This study is part of a larger effort to identify educational needs for adolescents with ABI from the perspectives of multiple stakeholders, including patients and their caregivers. [14] Our objective was to identify and prioritize urgent hazards in the home setting for adolescents with ABI. All study activities occurred remotely or in person at a large academic medical center or the local children’s hospital in the US Midwest. This study was approved by the local children’s hospital’s institutional review board (STUDY00001368) and registered through www.clinicaltrials.gov (NCT04768946).

We recruited providers using a convenience and snowball sampling method, via referrals from personal networks and posts to the children’s hospital listservs from May 1, 2021, to May 18, 2022. Our inclusion criteria were: 1) currently being a healthcare professional involved in discharge planning and community re-integration for adolescent patients (aged 12–19) with ABI, 2) able to participate in a 2-hour interview, 3) able to understand and communicate in English, and 4) able to attend interviews either in-person or online via their own means. Participants received a $25 gift card.

After enrollment, participants completed a demographic questionnaire and a questionnaire about home hazards electronically in REDCap (Version 10.0.28). The hazard questionnaire included these areas: slip or trip (11 items), lifting (6 items), electrical (10 items), fire (16 items), environmental (19 items), and pet or pest (5 items). For each item, participants indicated if 1) they were concerned about the hazard and 2) if the hazard could lead or had led to harm. Members of the research team developed the hazard questionnaire which lists potential home safety hazards that were previously identified and tested among home healthcare workers. [15] Thirteen participants completed the questionnaires. Between June 7, 2021, and May 18, 2022, twelve of the thirteen participants completed a 40-minute semi-structured individual or group interviews of 2–4 participants, in person (n=2) or over Zoom (n=10), facilitated by a group of 1–3 researchers [SA, SP, TS, JL, AD]. Next, participants completed the Hazard Prioritization Matrix activity, where they identified and rated no more than their five highest priority hazards on 1) frequency: rarely, sometimes, or often; and 2) severity: minor, moderate, or major consequences. Urgent hazards occurred often and had major consequences. The hazard prioritization activity was also used previously among home healthcare workers. [16] The entire session was audio and video recorded and transcribed using Express Scribe into Microsoft Word by trained researchers followed by independent verification by a second researcher. In-person participants completed this activity with sticky notes on a large whiteboard or paper template, and online participants completed this activity verbally or through Zoom chat while a facilitator populated a text document shared on the screen.

Study data were collected and managed using REDCap electronic data capture tools hosted by the Clinical and Translational Science (CCTS) at The Ohio State University Center. REDCap (Research Electronic Data Capture) is a secure, web-based application designed to support data capture for research studies. It provides an intuitive interface for validated data entry, audit trails for tracking data manipulation, automated export procedures for seamless data downloads to common statistical packages and procedures for importing data from external sources. [17]

Our overall approach was convergent mixed methods, aimed both at validating pre-identified hazards and risk factors (frequency and severity) with quantitative data, and discovering unexpected hazards and risk factors with qualitative data. [18] We generated descriptive statistics from the questionnaire and prioritization data using Microsoft Excel. We used a grounded theory approach [19] to guide the analysis of our interview transcripts and prioritization activity data. We used Dedoose software V9.0.107 to identify themes relevant to our aims on home safety hazards and associated risk factors. Three trained coders iteratively and independently created descriptive codes using constant comparison [19] and organized them into categories. A consensus on all codes and categories was achieved through discussion. One principal investigator [SA] used inductive thematic analysis to look at relationships across the data set [20] to generate themes and sub-themes related to concepts, which led to the conceptual model. We handled missing data by deleting the case with the missing values from that item’s data analysis – one participant did not complete the interview or hazard prioritization portions of the study and thus only twelve participants were included in those analyses.

No patients (i.e., individuals with similar characteristics to our participants) were involved in the development, procedural, or dissemination activities of this study. As part of our snowball sampling recruitment process, we did ask participants to refer others who may be eligible to participate in this study. Additionally, several co-authors currently hold and/or previously held similar professional roles to our participant population [SA, JL]. We plan to include representatives from our participant populations as clinical research partners in future endeavors to help us develop, deliver, and disseminate our research.

We obtained approval from the local children’s hospital’s institutional review board (STUDY00001368) to conduct this study. During the consent process, participants were instructed on the study goals, procedures, risks, and benefits. This included the voluntary nature of their participation and the collection of identifiable audio and video recordings. All enrolled participants provided their written or verbal informed consent prior to participating in the study.

Thirteen participants completed the demographic questionnaire, which included personal information (Table 1) and client caseload characteristics (Table 2). Twelve participants were recruited from the local children’s hospital, and one participant from a local hospital system that primarily served adults but whose caseload sometimes included adolescents. While our participants were primarily occupational therapists (OT), physical therapists (PT), and speech-language pathologists (SLP), one neuropsychologist and one rehabilitation case manager also participated.

Thirteen participants completed the hazards questionnaire and were included in the analysis. Figure 1 summarizes the responses to pre-identified hazard concerns and harms across 67 items in six categories. All participants agreed that seven (10%) hazards from three categories were concerns: 1) slip/trip: snow or ice on surfaces, slippery objects on the floor, large furniture/objects in the path of travel, 2) lift: inappropriate assistive equipment, 3) fire/burn/explosion: unattended open flames, smoke alarm not working, inappropriate use of cooking appliances. For most hazard categories, most participants believed the hazards to be concerns but not necessarily ones that could lead to harm.

Figure 2 examines the proportion of pre-identified hazard concerns as distributed across professions. By profession, OTs had the highest agreement with pre-identified hazards, PTs had the next highest, and SLPs had the lowest. OT providers agreed that more than 90% of the hazards in each category were concerning.

Twelve of the thirteen participants who completed the questionnaires participated in the interviews and were included in the analysis. The participant who did not complete the study was a PT and was lost to lack of follow-up. Figure 3 describes the proportion of pre-identified hazards discussed in the interviews. Few of the pre-identified hazards were self-identified by participants during interviews, and most of the fire (68.75%), electrical (80%), environmental (84.21%), and pet/pest (80%) hazards were not discussed. Of the seven items confirmed as hazards by all participants in the pre-identified hazard survey, participants self-identified four of them during the interviews: slippery objects on the floor, large furniture/objects in the path of travel, unattended open flames, and inappropriate use of cooking appliances.

Twelve of the thirteen participants who completed the questionnaires participated in the interviews and were included in the analysis. The participant who did not complete the study was a PT and was lost to lack of follow-up. Fifty-nine hazards were self-identified and assigned an urgency rating, with 12 (20.3%) urgent, 20 (33.9%) major, 19 (32.2%) moderate, and 8 (13.6%) minor risks. The urgent hazards, occurring often and with major negative consequences, were: 1) shower transfer/water, 2) getting up and down steps into home or bedroom, 3) risky behavior when spending time with family/friends, 4) ability to exit house in an emergency (emergency preparedness), 5) supervision in home/community, 6) getting into the home, 7) entryways, 8) independence/supervision, 9) stove use, hot appliances in kitchen, 10) stairs, 11) rugs, and 12) bathroom accessibility.

Table 3 lists the themes and sub-themes from the interviews with definitions, illustrative quotes, and cross-references to urgency ratings. The seven interactive themes were: hazardous activities, hazardous situations, hazardous objects, hazardous others, hazardous spaces, harms, and client factors. When the 59 self-identified hazards from the Hazard Prioritization Matrix Activity were qualitatively analyzed, we identified 116 distinct home safety hazard factors aligning with our seven themes, illustrating that these factors often interact to create hazards.

Figure 4 describes the conceptual model for home safety concerns – linking hazards, risks, and harms from the perspective of healthcare providers. Hazards are an integration of multiple factors and include how the adolescent interacts with objects and others within several contextual layers. Risks are a calculation of urgency based on the frequency of hazard occurrence and the severity of resulting harm. These determinations of concern can also change across settings and systems as the hazard factors and levels of risk change. Healthcare providers can intervene at every layer and step of this process.

This study was limited by its small sample size as fewer participants completed the study than we had planned. This was likely due to the ongoing impact of the COVID-19 pandemic adding additional strain to healthcare providers and hospital systems. The sample of participants who enrolled may not have been representative of the whole population of healthcare providers as they were all white, mostly female, and mostly from a large children’s hospital in large midwestern city. Other relevant professions were not represented in this sample (e.g., therapeutic recreation). Our location could also create bias as most children who experience TBI are not treated at a children’s hospital [33], and other regions may have a higher prevalence of different hazards. While we employed methodological rigor, due to the nature of our study design and qualitative exploratory approaches, there will always be risk of bias in the interpretation of the results, limiting their generalization to a more diverse population.