Linking job duties,functioning,and employment status using the Work-Disability Functional Assessment Battery (WD-FAB): An expert coding and quantitative analysis

Abstract

BACKGROUND:

An understanding of the link between specific occupational demands and individual worker functioning is limited, although such information could permit an assessment of the fit between the two in a manner that would inform national and state disability programs such as vocational rehabilitation and Social Security disability programs.

OBJECTIVE:

Our goal was to examine the utility of assessing physical and mental functioning relative to self-reported job duties to identify the domains of worker functioning most likely to create barriers to fulfilling an occupation’s specific requirements.

METHODS:

Through primary survey data collection, 1770 participants completed the Work-Disability Functional Assessment Battery (WD-FAB) instrument after reporting details on their occupations (or most recent occupation if not working). Expert coders evaluated the level of function expected to successfully carry out each self-reported job duty with respect to six scales of physical and mental function. Quantitative analysis is used to examine the relationship between functioning and job duties.

RESULTS:

Those not working due to disability were more likely to fall short of the threshold of the physical and mental functioning requirements of their last job’s three main job duties compared to those currently employed. Mental function scales were most likely to be the area experiencing a shortfall.

CONCLUSIONS:

Functional difficulties impede the ability to continue working in particular jobs that require that ability. This points to a need for specific accommodations to be implemented to bridge the gap between job requirements and functional capacity so that workers may remain engaged in their current work.

Keywords

Work disability physical health mental health employment

1 Introduction

In the U.S., the proportion of working-age adults with disabilities that work is alarmingly low compared to the proportion of working-age adults without disabilities that work. While approximately seven in ten working-age adults without disabilities are employed, only four in ten working-age adults with disabilities are employed [1]. This disparity limits the talent pool that could contribute in rich and unique ways to the success of workplaces as well as the national economy [2]. Since work disability is currently conceptualized as the outcome of the interaction (gap) between the functional abilities of individuals and the environment as characterized by job requirements [3,4], it is essential to be able to characterize both individual functional abilities and job requirements in order to evaluate the fit between individuals and specific jobs. Relating job requirements to worker functional abilities is critical to the design of effective work rehabilitation and employment support activities, and to the determination of eligibility for federal disability programs [5]. Further, preventing work disability ideally incudes early identification and assessment of barriers to work, as well as development of tailored stay-at-work/return-to-work (SAW/RTW) plans targeting specific job requirements [6]. In this context, the gap between worker’s functioning and job requirements is the focus of health care, vocational rehabilitation, and employment services. For the Social Security Administration (SSA)’s disability-related income support (Supplemental Security Income) and social insurance (Social Security Disability Insurance) programs, adjudicators must weigh medical, vocational factors, and other evidence to determine whether claimants are able to perform substantial gainful activity through sustained employment [7]. While the importance of the relationship between worker function and work requirements has been acknowledged [8], a systematic and scientifically validated process for relating job requirements to worker functioning has not been identified, and there is no standard of practice or consensus on methods to do so.

Efforts have been made to bridge the gap between functional requirements of jobs and the functioning of job seekers. For example, the current O*NET Resource Center provides a database of standardized descriptions of occupations in the U.S. economy [9]. O*NET provides information for a wide range of occupations within categories such as skills and abilities along with ratings of the level required to do the job. In addition, O*NET has developed a career exploration tool that assesses various abilities (e.g., arithmetic reasoning, computation, spatial ability, and manual dexterity) that can be linked to occupations in O*NET. However, the specificity of the information captured across these domains poses challenges to creating links between job demands and the requisite functional abilities needed to conduct them successfully. Li et al. developed a “proof-of-concept” web-based tool that linked job demands, as characterized by O*NET, with function as captured by a self-reported activity measure [10]. Work disability experts manually matched measurement constructs, job names, and item difficulties from the self-report measure with job demands. The tool was then piloted on fourteen adults with one or more activity limitations to yield a list of potential jobs that were aligned with their abilities. While demonstrating the feasibility of linking functional abilities with job demands, the tool lacked the ability to characterize individual abilities comprehensively and made some assumptions in the linking process that require further examination and validation. Currently, there are no tools that comprehensively link worker functioning with job demands that could inform an assessment of the fit between workers and jobs for use within work and vocational rehabilitation, or employment services.

In order to address this gap, we employed the Work-Disability Functional Assessment Battery (WD-FAB) which was developed [11,12] to comprehensively assess self-reported work-relevant functioning [3]. The WD-FAB measures physical and mental domains across eight scales developed using item response theory in large samples of individuals with and without work disability. Of particular relevance, levels of functioning and related score thresholds have been established [13]. We associated worker functioning, as assessed by the WD-FAB, with self-reported job duties via the O*NET. We hypothesized that the WD-FAB functional levels and score thresholds could be used to link person functioning to the ability to work at specific jobs.

Therefore, our objectives were to use existing WD-FAB functional levels to apply minimum score thresholds as functional requirements for specified job duties identified by individuals, and to test whether individuals with WD-FAB scores exceeding score thresholds for their job duties were more likely to be employed.

2 Materials and methods

2.1 Participant recruitment

We recruited a purposive sample of 2,224 respondents through a Qualtrics online opt-in internet panel conducted in April and May of 2020. Qualtrics utilizes third party partners to design samples meant to meet recruitment quotas. Qualtrics panel members receive incentives for participation (screening in and fully completing the survey), points which can be converted into a variety of monetary and non-monetary rewards over time [14]. We sampled from two Qualtrics panels: 1) general adults and 2) people with chronic health conditions to achieve a convenience sample consisting of 75% individuals not working due to a disability and 25% employed individuals. Participants had to meet age requirements for inclusion (working age, 18 to 67). Individuals who were employed at the time of enrollment were eligible for inclusion in the employed sample. Those who reported that they were not working due to a disability or health condition or retired due to a disability or health condition as one reason for retiring were eligible for inclusion in the work disability sample. Anticipating that some sample cases would be excluded due to insufficient data, we established a quota of a maximum sample size of 500 employed respondents (the remainder of the sample was allocated to those with a work disability) to ensure sufficient analytic power to detect small effect sizes when comparing differences in functioning scores to the work disability sample (with a power of 90% and adopting a Type I error rate of 5% and incorporating 12 or fewer independent variables as controls) [15]. The primary ethical concern with the involvement of human subjects was the risk of identifying subjects based upon their responses. While stored data did not include email addresses or location (ISP address), some respondents did provide detailed employment data, including employer name. We address this by presenting only summarized data. This study was approved by the University of New Hampshire Institutional Review Board (#8286) after review of data collection protocol.

2.2 Data collection instrument

The survey was conducted online and included three sections. One section asked for demographic characteristics, including age, gender, race and ethnicity. One section included open-ended questions about job title, industry and job duties using the following questions: 1) “What type of business or industry is this? (What do they make or do?)” 2) “What is your job called?” and 3) “What do you actually do at that job? What are three of your main activities or duties at this job? For example: typing, keeping accounting books, filing, selling cars, operating printing press, and laying brick.” Additional detail on the job questions is provided in Appendix A [note that those not employed reported about their most recent job]. In developing the specific wording for these items, we examined a variety of established questionnaires that included similar work-specific measures. Most of these studies employed identical or similar wording and our final selections utilized wording from the Medical Expenditures Panel Survey 2018 Employment Module for employment and occupation and the 2011-2018 American Time Use Survey for job duties.

The third portion of the survey entailed administration of the computer-adaptive WD-FAB. The WD-FAB characterizes work-relevant functioning in physical and mental domains across eight scales: Basic Mobility; Upper Body Function; Fine Motor Function; Mood & Emotions; Self-Regulation; Cognition & Communication, Resilience & Sociability, and Community Mobility, which includes driving and use of public transportation. A filter question is used to identify individuals who use a wheelchair for locomotion who are then administered the Wheelchair Mobility scale. Using item-response theory, the WD-FAB draws from item banks including over 300 items delivered as eight brief 6-10 item computer-adaptive tests. Several studies have reported on the validity and reliability of the WD-FAB in populations of SSA claimants and U.S. working-age adult population samples [11, 16]. Scores for each scale are standardized based on a working-age adult sample score distribution with a mean of 50 and standard deviation of 10, with lower scores indicating lower function.

We did not use the Resilience & Sociability, Community Mobility, or Wheelchair Mobility scales in this study. Functional levels have not yet been established for the Wheelchair Mobility or Community Mobility scales. The Resilience & Sociability scale characterizes one’s management of stressful situations and socializing. Data collection coincided with the start of nationwide stay-at-home orders resulting from the COVID-19 pandemic in 2020, a time during which people were avoiding social situations and experienced marked increase in anxiety, stress, and depressive symptoms [17]. The scores obtained from the Resilience & Sociability scale were dissimilar from previous WD-FAB data in similar samples, a result we attributed to the timing of data collection. For this reason, we excluded it from the analysis.

Note that in addition to the survey data, Qualtrics provided demographic data on these respondents from their sampling frame, including educational attainment, household and individual income, and geographic region. These are also based on self-reports.

2.3 Description of analytic approach

2.3.1 Data cleaning

Each respondent reported three main job duties performed at their current or immediate past work as an open-ended response. These write-in responses were manually reviewed to identify missing or unusable responses, and 75 cases were dropped prior to coding. Additional instances of unusable data were flagged for removal as appropriate by the coders.

2.3.2 Applying WD-FAB functional levels and score thresholds to job duties

Using previously developed WD-FAB functional levels [13], coders reviewed each job duty and for each of the six WD-FAB scales, assigned it a functional level based on expert judgement representing the minimal level of function required for that specific job duty in that occupation. For example, prior research has established five functional levels for the Basic Mobility scale with the following score ranges: lowest (0-17), low (18-30), average (31-40), high (41-53), and highest (54-100) (Table 1). If the minimal functional level required for a job duty (e.g., run cash register) was assigned as average, the minimum score threshold would be assigned in the analysis as the lowest score for the average functional level, which is 31. Based on the score thresholds and the descriptions of the functional levels in the WD-FAB, the coding team determined that every job duty would require at least a score of “average” in every scale in order to complete the task in the workplace. The full set of functional levels, score thresholds and descriptions are provided in Appendix B.

Table 1
Previously established functional levels, score ranges, and descriptions for the WD-FAB’s Basic Mobility scale (13)

Functional levels	Score range	Basic Mobility involves getting into and out of positions, staying in positions for periods of time, and walking and moving around from one place to another
Lowest	0-17	Persistent and significant limitations in ability to move around in everyday life. For example: Unable to perform the easiest activities such as getting in and out of bed, sitting for short periods, and walking short distances (e.g. inside the home). Unable to do mobility activities such as getting into, out of, and staying in kneeling, squatting, and bending positions. Unable to get around by walking, climbing, or running with or without assistive devices such as a wheel chair or walker.
Low	18-30	Periodic, significant limitations in ability to move around in everyday life. For example: Has a lot of difficulty performing the easiest mobility activities such as getting in and out of bed, standing and sitting for short periods (e.g. 20 minutes), walking short distance (e.g. inside the home). Unable to do more difficult mobility activities such as kneeling to tie a shoe, standing up from an armless chair, going down a flight of stairs with a handrail, walking for 30 minutes, sitting or standing for longer periods (e.g. 2 hours).
Average	31-40	Somewhat limited in ability to move around in everyday life. For example: Has some or a little difficulty with the easiest mobility activities such as getting in and out of bed, standing and sitting for short periods (e.g. 20 minutes), walking short distance (e.g. inside the home). Has a lot or some difficulty with harder mobility tasks such as walking for 30 minutes, kneeling to tie a shoe, standing up from an armless chair, going down a flight of stairs with a handrail, walking for 30 minutes, sitting or standing for longer periods (2 hours).
High	41-53	Some or mild limitations in ability to move around in everyday life. For example: Has no or a little difficulty with the easiest mobility activities such as getting in and out of bed, standing and sitting for short periods (e.g. 20 minutes), walking short distance (e.g. inside the home). Has little or no difficulty with more difficult tasks such as walking for 30 minutes, kneeling to tie a shoe, standing up from an armless chair, going down a flight of stairs with a handrail, walking for 30 minutes, sitting or standing for longer periods (2 hours). Has some difficulty with the hardest activities such as walking long distances (e.g. a mile without rest), running fast (e.g. for 2 minutes), standing for long periods (e.g. 6 hours), squatting for long periods (e.g. 20 minutes).
Highest	54-100	No limitations in moving around in everyday life. For example: Able to perform the most difficult activities such as walking long distances (e.g. a mile without rest), running fast (e.g. for 2 minutes), standing for long periods (e.g. 6 hours), squatting for long periods (e.g. 20 minutes).

Because we coded the functioning requirement of each job duty in each of the six scales, we were able to evaluate whether an individual’s functioning (WD-FAB scores) met the expected score threshold for carrying out the self-reported job duties specific to their occupation. (For our work disability sample, the job duty information was collected for the last job held.) We took the highest score threshold across the three job duties within each scale and evaluated whether an individual’s WD-FAB score for that scale met the threshold.

2.3.2.a Blanket coding of similar job duties

The coding team identified job duties that were identical or similar in nature (e.g., cash checks, cash handling, cash out customers) upon first identification, and established consensus on the functional requirement levels. Functional levels for these cases were then applied to all the similar job duties in the dataset. These cases were flagged so that coders could review the associated industry and occupation and confirm appropriateness of the functional level in the context of the participant’s job. Just under one-third (31.7%) of all job duties were coded using this method.

2.3.2.b Job duty coding calibration and reconciliation

The open-ended data were split among three teams, each with two coders for efficiency, such that each set of participant responses (occupation, industry, three job duties) was reviewed independently by two coders. The coders were either experts in measurement of function and the WD-FAB or were healthcare clinicians trained by them on the coding methods. To start, each team conducted preliminary calibration coding. They completed coding for 10 participants independently, then met to reach consensus on coding decisions. The result was an agreement on how to approach the coding of the remaining data.

To examine intercoder reliability of the full data, we compared values assigned by each coder pair for each scale within each job duty. Due to careful calibration at the onset of coding, 89% of coder pairs were in agreement. The coders conducted a consensus process for divergent cases (11% of pairs) to identify the main reason(s) for disagreement. For the majority of these divergent cases, insufficient detail was provided by the respondent about job duties to allow for reliable coding of the functional complexity of the task. As a result, we excluded these cases where coders did not agree on functional complexity required of a task. After examining the demographic profile of these cases compared to those that remained in the analytic sample, we found no significant differences with respect to socio-demographic characteristics that would suggest sampling bias. We examine the effect of this decision on our analysis in Appendix C.

2.4 Analytic approach

Using quantitative analysis, we summarized the demographic characteristics of our sample and present them alongside a weighted national sample to demonstrate the extent to which it represents the working-age population. Next, we compared the demographic characteristics between the analytic sample and the excluded cases.

We summarized the WD-FAB scale scores and analyzed WD-FAB data by employment status (employed and not employed due to disability) at both the scale score level and as a full profile of all scores, displaying also whether there are significant differences by employment status utilizing independent sample t-tests. First, we compared respondents’ scale scores with the coded job duty score thresholds. We determined the proportion of respondents that met the threshold for each job duty and identified patterns for when these thresholds were met. In order to analyze WD-FAB data across the six scales, we defined WD-FAB score profiles by collapsing the functional levels into categories of low, medium, and high scores. Using these categories, we defined five functional profiles such that each profile is distinct from all other profiles and any respondent can be represented by a profile. The five functional profiles are ranked ordinally and are defined as follows:

Profile 1 (P1): 2 or more low scores

Profile 2 (P2): Exactly 1 low score

Profile 3 (P3): 2 or more medium scores and no low scores

Profile 4 (P4): Exactly 1 medium score and no low scores

Profile 5 (P5): All high scores

Each respondent was assigned one of the five functional profiles based on their scale scores. The respondent’s occupation was also assigned an occupation score profile, using the profiles as defined above, based on the maximum functional requirements coded across the three job duties for the occupation. We then calculated the proportion of respondents in each sample (employed and not employed) where the functional profile matched the occupation score profile, examining whether there were statistically significant differences using independent sample t-tests (where the mean of employment, coded as 0 if not employed and 1 if employed permits this calculation). The coding for the functional profiles was performed in Python, and statistical analysis was performed in Stata, version 15.1.

Finally, we evaluated the ability of the continuous WD-FAB scores to predict whether a respondent was in our employed sample using a logistic regression model, controlling for age, gender, and occupational complexity. This method predicts the odds of someone being employed (value of 1) compared to not employed (value of 0). The WD-FAB scoring is normally distributed with a mean of 50 and a standard deviation of 10 [18]. While the WD-FAB items often are correlated with one another (low functioning in one physical functioning domain may be associated with low functioning in another), the standard errors for these scores are small in our model and we retain them all to answer our research question.

3 Results

3.1 Sample

After excluding cases with missing or ambiguous responses, the resulting analytic sample included 1,770 cases with complete employment and functioning scores. Of these, 421 (24%) were currently employed and 1349 (76%) were not employed or retired due to a disability. Employed respondents may or may not have a disability; their only criteria for screening in was to be employed and be between the ages of 18 and 67.

We show the demographic characteristics of our sample alongside weighted demographic data of adults from the 2018 National Health Interview Survey (NHIS) for context (Table 2) to allow consideration of the extent to which our sample compares to this nationally representative health-focused survey. We display the demographics separately for those employed and those not working due to disability. Our sample overrepresents white; non-Hispanic; high income workers; older workers; the highly educated; and women. Our health characteristics are similar to national estimates. The distribution of self-reported health is similar for both the employed and work disability samples. Slightly more chronic conditions were reported for those with work disability compared to employed respondents.

Table 2
Demographic characteristics for 1,770 respondents compared to U.S. population characteristics based on a weighted national sample from the National Health Interview Survey (NHIS) (column % unless otherwise noted)

	Not employed due to disability		Employed
	Qualtrics sample	NHIS 2018	Qualtrics sample	NHIS 2018
Race
White only, non-Hispanic	84.5%	64.1%	78.6%	65.0%
Black only, non-Hispanic	6.2%	2.7%	5.3%	12.6%
Hispanic (any race)	3.8%	17.8%	3.1%	15.2%
Asian only, non-Hispanic	0.7%	2.6%	11.1%	5.9%
Other/multiracial	4.8%	12.9%	1.9%	1.4%
Gender
Male	27.4%	49.1%	54.6%	50.9%
Female	72.6%	50.9%	45.4%	49.1%
Age (mean)	54.8	50.3	52.0	41.7
Self-reported health status
Excellent	1.5%	2.6%	24.7%	31.4%
Very good	8.2%	8.6%	40.1%	37.8%
Good	29.3%	27.4%	29.2%	24.9%
Fair	44.4%	38.1%	5.7%	5.4%
Poor	16.6%	23.2%	0.2%	0.5%
Chronic conditions^a
None	4.2%	13.3%	62.2%	58.3%
One	12.4%	8.4%	14.0%	11.2%
Two or more	81.0%	78.4%	21.9%	30.5%
DK/RF	2.4%	–	1.9%	–
Household income
Less than $50,000	71.9%	64.8%	16.3%	31.3%
$50,000-74,999	12.7%	11.0%	18.9%	16.5%
$75,000-99,999	5.3%	5.9%	15.6%	14.9%
$100,000+	5.6%	8.2%	43.5%	29.8%
DK/RF	4.5%	10.1%	5.7%	7.0%
Education
Less than high school	3.3%	17.8%	0.5%	5.5%
High school	20.5%	37.2%	7.4%	23.0%
Some college	53.0%	31.1%	24.2%	30.7%
4-yr college degree	15.4%	8.7%	30.4%	25.5%
Postgrad	7.5%	3.7%	37.3%	15.0%
DK/NA	0.4%	1.4%	0.2%	0.3%

^a On our survey, respondents were asked if they had a chronic health condition. If yes, they were asked if they had more than one chronic condition. We combined responses from these two questions to measure whether they had multiple chronic conditions. In the NHIS, we counted reports of the following as chronic conditions: hypertension, coronary heart disease, stroke, diabetes, cancer, arthritis, or hepatitis (diagnosed by doctor or health care provider); weak or failing kidneys (during the past 12 months); asthma (currently); or chronic obstructive pulmonary disease (defined as having emphysema or chronic bronchitis in the past 12 months).

3.2 WD-FAB scale scores

Table 3 shows the mean WD-FAB scale scores. Mean scores were significantly higher for those who were employed compared with those who reported work disability across all scales. The magnitudes of the differences between groups were larger for the physical functioning versus the mental functioning scales. The largest differences were found for the Basic Mobility, Upper Body Function, and Mood & Emotions scales.

Table 3
WD-FAB scale scores by employment status for 1,770 respondents.

WD-FAB Scale	Not employed due to disability mean (std err)	Employed mean (std err)	Independent sample t-test^a
Physical Functioning Scales
Basic Mobility	49.1 (.20)	61.7 (.36)	***
Upper Body Function	46.9 (.18)	57.2 (.24)	***
Fine Motor Function	63.9 (.18)	67.7 (.29)	***
Community Mobility	48.5 (.20)	53.5 (.30)	***
Wheelchair Mobility^b	50.5 (.44)	54.1 (1.2)	**
Mental Functioning Scales
Communication & Cognition	51.5 (.13)	54.9 (.18)	***
Resilience & Sociability	47.6 (.17)	50.1 (.30)	***
Self-Regulation	53.5 (.18)	55.4 (.35)	***
Mood & Emotions	54.2 (.24)	59.3 (.39)	***

^a* p < .0.5, **p < 0.01, ***p < 0.001. ^b Assessed only for respondents who report using a wheelchair (n = 259 work disability sample; n = 24 employed sample). The work disability group n = 1309 for every other scale except community mobility (n = 1066) and the employed sample has n = 413 for every other scale.

3.3 Ability to meet occupational functioning requirements

Table 4 shows the percentage of people whose individual functioning (WD-FAB scores) met the expected score threshold for carrying out the self-reported job duties specific to their occupation. (For our work disability sample, the job duty information was collected for the last job held.) Sixty-three percent of the employed respondents met the functioning requirement in all six scales compared to just 36.1% of the work disability sample. Over 95% of both the employed sample and work disability sample met the score thresholds for the physical functioning scales, though a significantly higher proportion of the employed sample met the requirements for Basic Mobility and Upper Body Function scales. A smaller proportion of respondents met each of the score thresholds for the mental functioning scales, and the employed sample was significantly more likely than those with work disability to meet the threshold in the Communication & Cognition and Mood & Emotions scales.

Table 4
Percentage meeting occupational requirements of all three of their specific job duties for each WD-FAB scale and across all six scales, by employment status (occupational requirements were for last job among work disability sample)

WD-FAB Scale	Not employed due to disability	Employed	Independent sample t-test^a
Physical Functioning Scales
Basic Mobility	96.7%	98.6%	*
Upper Body Function	95.6%	98.6%	***
Fine Motor Function	95.3%	96.9%
Mental Functioning Scales
Communication & Cognition	84.6%	94.3%	***
Self-Regulation	83.2%	86.7%
Mood & Emotions	81.1%	92.4%	***
Met functioning requirements across all six scales	36.1%	63.4%	***

^a* p <.0.5, **p < 0.01, ***p < 0.001

3.4 Individual and occupational functioning profiles

The way in which function across multiple scales collectively relates to the demands of work is complex. While some people may be unable to work due to a substantial decrease in one domain of functioning, other workers may experience smaller decreases in functioning in multiple large-scale areas, causing them to exit the labor force. For this reason, we examined the relationship between individual and occupational functioning profiles. Tables 5 (work disability respondents) and 6 (employed respondents) show the distribution of individuals in each functioning profile by their occupational profile. The cells shaded in grey are instances where an individual’s functioning profile did not meet their occupation’s functioning profile requirements. When an occupation’s duties are more demanding (requiring higher-level functioning across more scales), a greater proportion of workers do not meet the threshold. The work disability group fell below the necessary functioning profile more frequently than the employed group did. Among the group with work disability, 35% of those with the lowest occupational requirements (P3), 83.4% of those in the middle group (P4), and 91.5% of those in the most demanding occupational profile group (P5) did not have the functioning necessary to meet their job’s demands. By contrast, among those in the employed group, the percentage not meeting their occupation’s profile threshold for each category was 23.4% (P3), 39.5% (P4), and 79% (P5). Table 7 presents mean WD-FAB scores for each scale for each profile by employment status. Increasing score trends can be observed from low to high profiles for each scale.

Table 5
Distribution of individual functioning profiles by occupational demand profile, work disability sample

		Occupational demand profile
		P3: 3 or more scales medium (none low)	P4: 1 medium, all other scales high	P5: All scales high	Overall
Individual functioning profile	P1: 2 or more scales low	15.4%	14.4%	12.8%	14.6%
	P2: 1 domain low	19.6%	28.0%	28.0%	23.1%
	P3: 2 or more scales medium (none low)	37.0%	41.0%	34.7%	37.1%
	P4: 1 medium, all other scales high	18.1%	9.2%	16.1%	16.1%
	P5: All scales high	9.9%	7.4%	8.5%	9.1%
	Number of cases	791	229	329	1349

Table 6

Distribution of individual functioning profiles by occupational demand profile, employed sample

		Occupational demand profile
		P3: 3 or more scales medium (none low)	P4: 1 medium, all other scales high	P5: All scales high	Overall
Individual functioning profile	P1: 2 or more scales low	9.0%	3.7%	6.5%	7.6%
	P2: 1 scale low	14.4%	8.6%	16.1%	13.5%
	P3: 2 or more scales medium (none low)	28.1%	27.2%	30.7%	28.3%
	P4: 1 medium, all other scales high	28.4%	33.3%	25.8%	29.0%
	P5: All scales high	20.1%	27.2%	21.0%	21.6%
	Number of cases	278	81	62	421

Table 7

Mean score (standard error) for each scale by profile and employment status for 1,770 respondents

	Not employed due to disability (n = 1,349)					Employed (n = 421)
	> = 2	1	> = 2 MED	1 MED	All	> = 2	1	> = 2 MED	1 MED	All
	LOW	LOW	(no LOW)	(no LOW)	HIGH	LOW	LOW	(no LOW)	(no LOW)	HIGH
	P1	P2	P3	P4	P5	P1	P2	P3	P4	P5
Basic Mobility	47.9 (0.50)	48.8 (0.44)	49.0 (0.33)	49.5 (0.46)	51.5 (0.70)	56.1 (1.4)	59.5 (1.1)	60.3 (0.7)	62.8 (0.59)	65.3 (0.5)
Upper Body Function	45.6 (0.42)	46.3 (0.38)	46.6 (0.29)	48.2 (0.45)	49.2 (0.61)	52.5 (1.1)	55.4 (0.8)	56.2 (0.5)	58.5 (0.3)	59.7 (0.2)
Fine Motor Function	61.7 (0.45)	63.8 (0.35)	63.6 (0.32)	65.7 (0.37)	66.2 (0.48)	59.6 (1.5)	65.5 (1.0)	67.5 (0.6)	69.4 (0.2)	70.0 (0.1)
Communication & Cognition	46.2 (0.37)	50.1 (0.26)	52.0 (0.18)	54.9 (0.15)	56.0 (0.16)	48.9 (0.8)	52.9 (0.6)	54.4 (0.3)	56.5 (0.1)	57.0 (0.1)
Self-Regulation	44.6 (0.22)	50.1 (0.26)	54.5 (0.18)	58.2 (0.33)	64.0 (0.18)	43.6 (0.6)	48.0 (0.7)	53.8 (0.3)	56.9 (0.4)	64.0 (0.2)
Mood & Emotions	42.4 (0.30)	48.6 (0.38)	55.9 (0.19)	62.7 (0.30)	65.6 (0.28)	43.5 (0.7)	54.4 (1.2)	55.9 (0.3)	63.7 (0.4)	66.1 (0.3)

Note that all occupational job duties were assumed to require at least medium functioning on all scales. Medium functioning is not a middle or average rate of functioning but is a rather low requirement. Refer to Appendix A for description of what medium functioning levels require.

3.5 Predictive analysis

Logistic regression results that examine whether the WD-FAB scale scores predict whether respondents were employed or not after controlling for demographic characteristics, are summarized in Table 8. Younger respondents were more likely to be employed than older respondents. Men had 79% increased odds of being employed compared to female respondents. Those with at least a college degree had 21 times higher odds of being employed compared to those who did not complete high school. Respondents who met the functioning requirement for their job had 85% increased odds of being employed.

Table 8
Logistic regression predicting employed status for 1,770 respondents

	Log odds ratio (95% confidence interval)	P-value
Age	0.97 (0.95 - 0.99)**	.002
Gender (ref: female)	1.79 (1.27 - 2.53)***	.001
Education (ref: < high school)
High school	5.46 (0.66 - 45.15)	.115
Some college/tech school	4.58 (0.58 - 36.41)	.150
Bachelor’s degree or higher	21.09 (2.66 - 167.11)**	.004
WD-FAB scores (continuous)
Physical
Basic Mobility	1.09 (1.06 - 1.12)***	.000
Upper Body	1.21 (1.16 - 1.26)***	.000
Fine Motor	0.94 (0.91 - 0.98)**	.002
Mental
Communication & Cognition	1.10 (1.05 - 1.16)***	.000
Self-Regulation	0.97 (0.94 - 1.00)	.103
Mood & Emotions	1.00 (0.98 - 1.03)	.784
Individual met functioning threshold for all 3 job duties in all 6 scales	1.85 (1.29 - 2.65)***	.001
Intercept	0.00 (0.000.00)***	.000

Note: *p <.05, **p <.01, ***p <.001. Pseudo R2 = .5081.

Among the physical functioning scales, each point increase in Basic Mobility was associated with a 9% increased likelihood of being employed; each one-point increase in Upper Body Function was associated with a 21% increase in the odds of being employed. For Fine Motor Function, those with higher functioning had slightly decreased odds of being in the employed sample but note that almost everyone in our sample had very high fine motor functioning and the majority of our sample (75%) was not employed due to disability. Among the mental functioning scales, each one-point increase in Communication & Cognition was associated with a 10% increase in the odds of being employed. Self-Regulation and Mood & Emotions were not significant predictors of employment status after controlling for demographics, meeting occupational demands, and other types of functioning.

Figure 1 illustrates the relationship between mean categorical WD-FAB scores by employment status compared to the mean minimum required scale score for all respondents’ occupations. We note that the work disability group did not meet the threshold for the mental functioning scales, on average. While the work disability group did meet the average job duty requirements on the physical functioning scales, their scores were consistently lower than those in our employed sample, except for Fine Motor Function, where all respondents scored high and there was very little variation. The employed sample met the average threshold for all areas of functioning with the exception of Self-Regulation. The employed sample scored slightly higher than the work-disability group in this scale but still lower than hypothesized.

Fig. 1

Mean WD-FAB category by employment status compared to the mean minimum required scale score for all respondents’ occupations for n = 1,770 respondents.

4 Discussion

In this analytic sample of working-age adults, we found that WD-FAB scores reflecting physical and mental functioning were related to work disability after controlling for other important demographic factors. In uncontrolled analyses, we found that mean WD-FAB scores were higher for persons who were employed versus persons who were not working. In controlled analyses, each point increase in Basic Mobility, Upper Body Function, and Communication & Cognition were associated with a 9%, 21%, and 10% increased likelihood of being employed, respectively. In addition, uncontrolled analyses demonstrated a relationship between functional abilities and occupational demands relative to employment status. While 63% of the employed respondents met the functional requirements of their occupations in all six WD-FAB scales, only 36% of adults with work disability met these requirements. In controlled analyses, respondents meeting functioning requirements in all six scales for their occupations were 85% more likely to be employed than those who did not meet all requirements. On average, adults with work disability did not meet the threshold for mental functioning across all occupations.

Consistent with previous studies, we demonstrated an association between worker functioning and ability to work. In a study of 1,312 working-age adults with traumatic brain injury, functioning, as assessed by the WHO-DAS 2.0, predicted return to work with high accuracy using individual domain scores or summary scores [19]. Other studies that used different functioning assessment scales also predicted return to work with high accuracy among adults with TBI [20]. Self-reported functioning has been found to be consistently reduced among groups of adults with mental disorders such as schizophrenia, bipolar disorder, or depressive disorder, however work status varied based on the diagnostic group [21]. While very few adults in the group with schizophrenia worked, roughly half of those in the mood disorders group were working even though they reported greater functional impairment. The investigators attributed this to a well-documented under-estimation of functioning via self-report by adults with mood disorders compared to objective assessment. Measures of physical functioning have been shown to be far superior predictors of employment among adults with back pain compared to measures of pain or clinical diagnoses [22]. Finally, work instability has been defined as “a state in which the consequences of a mismatch between an individual’s functional and/or cognitive abilities and the demands of their job can threaten continuing employment if not resolved” [8]. Work instability instruments have been shown to predict future work transitions among adults with arthritis [23]. In a German study of geriatric nurses, the probability of long-term sick leave was eight times higher for those identified as being at high risk for work instability [24]. Thus, our finding of the association between better self-reported functioning and employment is well supported in the literature.

A unique finding of the current study is the association between specific functional difficulties in relationship to job demands. While the majority of the employed respondents met the functioning requirements of their occupations in all scales, only about one third of adults with work disability met these requirements. On average, adults with work disability did not meet the threshold for mental functioning across all occupations. This suggests that specific functional difficulties impede the ability to continue working in particular jobs that require that ability. These results mirror what is seen in SSA’s disability social insurance program (Social Security Disability Insurance) where those with mental disorders account for the highest proportion of beneficiaries [1]. It also points to a need for specific accommodations to be implemented in an attempt to reduce functional difficulties to allow workers to remain engaged in their current work. Commonly used workplace accommodations for persons with mental health conditions include flexible scheduling, reduced work hours, modified training and changes in supervisory practices [25]. Alternately, other jobs could be identified that better matched workers’ functioning profile, potentially allowing them to remain engaged in the workforce.

Contemporary return-to-work models depict the complexity of interactions and the variety of factors that impact work ability [26, 27]. While worker functioning is one component of these models, a multitude of other factors and system dynamics drive work outcomes. This is evidenced in the findings from the current study. We found that mental health functioning differentially affected a person’s ability to work. Literature suggests the development of mental health problems is complex and often represents a myriad of factors [28]. As such, the degree to which employment participation is restricted in relationship to a person’s mental health function limitations is equally complex [29]. The WD-FAB can be used to characterize multiple aspects of an individual’s mental health and functioning that may be associated with an individual’s ability to work but does not characterize their ability to manage these characteristics in a workplace setting nor does it characterize features of the workplace that may amplify or diminish the impact of these characteristics. While not included in our current study, the Resilience & Sociability scale focuses on interpersonal skills related to aspects of flexibility, openness, acceptance of feedback, and persistence in stressful situations when interacting with others. Often these features of one’s ability is reflected in more visible or externally facing behavioral manifestations of this construct of mental functioning and may be demonstrated through limitations in one’s ability work in teams or complete new or challenging tasks. In contrast, the Self-Regulation scale reflects more of one’s ability to regulate internal feelings or emotions including anger, frustration, or fear. These items are developed in a way where one may have difficulties interacting with others in general and may demonstrate inappropriate behaviors including outbursts, antisocial behaviors, or frequent worry that others are against them.

Similarly, the Mood & Emotions scale reflects a spectrum of items related to general anxiety and depression. This content however, is less targeted at specific external stimuli (another person) but represents a general underlying feeling along the depression-anxiety continuum that one may experience in any context. While the WD-FAB provides an innovative approach to describe a comprehensive profile of mental health function as well as physical function, this information alone does not capture the full complexity of environmental, contextual, or sociodemographic factors that also play a critical role in one’s ability to successfully engage in gainful employment.

5 Limitations

The sample collected is not designed to be representative of the adult population and certain demographic groups are overrepresented. The quotas established for this study were meant to include a large population of people not working due to a disability, as the intent of the WD-FAB was to assess the functioning of people with a work-limiting disability. Because women are substantially overrepresented in this sample (as is the case with many opt-in panel surveys), there is likely a bias in the types of occupations represented among respondents because women tend to hold different occupations than men. However, our focus is on the relationship between functioning and job demands and job demands should be gender-neutral. To that end, the demographic biases should not affect our ability to answer the research questions, but this should be considered.

A second limitation is the timing of data collection. Our survey was fielded just as stay-at-home orders were issued in 42 U.S. states due to the COVID-19 pandemic. Many of the WD-FAB items designed to assess resilience and sociability related to interactions with others. Given the national circumstances, we captured responses in this scale were uncharacteristically low and were therefore excluded from these analyses.

6 Conclusion

Functioning is an essential component of work disability but could better be understood within the context of an individual’s specific occupational demands. While standardized databases of occupations such as O*NET and the Occupational Requirement Survey report the typical requirements of specific jobs, self-reported job duties offer a unique picture of the demands each worker needs to meet. Our study characterized each job demand within the context of various types of physical and mental functioning. Doing so allowed us to identify who experienced a gap between functional ability and occupational requirement, the scales in which the gap led to an exit from the workforce, and offers considerations for potential interventions via workplace accommodations. Our findings indicate that levels of mental health functioning that are lower than the level necessary to carry out a worker’s required job duties are associated with work disability status. This offers employers, vocational rehabilitation professionals, and policy makers a motivation for considering new strategies for accommodating employee mental and physical health needs.

Funding

Funding for this study was provided by the Rehabilitation Research and Training Center on Employment Policy and Measurement at the University of New Hampshire, which is funded by the National Institute on Disability, Independent Living, and Rehabilitation Research (NIDILRR), in the Administration for Community Living (ACL), at the U.S. Department of Health and Human Services (DHHS) under grant number 9ORT5037-02-00. The contents do not necessarily represent the policy of DHHS and you should not assume endorsement by the federal government (EDGAR, 75.620 (b)).

Footnotes

Conflict of interest

None to report.

Acknowledgements

The authors wish to acknowledge the contributions of Jonathan Comacho at NIH and Melanie Cen at the University of Pittsburgh during coding, and Marisa Rafal at the University of New Hampshire during data collection.

Supplementary materials

The appendices are available from .

References

Paul

, Houtenville

Annual Disability Statistics Compendium: 2020. Durham, NH: University of New Hampshire, Institute on Disability; 2020.

Polaneczky

. These Philadelphia companies are changing lives by hiring employees with disabilities. The Philadelphia Inquirer 2019.

World Health Organization. International Classification of Functioning, Disability and Health (ICF). 2001.

Verbrugge

, Jette

. The disablement process. Social Science & Medicine. 1994;38:1–14. https://doi.org/10.1016/0277-9536(94)90294-1.

Brandt

, Houtenville

, Huynh

, Chan

, Rasch

. Connecting Contemporary Paradigms to the Social Security Administration’s Disability Evaluation Process. Journal of Disability Policy Studies. 2011;22:116–28. https://doi.org/10.1177/1044207310396509.

Stapleton

, Christian

. Helping Workers Who Develop Medical Problems Stay Employed: Expanding Washington’s COHE Program Beyond Workers’ Compensation. Washington, DC: Mathematica Policy Research; n.d.

Social Security Administration. Basic definition of disability. Code of Federal Regulations: Definition of Disability 2012. https://www.ssa.gov/OP_Home/cfr20/404/404-1505.htm (accessed August 13, 2021).

Gilworth

, Carey

, Eyres

, Sloan

, Rainford

, Bodenham

, et al. Screening for job loss: development of a work instability scale for traumatic brain injury. Brain Inj. 2006;20:835–43. https://doi.org/10.1080/02699050600832221.

U.S. Department of Labor. O*NET-SOC Taxonomy. O*NET Resource Center 2021. https://www.onetcenter.org/taxonomy.html (accessed August 23, 2021).

10.

C-Y

, Arthur

, Hong

, Velozo

. Piloting an empirical approach to link the international classification of functioning, disability, and health to job demand classification. Work. 2019;64:721–9. https://doi.org/10.3233/WOR-193034.

11.

Porcino

, Marfeo

, McDonough

, Chan

. The Work Disability Functional Assessment Battery (WD-FAB): Development and validation review. TBV – Tijdschrift Voor Bedrijfs- En Verzekeringsgeneeskunde. 2018. https://doi.org/10.1007/s12498-018-0247-0.

12.

Marfeo

, Ni

, Meterko

, Marino

, Peterik

, McDonough

, et al. Development of a New Instrument to Assess Work-Related Function: Work Disability Functional Assessment Battery (WD-FAB). Am J Occup Ther. 2016;70:7011500012p1–7011500012p1. https://doi.org/10.5014/ajot.2016.70S1-RP402B.

13.

Marfeo

, Ni

, Chan

, Rasch

, McDonough

, Brandt

, et al. Interpreting physical and behavioral health scores from new work disability instruments. Journal of Rehabilitation Medicine. 2015;47:394–402. https://doi.org/10.2340/16501977-1947.

14.

Qualtrics. ESOMAR 28: 28 Questions to Help Buyers of Online samples. Provo, UT: Qualtrics; 2019.

15.

Cohen

. Statistical Power Analysis for the Behavioral Sciences. Academic Press; 2013.

16.

Jette

, Ni

, Rasch

, Marfeo

, McDonough

, Brandt

, et al. The Work Disability Functional Assessment Battery (WD-FAB). Phys Med Rehabil Clin N Am. 2019;30:561–72. https://doi.org/10.1016/j.pmr.2019.03.004.

17.

Twenge

, Joiner

. U.S. Census Bureau-assessed prevalence of anxiety and depressive symptoms in 2019 and during the 2020 COVID-19 pandemic. Depress Anxiety. 2020. https://doi.org/10.1002/da.23077.

18.

Meterko

, Marfeo

, McDonough

, Jette

, Ni

, Bogusz

, et al. The Work Disability Functional Assessment Battery (WD-FAB): Feasibility and Psychometric Properties. Arch Phys Med Rehabil. 2015;96:1028–35. https://doi.org/10.1016/j.apmr.2014.11.025.

19.

Huang

S-W

, Chang

K-H

, Escorpizo

, Liou

T-H

. Accuracy of a modified World Health Organization Disability Assessment Schedule 2.0 as an assessment tool for predicting return-to-work among patients with traumatic brain injury. Disability and Rehabilitation. 2020;42:3370–6. https://doi.org/10.1080/09638288.2019.1594401.

20.

Rao

, Kilgore

. Predicting return to work in traumatic brain injury using assessment scales. Arch Phys Med Rehabil. 1992;73:911–6.

21.

Karpov

, Joffe

, Aaltonen

, Suvisaari

, Baryshnikov

, Näätänen

, et al. Level of functioning, perceived work ability, and work status among psychiatric patients with major mental disorders. Eur Psychiatry. 2017;44:83–9. https://doi.org/10.1016/j.eurpsy.2017.03.010.

22.

Baldwin

, Butler

, Johnson

, Cote

. Self-reported Severity Measures as Predictors of Return-to-work Outcomes in Occupational Back Pain | SpringerLink. Journal of Occupational Rehabilitation. 2007;17:683–700.

23.

Tang

, Beaton

, Boonen

, Gignac

MAM

, Bombardier

. Measures of work disability and productivity: Rheumatoid Arthritis Specific Work Productivity Survey (WPS-RA), Workplace Activity Limitations Scale (WALS), Work Instability Scale for Rheumatoid Arthritis (RA-WIS), Work Limitations Questionnaire (WLQ), and Work Productivity and Activity Impairment Questionnaire (WPAI). Arthritis Care & Research. 2011;63:S337–49. https://doi.org/10.1002/acr.20633.

24.

Harling

, Schablon

, Peters

, Nienhaus

. Predictive values and other quality criteria of the German version of the Nurse-Work Instability Scale (Nurse-WIS) - follow-up survey findings of a prospective study of a cohort of geriatric care workers. Journal of Occupational Medicine and Toxicology. 2014;9:30. https://doi.org/10.1186/s12995-014-0030-9.

25.

Sundar

. Operationalizing workplace accommodations for individuals with disabilities: A scoping review. Work. 2017;56:135–55. https://doi.org/10.3233/WOR-162472.

26.

Jetha

, Pransky

, Fish

, Hettinger

. Return-to-Work Within a Complex and Dynamic Organizational Work Disability System. J Occup Rehabil. 2016;26:276–85. https://doi.org/10.1007/s10926-015-9613-2.

27.

Jetha

, Pransky

, Hettinger

. Capturing complexity in work disability research: application of system dynamics modeling methodology. Disability and Rehabilitation. 2016;38:189–94. https://doi.org/10.3109/09638288.2015.1031291.

28.

Druss

, Marcus

, Rosenheck

, Olfson

, Tanielian

, Pincus

. Understanding disability in mental and general medical conditions. Am J Psychiatry. 2000;157:1485–91. https://doi.org/10.1176/appi.ajp.157.9.1485.

29.

Corbière

, Negrini

, Dewa

. Mental Health Problems and Mental Disorders: Linked Determinants to Work Participation and Work Functioning. In: Loisel

, Anema

, editors. Handbook of Work Disability: Prevention and Management, New York, NY: Springer; 2013, p. 267–88. https://doi.org/10.1007/978-1-4614-6214-9_17.

Linking job duties,functioning,and employment status using the Work-Disability Functional Assessment Battery (WD-FAB): An expert coding and quantitative analysis

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1 Introduction

2 Materials and methods

2.1 Participant recruitment

2.2 Data collection instrument

2.3 Description of analytic approach

2.3.1 Data cleaning

2.3.2 Applying WD-FAB functional levels and score thresholds to job duties

Table 1 Previously established functional levels, score ranges, and descriptions for the WD-FAB’s Basic Mobility scale (13)

2.3.2.a Blanket coding of similar job duties

2.3.2.b Job duty coding calibration and reconciliation

3 Results

3.1 Sample

Table 2 Demographic characteristics for 1,770 respondents compared to U.S. population characteristics based on a weighted national sample from the National Health Interview Survey (NHIS) (column % unless otherwise noted)

Table 3 WD-FAB scale scores by employment status for 1,770 respondents.

Table 4 Percentage meeting occupational requirements of all three of their specific job duties for each WD-FAB scale and across all six scales, by employment status (occupational requirements were for last job among work disability sample)

Table 5 Distribution of individual functioning profiles by occupational demand profile, work disability sample

Table 8 Logistic regression predicting employed status for 1,770 respondents

5 Limitations

6 Conclusion

Funding

Footnotes

Conflict of interest

Acknowledgements

Supplementary materials

References

Table 1
Previously established functional levels, score ranges, and descriptions for the WD-FAB’s Basic Mobility scale (13)

Table 2
Demographic characteristics for 1,770 respondents compared to U.S. population characteristics based on a weighted national sample from the National Health Interview Survey (NHIS) (column % unless otherwise noted)

Table 3
WD-FAB scale scores by employment status for 1,770 respondents.

Table 4
Percentage meeting occupational requirements of all three of their specific job duties for each WD-FAB scale and across all six scales, by employment status (occupational requirements were for last job among work disability sample)

Table 5
Distribution of individual functioning profiles by occupational demand profile, work disability sample

Table 8
Logistic regression predicting employed status for 1,770 respondents