This is a mixed-methods study that uses an explanatory sequential design [39]. During the first phase, we will perform quantitative data collection, and during the second phase we will perform a qualitative study to clarify, interpret, and describe the quantitative results using a complementary approach.
Quantitative phase
Sample size and participants
We aim to examine all potential participants (approximately 300 HP) in Palma de Mallorca, according to data from 2019 [40]. The inclusion criteria will be: (i) age of at least 18 years; (ii) living in Palma and meeting the European Typology of Homelessness and Housing Exclusion (ETHOS) classification (5); and (iii) agreeing to participate in the study and signing the informed consent document. For operational purposes, we define HP as those who, during last year, lived in the streets, or other public areas, or in an abandoned building, or in place that does not meet minimum conditions for habitability (i.e., without power or water supplies) and excludes HP. This is similar to the definition used previously [24]. The exclusion criteria will be: (i) staying overnight in a private or municipal shelter for more than three months during the previous year; (ii) having an acute episode of mental disease or being under the influence of alcohol or any drug during recruitment.
Data collection
We will collaborate with two NGOs that work on specific programs with HP in the city of Palma. Their facilities will be used for the individual interviews and blood tests. Moreover, a worker from the NGOs will collaborate in contacting potential participants and arranging appointments. A nurse from our research group will collect data using a questionnaire and will perform blood tests. The estimated time for both activities is 30 to 40 min. Later, another nurse will check the electronic health records of all participants. We will give financial compensation to all participating HP.
A questionnaire structured in four blocks will be administered to assess the following: (i) sociodemographic characteristics (including age, sex, nationality, educational level, last occupation, receipt of public subsidies, time of homelessness, and shelter visits during last year); (ii) challenges experienced during the COVID-19 pandemic, especially during the confinement period (March to June of 2020), difficulties in finding essential resources, and compliance with sanitary measures (use of surgical masks, hand washing, and social distancing); (iii) self-reported use of healthcare servicesdiagnoses and pharmacologic treatments, and assessments of substance abuse (DAST-10), diet quality (IASE), and depression symptoms (PHQ-9); (iv) human basic needs assessment, adapted from Virginia Henderson’s Needs theory [41]. We will also ask participants about feeding, safety, hygiene, and sleep, and will assess social support using the SSQ-6 questionnaire [42].
Drug Abuse Screening Test (DAST): The DAST is a questionnaire used to identify drug abuse in adults, and a short form (DAST-10) was validated for the Spanish population [43]. The possible answers for each item are “yes” or “no”, the maximum score is 10 points, and a score of 3 or more is used to define substance abuse.
Healthy feeding index (IASE): The IASE measures 10 items to determine diet quality, and was validated for the Spanish population [44]. Nine measures are related to food groups and one is about diet variety. The score for each item is based on the adequacy of consuming the appropriate amount of each food group. Each item has a score between 0 and 10, and the total score is the sum of all ten items (range: 0–100 points). The different classifications are “healthy diet” (>80 points), “diet that needs changes” (50–80 points), and “unhealthy diet” (<50 points).
Patient Health Questionnaire (PHQ-9): The PHQ-9 is a tool used to diagnose and assess the severity of depression, and was validated for the Spanish population [45]. The 9 items ask about symptoms during the previous two weeks. Each answer is scored using a Likert scale, and varies from 0 to 3 points (total range: 0–27). The total score is used to characterize a participant as having “no depressive symptoms” (0–4), “mild depressive symptoms” (5–9), “moderate depressive symptoms” (10–14), “moderately severe depressive symptoms” (15–19), or “severe depressive symptoms” (20–27). A previous study of HP established a cut-off of 10 points [46].
Social Support Questionnaire (SSQ-6): The SSQ-6 is a 6-item questionnaire derived from a longer questionnaire (SSQ) that assesses perceived social support, and was validated for the Spanish population [42]. Each item asks the study participant about different occurrences that lead to stress or the need for assistance, the number of people that can be relied upon during that time, and their satisfaction with the perceived social support.
Blood test
Blood will be collected for serology testing of SARS-CoV-2, HIV, syphilis, and hepatitis B and C, and for measurements of glycosylated hemoglobin. Blood samples will be managed following the usual procedures for primary health care facilities, and analyzed by the referral hospital (Hospital Universitario Son Espases).
Health record review
We will also analyze the electronic health records and examine information about the number of visits to different services during the previous two years (Primary Healthcare, Primary Healthcare Emergency Department, Addictive Behaviors Unit, Mental Health Unit, Oral Health Unit), and will also assess the use of hospital outpatient consulting, hospitalization, and admission to emergency departments. All diagnoses and pharmacologic treatments during the previous year will also be assessed.
Data analysis
The survey will be prepared and administered using the TeleForm program (Cardiff Software, Vista, CA, USA), which allows automated data entry and subsequent verification. All continuous variables will be presented as means and standard deviations or medians and interquartile ranges, depending on variable distribution. The categorical variables will be presented as absolute numbers and relative frequencies. We will calculate the prevalences of different diagnoses in HP using percentages and a 95% CIs. The chi-Square test will be used to compare categorical variables, and Student’s t-test to compare continuous variables. All statistical analyses will be performed using Statistical Package for Social Science software (SPSS) version 24 (IBM, NY, Illinois, USA) and a p value below 0.05 will be considered significant.
Qualitative phase
Theoretical-methodological framework
We will use a critical theory framework and an ethnosociological approach to examine the relationship of homelessness and health. This perspective considers homelessness to be a consequence of the power relationships in society and in a given context. We will examine the social determinants of health using Nussbaum’s capability approach [2, 38]. This framework will be used to analyze and examine the relationships of human dignity and social justice with homelessness and health [47].
Moderate inductive and deductive logic will be used in a predominantly qualitative approach that is developed from a quantitative approach [48, 49]. This will highlight different experiences and the interactions of homelessness and health in an effort to characterize them as inseparable. This phenomenological description places the researchers and participants at the convergence of subjective testimony and social reality [50].
Sampling methods and participants
A purposive sampling design will be used in an effort to achieve high variability among the enrolled HP. Thus, the sample will be stratified by sex, age, sexual orientation, duration of homelessness, sleeping location, and administrative status. The saturation and representativeness of the data will limit the number of participants to a manageable level.
Data collection
Semi-structured interviews will be performed. More specifically, “life stories” will be used for a biographical analysis of an individual’s life trajectory before and after homelessness [51, 52]. This technique will allow us to explore and know how different personal, family, and structural factors contributed to homelessness and health of HP. The interviews will be structured and defined through Nussbaum's capability approach [2] and will address the following blocks: (i) current health situation: COVID-19 pandemic, health, and access to healthcare services; (ii) circumstances that led to homelessness and previous way of life; and (iii) current situation and life prospects. The first block will allow concise analysis of the experience during the COVID-19 pandemic. The second block will review the life trajectory that defines an individual’s current situation in terms of structural, personal, and emotional aspects, and will allow examination of the relationship between homelessness and health. The third block will assess the individual’s perspective and will record hopes, life prospects, and perception of control over the environment.
We expect to perform 15 to 20 interviews at a location agreed upon with the participant. Interviews will be administered until data saturation is achieved. The participants will be contacted directly by telephone or through NGOs, each interview will last 60 to 90 min, and the participants will be remunerated. A second meeting with each participant will be scheduled when necessary. Interviews will be audio-recorded for subsequent transcription. During the interview, researchers will also record notes in a field diary.
Data analysis
An analysis of the thematic and comprehensive content of the corpus of life stories will be performed. Given the study design, this will be a cumulative analysis of the different life stories.
An analysis of all transcripts will be performed by reading the life story of each participant. The richest life story, in terms of the description of the experience, will be selected and deconstructed into units and themes. Furthermore, a thematic analysis will be used to identify the transversal elements of the different stories.
After the topics are categorized, we will perform a comprehensive analysis of the corpus of stories. We will also analyze life trajectories to identify different factors involved in the relationship between homelessness and health status. Thus, a diachronic analysis will be used to establish a connection of the periods before and after homelessness and to determine how this connection developed over time. The fieldwork notes will be used to complement the transcripts by providing records of non-verbal cues, context, and/or emotions.
To guarantee methodological rigor, data triangulation will be used to achieve saturation. Thus, two researchers will perform these methods using reflection and discussion. Then, a third researcher will discuss the results with health and social professionals. ATLAS.ti version 9.0 will be used to manage these data.
Integrating quantitative and qualitative data
First, we will integrate the results from quantitative data to design the qualitative interviews. In particular, we will select specific issues that were not well described in the quantitative phase for inclusion in the qualitative phase. Secondly, we will triangulate the quantitative and qualitative data to improve our understanding of homelessness and health, will increase the validity of the overall findings by using different perspectives, and will capture convergence and divergence using clear and reproducible rules [53]. Finally, we will discuss the main findings from both designs by integrating the qualitative and quantitative data.