Measuring Four Personal Opportunities for Adults With Intellectual and Developmental Disabilities

Abstract

This study tests an empirically derived model for measuring personal opportunities for people with intellectual and developmental disabilities (IDD) using National Core Indicators In-Person Survey (NCI-IPS) state and national datasets. The four personal opportunities measured, (a) privacy rights, (b) everyday choice, (c) community participation, and (d) expanded friendships, were informed by existing conceptualizations of service as well as NCI-IPS measures. Analyses confirmed the fit of a four-factor model and demonstrated that factors were significantly and positively correlated. To demonstrate the relationships between personal opportunities and personal and environmental characteristics, we estimated a structural equation model that regressed personal opportunities on age, gender, place of residence, and level of intellectual disability. Implications for using personal opportunities for evaluating service quality of IDD systems are discussed.

The Centers for Medicare and Medicaid Services (CMS) partners with states to assure and improve the quality of Medicaid Home and Community Based Services (HCBS) for adults with intellectual and developmental disabilities (IDD). The primary service goals are to integrate people with IDD into their communities, maximizing the quality of life, independence, health and well-being of people who use HCBS (CMS, 2020) Depending on their needs and the service array received, people can experience a range of positive outcomes related to health, safety, relationship building, and community participation.

Definitions of quality services often vary throughout the United States and consequently the ways quality is monitored remain relatively diverse. The National Quality Forum has guided ongoing processes to improve clarity about the conceptualization of quality services (2015; 2016). Quality services enhance human rights such as privacy, dignity and respect while providing opportunities for choice, social connectedness, and community participation (Friedman, 2018a; National Quality Forum, 2015). Quality services are shaped by and fulfill the needs and preferences of the service users.

People with IDD experience persistent barriers in their ability to exercise personal rights, make everyday choices, engage in community activities, and develop meaningful relationships. These barriers have been addressed at a systems level via litigation, policy, and regulation that promote independence, full participation, inclusion, and equal opportunity for people with IDD (Hewitt et al., 2016; Rich et al., 2002). A recent example of this can be seen in the Medicaid Home and Community-Based Services (HCBS) Final Rule (CMS 2249-F/2296-F). This regulation is specifically focused on the development and implementation of “innovative strategies to increase opportunities [emphasis added] for Americans with disabilities and older adults to enjoy meaningful community living” (CMS, 2014, para. 1). In other words, the Final Rule mandates that states work to enhance HCBS service quality.

Opportunities

People are more aware of their capabilities when given opportunities to assess and enact their potential. The administration of high-quality services and supports should accentuate opportunities whereas lower quality supports can extinguish opportunities (Bradshaw et al., 2018). Health and safety are often provided as a rationale for limiting opportunity, but only through opportunities can people with IDD prove to themselves and others that they can remain safe when making choices regarding their privacy, activities, community participation or relationships (Petner-Arrey et al., 2015). Reinders and Schalock (2014) proposed that development is realized through access to opportunities that are built on personal strengths and permeated with human and legal rights.

Opportunity expands as people with IDD are provided with a wider range of options to fully participate in decision making and community life. For example, young people with IDD participating in paid work experiences during high school are more likely than their peers that have no or unpaid work training to attain paid, integrated employment in early adulthood (Carter et al., 2012; Simonsen & Neubert, 2012). Likewise, when people with IDD are given opportunities to make more choices they develop greater self-determination (Wehmeyer & Garner, 2003; Williams-Diehm et al., 2008). More commonly though, people with IDD have limited opportunities to make their own choices, to participate in desired activities, or to fully engage in their communities in ways that are typical for people without IDD, making opportunity, in its own right, an important outcome for people with IDD (Bradley & Moseley, 2007; Shogren et al., 2009).

Measuring Quality and Opportunity

Ideally, measures of service quality should capture both the nature of the services provided and the opportunities that the services produce. Many existing measures, such as the National Core Indicators In Person Survey (NCI-IPS; Human Service Research Institute [HSRI] & National Association of State Directors of Developmental Disabilities Services [NASDDDS], n.d.), the Personal Outcome Measures (Friedman, 2018b), and the Money Follows the Person Quality of Life Survey (Sloan & Irvin, 2007), address core domains and subdomains of quality and opportunity including privacy rights, choice, community participation and social relationships. Therefore, efforts to improve the measurement of quality services should use existing measures to construct standardized scales (National Quality Forum, 2016).

The Center for Medicaid and Children’s Health Insurance Program (CHIP) Services recognized the NCI-IPS as a core measure of quality for adults enrolled in Medicaid services, including HCBS (NASDDDS, 2019). The present study used national NCI-IPS data from 2017–2018 to test measurement models for four empirically generated domains of personal opportunities: (a) privacy rights, (b) everyday choice, (c) community participation, and (d) expanded friendships. These factors were chosen because they represent five subdomains of quality: (a) privacy, (b) personal freedoms and dignity of risk, (c) personal choice and goals, (d) social connectedness and relationships, and (e) resources and settings to facilitate inclusion. These subdomains are among 40 deemed as “the most important areas for quality measurement in HCBS” (National Quality Forum, 2016, p. 10). Also, the factors build upon previously operationalized NCI-IPS measures, including Everyday Choice (Lakin et al., 2008), Social Participation and Relationships (Mehling & Tasse 2014a; 2014b), and Home Privacy (Houseworth et al., 2019).

This study expands on previous work by describing the relationship between these factors and providing examples of personal and service characteristics that predict opportunities. The analytical steps that were employed provide a simple heuristic for developing and using quality measures from the NCI-IPS to support quality monitoring and improvement.

This study addresses the following three research questions:

1. Does a state-derived model of personal opportunities fit data from a national sample?

2. Are personal opportunities correlated and how?

3. In what way do basic contextual variables predict personal opportunities?

Method

All methods described in the following section were reviewed and approved by the Virginia Commonwealth University Institutional Review Board.

Data Source

The National Core Indicators (NCI) are a collection of nationally validated surveys of people with IDD, their family members, and support workers that was developed by NASDDDS and HSRI. This study used the NCI-IPS, formerly known as the Adult Consumer Survey, which was administered in 35 states and the District of Columbia in 2017/2018.

The NCI-IPS is administered in an interview format with people with IDD aged 18 and older who use at least one public IDD service in addition to case management. The survey has multiple sections. A background section collects demographic and basic health status information and is often completed by a case manager or other individual with access to administrative file information. The in-person portion of the NCI-IPS is administered directly with the person with IDD by a trained interviewer. The first part of the in-person section of the NCI-IPS must be answered by the person with IDD directly, but the second part permits proxy response, if needed.

Procedure

The development of the personal opportunities measures began by examining NCI-IPS data from Virginia. Items from previously developed NCI-IPS factors were used to build measurement models. Included were various measures of (a) choice: Everyday Choice (Lakin et al., 2008), Choice (Neely-Barnes et al., 2008), and Social Determination (Mehling & Tasse, 2014b); (b) rights: Rights (Neely-Barnes et al., 2008), Home Privacy (Houseworth et al., 2019); and (c) community and social inclusion: Social Participation and Relationships (Mehling & Tasse, 2014b), Relationships, Community Inclusion (Neely-Barnes et al., 2008). Confirmatory factor analysis was used to test item loadings and the fit of models to state NCI-IPS data. The process and results used for arriving at the four-factor model of personal opportunities are outlined in the supplementary information (https://osf.io/5jk7h/wiki/home/?view_only=b821d9a3f906455f88f2036879f2612e). Permission was sought and granted from HSRI and NASDDDS to determine the fit of the four-factor model to the 2017/2018 national NCI-IPS data.

Variables

Privacy Rights

Two variables were used to construct the privacy rights factor, both of which came from Part II of the IPS, a section completed by the person with IDD or by a proxy respondent (i.e., friend/family, staff, other). The two variables in this grouping asked (a) whether a person with IDD had a key to their home; and (b) whether the person could lock their bedroom door when desired. Both had three response options (yes; maybe; no). We dichotomized responses to 1 = no and 2 = yes and maybe. Uneven group sizes across response categories drove the choice to dichotomize responses. Dichotomization was used in order to facilitate more robust analyses while maintaining the overall meaning of the variables.

Everyday Choice

The daily choices variables all come from Section II and were answered either by a person with IDD or a proxy respondent. Three variables exist in this grouping: (a) whether the person with IDD reported having choice over their daily schedule, (b) whether the person can choose how to spend free time, and (c) whether the person can choose what to buy. These are the same items used previously to measure short-term decision making (Bush & Tasse, 2017) and everyday choice (e.g., Lakin et al., 2008). Each item relied on three primary response options which we rescaled as 1 = Someone else chose, 2 = Person had some input, and 3 = Person made the choice. The some input option for choosing what to buy with one’s spending money varied slightly and was person has help choosing what to buy, or has set limits (such as can buy small items, but not big items). Responses were dichotomized under the same rationale that was applied to Privacy Rights (1 = someone else chose and 2 = made choice or had input in the choice).

Community Participation

Many of the variables in the community participation grouping have been included in previously formed measures such as “community inclusion” (e.g., Neely-Barnes et al., 2008) and “social participation and relationships” (Mehling & Tasse, 2014b). Measured in this way, community participation is akin to physical integration and is not the same as terms such as community inclusion, social inclusion, or social participation, though they are sometimes used interchangeably.

The four variables in this section all come from Part II of the IPS. The four variables were how often: (a) did the person with IDD go out shopping in the past month, (b) did the person with IDD go out for errands in the past month, (c) did the person with IDD go out for entertainment in the past month, and (d) did the person go out to eat in the past month. All responses related to frequency of participation in a community activity in the past month and ranged from 1 (0 times) to 4 (More than 5 times).

Expanded Friendships

Friendship items comprised the final group in the Personal Opportunities model. In addition to frequency of participation in community activities, the IPS (Part II) included items to capture with whom respondents usually participated in community activities in the month prior. Response options included alone, friends, family, housemates or co-workers, staff or others. We used the dichotomous response (1 = no; 2 = yes) of “friends” for seven community activities. The activities considered were (a) shopping, (b) errands or appointments, (c) going to community groups, (d) going out for entertainment, (e) going to restaurants or coffee shops, (f) attending religious services or spiritual practices, and (g) going on vacation.

This factor also included a friendship item used previously to determine if friendship of NCI respondents were “limited” or “expanded.” (Giordano et al., 2016). This item asked “Do you have friends you like to talk to or do things with?” and we recoded the response options to be 1 = No, does not have friends, 2 = Yes, all friends are staff or family, or 3 = Yes, has friends who are not staff or family. A response of 1 indicated limited friendships and a response of 2, along with the other items in the factor, when affirmative, were indicative of expanded friendships.

Personal Characteristics and Residential Context

Gender, age, level of intellectual disability and type of residence were collected using the background information section of the NCI-IPS which was completed by referring to existing records. Age was calculated from date of birth. Gender response options were male, female, or other. We recoded this variable so female was 1 and male was 0. The “other” category had only three responses nationally and were not included in analyses.

Level of intellectual disability was a follow-up question to an item which asked whether the person had been diagnosed with ID (1 = No; 2 = Yes; 99 = Don’t know). When initial responses were affirmative then the level of ID noted in personal records were recorded as either (1) mild, (2) moderate, (3) severe or (4) profound. Additional responses included “no ID diagnosis,” “unspecified level of ID” or “ID diagnosis unknown.”

The NCI-IPS included 15 residence types, including “other” and “unknown.” Our data did not include people living in institutions such as intermediate care facilities. We combined host homes for one individual with a disability with host homes for two or more individuals with disabilities to create a single host home category. Final categories included 2–3 bed group homes, 4–6 bed group homes, 7–15 bed group homes, living with family, living independently, and host homes.

Sample

As highlighted earlier, the development of the Personal Opportunities Scale began by examining NCI-IPS data from Virginia. To validate the scale with a larger, more diverse sample we used the national 2017/2018 NCI-IPS dataset, which contained a total of 25,671 participants from 35 states and the District of Columbia. Many states contributed approximately 400 IPS responses, because this was the minimum number recommended in guidance from NASDDDS and HSRI.

Nationally, participants in the 2017/2018 NCI-IPS dataset had a mean age of 42 years and 59% were male. The sample’s racial composition was 67% White, 16% Black, and 10% Latinx, with other racial identities present in smaller numbers. The majority of participants in the national sample ranged in age from 26 to 58 years. The youngest participants were 18 years old and the oldest was 95 years old. Additional information about the NCI-IPS sample, both nationally and in Virginia is available in the 2017/2018 IPS Report (HSRI and NASDDDS, 2019).

Analysis

Data analysis was carried out in three phases, corresponding to the three primary research aims of the study. First, confirmatory factor analysis (CFA) was used to test the relative fit of a series of factor models for personal opportunities. For each CFA, we tested a tau equivalent model, where item loadings were fixed to be equal, and a congeneric model, where all item loadings were allowed to freely vary except the first item, which was fixed at one to scale the latent variable. This approach, recommended by Brown (2015), allowed us to test the relative fit of models with increasing complexity, both in terms of the dimensionality of the latent construct (one vs. four factors) and in terms of the weighting of each item.

Data analysis was conducted using Mplus Version 7.11 (Muthén & Muthén, 1998–2017) with the mean- and variance-adjusted weighted least squares estimator (WLSMV) for all CFA SEM models. To explore missing data patterns, Little’s MCAR test (Little, 1988) was used as an omnibus test to determine whether the data could be assumed missing completely at random (MCAR). The results of this test suggested that the data could not be treated as MCAR, thus requiring estimation procedures that are more robust to missing data.

To address potential bias due to missing data, we compared our preferred estimates as discussed previously with model results estimated via full information maximum likelihood (FIML) using the MLR estimator in Mplus. Overall, rates of missingness were quite low for most survey questions, with more than 90% of individuals recording a response for community participation, everyday choice, and privacy rights. Missingness was higher for “Do you have friends you like to talk to or do things with?” which necessitated that we compare our preferred results with those generated via FIML in order to ensure that potential bias due to missing data had not occurred. Estimates did not differ in terms of magnitude or significance and are available from the authors upon request.

Data Assumptions and Model Fit/Parsimony

We used the mean- and variance-adjusted weighted least squares estimator (WLSMV in MPlus) and Satorra and Bentler’s (2001) scaled chi-square (S-B χ 2) to examine model fit for all CFA and SEM models. In addition to the S-B χ 2, we also used the comparative fit index (CFI), standardized root-mean-square residual (SRMR), and root-mean-square error of approximation (RMSEA) with a 90% confidence interval to assess model fit. We adopted Hu and Bentler’s (1999) guidelines for assessing acceptable model fit, which include CFI ≥ .95, SRMR ≤ .08, and RMSEA ≤.06. In cases where the CFI value was below the suggested threshold, we then used Hu and Bentler’s (1999) recommendations to simultaneously evaluate SRMR and RMSEA, with RMSEA ≤ .06 and SRMR ≤ .10 indicating an acceptable model fit.

Latent Correlations

To estimate the latent correlations between our personal opportunity factors, we used the TECH4 procedure in MPlus (Muthén & Muthén, 1998–2017). This is a diagnostic routine that provides a covariance matrix, correlation matrix, and corresponding p values for all latent variables specified in a given model.

Impact of Personal Characteristics and Residential Contexts

To determine the relationship among personal characteristics, residential contexts, and the personal opportunity scales, we estimated a structural equation model that included factors for all four personal opportunities and regressed those factors on to a set of predictors including the demographic and personal characteristics described previously. This is similar to running a multivariate regression model with four simultaneous outcomes, in that we were able to account for each individual’s correlation between each of the four measures of personal opportunity. Further, by fitting this model in an SEM framework, we were able to account for measurement error at the item and construct levels while obtaining our regression estimates.

Results

Correlations

Table 1 presents item-level polychoric correlations among all items used in this study. Polychoric correlation is used to determine the strength of association between two ordinal variables with two or more categories (Jöreskog 1994; Olsson 1979). This approach is especially preferred when conducting CFA with ordinal variables composed of less than five categories (Rhemtulla et al. 2012), which is often the case when working with NCI data. A number of individual correlations occurred in the .30–.50 range. No items demonstrated consistently weak correlations below .10, and no items demonstrated exceedingly large correlations above .80 that may hinder the detection of distinct and interpretable factors.

Table 1.

Polychoric Correlations Among Variables of Interest

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16
1. Key	--
2. Lock	0.48	--
3. Shopping	0.20	0.15	--
4. Entertainment	0.06	0.12	0.36	--
5. Eat Out	0.14	0.16	0.50	0.45	--
6. Errands	0.17	0.14	0.46	0.32	0.36	--
7. Schedule	0.54	0.37	0.14	<.01	0.12	0.12	--
8. Free Time	0.46	0.37	0.15	0.05	0.16	0.11	0.79	--
9. Choose Buy	0.44	0.34	0.15	0.04	0.14	0.15	0.60	0.67	--
10. Has Friends	0.14	0.10	0.12	0.17	0.14	0.08	0.11	0.17	0.14	--
11. Shop w/ Friends	0.20	0.09	0.31	0.13	0.17	0.09	0.20	0.13	0.16	0.22	--
12. Errands w/ Friends	0.19	0.07	0.20	0.10	0.12	0.26	0.18	0.12	0.13	0.14	0.85	--
13. Comm. Grp. w/ Friends	0.23	0.07	0.22	0.22	0.22	0.18	0.16	0.17	0.15	0.30	0.54	0.50	--
14. Ent. w/ Friends	0.33	0.14	0.22	0.42	0.21	0.12	0.28	0.24	0.22	0.36	0.77	0.70	0.67	--
15. Eat w/ Friends	0.32	0.14	0.22	0.19	0.33	0.13	0.29	0.25	0.22	0.34	0.79	0.72	0.61	0.88	--
16. Rel. Ser. w/ Friends	0.20	0.10	0.15	0.09	0.13	0.10	0.17	0.14	0.14	0.18	0.60	0.61	0.67	0.62	0.66	--
17. Vacation w/ Friends	0.19	0.06	0.18	0.17	0.15	0.13	0.14	0.08	0.11	0.16	0.57	0.56	0.47	0.6	0.58	0.56

Notes. N = 25,670

Investigating Alternative Factor Structures

Having examined the polychoric correlation matrix and finding it suitable for factor analysis, we then used CFA to investigate the underlying factor structure of the selected items. The initial working hypothesis was that the items would best be represented by four latent factors, privacy rights, everyday choice, community participation, and expanded friendships. To confirm this, we began by testing a model that included all factors in one latent factor (a plausible rival hypothesis), then tested the four-factor model. For each class of model, we compared a tau-equivalent model with a congeneric model in order to best represent the underlying item loading patterns.

Four separate CFA models were tested. Model fit and parsimony statistics are available in Table 2. Model 4, the four-factor model with congeneric item loadings demonstrated the best model fit and parsimony according to the guidelines suggested by Hu and Bentler (1999). In contrast, the one-factor model with tau-equivalent items demonstrated the worst fit and parsimony. Chi-square difference tests were significant for the comparisons between Models 1 and 2, 2 and 3, and 3 and 4, respectively (all ps < .001). Therefore, the four-factor model with congeneric items appears to be a more accurate representation of the underlying constructs. Figure 1 shows a path diagram representing Model 4, the best-fitting model that was tested.

Table 2.

Model Fit and Parsimony Statistics for Competing Models of Everyday Choice, Community Participation, Privacy Rights, and Expanded Friendships

Model No.	Model Description	S-B χ2	df	p	CFI	TLI	RMSEA	90% LB	90% UB	WRMR
			13	<	0.6	0.6
1	1 Factor Model w/ Tau Equivalent Items	54354.28	5	.001	3	3	0.120	0.120	0.130	20.80
			11	<	0.7	0.6
2	1 Factor Model w/ Congeneric Items	39685.24	9	.001	3	9	0.110	0.110	0.120	15.92
			12	<	0.9	0.9
3	4 Factor Model w/ Tau Equivalent Items	10998.01	6	.001	3	3	0.058	0.057	0.059	8.24
				<	0.9	0.9
4	4 Factor Model w/ Congeneric Items	3724.85	54	.001	8	7	0.035	0.034	0.036	4.43

Notes. N = 25,670 for all models

Figure 1. Personal Opportunity Confirmatory Factor Analysis for National Data.

Note. Items are numbered in the order they are presented in Table 2. Friends= expanded friendships; choice= everyday choice; compar= community participation; rights= privacy rights. All indicated correlations and path coefficients are significant (p < .05). Estimates are presented as standardized coefficients. S-B χ² (113) = 3724.85, p < .001; Root-mean-square error of approximation (RMSEA) = .035, 90% CI [.034, .036]; Comparative fit index (CFI) = .98; Tucker Lewis Index (TLI) = .97; Weighted root mean square residual (WRMR) = 4.43. N = 25,549.

Factor Correlations and Regressions

Factor Correlations

Correlations between the four personal opportunities are shown in Table 3. All of the factors were significantly and positively correlated as expected for factors conceptually related. The relationships between latent factors varied in strength. Some correlations were weak (everyday choice with community participation; choice with expanded friendship), some were moderate (privacy rights with community participation; privacy rights with expanded friendship; community participation with expanded friendships) and one was strong (privacy rights with everyday choice).

Table 3.

Correlations Between Latent Measures of Personal Opportunity and Wellness, National Data

	1	2	3	4
1. Privacy Rights	--
2. Community Participation	0.302***	--
3. Everyday Choice	0.698***	0.212***	--
4. Expanded Friendships	0.331***	0.372***	0.285***	--

Notes. N = 25,670

^***p <.001.

Personal Characteristics and Residential Contexts

Basic personal characteristics (age, gender, level of impairment) and place of residence were included as predictors of personal opportunities in four multiple regression models. Complete results are displayed in Table 4.

Table 4.

Estimates for Personal Opportunity Regression Models

	Privacy Rights Model				Community Participation Model				Everyday Choice Model				Expanded Friendships Model
Variable	B	SE	z	p	B	SE	z	p	B	SE	z	p	B	SE	z	p
Group 4–6	−0.400	0.032	−12.524	<.001	−0.023	0.023	−1.021	0.307	−0.069	0.026	−2.713	0.007	0.070	0.014	5.164	<.001
Group 7–15	−0.295	0.055	−5.349	<.001	−0.057	0.038	−1.479	0.139	0.150	0.043	3.488	<.001	0.097	0.022	4.329	<.001
Live w/ Fam.	0.087	0.029	2.947	0.003	0.048	0.021	2.254	0.024	−0.090	0.024	−3.701	<.001	0.151	0.013	11.188	<.001
Live Ind.	0.826	0.035	23.446	<.001	0.082	0.023	3.567	<.001	0.439	0.028	15.596	<.001	0.186	0.015	12.529	<.001
Live w/ Host	0.126	0.042	2.966	0.003	0.316	0.031	10.21	<.001	0.100	0.036	2.753	0.006	0.109	0.017	6.382	<.001
Age	0.028	0.009	3.015	0.003	−0.100	0.009	−11.164	<.001	0.056	0.008	7.034	<.001	−0.063	0.011	−5.684	<.001
Female	−0.050	0.018	−2.818	0.005	0.035	0.016	2.11	0.035	−0.006	0.015	−0.419	0.676	0.076	0.019	3.98	<.001
Moderate ID	−0.344	0.020	−17.088	<.001	0.032	0.014	2.27	0.023	−0.452	0.016	−27.658	<.001	−0.099	0.009	−11.369	<.001
Severe ID	−0.872	0.028	−31.293	<.001	−0.148	0.018	−8.035	<.001	−1.047	0.022	−48.639	<.001	−0.191	0.014	−13.209	<.001
Profound ID	−1.079	0.035	−30.775	<.001	−0.431	0.022	−19.336	<.001	−1.359	0.026	−52.038	<.001	−0.277	0.020	−13.635	<.001

Notes. N = 20,314.

Group 2–3 is reference category for residence. Male is the reference category for gender. Mild ID is the reference category for Level of ID. All coefficients are standardized and interpreted in terms of standard deviation units

Gender.

Participants’ gender predicted personal opportunity factors, except for everyday choice. Gender had a small but significant contribution to models for privacy rights, community participation, and friendship. Women had fewer privacy rights than men but had more community participation and friendship.

Age.

In the regression models, age significantly predicted all personal opportunities. Older participants reported more rights and everyday choice, factors which tended to vary together. However, older adults were also more isolated than their younger counterparts; younger participants reported more friendships and community participation.

Level of Impairment.

Disability status was the strongest predictor of everyday choice and privacy rights, which are two closely related personal opportunities. Between each level of IDD–from mild to profound–everyday choice decreased pointedly. Decreases in opportunity for everyday choice and privacy rights were most pronounced between mild and moderate IDD and moderate and severe IDD, with smaller decreases between severe and profound IDD. For community participation, differences between mild and moderate IDD were significant. The differences in community participation for people with mild IDD compared with severe and profound IDD were much larger. The relationship between level of IDD and friendship differed from the other variables. Those with mild IDD had more expansive friendships, while those with moderate, severe, and profound IDD had friendships that were similarly constricted.

Type of Residence.

Residence type broadly predicted personal opportunities. Those living independently showed the strongest outcomes for privacy rights, everyday choice, and friendships, though friendships were somewhat similar among people who lived with their families. People living in host homes had the most community participation. There were no statistical differences in community participation when comparing the different group home sizes (e.g., 2–3 bed, 4–6 bed, or 7–15 bed). People living with their families or living independently in their own home or apartment participated in the community at greater frequencies than people living in group homes of any size, although not as frequently as those living in host homes. Those with the least everyday choice lived with their families or in 4–6 bed group homes and those with the fewest privacy rights also lived in 4–6 bed group homes. People living in larger group homes had more expansive friendships than those in 2–3 bed group homes.

Discussion

This research responded to a national call to configure components of existing measures to monitor fundamental characteristics of quality HCBS (National Quality Forum, 2016). We used national NCI data to evaluate the factorial validity of a state-constructed model of personal opportunities consisting of four latent factors: privacy rights, everyday choice, community participation, and expanded friendship. We confirmed that the state-level structure of personal opportunities demonstrated good fit with national NCI-IPS data (RQ1) and the model demonstrated better fit than a single-factor model. These results add to ongoing efforts at quality measure development, and can be used to better understand, evaluate, and strengthen personal opportunities for HCBS users with IDD.

Latent Factor Correlations

The study also examined relationships between four personal opportunities. Pairwise correlations between latent factors, each representing a personal opportunity, were all positive and significant (RQ2). These results support the possibility that personal opportunities share common contextual origins, even when opportunities represent seemingly distinct quality domains. These contexts can range from a person’s adaptive behaviors to the training that direct support professionals receive to broaden rather than constrain opportunities (Qian et al., 2015).

For example, privacy rights and everyday choice were the two most strongly correlated opportunities. Similar findings have been reported elsewhere (e.g., Jones et al., 2018). Although there are many possible reasons why choice and privacy rights are highly correlated, one potential explanation can be attributed to the culture of one’s home environment. In the context of IDD services and supports, Bigby and Beadle-Brown (2016) describe culture simply as the way things are done within the home environment. Both the privacy rights and choice factors are concerned with issues of access, such as whether people are given options to choose their daily schedule or to lock their bedroom door. When high-quality residential services promote a culture of respect, autonomy, or empowerment, then greater opportunities for everyday choice and privacy rights may co-develop in parallel. Conversely, more restrictive living environments may limit access to everyday choice and privacy rights.

A significant relationship between community participation and everyday choice is supported by existing literature (Heller et al., 2002; Verdonschot et al., 2009). However, in our analysis, this relationship, although statistically significant, was fairly weak. Differences in individual preferences may account for the relatively weak relationship between community participation and choice (Carter et al., 2013). Although explanations for this are not entirely clear, it is possible that some people may exert choice by opting not to engage in community activity. Another possible explanation is that community participation often requires additional resources (i.e., transportation, spending money, supporter’s time) necessary to engage in activities such as shopping or going out to eat (Simplican et al., 2015).

External factors may similarly account for the relatively weak, positive relationship between expanded friendship and everyday choice. Supporters, including paid staff and family, often play a pivotal role in helping people with IDD to foster and maintain friendships (Clifford et al., 2016; Friedman & Rizzolo, 2018). As seen with community participation, additional support requires increased external investment. Although quality HCBS should advance social inclusion through supporting opportunities for relationships to develop between people, friendships often survive or fail by the mutual efforts of two parties. Investments of time, interest, respect, and support are needed to build and maintain friendship outcomes.

Contexts and Personal Opportunities

The degree to which contextual factors impact personal opportunities was tested empirically. We modeled how systems can use latent scores from personal opportunities measures to explore differences in opportunities among HCBS users. This is important because quality services should improve rights and opportunities. However, we were able to show that service contexts (i.e., type of residence) and personal characteristics (e.g., age, gender, and level of IDD) were predictive of greater opportunities (RQ3).

Type of residence predicted most outcomes, but the variable’s relationship to privacy rights showed particularly discernable patterns. Smaller residences that best approximated home environments were the best predictors of more privacy rights. People who lived independently had the most control of their residential privacy and people living with their families or in host homes also had more privacy rights than people living in group homes. Privacy among people with IDD living in group homes may be influenced by a combination of personal characteristics and contextual elements, such as changes in support persons, which may be more stable for people living in host homes or with their families (Qian et al., 2019).

Many of these patterns held across personal opportunities. Findings echoed Friedman’s (2017; 2019) studies, which found that people living in group homes had less community, friendships, and social capital than those living independently or with families. People living in group homes of all sizes had lower levels of community participation and friendships compared to those living independently or with parents. People living independently had more personal opportunities than those in group homes, and like previous studies (Lakin et al, 2008; Ticha et al, 2012), we found everyday choice was greatest for those living independently. However, there was not a direct linear relationship between size of residence and everyday choice. We found that people living in 7–15 bed group homes reported more everyday choice than people who lived in smaller group homes and people living with family.

Overall findings indicated that level of IDD is a powerful predictor of personal opportunities, which is in line with extant literature (Houseworth et al., 2018; Lakin et al., 2008; Nord et al., 2018; Ticha et al., 2012). With each increase in level (e.g., mild compared to moderate IDD), indicators of personal opportunity typically decreased. This inequitable access to opportunities between people with different levels of IDD demands that HCBS and other service systems review their procedures and make direct efforts to minimize disparities. Opportunity measures provide systems with a mechanism to further identify and address disparities based on a range of characteristics such as race, urbanicity, co-occurring behavior health conditions, to assure increased equity among people who use supports and services to participate in their communities.

Limitations and Future Research

Our recommended measures for assessing personal opportunities capture many important outcomes, but our model of personal opportunities is not an exhaustive encapsulation of all forms of opportunity related to HCBS. The four factors that were the focus of this study–privacy rights, everyday choice, community participation, and expanded friendships–represent just a modest proportion of HCBS quality areas that have been identified as critical to measure (National Quality Forum, 2016). Future work ought to address and include meaningful activities (e.g., employment), self-direction, and long-term choices related to supports and services.

Additionally, systems intending to use the measures of personal opportunities presented for planning and evaluation are advised to examine whether the measure maintains acceptable fit with their state’s NCI-IPS data. Nationally validated measures that have appeared in the literature were evaluated as part of this research team’s state-based analyses (e.g., Mehling & Tasse, 2014b). Most of these measures showed poor fit to our state’s data over multiple years, driving our development and validation of measures that fit our state’s data and national data. The effects of proxy responses in Section II of the NCI-IPS should also be further explored because it is possible that the factor structure of certain personal opportunities could vary depending on respondent.

Like previous studies, level of impairment was often related to personal outcomes. People with mild IDD always fared better than people with greater support needs. Dinora and colleagues (2020) recently provided a more thorough examination of relationships between outcomes and support needs by merging NCI-IPS data with more sensitive measures of support needs, such as the Supports Intensity Scale (SIS; Thompson et al., 2004). By merging NCI data with SIS data researchers can support their systems in identifying adaptive behaviors, rather than level of IDD, that best predict improved personal outcomes.

Implications for Inclusion

The results of this work extend a growing body of research that uses NCI-IPS datasets to measure important constructs related to service quality (see Houseworth et al.2019; Lakin et al., 2008; Mehling & Tasse, 2014a; Neely-Barnes et al., 2008), but stakeholders need access to more reliable and valid measures to evaluate IDD service system performance (Tichá et al., 2013). Policymakers are also in need of these data to inform decisions regarding the considerable public investment in HCBS. As stated by Owen and colleagues (2015), data must form the basis for policy decisions in long-term services and supports so that limited resources can be used to achieve the best outcomes for people with IDD. Our measure of personal opportunities is intended to provide reliable, sound measurement about a fundamental expectation of IDD service systems that can be used in conjunction with other compliance and personal outcome measures to inform programmatic and policy decision making.

Because opportunities influence a variety of personal outcomes for people with IDD (see Carter et al., 2015; Wehmeyer, 1997), this measure may have particular utility as a macro-level indicator in the context of evaluating state-based policy or service system change over time. The HCBS final settings rule and other recent regulations have a particular mandate to increase opportunity for people with IDD to be full, contributing members of their communities. Measures of personal opportunities can also provide high-level indicators of progress towards primary objectives of U.S. policy, such as the Olmstead decision (1999) and the Integration Mandate of Title II of the Americans with Disabilities Act (1990). The enforcement of these regulations assures that people with IDD can live, work, and fully participate in their communities, much like people without disabilities, and measures such as the ones presented may be used to benchmark progress made when people are given a chance.