Predictors of employment outcomes among supported employment program participants with spinal cord injury

Abstract

OBJECTIVE:

To examine factors associated with employment outcomes of Individual Placement and Support (IPS) Supported Employment (SE) among participants with spinal cord injury (SCI).

METHOD:

A secondary analysis of data from a 24-month study of IPS participants with SCI was conducted. Factors from three broad domains of sociodemographic, clinical, and supports were examined for potential impact on employment outcomes. Regression analysis was used to model the probability of obtaining employment.

RESULTS:

Univariate models showed significant predictors by domain as follows: Sociodemographic – college education, employment at time of SCI, and post-SCI work experience; Clinical – tetraplegia, number of inpatient days, and traumatic brain injury history; Supports – transportation and motor function/mobility. In the multivariate model, only number of inpatient hospital days during IPS and post-SCI work experience remained significant.

CONCLUSION:

Longer hospital admissions were associated with a reduced likelihood for IPS participants to obtain employment. Full participation in IPS leading to job acquisition is most likely to occur when health is optimized and prolonged hospitalization is minimized. To understand how past experience working with a SCI influences participation in IPS services and employment outcomes, including job sustainability, further examination is needed.

Keywords

Disability individual placement and support (IPS)outcomes predictors spinal cord injury supported employment veterans vocational rehabilitation

1 Introduction

In 2016, spinal cord injury (SCI) affected approximately 282,000 persons (range 243,000–347,000) in the United States, with about 17,000 new injuries per year (National Spinal Cord Injury Statistical Center (NSCISC), 2016). Since the 1970s, the age at injury has been trending upwards, with average age increasing from 29 to 42 years of age at the time of injury. Although less common than some other disabling neurologic conditions such as cerebrovascular accident, SCI has a disproportionately large economic impact for several reasons. A major factor is that most persons with SCI are of working age at the time of injury. The physical sequelae of SCI are numerous and co-occurring with permanent, life-altering physiological changes, including pressure wounds, neurogenic bowel and bladder, spasticity, and neurogenic pain in addition to paralysis. Furthermore, both mental health diagnoses and traumatic brain injury are frequent co-morbidities in those with SCI (Budisin et al., 2016; Elovic & Kirschblum, 1999; Macciocchi, Seel, Warshowsky, Thompson, & Barlow, 2012). Despite these catastrophic consequences, many persons with SCI want to work and judge themselves capable of work (Tomassen, Post, & van Asbeck, 2000; Young & Murphy, 2002). Like most, persons with SCI have much to gain through employment, including increased economic self-sufficiency, higher life satisfaction, and improved health status and psychological well-being (Meade, Reed, Saunders, & Krause, 2015). In fact, occupation is one of the highest rated needs by people in the community living with SCI (Kennedy, Lude, & Taylor, 2006).

Despite the known benefits of work, many of which maximize social participation, employment rate following SCI has been historically low. Although there are many studies of factors related to employment following SCI, little is known about what factors predict outcomes of employment interventions for persons with SCI. A recent systematic review of factors associated with employment outcomes following SCI identified 20 modifiable and 12 non-modifiable factors (Trenaman, Miller, Querée, & Escorpizo, 2015). The following modifiable factors have been consistently and independently associated with employment outcome: education, vocational rehabilitation, functional independence, social support, and financial disincentives. In a retrospective evaluation of employment history of 238 veterans with SCI, Mobility and Occupation subscale scores on the Craig Handicap Assessment and Report Technique (CHART) were higher for those who were employed, however, depressive symptoms and self-assessed, health-related quality of life did not differ significantly between employed and unemployed veterans (Ottomanelli, Sippel, Cipher, & Goetz, 2011). Receipt of Social Security benefits was a disincentive for employment, but military service disability benefits (i.e., Veterans Affairs benefits) were not. Notably, in this group, neither education, gender, ethnicity, nor severity of injury predicted employment status. In other studies, race was not a predictor nor was severity of injury (Krause & Anson, 1996; Meade et al., 2015). Collectively these findings suggest personal factors may outweigh factors related to injury severity.

While factors associated with employment following SCI are well identified, the field lacks empirical data on what factors predict positive employment outcome from vocational rehabilitation interventions. This knowledge gap is, in large part, due to the lack of research on intervention effectiveness in SCI (Trenaman, Miller, & Escorpizo, 2014). Indeed, to date there has only been one randomized controlled trial of vocational interventions in SCI. The results of that randomized clinical trial (RCT), referred to as the Spinal Cord Injury Vocational Integration Program (SCI-VIP) study, demonstrated that the Individual Placement and Support (IPS) model of supported employment (SE) significantly improved employment rates of persons with SCI when compared to conventional vocational rehabilitation practices (Ottomanelli et al., 2012; Trenaman et al., 2014). Participants in the study who obtained employment had higher levels of social integration, community mobility, and time spent in productive roles (Ottomanelli, Barnett, & Goetz, 2013). A longitudinal, observational follow up study of IPS in SCI centers of the Veterans Health Administration (VHA) showed 24 months of IPS yielded a 43.2% employment rate (Ottomanelli et al., 2017). The study sample was heterogeneous and reflected typical veterans with SCI treated as either inpatients or outpatients. Veterans with a lifetime or current history of psychiatric diagnoses or traumatic brain injury (TBI) were also included. Among a sub-sample of persons who started IPS as outpatients and who had no history of TBI, the employment rate was greater than 50%. To date, there are no published studies that examine factors associated with positive employment outcomes among persons with SCI receiving IPS services.

IPS was originally developed for those with serious mental illness (SMI), and several studies identified factors associated with outcomes in this population. In an international, 6-center RCT of employment for people with SMI, those with a history of work, unmet social needs, and better relationships with their vocational workers were more likely to both obtain and sustain employment (Burns et al., 2007, 2009). These data suggested that satisfaction with current life circumstances may decrease motivation to work. In a regression analysis of four recent RCTs of IPS among persons with mental illness, work history was the only significant predictor for obtaining a job among those receiving IPS. In this study, Supplemental Security Income (SSI), with or without Social Security Disability Insurance (SSDI), was associated with fewer total weeks worked (Campbell, Bond, Drake, McHugo, & Xie, 2010). With other types of vocational services (not IPS), persons with less severe psychiatric disorders (e.g., mood disorders or fewer psychotic symptoms) were more likely to obtain competitive employment. Work history being the only significant predictor of employment outcomes with IPS challenges the notion of using mental health diagnosis, current psychiatric symptoms, or substance use disorders as exclusion criteria for IPS services. That is, perceived “high-risk” clients can benefit from IPS and should be eligible for services. As the sample excluded persons with SMI who had medical conditions “that would preclude working or participating in assessment interviews,” this study does not shed light on how physical symptoms impact IPS outcomes either alone or in combination with mental health disorders.

A recent secondary analysis of unemployed SSDI beneficiaries with mental illness also found work history to be the strongest predictor of employment (Metcalfe, Drake, & Bond, 2017). Interaction between treatment group and work history indicated IPS benefited those without a recent job history (previous 2 years) more than those with a recent work history. The study also showed that fewer physical problems, as measured by the Short Form Health Survey (SF-12), was a modest predictor of securing competitive employment. Other modest predictors of obtaining competitive employment were fewer years of receiving disability benefits and Hispanic ethnicity. This research suggests that IPS may compensate for poor work history and supports the principle of basing eligibility for services on desire to work rather than factors related to disability. In a brief report of a similar population, the top three provider-rated barriers to employment were general medical problems, uncontrolled psychiatric symptoms, and poor engagement in SE services (Milfort, Bond, McGurk, & Drake, 2015). Therefore, work history is the most consistent predictor of employment outcome in the mental health literature. Physical symptoms may also play a role in outcomes but the significance of their impact remains understudied and merits furtherattention (Metcalfe et al., 2017).

Predictors of response to employment interventions have not been studied among a medical population of persons with physical disabilities and other co-morbidities. The objective of the present study is to explore whether there are client factors associated with employment outcomes in persons with SCI who participated in a 24-month IPS program. We hypothesized that work history and factors related to medical issues and handicap (supports) would be associated with employment outcomes.

2 Methods

A secondary analysis was conducted on data from a longitudinal observational study of IPS for veterans with SCI that extended the work of a previous RCT with the same population (Ottomanelli et al., 2012, 2013). The Predictive Outcome Model Over Time for Employment (PrOMOTE) was a prospective, multi-site project evaluating employment history, quality of life, and economic outcomes in SCI. The study had local institutional review board approval. As details of both studies have been previously reported (Ottomanelli et al., 2009, 2012, 2017), this report provides only an overview of those methods. For the purposes of this analysis, examined factors included:

Sociodemographic: age, gender, educational history, and work experience;

Clinical: Spinal cord injury characteristics, hospital inpatient days, medical comorbidities, TBI history, and mental health;

Supports: transportation, social support, mobility, and financial.

2.1 Subjects and settings

The study population was veterans receiving inpatient or outpatient medical or rehabilitative care at SCI centers within Veterans Affairs Medical Centers (VAMCs) (n = 7) from interdisciplinary SCI care teams. That is, veterans with new spinal cord injuries receiving acute inpatient rehabilitation, veterans with chronic SCI receiving inpatient medical care, or veterans with chronic SCI who received outpatient healthcare were eligible to enroll in the study. Inclusion criteria for a baseline interview on employment, health, and quality of life were (1) 18 to 65 years of age and (2) medically and neurologically stable. To qualify for enrollment in the IPS supported employment program, veterans had to meet the following additional criteria: (1) not currently have a paying job in the community, (2) express a desire to be employed, and (3) live≤100 miles from the hospital. Veterans with untreated psychosis, untreated alcohol or drug dependence, or a terminal illness were ineligible for IPS. Of the 1047 veterans who enrolled in the study and completed baseline interviews, 279 met eligibility and enrolled in the IPS SE program. Of these, 213 had not participated in the previous RCT and are included in this outcomes analysis (n = 213) (Fig. 1).

Fig.1

PrOMOTE study IPS SE enrollment.

2.2 Intervention

The SE program consisted of IPS applied in SCI centers per standardized principles: Integration of IPS into medical care, rapid engagement in IPS, goal of competitive employment, zero exclusion from IPS, ongoing follow-along support, client–centered choice of job, community-based services, and personalized benefits counseling. The IPS program was delivered by a vocational rehabilitation specialist (VRS) who received training in both IPS and SCI-specific medical management and rehabilitation. The VRS was integrated as a provider into the SCI interdisciplinary care team in the SCI center. Throughout the study period, implementation of IPS was supported by ongoing trainings, written manuals, biannual fidelity monitoring (i.e., adherence to IPS principles), and a National IPS Clinical Coordinator, who provided ongoing technical assistance to the VRS staff across the 7 sites. The primary endpoint was competitive employment (CE), meaning a paying job earning at least minimum wage in the community. Volunteer work and sheltered employment did not qualify as employment for the purposes of the present study.

2.3 Measures

Age, gender, educational history (total education aend post-SCI education), and work experience were obtained by interview and extraction from the electronic medical record. Clinical data obtained were neurologic level of injury and severity (completeness), time since injury, and medical comorbidities were obtained by medical record review. Number of hospital inpatient days during the study period was extracted from administrative data. History of traumatic brain injury (TBI) was determined by use of the Ohio State University TBI Identification Method-Short Form (OSU TBI-ID-SF) (Spitzer, Kroenke, & Williams, 1999), a structured interview based on recommendations of the Centers for Disease Control. The OSU TBI-ID elicits self- or proxy-reports of TBI occurring over a person’s lifetime. The published inter-rater reliability intraclass correlation coefficients for the seven domains range from 0.84 to 0.95.

Mental health measures consisted of the following: The Patient Health Questionnaire-9 (PHQ-9) (Spitzer et al., 1999) and the Satisfaction with Life Scale (SWLS) (Diener, Emmons, Larsen, & Griffin, 1985). The PHQ-9, a brief measure of the severity of depressive symptoms, consists of 9 items that rate the frequency of depressive symptoms and signs experienced over the last 2 weeks. Items are based on the Diagnostic and Statistical Manual diagnostic criteria, and agreement with independently derived mental health professionals’ depression diagnoses is excellent (overall accuracy 85%). The SWLS is a self-report measure designed to measure global cognitive judgments of satisfaction with life. The SWLS has well-established psychometric properties (Pavot & Diener, 1993) and consists of 5 items that all load onto a single factor of well-being.

The effect of transportation as a measure of support was determined from a combination of 4 items assessing adequacy of transportation from the Craig Handicap Assessment and Reporting Technique (CHART) (Whiteneck, Charlifue, Gerhart, Overholser, & Richardson, 1992) and items from the motor score from the Functional Independence Measurement Scale (FIM) (Linacre, Heinemann, Wright, Granger, & Hamilton, 1994). The CHART assesses 6 dimensions of handicap identified by the World Health Organization: orientation, physical independence, mobility, occupation, social integration, and economic self-sufficiency. Published test-retest coefficients range from 0.80 to 0.95. The FIM is a measure of disability designed to operationally measure functional independence in self-care, mobility, and cognition/communication. Higher scores represent a greater level of independence. Published inter-rater reliability kappa coefficients were 0.53 (memory) to 0.66 (stair climbing) in a sample of medical rehabilitation inpatients (Linacre et al., 1994). Level of social support was determined from a combination of the CHART social integration domain score and marital status. Mobility was measured by the CHART mobility domain. Financial support was measured from a total monthly benefits parameter, in US dollars, created by summing all received monies from Social Security, Social Security Disability, and Department of Veterans Affairs disability payments.

2.4 Attrition and missing data

The 2-year rate of attrition was 21.6% (n = 46), most whom of were lost during year one (71.7%, n = 33). Reasons for attrition were subject withdrew (39.1%, n = 18), death (19.6%, n = 9), moved (13.0%, n = 6), discharged or ineligible for VHA Compensated Work Therapy and/or SCI services (10.9%, n = 5), incarcerated or facility safety and security restrictions (8.7%, n = 3), terminal illness (2.2%, n = 1), lost to follow-up (4.4%, n = 2), and unknown (4.4%, n = 2).

Minimal missing data were present due to forced data entry capture during participant interview. For parameters such as Marital Status or Race, discrete data collected as ‘N/A’ or ‘Missing’ were recoded as ‘No.’ Participant missing data for CHART Social Support (n = 4), CHART Mobility (n = 3), and FIM Mobility (n = 12) were recalculated to represent the average of each overall parameter.

2.5 Statistical analysis

Continuous parameters were reported as mean ± SD, and discrete parameters were reported as percent (%). Data were explored for departures from normality by standard descriptive statistics. In the event data were observed to not reflect a normal distribution, group comparisons were made with Student’s t-test or Wilcoxon rank sum tests with normal approximation, where appropriate, for continuous data and with Pearson chi-square test, or Fisher’s Exact Test, where appropriate, for categorical data.

To model the probability of obtaining CE, we first dichotomized CE as ‘yes’ or ‘no.’ We then used unconditional logistic regression to model the probability of obtaining CE through a univariate modeling approach to determine statistically significant predictors of CE. Statistically significant predictors at the p < 0.10 criterion level were then explored in a final multivariate model. The final model included all parameters into a final multivariate model. Included terms were: Sociodemographic (age [continuous], race [discrete], marital status [discrete], gender, education [discrete], education history [discrete], work experience prior to SCI [discrete], work experience following SCI [discrete]), Clinical (injury level [discrete], time since injury [continuous], inpatient days [continuous], number of comorbidities [continuous], TBI [discrete], PHQ-9 [continuous], SWLS [continuous]), and Supports (CHART Transportation [continuous], CHART Social Integration [continuous], FIM Motor [continuous], CHART Mobility [continuous], total monthly benefits received [continuous]). To avoid potential collinearity, FIM mobility scores were excluded from the final model given the high correlation with CHART Mobility (r = 0.44). For all models, we present odds ratios (OR) and 95% confidence intervals (CI) in addition to Nagelkerke’s pseudo R² (Nagelkerke, 1991). All analyses were performed with SAS (Ver. 9.4, Cary, NC).

3 Results

3.1 Subject characteristics

Participant characteristics are presented in Table 1. Participants (n = 213) were primarily male (96.7%), age 51(±10.1) years, and Caucasian (55.4%). Less than half of participants received any VA benefits (47.9%) or service-connected benefits for SCI (22.5%). On average, participants received less than $1,000 per month ($945.9±640.2) in non-service connected benefits, $837.0±463.3 in SSI and $1240.0±460 in SSDI. Nearly half of participants reported cause of injury as either vehicular accident (28.3%) or fall (23.9%). When neurologic level of injury and completeness were combined to reflect varying impairment levels, 27.2% (n = 58) of the sample had tetraplegia (C1-8) with American Spinal Injury Association Impairment Scale (AIS) grades A, B, or C; 27.2.0% (n = 58) had paraplegia with AIS grades A, B, or C; and 45.5% (n = 97) had AIS grade D (n = 93) or E (n = 2) regardless of neurologic level. More than half the participants had a history of TBI.

Table 1
Demographic characteristics at baseline (n = 213)

Parameter Competitive employment (N = 92) No competitive employment (N = 121) Total (N = 213) p-value

Age, yr 50.4±10.6 51.4±9.8 51.0±10.1 0.419

Education, yr 14.2±2.2 13.7±1.9 13.9±2.0 0.077

Male 88 (95.7) 118 (97.5) 206 (96.7) 0.449

College (vs. other) 22 (24.2) 16 (13.2) 38 (17.8) 0.044

Race

Caucasian 54 (58.7) 64 (52.9) 118 (55.4) —

African American 29 (31.5) 54 (44.6) 83 (39.0) —

Asian 3 (3.3) 1 (0.8) 4 (2.0) —

Native American 0 (0.0) 1 (0.8) 1 (0.5) —

Other/Unknown 6 (6.5) 1 (0.8) 7 (3.3) 0.038

Marital status

Married 28 (30.4) 42 (34.7) 70 (32.9) —

Divorced 37 (40.2) 35 (28.9) 72 (33.8) —

Other 27 (29.5) 44 (36.4) 71 (33.4) 0.219

VA benefits 44 (47.8) 58 (47.9) 102 (47.9) 0.988

SC benefits for SCI 24 (26.1) 24 (19.8) 48 (22.5) 0.281

If yes, SC 100% 21 (22.8) 23 (19.0) 44 (20.7) 0.495

Non-SC pension 8 (8.7) 12 (9.9) 20 (9.4) 0.762

Monthly non-SC $1,474±1,325 $739±563 $946±640 0.001

Monthly SSI amount $726±379 $899±501 $837±463 0.006

Monthly SSDI amount $1,303±468 $1,198±453 $1,242±460 0.729

Monthly total amount $1,167±489 $1,261±515 $1,205±500 0.244

Neither SSI/SSDI 30 (32.6) 27 (22.3) 57 (26.8) 0.093

Post SCI education 23 (25.3) 27 (22.3) 50 (23.5) 0.647

CE at time of SCI 61 (67.0) 41 (33.9) 102 (47.9) 0.001

Post-SCI work experience 26 (28.6) 9 (7.4) 35 (16.5) 0.001

Injury level

Tetraplegia 17 (18.7) 41 (33.9) 58 (27.2) —

Paraplegia 29 (31.9) 29 (24.0) 58 (27.2) —

AIS D/E 46 (50.0) 51 (42.2) 95 (44.6) —

Unknown/not reported 1 (1.1) 1 (0.8) 2 (0.9) 0.042

Time since injury, yr 12.1±11.6 10.0±10.6 10.9±11.1 0.001

Inpatient days following enrollment 19.4±37.2 59.6±126.7 42.2 100.4 0.001

Total number of medical comorbidities 5.1±2.2 5.3±2.3 5.5±2.2 0.501

OSU-TBI ID history of TBI 47 (48.4) 79 (65.3) 126 (59.2) 0.037

PHQ-9 3.3±4.6 3.0±3.5 3.11±4.02 0.569

SWLS 20.6±8.0 19.2±7.8 19.8 7.9 0.216

CHART transportation 3.3±1.1 2.9±1.3 3.1 1.2 0.038

CHART social 89.5±12.8 8a.1±15.5 87.5±14.4 0.090

FIM motor 60.2±21.7 52.8±23.3 56.0±22.9 0.023

CHART mobility 82.4±19.3 76.6±19.8 79.1±19.7 0.034

Parameter	Competitive employment (N = 92)	No competitive employment (N = 121)	Total (N = 213)	p-value
Age, yr	50.4±10.6	51.4±9.8	51.0±10.1	0.419
Education, yr	14.2±2.2	13.7±1.9	13.9±2.0	0.077
Male	88 (95.7)	118 (97.5)	206 (96.7)	0.449
College (vs. other)	22 (24.2)	16 (13.2)	38 (17.8)	0.044
Race
Caucasian	54 (58.7)	64 (52.9)	118 (55.4)	—
African American	29 (31.5)	54 (44.6)	83 (39.0)	—
Asian	3 (3.3)	1 (0.8)	4 (2.0)	—
Native American	0 (0.0)	1 (0.8)	1 (0.5)	—
Other/Unknown	6 (6.5)	1 (0.8)	7 (3.3)	0.038
Marital status
Married	28 (30.4)	42 (34.7)	70 (32.9)	—
Divorced	37 (40.2)	35 (28.9)	72 (33.8)	—
Other	27 (29.5)	44 (36.4)	71 (33.4)	0.219
VA benefits	44 (47.8)	58 (47.9)	102 (47.9)	0.988
SC benefits for SCI	24 (26.1)	24 (19.8)	48 (22.5)	0.281
If yes, SC 100%	21 (22.8)	23 (19.0)	44 (20.7)	0.495
Non-SC pension	8 (8.7)	12 (9.9)	20 (9.4)	0.762
Monthly non-SC	$1,474±1,325	$739±563	$946±640	0.001
Monthly SSI amount	$726±379	$899±501	$837±463	0.006
Monthly SSDI amount	$1,303±468	$1,198±453	$1,242±460	0.729
Monthly total amount	$1,167±489	$1,261±515	$1,205±500	0.244
Neither SSI/SSDI	30 (32.6)	27 (22.3)	57 (26.8)	0.093
Post SCI education	23 (25.3)	27 (22.3)	50 (23.5)	0.647
CE at time of SCI	61 (67.0)	41 (33.9)	102 (47.9)	0.001
Post-SCI work experience	26 (28.6)	9 (7.4)	35 (16.5)	0.001
Injury level
Tetraplegia	17 (18.7)	41 (33.9)	58 (27.2)	—
Paraplegia	29 (31.9)	29 (24.0)	58 (27.2)	—
AIS D/E	46 (50.0)	51 (42.2)	95 (44.6)	—
Unknown/not reported	1 (1.1)	1 (0.8)	2 (0.9)	0.042
Time since injury, yr	12.1±11.6	10.0±10.6	10.9±11.1	0.001
Inpatient days following enrollment	19.4±37.2	59.6±126.7	42.2 100.4	0.001
Total number of medical comorbidities	5.1±2.2	5.3±2.3	5.5±2.2	0.501
OSU-TBI ID history of TBI	47 (48.4)	79 (65.3)	126 (59.2)	0.037
PHQ-9	3.3±4.6	3.0±3.5	3.11±4.02	0.569
SWLS	20.6±8.0	19.2±7.8	19.8 7.9	0.216
CHART transportation	3.3±1.1	2.9±1.3	3.1 1.2	0.038
CHART social	89.5±12.8	8a.1±15.5	87.5±14.4	0.090
FIM motor	60.2±21.7	52.8±23.3	56.0±22.9	0.023
CHART mobility	82.4±19.3	76.6±19.8	79.1±19.7	0.034

NOTE: Values expressed are mean±SD or n (%). Reported p-values are a result of either Students’ t-test or χ². VA, Veterans Administration; SCI, spinal cord injury; SC, service connected; SSI, Social Security Income; SSDI, Social Security Disability Income; AIS, American Spinal Injury Association Impairment Scale; OSU-TBI ID, Ohio State University Traumatic Brain Injury Identification Method; TBI, traumatic brain injury; PHQ-9, Patient Health Questionnaire-9; SWL, Satisfaction with Life Score; CHART, Craig Handicap Assessment and Reporting Technique; FIM, Functional Independence Measure.

3.2 Employment rate, duration, and wages and job classifications

For 92 (43.2%) participants who obtained CE, length of employment averaged 38.2±29.7 weeks during the 2-year period. On average, subjects worked 15.1 hours per week (±13.1). While the majority (n = 77, 83.6%) worked part time, 16.3% (n = 15) achieved full time employment (>32 hours/ week). On average, subjects earned $264.8 (±$348.2) per week, or an annualized salary of $13,234 (range: $28.50–99037.00).

3.3 Predictors of CE

Potential predictors of CE were categorized within the 3 domains of Sociodemographics, Clinical, and Supports (Table 2). In univariate models, 9 parameters were significant, 3 in each domain: Significant predictors by domain were as follows: Sociodemographics – college education (OR: 2.06; 95% CI: 1.01–4.20), CE at time of SCI (OR: 3.89; 95% CI: 2.16–6.81), and post-SCI work experience (OR: 1.31; 95% CI: 1.18–1.46); clinical – tetraplegia (95% OR: 0.46; 95% CI: 0.23–0.92), number of inpatient days during study (OR: 0.99; 95% CI: 0.99–1.00), and TBI history (OR: 0.56; 95% CI: 0.32–0.97); supports – CHART Transportation (OR: 1.29; 95% CI: 1.01–1.64), FIM Motor (OR: 1.02; 95% CI: 1.00–1.03), and CHART Mobility (OR: 1.02; 95% CI: 1.00–1.03). When all terms were combined into a multivariate model, only post-SCI work experience (OR: 1.25; 95% CI: 1.08–1.42) and number of inpatient days during study (OR: 0.99; 95% CI: 0.99–1.00) remained significant after adjustment. Participants with work experience post-SCI and prior to study enrollment were 24% more likely to obtain CE. Lastly, results from the multivariate model suggest every 1 day of hospitalization led to a 1% decrease in likelihood of obtaining CE, which became a 7% decrease for every 10 days.

Table 2
Odds ratios and 95% confidence intervals for the prediction of competitive employment (n = 213)

Univariate Multivariate

Domain Parameter Description OR 95% CI p-value R² OR 95% CI p-value R²

Socio-demographic Age Age, yr 0.99 0.96–1.02 0.486 0.002 0.97 0.94–1.01 0.182 —

Gender Male 0.55 0.12–2.56 0.454 0.003 1.81 0.32–10.40 0.651 —

Education College (vs. other) 2.06 1.01–4.20 0.046 0.019 2.36 0.89–5.54 0.089 —

Education history Post-SCI education 1.16 0.61–2.19 0.647 0.001 0.82 0.37–1.82 0.605 —

Work experience CE at time of SCI 3.89 2.16–6.81 0.001 0.100 1.54 0.67–3.56 0.301 —

Post-SCI work experience 1.31 1.18–1.46 0.001 0.141 1.25 1.08–1.42 0.003 —

Clinical Injury level Tetraplegia 0.46 0.23–0.92 0.013 — 0.33 0.11–1.01 0.078 —

Paraplegia 1.11 0.58–2.13 0.120 — 0.75 0.32–1.69 0.682 —

AIS D/E 1.00 — 0.120 0.030 1.00 — — —

Time since injury Time since injury, days 1.02 0.99–1.04 0.166 0.009 1.02 0.99–1.06 0.147 —

Inpatient days Inpatient days following enrollment 0.99 0.99–1.00 0.012 0.053 0.99 0.99–1.00 0.044 —

Medical comorbidities Total number of medical comorbidities 0.96 0.85–1.08 0.499 0.002 1.02 0.88–1.20 0.765 —

TBI OSU-TBI ID history of TBI 0.56 0.32–0.97 0.038 0.020 0.57 0.30–1.14 0.115 —

Mental health PHQ-9 1.02 0.95–1.09 0.568 0.002 1.10 0.99–1.21 0.066 —

SWLS 1.02 0.99–1.06 0.215 0.007 1.03 0.98–1.08 0.202 —

Supports Transportation CHART transportation, avg 1.29 1.01–1.64 0.040 0.021 0.92 0.63–1.32 0.625 —

Social support CHART social, avg 1.02 1.00–1.04 0.093 0.014 1.02 0.99–1.04 0.157 —

Mobility FIM motor 1.02 1.00–1.03 0.025 0.026 — — — —

CHART mobility, avg 1.02 1.00–1.03 0.036 0.021 1.00 0.98–1.03 0.835 —

Financial Total monthly benefit, $ 1.00 1.00–1.00 0.245 0.006 1.00 1.00–1.00 0.270 0.257

			Univariate	Multivariate
Socio-demographic	Age	Age, yr	0.99	0.96–1.02	0.486	0.002	0.97	0.94–1.01	0.182	—
	Gender	Male	0.55	0.12–2.56	0.454	0.003	1.81	0.32–10.40	0.651	—
	Education	College (vs. other)	2.06	1.01–4.20	0.046	0.019	2.36	0.89–5.54	0.089	—
	Education history	Post-SCI education	1.16	0.61–2.19	0.647	0.001	0.82	0.37–1.82	0.605	—
	Work experience	CE at time of SCI	3.89	2.16–6.81	0.001	0.100	1.54	0.67–3.56	0.301	—
	Post-SCI work experience	1.31	1.18–1.46	0.001	0.141	1.25	1.08–1.42	0.003	—
Clinical	Injury level	Tetraplegia	0.46	0.23–0.92	0.013	—	0.33	0.11–1.01	0.078	—
		Paraplegia	1.11	0.58–2.13	0.120	—	0.75	0.32–1.69	0.682	—
		AIS D/E	1.00	—	0.120	0.030	1.00	—	—	—
	Time since injury	Time since injury, days	1.02	0.99–1.04	0.166	0.009	1.02	0.99–1.06	0.147	—
	Inpatient days	Inpatient days following enrollment	0.99	0.99–1.00	0.012	0.053	0.99	0.99–1.00	0.044	—
	Medical comorbidities	Total number of medical comorbidities	0.96	0.85–1.08	0.499	0.002	1.02	0.88–1.20	0.765	—
	TBI	OSU-TBI ID history of TBI	0.56	0.32–0.97	0.038	0.020	0.57	0.30–1.14	0.115	—
	Mental health	PHQ-9	1.02	0.95–1.09	0.568	0.002	1.10	0.99–1.21	0.066	—
		SWLS	1.02	0.99–1.06	0.215	0.007	1.03	0.98–1.08	0.202	—
Supports	Transportation	CHART transportation, avg	1.29	1.01–1.64	0.040	0.021	0.92	0.63–1.32	0.625	—
	Social support	CHART social, avg	1.02	1.00–1.04	0.093	0.014	1.02	0.99–1.04	0.157	—
	Mobility	FIM motor	1.02	1.00–1.03	0.025	0.026	—	—	—	—
		CHART mobility, avg	1.02	1.00–1.03	0.036	0.021	1.00	0.98–1.03	0.835	—
	Financial	Total monthly benefit, $	1.00	1.00–1.00	0.245	0.006	1.00	1.00–1.00	0.270	0.257

Note. Univariate predictors were considered significant in initial model for p < 0.10. OR, odds ratio. AIS, American Spinal Injury Association Impairment Scale; CI, confidence interval; CE, competitive employment; SCI, spinal cord injury; CHART, Craig Handicap Assessment and Reporting Technique; OSU-TBI ID, Ohio State University Traumatic Brain Injury Identification Method; TBI, traumatic brain injury; PHQ-9, Patient Health Questionnaire-9; SWL, Satisfaction with Life Score; FIM, Functional Independence Measure. Overall multivariate model C-Statistic = 0.79.

4 Discussion

The purpose of this paper was to examine factors that predicted employment outcomes among IPS participants with SCI who were receiving VHA healthcare, a population with the primary physical disability of SCI but which also included comorbidities such as TBI and mental health conditions. Individual factors that were significant predictors of employment included education, work experience, injury level, inpatient days, TBI history, transportation, and mobility. When these factors were examined in combination, only a work history after the SCI and number of inpatient hospital days remained significant predictors of competitive employment with IPS. Hence, a careful examination of broad domains of sociodemographic, clinical, and social supports did not find many limitations that significantly impacted employment outcome except for two specific factors. This finding supports the basic SE IPS tenet of zero exclusion, which asserts that persons with significant disabilities who want to find competitive employment in the community can benefit from IPS, regardless of level of disability.

In terms of the specific factors that influenced outcomes, the results are consistent with findings from the mental health field – work history is a strong and consistent predictor of IPS employment outcomes. In our sample, work experience after injury predicted positive outcomes, more so than being employed at the time of injury. While the co-morbidity between mental health and medical conditions is acknowledged (Craig, Tran, & Middleton, 2009; Druss, Zhao, Von Esenwein, Morrato, & Marcus, 2011), this was the first study to examine response to IPS employment interventions among a medical rehabilitation population. Among this sample, the longer IPS participants spent in the hospital, the less likely they were to obtain employment. Thus, prolonged hospitalization interferes with the ability to explore opportunities for competitive employment and participate in job development. While hospitalization to optimize medical and rehabilitation goals should not be curtailed, minimizing hospital stay once goals are achieved may be an appropriate goal of IPS services for persons with SCI. Inpatient hospitalization had greater impact on employment outcomes than severity of injury, number of medical problems, or mental health indicators. However, the number of medical complications may not fully reflect the intensity and impact of complications on well-being; an alternative explanation of this finding could be that inpatient hospitalization is a surrogate for severity of medical complications that interfere with obtaining competitive employment.

It is unclear why working at the time of injury was not significant when all predictors were combined, yet post-SCI work experience was a unique predictor of employment outcomes from IPS. Research suggests that after SCI, occupational interests change and become more compatible with the limitations of SCI (Krause & Clark, 2014). That is, post SCI rehabilitation, more realistic appraisals emerge and people may be more open to other occupational fields rather than maintaining interest in previously held positions, which in many cases may have been more physically demanding or traditional jobs. On average, our sample had been living with SCI for over a decade. From a clinical perspective, vocational providers used knowledge gained from post-SCI work experience to inform vocational plans and job searches. Hence, employment interventions were better targeted to an individual’s disability-specific barriers. Another explanation is some individuals may acquire employment post-SCI but not retain it on their own or with limited vocational assistance and need more intensive IPS follow along support services to sustain employment successfully. As this study did not examine factors such as duration of employment or prior vocational rehabilitation, more research is needed to explore this possibility. Clearly, hospital inpatients are less able to participate in community-based employment services such as job development and placement, which are most often linked to positive employment outcome (Ottomanelli, Barnett, Goetz, & Toscano, 2015).

As expected, physical functioning, whether reflected by level of motor and/or sensory impairment (clinical) or by mobility (supports), influenced IPS outcome. Inpatient hospitalization having a stronger effect emphasizes the importance of vocational and healthcare teams working together to optimize health so that individuals can participate in outpatient care and address employment in the community. In term of cognitive factors, TBI history was a modest predictor in the initial model but did not remain significant when combined with other factors. In our earlier report of employment outcomes, participants with SCI who enrolled in IPS as inpatients with a TBI history had worse outcomes than those enrolled as outpatients without a history of TBI (Ottomanelli et al., 2017). Interventional employment studies that include individuals with both conditions are rare, and the complex relationship awaits better elucidation to help inform employment interventions. Certainly, TBI alone exerts a serious impact on return to work, and SE is recognized as a best practice in the field for the TBI population (Wehman, Targett, & Avellone, 2017). Further examination of the SCI and TBI dual diagnosis population with measures of cognitive impairment rather than history of brain injury or examining the severity of the TBI in relation to outcomes may produce more significant and clinically relevant findings for these complex co-occurring conditions.

4.1 Study limitations

The significant interaction of work history and treatment was well described in the mental health research reported above. A notable limitation of this study was that the single arm study design did not allow for the analysis of treatment interaction effects as there was no comparison condition for the IPS intervention. As was previously mentioned, the analysis included the presence of TBI history as a discrete variable rather than severity of injury or assessment of current cognitive impairments. Potentially, TBI severity and/or clinical impairment may be a stronger predictor then these results suggest. This analysis narrowly focused on the single employment outcome of job acquisition, hence, these data do not speak to predictors for job maintenance. As a final note, male veterans were an even larger majority than in the general population of persons with SCI, hence the predictors found in this sample may not be reflective of the general SCI population.

4.2 Conclusions and implications

In the final analysis, only post-SCI work experience and inpatient hospitalization during the study were significant predictors of competitive employment for IPS participants with SCI. The lack of other significant factors emphasizes that, despite many limitations, individuals with SCI can benefit from IPS. Full participation leading to job acquisition is most likely to occur when health is optimized and prolonged hospitalization is minimized. When IPS services are integrated with healthcare and collaboratively focus on staying healthy and engaged in the community, positive employment outcomes can be expected. The less one is hospitalized, the greater the ability to participate in a job development or career planning process. To better target individual supports for ongoing employment of individuals living with disability, programs need to prioritize maintaining health and gaining work experience after spinal cord injury.

Conflict of interest statement

None to report.

Footnotes

Acknowledgements

This material is based on work supported by the Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, Rehabilitation Research and Development Service, Project #O7824R. Clinical Trial Number: NCT01141647. Contents of this article do not represent the views of the Department of Veterans Affairs or the United States Government. Members of the Predictive Model Over Time for Employment (PrOMOTE) group are as follows: Herb Ames, PhD; Melissa Amick, PhD Maggie Budd, PhD, ABPP; Thomas M. Dixon, PhD, ABPP; Kirsten Fisher, MD; Sally Ann Holmes, MD; Virginia “Jennie” Keleher, MSW; Anthony Kerrigan, PhD; B. James LePage, PhD; B. Jenny Kiratli, PhD; Scott McDonald, PhD; Charles McGeough, MS; Doug Ota, MD; Fides Pacheco, MD; Mary Ann Richmond, MD; Sunil Sabharwal, MD; Richard Toscano, MEd; Kevin T. White, MD; and Catherine Wilson PsyD, ABPP. We appreciate the editorial assistance of Lynn Dirk, MAMC, in the preparation of this manuscript.

References

Budisin, B., Bradbury, C. C. L. B., Sharma, B., Hitzig, S. L., Mikulis, D., Craven, C., & Green

R. E.

(2016). Traumatic brain injury in spinal cord injury: Frequency and risk factors. The Journal of Head Trauma Rehabilitation, 31(4), E33–42. 10.1097/HTR.0000000000000153

Burns, T., Catty, J., Becker, T., Drake, R. E., Fioritti, A., Knapp, M., & Wiersma

(2007). The effectiveness of supported employment for people with severe mental illness: A randomised controlled trial. Lancet, 370(9593), 1146–1152. 10.1016/S0140-6736(07)61516-5

Burns, T., Catty, J., White, S., Becker, T., Koletsi, M., Fioritti, A., & Lauber

(2009). The impact of supported employment and working on clinical and social functioning: Results of an international study of individual placement and support. Schizophrenia Bulletin, 35(5), 949–958. 10.1093/schbul/sbn024

Campbell, K., Bond, G. R., Drake, R. E., McHugo, G. J., & Xie, H. (2010). Client Predictors of Employment Outcomes in High-Fidelity Supported Employment: A Regression Analysis.The Journal of Nervous and Mental Disease, 198(8), 556–563. 10.1097/NMD.0b013e3181ea1e53

Craig, A., Tran, Y., & Middleton, J. (2009). Psychological morbidity and spinal cord injury: A systematic review. Spinal Cord, 47(2), 108–114. 10.1038/sc.2008.115

Diener, E., Emmons, R. A., Larsen, R. J., & Griffin, S. (1985). The Satisfaction With Life Scale. Journal of Personality Assessment, 49(1), 71–75. 10.1207/s15327752jpa490113

Druss, B. G., Zhao, L., Von Esenwein

, Morrato, E. H., & Marcus, S. C. (2011). Understanding excess mortality in persons with mental illness: 17-year follow up of a nationally representative US survey. Medical Care, 49(6), 599–604. 10.1097/MLR.0b013e31820bf86e

Elovic, E., & Kirschblum, S. (1999). Epidemiology of Spinal Cord Injury and Traumatic Brain Injury: The Scope of the Problem. Topics in Spinal Cord Injury Rehabilitation, 5(2), 1–20.

Kennedy, P., Lude, P., & Taylor, N. (2006). Quality of life, social participation, appraisals and coping post spinal cord injury: A review of four community samples. Spinal Cord, 44(2), 95–105. 10.1038/sj.sc.3101787

10.

Krause, J., & Anson, C. A. (1996). Self-perceived reasons for unemployment cited by persons with spinal cord injury: Relationship to gender, race, age, and level of injury. Rehabilitation Counseling Bulletin, 39(3), 217–228.

11.

Krause, J. S., & Clark, J. M. R. (2014). Stability of vocational interests after recent spinal cord injury. Rehabilitation Psychology, 59(3), 321–328. 10.1037/a0037265

12.

Linacre, J. M., Heinemann, A. W., Wright, B. D., Granger, C. V., & Hamilton, B. B. (1994). The structure and stability of the Functional Independence Measure. Archives of Physical Medicine and Rehabilitation, 75(2), 127–132.

13.

Macciocchi, S., Seel, R. T., Warshowsky, A., Thompson, N., & Barlow, K. (2012). Co-occurring traumatic brain injury and acute spinal cord injury rehabilitation outcomes. Archives of Physical Medicine and Rehabilitation, 93(10), 1788–1794. 10.1016/j.apmr.2012.01.022

14.

Meade, M., Reed, K., Saunders, L., & Krause, J. (2015). It’s All of the Above: Benefits of Working for Individuals with Spinal Cord Injury. Topics in Spinal Cord Injury Rehabilitation, 21(1), 1–9. 10.1310/sci2101-1

15.

Metcalfe, J. D., Drake, R. E., & Bond, G. R. (2017) Predicting Employment in the Mental Health Treatment Study: Do Client Factors Matter?Administration and Policy in Mental Health and Mental Health Services Research, 44(3), 345–353. 10.1007/s10488-016-0774-x

16.

Milfort, R., Bond, G. R., McGurk, S. R., & Drake, R. E. (2015). Barriers to Employment Among Social Security Disability Insurance Beneficiaries in the Mental Health Treatment Study. Psychiatric Services, 66(12), 1350–1352. 10.1176/appi.ps.201400502

17.

Nagelkerke, N. (1991). A note on a general definition of the coefficient of determination. Biometrika, 78, 691–692.

18.

National Spinal Cord Injury Statistical Center (NSCISC).(2016). Spinal cord injury (SCI) facts and figures at a glance. Retrieved May 23, 2017 from https://www.nscisc.uab.edu/Public/Facts%202016.pdf

19.

Ottomanelli, L., Barnett, S. D., & Goetz, L. L. (2013). A prospective examination of the impact of a supported employment program and employment on health-related quality of life, handicap, and disability among Veterans with SCI. Quality of Life Research: An International Journal of Quality of Life Aspects of Treatment, Care and Rehabilitation, 22(8), 2133–2141. 10.1007/s11136-013-0353-5

20.

Ottomanelli, L., Barnett, S., Goetz, L., & Toscano, R. (2015). Vocational Rehabilitation in Spinal Cord Injury: What Vocational Service Activities Are Associated with Employment Program Outcome?Topics in Spinal Cord Injury Rehabilitation, 21(1), 31–39. 10.1310/sci2101-31

21.

Ottomanelli, L., Goetz, L. L., Barnett, S. D., Njoh, E., Dixon, T. M., Holmes, S. A., White

K. T.

(2017). Individual Placement and Support in Spinal Cord Injury: A Longitudinal Observational Study of Employment Outcomes. Archives of Physical Medicine and Rehabilitation, 10.1016/j.apmr.2016.12.010

22.

Ottomanelli, L., Goetz, L. L., Suris, A., McGeough, C., Sinnott, P. L., Toscano, R., & Thomas

F. P.

(2012). Effectiveness of supported employment for veterans with spinal cord injuries: Results from a randomized multisite study. Archives of Physical Medicine and Rehabilitation, 93(5), 740–747. 10.1016/j.apmr.2012.01.002

23.

Ottomanelli, L., Goetz, L., McGeough, C., Suris, A., Sippel, J., Sinnott, P., & Cipher

D. J.

(2009). Methods of a multisite randomized clinical trial of supported employment among veterans with spinal cord injury. Journal of Rehabilitation Research and Development, 46(7), 919–930.

24.

Ottomanelli, L., Sippel, J., Cipher, D., & Goetz, L. L. (2011). Factors associated with employment among Veterans with spinal cord injury. Journal of Vocational Rehabilitation, 34, 141–150. 10.3233/JVR-2011-0542

25.

Pavot, W., & Diener, E. (1993). Review of the Satisfaction With Life Scale. Psychological Assessment, 5(2), 164–172.

26.

Spitzer, R. L., Kroenke, K., & Williams, J. B. (1999). Validation and utility of a self-report version of PRIME-MD: The PHQ primary care study. Primary Care Evaluation of Mental Disorders. Patient Health Questionnaire. JAMA282(18), 1737–1744.

27.

Tomassen, P. C., Post, M. W., van Asbeck

F. W.

(2000). Return to work after spinal cord injury. Spinal Cord, 38(1), 51–55.

28.

Trenaman, L. M., Miller, W. C., & Escorpizo, R. (2014). Interventions for improving employment outcomes among individuals with spinal cord injury: A systematic review. Spinal Cord, 52(11), 788–794. 10.1038/sc.2014.149

29.

Trenaman, L., Miller, W. C., Querée, M., & Escorpizo, R. (2015). Modifiable and non-modifiable factors associated with employment outcomes following spinal cord injury: A systematic review. The Journal of Spinal Cord Medicine, 38(4), 422–431. 10.1179/2045772315Y.0000000031

30.

Wehman, P. H., Targett, P. S., & Avellone, L. E. (2017). Educational and Vocational Issues in Traumatic Brain Injury. Physical Medicine and Rehabilitation Clinics of North America, 28(2), 351–362. 10.1016/j.pmr.2016.12.010

31.

Whiteneck, G. G., Charlifue, S. W., Gerhart, K. A., Overholser, J. D., & Richardson, G. N. (1992). Quantifying handicap: A new measure of long-term rehabilitation outcomes. Archives of Physical Medicine and Rehabilitation, 73(6), 519–526.

32.

Young, A. E., & Murphy, G. C. (2002). A social psychology approach to measuring vocational rehabilitation intervention effectiveness. Journal of Occupational Rehabilitation, 12(3), 175–189.