Study site
Within Portland, Oregon city limits, there are over 4000 people who are houseless or marginally housed individuals per night; the city has a population of just over 647,000 people [34]. Oregon Health & Science University (OHSU) works to serve people who are houseless through collaborations with Central City Concern, a large non-profit medical home for people who are low-income or houseless in the city, though students are not routinely involved in care delivery settings through this clinic. OHSU itself is located on top of a hill in Portland; most people going to the hospital either take an air-tram from the base of the hill to the hospital or drive or bus to the top. OHSU School of Medicine offers an 8-week continuity elective at Central City Concern that is available to a very limited number of students. Students at OHSU also operate an interprofessional free clinic in partnership with Transition Projects, a shelter and community-based organization, and may volunteer or receive credit for their involvement. While there are limited opportunities for student engagement through OHSU, there are many community-based organizations in Portland that provide extensive support services to houseless community members, the majority of which are not affiliated with OHSU. For this reason, a decision was made to choose study sites in community spaces outside the OHSU network.
This study was completed by medical students at OHSU in Portland, Oregon (CK, CF, AC, JJ). OHSU medical students began collaborating with people who are houseless and the organizations that work to serve them in 2017. This collaboration led to the formation of two advisory groups: the Houseless Neighbor Advisory Team (HNAT), comprised of current, formerly houseless, or marginally housed individuals; and the Community Partner Advisory Team (CPAT), comprised of community partners and medical providers. These two research panels directly oversaw the research described in this paper. Both HNAT and CPAT meet with students to provide feedback at each step of the research process.
Study design and setting
This mixed-methods study took place at two community locations in Portland, Oregon, during the winter and spring of 2018. With community input, we developed an interview guide that included both discrete choice (quantitative) and open-ended (qualitative) questions (Additional file 1), with the goal of understanding how medical students and institutions can serve people who are houseless.
Two sampling locations for interviewing people who are houseless were chosen in collaboration with HNAT and CPAT. The first location was Street Roots (SR), a community-run newspaper and media organization that allows people who are houseless to generate income through newspaper sales [35] . Street Roots is located in Portland’s Old Town/Chinatown neighborhood, and employs over 160 vendors. Medical students partnered with the Vendor Coordinator to identify if there was potential interest in having medical students talk to vendors visiting the site during morning coffee hours. Medical students interviewed participants on Saturday mornings from 7:30 am–9 am weekly.
The second data-collection location was Operation Nightwatch (ONW), a community-based hospitality center for people who are houseless that runs out of a local church on weekends [36]. Medical students worked with two Volunteer Coordinators (as one employee left, and another was hired) to identify if it was possible and appropriate to recruit study participants at ONW. Unlike SR, whose coffee hours were open exclusively to Street Roots Vendors, ONW is open to any guest visiting while open. Medical students interviewed participants at ONW on Saturday evenings from 7 to 10 pm.
We approached all individuals for consent until we had interviewed no more than half of our total sample goal (20 per site) in total. As medical students ourselves, we also wanted to understand how engagement that prioritizes community voices within our role as students might look. Participants were offered a coffee voucher and/or a snack at the beginning of the interview, and they were permitted to complete the interview alone or in small groups of two to three participants, if desired.
Participants
Study participants were eligible to participate if they were 18, considered themselves houseless or marginally housed, and were English speaking. Our outreach teams, comprised of two to four medical students, approached and enrolled potential participants at the two sites (SR and ONW), and conducted the interviews.
Study sample
We completed 38 interviews total at both ONW (n = 20) and SR (n = 18). This sample size was determined a priori in consultation with mixed-methods researchers at OHSU.
Variables
Quantitative
The interview guide included questions adapted from the 2016 Seattle, Washington Houseless Needs Assessment [37] . A list of discrete-choice questions was presented to CNAT and HPAT in January of 2018. We used a mixed-methods approach at the suggestion of the panels, who asked us to incorporate both quantitative and qualitative questions to answer our research question. These questions were modified based on feedback from both panels.
Binary and categorical covariates include gender, sexual preference, race/ethnicity, highest education level, employment status, primary language, born in the United States, veteran status, types of places of residence in past month, history of using alcohol and drugs, history of comorbid conditions, comorbid conditions contribute to houselessness, history of receiving healthcare from provider in last year, interest in foot care information, interest in naturopathic medicine information, access to nutritious meals, history of pregnancy and current pregnancy status, interest in receiving information about pregnancy, access to clean needles, frequency of clean needle use, length of time houseless in years, and age at first time of houselessness. Our only continuous variable was age.
Qualitative data
A set of open-ended questions was drafted based on feedback from initial conversations with CNAT and HPAT in fall 2017. These questions were reviewed by CNAT and HPAT in January 2018 and modified based on feedback. We explored ideas related to the experience of participants engaging with healthcare systems before, during and after clinic or hospital visits, which we believed may shape the asks of healthcare providers and systems. We then asked what medical students and separately, healthcare systems can do to meet the goals of people who are houseless. Qualitative data was not recorded at the request of community members during study-planning phases; medical students took notes during interviews by hand. The interview guide included both discrete choice and open-ended questions (Additional file 1) and data were collected at the same time, but we prioritized qualitative information from the survey in analysis for this paper.
Data analysis
Primary analysis
Univariate analyses displaying frequencies and means describe the quantitative sample. Quantitative data was entered into REDCap; basic frequencies were analyzed using STATA 14. Qualitative data was analyzed by four medical students trained in qualitative methods using thematic analysis rooted in an inductive process [38]. First, all four students (three female, one male) reread interview notes and transcripts to familiarize themselves with the data. Using a subset of interviews, we coded all interviews, and included codes that were related to answering our research question. We met regularly to discuss codes, and upon completion of coding, to identify themes and group codes within these themes. We then reviewed all codes and quotes within the context of the themes we constructed, and reorganized until consensus was met. Data analyzed was stored in the web-based program Dedoose. Mixing of quantitative and qualitative results occurred during the interpretation phase using a convergence study design.
Study results
Upon study completion, study results were shared with CNAT and HPAT for feedback on the analysis and to continue discussions of future planning for community-driven outreach and advocacy.
Consent and ethical considerations
Study participants provided verbal consent to participate. This project was approved by Portland State University’s Institutional Review Board. We applied for a waiver of written consent for this population, as the study was of minimal risk to participants. Participants gave verbal consent before beginning the study.