/22) achieved a mean weight loss of 5.03 ± 8.58 kg and their average percent 12-month weight change was −4.14% ± 6.53%. The final adjusted regression model for weight loss included barriers of other health issues (i.e., not SCI-related), money problems, and lack of motivation and significantly explained 56.6% of variance (F = 7.19, p < .01, adjusted R² = 0.57, R² = 0.66). Only lack of motivation was significantly associated with weight change (p < .01). Those reporting lack of motivation achieved less weight loss or gained weight, with this factor accounting for 24% of the variance (sr² = 0.24).

Discussion

Our study is one of the first to identify key barriers and facilitators of engagement and weight loss in an ILI among individuals with SCI. Notably, those who completed the GLB-AIM achieved an average weight loss of approximately 4% at 1 year, which was significant but below the recommended 5%. Three significant factors emerged as barriers or facilitators — nutrition self-efficacy, relationship issues, and lack of motivation. Notably, all three factors are commonly reported in studies of weight loss in the general population.^46–50 This is consistent with focus group findings in participants with SCI after a dietary intervention, who reported lack of motivation as a key barrier to diet adherence,²⁸ and findings of individuals with amputation,⁵¹ which highlighted weight loss barriers of lack of money and unsupportive people. This suggests that the GLB-AIM successfully addressed many disability-specific concerns and reflects the reality that weight loss is challenging for everyone. The results support modifying the intervention to promote greater weight loss and further investigating how SCI individuals attempting weight loss experience these barriers and facilitators while participating in ILI.

Participants with greater confidence for healthy eating at baseline had significantly higher intervention engagement, yet this factor was not significant among the completers at 12 months, suggesting that low self-efficacy may have contributed to dropout. The GLB-AIM program relies heavily upon participants engaging in daily self-monitoring of dietary intake. As skill-building may increase self-efficacy,^36,52,53 participants who report low initial self-efficacy may benefit from additional early support for dietary change to improve adherence. Support could include supplemented meals, meal plans, facilitated cooking or grocery shopping experiences, and/or a stepped approach to self-monitoring. Notably, self-monitoring has been reported as the most common barrier to weight loss by half^54,55 to three-quarters⁵⁶ of participants in ILI programs.

More than a third of participants who completed the 1-year study reported relationship issues as a barrier to program engagement. Relationship issues were a barrier to weight loss in bivariate analysis (ρ = .40, p = .08, data not shown), though it was not included in the weight change model because it correlated with lack of motivation (ρ = 0.52, p = .01), an even stronger barrier to weight loss. Relational issues are common in weight loss,^46,47,57,58 and focus groups with individuals with SCI and caregivers have identified relationships as a key factor in promoting long-term health.⁵⁹ The GLBAIM addresses relational challenges, including those with caregivers, and family members were invited to attend sessions. Further attention to this topic, such as reducing burnout in caregivers and facilitating access to counseling, may be appropriate for SCI participants.

Lack of motivation was the strongest barrier to weight loss, accounting for almost a fourth of the variance. This 12-month ILI is longer than others conducted in people with SCI.^25–27,60 Sustained weight loss is challenging; national studies show that only roughly 20% of those who lose 10% of their body weight maintain the loss for 12 months.⁶¹ Qualitative research suggests that internal motivation is key to maintaining an initial loss for 6 months or longer.⁴⁸ The GLB-AIM addresses motivation through theory-based methods such as personal reflection, enlisting social support, planning rewards, and public commitment. Monitoring participants’ motivation toward program goals throughout the 12-month intervention may provide additional insight into motivation change over time and facilitate timely, tailored implementation of strategies to enhance motivation, such as motivational interviewing or peer accountability. In our sample, lack of motivation was associated with adjusting to diagnosis (ρ = 0.53, p = .01) and negative thoughts (ρ = 0.43, p = .04) in bivariate analyses (data not shown), in addition to relationship issues. These barriers may drive motivational decline and require additional attention.

Several study limitations should be noted. The small sample of predominantly non-Hispanic White participants limited statistical power and generalizability. Results may not be generalizable to minorities or weight loss programs that are less intensive, shorter duration, or do not address common barriers facing persons with mobility impairment (e.g., transportation). Barriers and facilitators assessed at 12 months could only be examined in the subset who completed surveys (n = 22/31). Yet because these factors were assessed at the same time as 12-month weight loss, knowledge of the outcome may have influenced participant responses. We dichotomized responses to maximize use of available data in our small sample; however doing so may have decreased the ability to detect the influence of some barriers that exist on a continuum.⁶² Additionally, exit survey items were not validated measures, thus our exploratory findings should be viewed as preliminary. Given our focus on individually targeted intervention, potentially important ecological factors (e.g., sidewalk accessibility, distance to a fitness center) were not assessed.⁶² However, including emergent barriers and facilitators that occurred during intervention implementation helped ensure that the most salient factors in this context were included.

Conclusion

This study identified nutrition self-efficacy, relationship issues, and lack of motivation as critical barriers and facilitators for people with SCI who enrolled in a 1-year ILI. Future research should assess dimensions and drivers of these constructs in people with SCI and explore acceptability of evidence-based approaches to address them.