People with intellectual disability (ID) engage in insufficient physical activity which negatively affects their health. In accordance with the present conception of ID and the supports paradigm, the current study aimed to develop and psychometrically assess an instrument examining the support needs and strategies regarding physical activity by using individuals with ID (n = 529), service providers (n = 522), and family members (n = 462) as information sources. The analysis revealed adequate reliability for the proposed instrument, with α values between .70 and .80, and adequate construct validity for the versions of the scale for the 3 information sources, particularly for service providers. The assessment information can be included in Individualized Support Plans and could be used to design and implement the strategies for facilitating a person's physical activity in their community.
People with intellectual disability (ID) have a high risk of having cardiovascular diseases (Draheim, 2006) and health problems such as obesity (Rimmer, Braddock, & Fujiura, 1993; Rimmer & Yamaki, 2006), which are related to insufficient physical activity (PA). Compared with individuals in the general population, people with ID engage in lower levels of PA, which negatively affects their health and contributes to a higher risk of health problems (Fernhall & Pitetti, 2001; Finlayson et al., 2009; Pitetti, Yarmer, & Fernhall, 2001). The literature consistently identifies the benefits of PA: (a) a reduction of anxiety levels (Carmeli, Barak, Morad, & Kodesh, 2009; Carraro & Gobbi, 2012), (b) good physical appearance and the establishment of social relationships (Frey, Buchanan, & Rosser Sandt, 2005), and (c) a decrease in maladaptive behaviors and improvements in the perception of well-being and functional skills (Carmeli, Zinger-Vaknin, Morad, & Merrick, 2005; Heller, McCubbin, Drum, & Peterson, 2011).
Persons with ID should have the opportunity to engage in the recommended minimum level of PA, defined as five or more 30-min sessions of moderate PA per week according to the World Health Organization (2009). However, despite all these recommendations, previous studies have shown that participation of this group in PA options is insufficient (Draheim, Williams, & McCubbin, 2002; Frey et al., 2005).
The literature identifies common factors that limit PA: transportation difficulties, economic cost, and lack of personalized support, choices, or availability of community PA programs (Frey et al., 2005; Howie et al., 2012; Lante, Walkley, Gamble, & Vassos, 2011; Mahy, Shields, Taylor, & Dodd, 2010; Messent, Cooke, & Long, 1999; Temple & Walkley, 2007; van Schijndel-Speet, Evenhuis, van Wijck, van Empelen, & Echteld, 2014).
Thus, PA should be approached from the viewpoint of current conceptions of ID. What was once called “mental retardation” and considered an absolute trait of the person is now conceived as an interaction between the person and their environment (Luckasson et al., 1992). This is an ecological view that conveys the relevance of environment and the role of supports. Supports function as a bridge the person needs to reduce the gap between abilities and specific environmental demands. This is the approach that has guided development of the definition manuals of the American Association on Intellectual and Developmental Disabilities (AAIDD; Luckasson et al., 2002; Schalock et al., 2010). Based on this conception of ID, the Supports Intensity Scale was developed (SIS; Thompson et al., 2004) to evaluate the support needs of individuals across different life environments (home living, community living, lifelong learning activities, employment, health and safety, social activities, protection and advocacy).
Based on the impact of the SIS (translated and adapted into 13 languages and used in 17 different countries) our goal was to adapt a scale with the same theoretical foundations, but focusing exclusively on the domain of PA. Such a scale would be able to identify the main support needs and provide the necessary strategies to facilitate access and participation on PA opportunities in the community.
Within the SIS (Thompson et al., 2004), the evaluation of supports is made by individuals with ID and the people who know them well. Thus, the scale developed in this study includes three versions to facilitate assessment: the person with ID and proxy respondents that are staff or family. The inclusion of these three perspectives allowed us to examine agreement across perspectives and to identify the most important aspects from each.
The main goal of this research is to develop and examine the psychometric properties of the previously developed Support Needs and Strategies for Physical Activity scale, while modifying and evaluating items that prior testing indicated required revision (Carbó-Carreté, Giné, & Guàrdia-Olmos, 2013). This instrument assesses the support needs and strategies of people with ID regarding PA in order to include the resulting information into Individualized Support Plans (ISP) along with other data (e.g., data obtained from the SIS).
In addition, two complementary descriptive scales were developed, one assessing the current level of PA and the other focused on the person's perception of PA. Both types of information are meaningful if we take into consideration their use in the design and implementation of the ISP process. The Level of PA scale is used to identify the amount and characteristics of PA that the individual with ID practices in their daily life. The Perceptions scale provides information about the individual's perspectives on PA. In the ISP process, it is essential to know the individual's goals and desired life experiences in order to design ISP to achieve personal outcomes, and these have to be examined according to the support needs (Schalock, Bonham, & Verdugo, 2008; Thompson et al., 2009). Thus, the purpose of these additional scales was to contrast activity levels and perceptions with data obtained from the Support Needs and Strategies for Physical Activity scale.
To recruit participants, service organizations were asked to participate through the Spanish Confederation of Organizations for the Persons With Intellectual Disability (FEAPS) and with the logistical support of each Autonomous Community (formal government regions of Spain) delegation. The organizations that consented to participate offered disability day services (e.g., special work centers, occupational therapy, and day centers), and most operated residential services (e.g., supervised flats and residences) for adult individuals with ID.
The sample consisted of 529 participants with ID (296 men and 233 women; Mage = 35.03 years, SD = 10.82, age range: 16-66). We conducted a convenience non-random sampling in the following seven Autonomous Communities: Andalusia (20.9%), Aragon (4%), Catalonia (25%), Castile and León (6.6%), Castile–La Mancha (14.8%), Community of Madrid (17.4%), and Galicia (11.7%). Of the total sample, 84.9% engaged in PA, and of those who did not, 73% admitted that they had no health or mobility problems preventing them from engaging in PA. Regarding the participants' environments, 13.9% lived in rural areas, 33.8% in semi-urban areas, and 52.4% in urban areas.
Spanish law assigns a “handicap” percentage to every person with a disability to represent the severity of the condition. The law stipulates that those with a percentage of 33% or higher receive an economic subsidy or support to facilitate everyday life. The disability percentage is assigned administratively based on all types of impairments (e.g., intellectual, physical, sensorial). Half of the participants (51.2%) were at 65%-74% (high level of dependency). The second largest group (36.8%) were at 75% (very high level of dependency), and the smallest group (12.1%) of participants were at 33%-64% (moderate level of dependency). The participants were predominately at the ID in the middle range of the spectrum (47.3%) or ID at the upper end of the spectrum (33.3%) of ID, whereas the smallest groups consisted of individuals with ID in the lower range of the spectrum or ID at the lowest end (11.3%) or at the borderline level of intellectual impairment (8.1%). The majority of the participants attended occupational therapy services (71.7%), followed by day centers (11.8%) and special work centers (9.7%). Regarding residential services, most participants did not utilize them. Most of them lived with their families (79.9%), whereas a minority lived in supervised flats (9.5%) or residences (7.5%).
A person from each organization contacted families and service providers and asked them to participate. A letter was sent to each family and service provider who accepted to explain the project and to thank them for collaborating.
Among the service providers (n = 522) the majority (64.4%) were direct assistance workers at day centers or a specific professional such as a psychologist (21.2%). A number of providers who designed and developed PA programs at the day centers (9%) also participated. In terms of training, over half of the service providers (56%) had a university degree, 9.2% of which had some high level of graduate education (e.g., masters or postgraduate degree). The remaining providers had some level of secondary education (12.9%) or other training (21.9%).
Among the families (n = 462), the parents were the most common participants (71.9%). In some cases, siblings (21.6%) or legal tutors (3.2%) were the respondents. Of the family members, 48.3% lived in urban areas, 36.3% in semi-urban areas, and 15.4% in rural areas. Most of the family member respondents had an elementary education (40.7%), some had a secondary education (22.8%), and some had a university education (18.4%). The remaining 11.6% of participants had no formal education. Informed consent documents were provided for all of the participants to read and sign.
The three instruments, the Support Needs and Strategies scale, Level of PA, and Perceptions of PA, are described separately in the following sections.
Support Needs and Strategies for Physical Activity Scale
Description and factors of the scale
The Support Needs and Strategies for Physical Activity scale (hereafter referred to as the Support Needs and Strategies scale), assesses two factors: (1) the support needs of people with ID regarding PA and (2) the strategies provided to support these needs. The intent is to identify the main support needs for adequately engaging in PA and, given the presence of a need, to assess whether this need is met or not. This scale was developed based on the supports paradigm that assumes that supports (a) are defined in terms of resources and strategies; (b) allow the persons to access the resources, information, and relationships in integrated environments; (c) bring about an increase of integration and an improvement of personal growth and development; and (d) are assessed in terms of results.
Support needs are defined as a “psychological construct referring to the pattern and intensity of supports necessary for a person to participate in activities linked with normative human functioning” (Thompson et al., 2009, p. 135). The second factor, support strategies, aims to assess “[t]he planned and integrated use of individualized support strategies and resources that encompass the multiple aspects of human performance in multiple settings” (Schalock & Luckasson, 2013, p. 91). Items are based on elements within individuals' supports systems related to skills and knowledge, environmental accommodation, incentives, personal characteristics, natural supports, professional services, positive behavior supports, organizational, and societal policies and practices (Schalock & Luckasson, 2013; Schalock & Verdugo, 2012). The supports strategies evaluated in the scale includes a broad vision beyond just those provided by a service agency, as PA programs are more successful and effective when they are developed within community settings (Howie et al., 2012; Rimmer & Rowland, 2008).
Developing and selecting items was based on the literature and a panel of experts: procedures and results are explained in a previous article (Carbó-Carreté et al., 2013). The instrument evaluated here includes modifications based on the previous study, for example, removal of items and modification of wording. Reliability was acceptable (n = 42) for each of the three versions. The reliability coefficients for the Support Needs and Strategies scale were .80, 95% CI [.74, .86] for the individual with ID version, .82, 95% CI [.77, .87] for the family member, and .84, 95% CI [.79, .89] for the service provider version of the scale.
Support needs and strategies were assessed using fifteen items for each scale. When administered items were directly and explicitly linked to the factor (e.g., “If you want to engage in PA or a sport activity during your leisure time, do you need someone to go with you? If so, is there someone like staff, family members, or a friend who can go with you?”).
Each of the three types of participants in the ISP process were assessed (the individual with ID, the family member, and the service provider). Affirmative responses on the Support Needs scale indicate the resources are needed to engage in PA. Similarly, affirmative responses on the Strategies scale indicate satisfaction with how the needs identified in the Support Needs scale are addressed. In other words, support strategies are assessed only when a support need is identified. Accordingly, when no support need is identified (for an item with a negative answer), the corresponding strategy item is not assessed.
Level of PA scale
The Level of Physical Activity scale consists of 11 items that assess the dimensions (mode, frequency, duration, and intensity) and domains (occupational, leisure time) of PA (Strath et al., 2013). Items were based on previous research on PA levels among people with ID (Draheim et al., 2002; Frey, 2004; Messent, Cooke, & Long, 1998) and revised by the expert panel in order to capture the unique characteristics of this population. For example, the scale assesses how disability day services influence PA activities (e.g., “Generally, how does he/she engage in physical activity? (a) on his/her own; (b) as a planned activity at a day center and at a scheduled time; (c) as a planned activity by a service provider from the day center but performed during leisure time; (d) as a shared activity with his/her family”).
Based on the pilot test (Carbó-Carreté et al., 2013), we revised the composition of the items to facilitate comprehension. Because the scale is intended to measure objective indicators, it is filled out by a service provider who knows the person well.
Five of the 11 items represent the global PA scale score (items 2, 5, 7, 8, and 9), and the remaining items are descriptive variables for use by administrators. A high scale value (3) reflects greater PA participation and the lowest value (1) represents less PA. Item 7 (“When does he/she engage in more PA or sports?”) is scored differently with the same value (1) assigned to responses indicating weekdays as well as during the whole week. This decision was taken on the basis that the main difference we identified is between weekdays and weekends. It also agrees with previous research (Stanish, 2004; Stanish & Draheim, 2005) that shows that individuals with ID engage in more PA on weekdays than on weekends. The global score for the Level of PA is the sum of corrected values for the five PA items.
The Perceptions scale consists of nine dichotomously scored items that examine the perception, knowledge, and motivation of the individuals regarding their PA (e.g., “Does PA help you lose or manage your weight?”). Items were developed based on previous research (Heller, Hsieh, & Rimmer, 2004; Hutzler & Korsensky, 2010; Salaun, Reynes, & Berthouze-Aranda, 2013; Temple, 2007; Temple & Walkley, 2007) and feedback from the expert panel. The intent of this scale was to assess the role of perception and motivation because previous research has shown that training people with ID about health and exercise concepts also improves their level of PA (Bazzano et al., 2009; Mann, Zhou, McDermott, & Poston, 2006). In the pilot study, the Perceptions scale yielded a Cronbach's α of .85, 95% CI [.81, .89]. Items are scored during interviews with the individuals with ID. The items are dichotomous, and the sum of affirmative responses is an estimate of the extent to which individuals have a positive perception of PA. Figure 1 is a schematic of the Level of PA scale and Perceptions scale.
Professionals in the participating organizations who were interested in serving as interviewers were provided training on administering the instrument. Interviewers administered the Support Needs and Strategies Scale and the Perception scale. For the former scale, the same interviewer administered the scale to the respondent, service provider who knew the respondent, and the respondent's family member. The service provider and family interviews lasted about 30 min, and the individual with ID interview took about 40 min since each item was covered in more detail. The administration of the Perception scale lasted about 15 min. In instances where the respondent had problems in communication or in understanding an item, the scale was not applied. The Level of PA scale was administered to the service provider in less than 10 min. From a total of 670 eligible participants, 529 were interviewed.
IBM SPSS Statistics 21 software was used for the majority of analyses, and MPlus for the construct validity estimations. First, we estimated Cronbach's α reliability coefficients for all the scales and factors. For each estimate 95% confidence intervals were reported. Second, construct validity was assessed for each scale through Confirmatory Factor Analysis (CFA).
Finally, in order to facilitate the use and interpretation of these results, we generated scores based on the needs and strategies totals of the Support Needs and Strategies scale. Difference scores were generated for each kind of the three types of respondents. The scores were used to estimate the extent to which the three informants agree or disagree with each other regarding the required support needs, and the extent to which actual support is received. In the final analysis, we estimated the correlation coefficients between the scales' global values and the informant specific scores as evidence of the convergent validity between the different informants.
The reliability of the Support Needs and Strategies scale was assessed for each of the three informants. As shown in Table 1, the values obtained from these analyses were acceptable. In all cases, the results show adequate reliability regarding the internal consistency of the scales.
Construct Validity Estimation
To evaluate construct validity, we assumed a two-factor measurement model in a CFA for the respondent, service provider and family member versions of the Support Needs and Strategies scale and a one-dimensional model (one factor) for the Level of PA and Perceptions scales. Given the ordinal scaling of the items, we estimated Spearman's rank correlation coefficients and used Maximum Likelihood (ML) estimation, in which we assumed that the total factorial scores were distributed according to the normal model of μ = 0 and σ2 = 1 (see Ory & Mokhtarian, 2010; Palomo, Dunson, & Bollen, 2007; Poon & Lee, 1994 for more details about the estimation process). The results regarding the measurement model fit are presented in Table 2.
According to the indices, a good fit was obtained for each of the models. In general, the models exhibited somewhat high χ2 values, with a level of significance (p) below .10. and the fit indicators (GFI, AGFI, BBNFI, BBNNFI, TLI, and CFI) exhibited values well over .95. Moreover, the standardized residuals (RMSR) are below .05, and both AIC and BIC have clearly negative values, indicating that an acceptable level of fit for the models. For the Support Needs and Strategies scale, the model for service providers exhibited the best fit (GFI = .972; AGFI = .970; BBNFI = .975; BBNNFI = .971; TLI = .972; CFI = .973; and RMSR = 0.02), although the models for individuals with ID and family members were also satisfactory. Factor loadings were also examined for each model. For the Support Needs and Strategies scale, the ranges for the loadings for the three models based on the different informants were included (Table 3).
As shown in Table 3, the values for the factor loadings linked to each model and source of information were all statistically significant (p < .001). The values obtained indicate that the factor loadings at the theoretically assigned factor are important for each item. In addition, the coefficients of determination (R2) indicate a highly acceptable level of explained variance for the observed variables and the general model.
Scoring Support Needs and Strategies
A series of scores indicating the different dimensions of the Support Needs and Strategies were generated: (a) INID (Total Needs according to the individual with ID); (b) INSP (Total Needs according to the service provider); (c) INF (Total Needs according to the family member); (d) ISID (Total Support Strategies according to the individual with ID); (e) ISSP (Total Support Strategies according to the service provider); and (f) ISF (Total Support Strategies according to the family member).
In addition to these scores, we developed three measures of the discrepancies between the perceived support needs and the actual support received. Ideally, a discrepancy should be small or equal to zero. Higher scores are an indication of unmet support needs. As shown in Table 4, indicators are separately computed for each of the three types of respondents.
A total congruence index (TCI) indicating the degree of agreement between the three types of informants was computed for support needs but not estimated support strategies. Since the strategies depend on identified needs, any such index would likely be strongly correlated with the TCI and, therefore, of little diagnostic value. The index is a simple statistical value based on the average of the discrepancies between the judgments of the three types of respondents. High TCI values indicate poor congruence. The TCI value is expressed as follows:
Table 5 presents the descriptive statistics for the corrected items and indicators.
The TCI scores indicated discrepancy in judgments across the three informants. All three informants believed the provided support strategies did not meet needs, with the service providers perceiving the highest level of need as well as placing the greatest value on the presence of strategies.
The distribution of TCI scores, which indicates the degree of congruence across the informants, is shown in Table 6 in terms of percentiles in order to facilitate an easy interpretation of lower and higher values of discrepancy.
Although the distribution shows few support needs for which the congruence value tended toward zero, the majority were relatively modest. The discrepancy indicators and the TCI suggest that planning teams for individuals with ID have the necessary information to include support strategies in individualized plans. The proposed instrument is based on the idea that optimal assessment is obtained from using the three informants' evaluations, who collectively contribute perspectives from the different contexts of the person with ID.
Convergent Validity Estimation
Table 7 summarizes the correlations among the different scales and indicators.
Results were consistent with the proposed models underlying the article. There were significant correlations between support needs and each informant's report of the strategies they received (r = .90, p < .001 for individuals with ID and family members, and r = .85, p < .001 for service providers). Second, the results for the discrepancy indicators were linearly related to support needs (r = .29, p < .001 for individuals with ID, r = .27, p < .001 for service providers, and r = .32, p < .001 for family members), and their values provide information regarding the needs that the provided support strategies currently fail to meet. Similarly, the strategy indicators were negatively correlated (r = −.15, p < .001 for individuals with ID, r = −.27, p < .001 for service providers, and r = −.12, p < .001 for family members), which is consistent with the notion that support needs will tend to diminish and the level of discrepancy between informants will decrease as the provision of strategies increases.
The correlations between the discrepancy indicators and the scores for the Level of PA and Perceptions scales are notable, as are the correlations among the indicators themselves. The estimated correlation between the scores for the Level of PA scale and the scores for the Perceptions scale (r = .16, p < .001) corroborates existing findings that increased knowledge and a positive perception of PA foster engagement in PA. For the Level of PA scale, scores for the scale were negatively correlated with the discrepancy indicators for all three informants (r = −.25, p < .001 for individuals with ID and family members, and r = −.26, p < .001 for service providers). This result may be due to the unresolved support needs that restrict the ability of individuals with ID to engage in PA. Regarding the Perceptions scale, which focused on individuals with ID only, scores on this scale were correlated with both support needs (r = .27, p < .001) and support strategies (r = .29, p < .001). This finding indicates that although individuals with ID acknowledge their need for some support in engaging in PA, they still have positive perceptions of PA. Similarly, the strategies perceived by individuals with ID were positively correlated, indicating that such strategies are relevant for PA-related programs for individuals with ID. Therefore, these strategies should be maintained and encouraged by service organizations that assist people with ID.
The current article aimed to validate an instrument to measure support needs and strategies for engaging in PA for people with ID. Additionally, we developed two descriptive scales that can provide useful supplemental information for the ISP.
The assessment of support needs and strategies regarding PA was based on three informants. To determine the difference between the support needs perceived and the strategies actually received by individuals with ID, discrepancy indicators were developed for each informant. We also designed a TCI, which indicates the degree of agreement among the three informants with regard to support needs, and this index revealed a low level of congruence among the informants regarding support needs. Currently, research on the degree of agreement among the three examined informants regarding support needs for people with ID is lacking. Nevertheless, a number of studies using qualitative assessments have shown that parents and service providers have similar perspectives regarding PA among individuals with ID. Specifically, within the focus groups studied by Temple and Walkley (2007), parents and family members identified psychological and emotional factors as the underlying determinants of the participants' inactivity, whereas the individuals with ID did not mention a lack of motivation or report a preference for sedentary activities.
The Support Needs and Strategies scale and the two complementary scales may be useful for service organizations assisting people with ID in the ISP process. The convergent validity analysis revealed that individuals' level of PA and perceptions covary with their support needs and available strategies regarding PA. Therefore, we should identify and address the elements that are necessary for people with ID to reach adequate levels of PA and thus achieve better general health. In this sense, the proposed instrument is a valid, reliable tool to assess support needs regarding PA for people with ID, and it can be used to complement data obtained from the SIS (Thompson et al., 2004). Providing necessary support for satisfactory functioning is one of the main goals of service organizations that assist people with ID, and the proposed instrument can contribute to identifying such individuals' PA-related support needs. Moreover, if we bear in mind that ISP rely on different sources of information; this instrument stands out because it directly involves the individual with ID. In addition, a congruence index was designed to facilitate the interpretation of the data provided by the three sources of information (i.e., individual with ID, service provider, and family member) and to aid in planning and providing support.
Accordingly, our instrument is notable because of the direct involvement of the individuals with ID. For both the Support Needs and Strategies scale (the individual with ID version) and the Perceptions scale, we likely obtained adequate reliability values because we followed Finlay and Lyons' (2002) recommendation to assess this population and thus reduce their tendency toward acquiescence. It should be noted that when individuals with ID cannot complete these scales themselves, the responses do not apply, and in no case is the participation of a referent accepted.
This article has a number of limitations. First, the participants did not include individuals with ID in the lower range of the spectrum. The scale was administered during an interview with the individuals with ID, which would have been complicated for individuals with ID in the lower range, as they have severe comprehension difficulties; therefore, the data may not be representative of the whole population of individuals with ID. Another issue concerning the sample's characteristics is that a large number of participants with ID lived with their families. Thus, few participants lived in supervised or independent apartments, which is an issue given that family participation was necessary to participate in the current study. When the participants lived at home, the organizations had easier access to the parents living with them. By contrast, when the individuals with ID lived in supervised apartments, their legal tutors were accepted as family members, given that these individuals with ID often had no family or their parents were too old to participate. Future studies should examine the psychometric properties of our instrument with individuals with ID in the lower range of the spectrum and should adapt the item presentation to facilitate their comprehension.
Future research should also focus on the covariation between the discrepancy indicators for the three informants and the TCI. Furthermore, as shown in previous studies (Johnson, Yun, & McCubbin, 2014), having a support person collaborate with participants with ID who have limited comprehension should be considered when examining the discriminant validity of our self-report scales (i.e., the Perceptions scale and the version of the Support Needs and Strategies scale for individuals with ID). Despite these limitations, this study opens new research paths regarding the impact of PA on quality of life for people with ID, as it examines the extent to which available support strategies help improve both the level and perceptions of PA among people with ID.