A comprehensive set of items related to barriers and facilitators to healthy, active living for adults and children with disabilities were first identified in existing literature, building on the extensive review of built environment instruments completed by Gray [20]. Instruments were selected that either reported psychometric properties, or were considered highly influential and currently used in public health practice. In addition, items were gathered from published studies that examined barriers and facilitators to healthy living for people with disabilities. Database searches included NARIC, PubMedAHL, and SportsDiscus, and books and grey literature identified from reports and non-peer reviewed magazines. Key search terms were used and articles screened for their relevance to people with disabilities. Among articles that met these criteria, the specific barriers and facilitators described in the manuscripts were extracted for potential use in the CHII.

Twenty semi-structured focus groups were conducted in geographically diverse communities across the U.S. from the East Coast, Midwest, Southeast, and Northwest. The focus groups were conducted in a diverse set of communities, with most coming from low and medium income communities. Eight were in rural and 12 in urban communities. The interviews were evenly split among individuals with disabilities (group 1) and professionals (group 2) who work with people with disabilities. The inclusion criteria for people with disabilities was having a physical, sensory, or cognitive disability and being their own guardian (capable of consenting for themselves). They had to be ≥18 years old for adult focus groups, and between 13–18 years for the youth focus group. For service providers, the inclusion criterion was they had to provide services to people with disabilities. All participants had to speak and understand English.

A purposive sampling framework was used to ensure that in each focus group of people with disabilities, there was good representation from various disability types. The professional focus groups included a diversity of public/private and small/large organizations providing services to a range of people with physical, cognitive and sensory disabilities. Both groups were recruited through local partners and their existing networks through email and website postings.

The major theme of the focus groups was a discussion of the facilitators and barriers to community health inclusion structured around five domains: Built Environment, Equipment, Programs/Services, Staff and Policies. These domains had been established in a previous study by Rimmer et al. as being associated with community health inclusion [24]. The focus groups were stratified so that sectors of the community, previously identified by the CDC [22] (schools, health care facilities, workplaces, community institutions/organizations and the community-at-large), were discussed by each of the two respondent groups. Participants were asked about barriers and facilitators they experienced to being physically active or accessing healthy foods within their assigned sector. For instance, young adults with disabilities in high school provided feedback about school-related issues, while working-age adults with disabilities provided feedback about work sites.

Focus group recordings were transcribed verbatim and coded by two trained coders using content analysis into barriers and supports for healthy, active living. Content analysis can be used to organize and distil content in to a few categories that share similar meanings and explain the phenomena under study. Inductive content analysis (used in this paper) involves open coding and development of code sheets of the categories the coders have decided on based on their interpretation of similar meanings [25]. Inter-rater agreement was assessed by dividing the agreed upon codes obtained from the coded transcripts by the total number of codes [26]. Written informed consent for participation in the focus groups was obtained from participants or, where participants were youth, a parent or guardian.

Items from focus groups, existing instruments, and articles on barriers/facilitators were compiled into a master item bank. Two trained research staff reviewed each item to determine if they met two criteria: 1) could it be objectively measured? and 2) did it relate to physical activity, nutrition or obesity at the PSE and not personal level? Items that did not meet these criteria were excluded. Duplicate items were combined. The research team grouped these items into a set of constructs that measured a common theme, which could be representative of a physical feature (e.g. sidewalks), program (e.g. nutrition class) or a policy (e.g. staff training). For instance, items on sidewalk length, slope, condition, materials, and connectedness were aggregated into a construct on sidewalks.

Twenty national experts, identified from the literature and through professional networks, were recruited to review the item bank. One fourth of the experts identified as having a disability. Eligibility criteria required all experts to have a background working with people with disabilities in one or more content areas related to physical activity or nutrition. The experts were organized into four panels based on their area of expertise (physical activity, nutrition, general accessibility and community design). Panels were charged with reducing the number of items to those most critical for measuring community health inclusion. An item was considered critical if it was necessary for people with disabilities to 1) independently access the physical activity or food environment, and 2) be able to participate in health promotion activities such as classes, programs, and services.

Review of items was conducted in two rounds using electronic surveys. Each expert panel had five members who reviewed materials related to their area of expertise. During the first review, experts were asked to decide on importance and level of measurement (narrow focus or broad) of each construct. In the second review, they decided on which items best served as indicators most representative of accessibility for each category. Items with less than two votes, out of a possible score of 5, were removed unless an expert had a strong rationale for keeping it, in which case it went through an additional review by the research team to determine its distinctiveness from other items.

The structure of the instrument is hierarchical in nature (sector-venue-domain) and was designed so that comparisons could be made across the sectors illustrated in Fig. 2. These sectors have been used in other national instruments [21, 22] and represent a common framework for examining a community’s healthy living resources. The CHII was organized by venues within each sector. For example, a cafeteria is considered a venue that is located in three different sectors: hospitals, schools and work sites. Each sector had a few additional items relevant to the specific populations that used the sector (e.g., children in schools). The Community-at-Large sector incorporated items that were about the community as a whole. Within each sector, there are items across five domains: Built Environment, Equipment, Programs/Services, Staff and Policies. The items under the built environment and equipment domains were conducted through an observational audit of the facility, while the items under the program, staff and policy domains required an interview with the management of the organization.Fig. 2

Structure of the CHII

Pilot testing of the initial instrument was carried out at eight sites from different sectors (3 schools, 1 work site, 1 hospital, 1 grocery store, and 2 park district recreational sites) around the Chicago area. The focus of this testing was to determine how well the order and protocol worked for the observational audit and to obtain feedback from respondents on the interview portion of the survey related to content and clarity of questions. Feedback from respondents was used to modify the interview section of the CHII. Feedback from raters on the observational assessment was used to reorganize items and ensure protocol clarity.

Academic partners within five states (Illinois, North Carolina, Arizona, Alabama and Montana) were involved in the field testing of the CHII. A purposive sampling methodology, whereby the sample is selected based on knowledge of the population of all communities, was used to select communities with varying levels of income, urbanity and geographic distribution using data from the American Community Survey and the United States Department of Agriculture [27]. Fifteen communities were identified based upon each combination of income and urbanity type. Within each community, raters assessed Schools, Work Sites, Health Care Facilities, Community Institutions/Organizations and the Community-at-Large.

An academic partner in each state coordinated the field testing locally. Raters were recruited who had knowledge of accessibility and inclusion and some experience with conducting assessments. Additionally, partners were asked to ensure that some of the raters were people with disabilities. Most of the raters (60 %) lived in the community that they were rating. Local knowledge of the community proved to be important for recruitment of field testing sites. Some raters (47 %) were recruited from Centers for Independent Living, organizations devoted to providing improved opportunities for people with disabilities within communities [28].

All raters were trained by the project director (YE) during a 2-day workshop in: (1) using the CHII in field-based sites; (2) learning how to conduct formal interviews for collecting data from management of organizations; and (3) practicing the use of two measurement tools related to ADA accessibility items (door pressure gauge and a smart tool to measure slope), which are commonly used in accessibility assessments and based on Americans With Disabilities Act Accessibility Guidelines, or ADAAG [29].

While field testing the CHII, raters were asked to sample three sites from each of the following sectors: schools, work sites, health care facilities, community institutions and food sites using inclusion criteria of: public schools, large employers, hospitals or clinics, community institutions/organizations, and either grocery stores, farmers markets or community gardens. Raters were instructed to identify sites that had physical activity and/or nutrition venues or programs, which were defined in the CHII structure (Fig. 2). Two raters went to each site to obtain inter-rater agreement for each type of venue (Level 2 in Fig. 2). Raters pre-identified transit stops, paths and venues and then proceeded to independently rate each site.

Focus group methods and field testing procedures were approved by the University of Illinois at Chicago’s Institutional Review Board (FWA #00000083) under expedited review. Raters worked with state coordinators to select and recruit sites that met the criteria using approved scripts and recruitment materials. Recruitment occurred by phone, email and in-person, where applicable. Permission was requested for all sites before conducting an assessment. No informed consent was needed for field testing (per IRB) as there were no human subjects; the subject was the site being assessed.