Conceptual framework

The study is informed by the Behavioral Model for Vulnerable Populations (BMVP), which is the primary conceptual model in the field of health care for PEH ([17]). The model includes individual level factors and contextual factors that highlight structural and enabling resources and has several domains. MRP-relevant variables at the individual level might comprise the following: the Predisposing domain includes demographic characteristics and housing/homelessness history, while the Enabling domain includes self-efficacy and level of social support ([17]). The Need domain includes current functional, physical and mental health status and substance use, while Behaviors include engagement with health care services and chronic disease self-management ([17]).

At the contextual level (e.g., organizational, larger social environment) MRP-relevant variables comprise the following: the Predisposing domain includes geographic location while the Enabling domain includes the structure and organization of health care services, such as the provision of mental health and substance use treatment services, processes of care, and the characteristics of health care personnel ([17]). The Need domain includes the area availability of housing resources. MRP-related Outcomes include treatment completion and housing disposition at discharge from the MRP.

Identification of study site sample and recruitment

The sampling frame was a convenience sample of three sites that were nominated by the National Health Care for the Homeless Council Respite Care Providers’ Network (RCPN) ([18]), had organizational stability, and agreed to participate. To account for variability in MRP settings, we sampled from different MRP practice types (e.g., freestanding medical respite unit; shelter-based model; and home and apartment-based. The administrator at candidate MRP sites was contacted via email, followed within 1 week by a video/telephone call to explain the study and assess interest in participation.

MRPs that agreed to participate had a video/telephone call to discuss the study methods, Internal Review Board (IRB) concerns, and to identify a facility liaison for the study. The liaison was responsible for disseminating the recruitment brochure to patients and staff, and coordinating communication between the research team and survey participants as needed. Facility liaisons did not ask for consent, answer questions about the study, or otherwise act as a representative of the study, but were allowed to participate in the study as a site administrator, provider, or staff member as self-designated by their responsibilities at the MRP.

Study instruments

The BMVP informed item and measure selection for the data collection instruments. We developed and maintained self-administered Qualtrics web-based surveys for MRP providers (e.g., licensed medical providers) and staff. We used computer assisted telephone surveys for MRP administrators and patients due to potentially sensitive items (e.g., MRP budget, patient reported substance use disorder). In addition, we developed a monthly program summary data census report of all patients, and a patient chart review document to abstract data from the MRP medical record of site patients who participated in the study. Tables 1–3 represents the data tables.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Medical respite program data elements: Individual characteristics.

https://doi.org/10.1371/journal.pone.0295543.t001

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Medical respite program data elements: Contextual characteristics.

https://doi.org/10.1371/journal.pone.0295543.t002

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Medical respite program data elements: Behaviors and outcomes.

https://doi.org/10.1371/journal.pone.0295543.t003

Data collection

After agreeing to participate, we mailed the facility liaison a packet of study brochures and copies of consent forms, and study incentives to distribute to patients. A research assistant (RA) was designated as the primary contact for the study site. After verbal, informed consent by the facility liaison, the RA collected facility-level measures via phone interviews with the site administrator (i.e., individual self-identified as responsible for MRP operations) and entered the data into a web-based survey. We sent the participating facility liaison the survey questions and consent form one week before the phone interview.

The facility liaison provided the research team with the names and email addresses of providers and staff at their facility who agreed to participate after reviewing consent forms and study brochures. The providers and staff (approximately 3–5 per site) received an email with a link to informed consent and the web-based survey (i.e., provider and staff survey) from the RA and confirmed via email when completed. The facility liaison also identified a convenience sample of 10 MRP patients as candidates to be interviewed. Responses were coded with identification numbers, not names, and the list linking numbers and names were kept in a separate file. No information provided was given to the treating provider or MRP staff. Since not all patients had access to individual phones, the interviews were conducted on a secure phone line at the facility. After verbal informed consent, responses to the computer facilitated telephone interview were entered into the database.

Patient consent also included a HIPAA authorization form granting permission for study investigators to access medical records for research purposes. After completion of patient interviews, the research team worked with the facility liaison in collecting, copying, and transferring medical records, from the time of admission to the time of data collection, to a HIPAA compliant fax line, which was entered into a web based Qualtrics database. At four and eight weeks after completion of patient surveys, the facility liaison was asked to complete an electronic monthly facility-level report that included total MRP census and patient discharge disposition.

Ethics

The study was reviewed and approved by the University of North Carolina Chapel Hill Internal Review Board (IRB #21–0196). All participating sites received a $2000 honorarium and staff did not receive an incentive to participate. All patients who participated in the study received a $10 gift card.

Data analyses

We will use descriptive statistics to determine the feasibility of the data collection methodology by comparing the participation and completion rates with a prior study of PEH, which is approximately 22% for patients ([19]). To our knowledge, there are no prior studies that have reported MRP study participation rates. We will use descriptive statistics to describe the study samples, Chi-square tests of independence and employ Fisher’s exact tests in cases of sparse data to examine the association between item response/nonresponse and facility characteristics, MRP provider/staff characteristics, and PEH characteristics.

We will examine the unadjusted associations between MRP organizational (based on the Administrator Survey), provider and staff characteristics (based on the Provide Survey), patient characteristics (based on the Patient Survey) and the outcomes of: (1) treatment completion and (2) enrollment into housing programs using chi-square test of independence and Fisher’s exact test. We will use Odd Ratios (ORs) to describe the magnitude of these associations. The primary goal of these future analyses will be to identify plausible relationships between selected facility and individual-level characteristics and outcomes of interest.