Perceptions, needs and preferences of chronic disease self‐management support among men experiencing homelessness in Montreal

Abstract Objective This study explored the perceptions, needs and preferences for chronic disease self‐ management (SM) and SM support among men experiencing homelessness. Design A qualitative interpretive approach was used. Eighteen semi‐structured interviews were conducted with 18 homeless men with a chronic disease at an emergency overnight shelter of Welcome Hall Mission (WHM) in Montreal, Quebec. Interviews were audio‐recorded, transcribed verbatim and thematically analysed. Results The majority of participants perceived SM as important, described confidence to perform medical SM behaviours, and creatively adapted their SM behaviours to homelessness. Emotional SM was described as most challenging, as it was intertwined with the experience of homelessness. Three vulnerable groups were identified: (a) those with no social networks, (b) severe physical symptoms and/or (c) co‐morbid mental illness. The preferred mode of delivery for SM support was through consistent contacts with health‐care providers (HCPs) and peer‐support initiatives. Discussion and Conclusions Despite competing demands to fulfill basic needs, participants valued chronic disease SM and SM support. However, SM support must address complex challenges relating to homelessness including emotional SM, multiple vulnerabilities and barriers to forming relationships with HCPs.


| INTRODUC TI ON
Characterized by the World Health Organization (WHO) as an invisible global epidemic, chronic disease is a leading cause of mortality worldwide. 1 In addition to premature loss of life, a chronic disease diagnosis is associated with adverse health outcomes including reduced quality of life, increased emotional distress, loss of productivity and increased health-care costs. 2 A substantial burden to the Canadian health-care system, 51.6 % of Canadians have at least one chronic disease, and 14.8% have two or more. 3 Populations with low socio-economic status (SES) have higher rates of chronic disease due to limited access to protective factors like nutritious food, safe housing, secure employment and health services. 4 Most at risk among low SES populations are those experiencing homelessness, with as many as 85% reporting a diagnosis of chronic disease. 5 Among this population, risks for developing a chronic physical disease include a higher prevalence of substance abuse, mental illness, exposure to the elements, poor sleeping and eating conditions, increased exposure to violence, and social stigma. 6,7 Self-management (SM) is one of the six components of the Chronic Disease Model, 8 and 'relates to the tasks that an individual must undertake to live well with one or more chronic conditions'. 9 Hence, SM is not only a series of tasks, but includes the notion of 'confidence'. SM tasks can be categorized as medical management (eg monitoring symptoms, managing side effects), role management (eg maintain activities of daily living, creating a community of peers) and emotional management (eg dealing with shock, making sense of illness). [9][10][11] Key patient SM behaviours in carrying out these tasks include the following: problem-solving, action planning, decision-making, locating and using resources, forming partnerships with HCPs and self-monitoring. 12 SM support is 'the systematic provision of education and supportive interventions to increase patients' skills and confidence in managing their health problems, including regular assessment of progress and problems, goal setting and problem-solving support'. 13 SM support involves the collaboration between individuals with chronic diseases and a multidisciplinary team of health-care providers through structured programmes and interventions to develop techniques and capacities to improve the management of chronic diseases. 14 SM support is an effective strategy for improving disease outcomes, including appropriate use of health-care services, medication compliance, symptom control, quality of life, self-efficacy and motivation. [15][16][17][18][19] Yet, despite evidence of the efficacy of SM support, research among low SES populations remains relatively scant.
Scarcer still is research on chronic disease SM among populations experiencing homelessness.
A search of the literature by the authors yielded only three non-randomized, pilot or feasibility intervention studies examining chronic disease SM among populations experiencing homelessness. [20][21][22] The one-group, pre-post-test pilot study by Hulton et al 20 examined the feasibility of implementing the Stanford Chronic Disease Self-Management Program (CDSMP) 23 combined with nursing case management among ten sheltered homeless adults with a chronic illness. Six out of the ten participants completed the programme, with positive changes in health outcomes noted (notwithstanding the small sample size). The feasibility study by Henwood et al 21 also explored the acceptability and feasibility of the CDSMP 23 for health promotion and this among 15 men experiencing homelessness. All participants were retained, which can potentially be attributed to the highly participative study design. Qualitative interviews with participants revealed that the programme was helpful in learning the skills needed to address health concerns and feeling more empowered to do so.
Also, peer-support and participatory elements emerged as central components in ensuring the intervention's success. Last, the pilot study by Savage et al 22 to assess retention of adults with diabetes and experiencing homelessness in a 12-week SM support intervention (focused on managing diabetes and obtaining needed recourses) combined with nursing case management found that of the three participants assigned to the intervention, two completed the intervention. A pre-post survey revealed improvement in some outcomes (eg cognitive symptom management); however, the very small sample size precludes any definitive conclusion.
Whereas the outcomes of these three studies are promising, their challenges with recruitment and retention may be partly explained by the dearth of exploratory research around the appropriateness of SM and SM support interventions among homeless populations. Furthermore, there is mixed evidence of the efficacy of adapting evidence-based SM interventions to divergent contexts, such as that of chronic disease SM interventions among populations experiencing homelessness. 24 Thus, further exploration of the basic acceptability and experience of SM support among this population is warranted. This study aimed to fill this knowledge gap by answering the question: what are the perceptions, needs and preferences of chronic disease SM support among men experiencing homelessness? Perceptions are defined as the ways in which something is regarded, understood or interpreted. 25 Needs are defined as a lack of something requisite, desirable or necessary for avoiding harm and can be revealed through barriers and challenges to accessing services. 26,27 Preferences are defined, in the context of this study, as that which is valued as the ideal choice in regards to health-care goods, services or interventions received. 28 Given these parameters, the objectives of this study were to explore: (a) the perceived importance of SM and SM support among participants, (b) the barriers and challenges participants experience in developing SM behaviours or receiving SM support and (c) participants' opinions about ideal SM support services and programmes.

| ME THODS
Ethics approval for the study was obtained from the McGill University Faculty of Medicine Institutional Review Board and the shelter where this study took place.

| Study design and setting
This study used a qualitative interpretive approach. 29 No other methodological orientation underpinned the study. Interviews were conducted between August and December 2016 at the emergency overnight shelter of a non-profit organization 30

| Sample and sample size
This study sought to recruit at least 12 participants, as this sample size has been found to often be sufficient in achieving data saturation in qualitative studies. 31 Inclusion criteria for participants consisted of the following: (a) men over the age of 18 accessing the overnight shelter; (b) who self-identified as living with a chronic physical disease, broadly defined as a non-communicable disease diagnosed for at least one year; (c) and spoke English or French well enough to participate in the study; and (d) had no known history of violence or inappropriate behaviour. There were no restrictions based on time spent homeless, type of chronic disease, or physical or mental co-morbidities. As a starting point, individuals' chronic physical diseases were categorized according to the types suggested by the Public Health Agency of Canada (A-Z Chronic Diseases). However, given our broad definition, we accepted what participants' identified as the chronic disease they were coping with, which included more than the typical chronic physical diseases.

| Recruitment procedures
Recruitment from the overnight shelter was coordinated by an intervention worker who identified which men were potentially eligible and then approached them (face-to-face) to provide a brief introduction to the study. Recruitment took place on nights where this was feasible for the intervention worker, given competing priorities. Individuals who agreed to participate were invited to meet with a Master's nursing student (first author) or a Research Assistant (RA) in a private room at the shelter, at which point more information about the study was provided. If participants were still interested in participating, the consent form was reviewed and participants were given time to ask questions before signing the consent form and commencing the interview. All participants were offered a choice between $5-10 gift certificates to a local coffee shop or grocery vendor.

| Data collection
Face-to-face, semi-structured interviews were conducted by the first author (female master's student in nursing) with each participant in either English or French. This was the first author's qualitative study. There were no a priori relationships between the interviewer and participants (participants did not know anything about the interviewer before the interview). Interviews took between 45 and 90 minutes, including a sociodemographic questionnaire administered at the end of each interview. All interviews took place at the overnight shelter between the 'check-in' hours of 15:30 and 18:00 and all (with the exception of one) were audio-recorded. Field notes were kept to record the interviewers' impression of the interviews.

| Data analysis
Data collection and analysis were conducted concurrently. All audiorecordings were transcribed verbatim and transcriptions were verified against original audio-recordings. Inductive thematic analysis was used as a method for identifying, analysing, and reporting patterns within the interview data through a process of coding. 32 The 1st author coded all the transcripts, and codes were discussed at regular meeting with the 2nd and 3rd author.
Codes, the most basic segment of raw data that can be assessed in a meaningful way regarding a phenomenon, 32 were identified by reading and re-reading transcripts and applying short sentences to describe the data extracted. Codes were then organized into categories according to their shared similarities. Themes were derived from the data by analysing patterns that emerged. 32 Transcripts were then re-reviewed to refine and clarify themes, ensuring coherence across interviews and generating clear definitions. 33 Transcripts were not returned to participants for comment and findings were not reviewed by participants. Figure 1 provides an example of the data analysis process. A summary of strategies used to ensure study rigour is available in Table 1. Microsoft Office Word and Transcribe were used as assistive software for data analysis and management.

| Sample description
The number of people approached by the shelter's staff is not known, but all men then referred participated in the study. Data saturation was reached with 18 individual interviews. Table 2

| SM important, despite competing demands to fulfil basic needs
Most participants (72.2%) agreed that chronic disease SM was important, despite the challenges to SM presented by homelessness (see Figure 2, perceptions). When comparing transcripts, no patterns were noted in the importance attributed to chronic disease SM according to disease, number of co-morbidities, symptom severity or presence of co-morbid mental illness. However, participants described daily life experiences that impinged on their ability to SM, including exposure to physical danger, harassment, theft, a lack of privacy, limited access to communication technologies, limited food choices, poor sleeping conditions, limited income and a lack of storage. As one participant said, 'It's not an easy life. People say that you're living for free, you're not … it's a lot of work (participant 8)'.
Several participants described having to choose between attending medical appointments and accessing a meal serving, or an employment opportunity.

| Participants confident to carry out and adapt SM behaviours
Almost all participants (88.9%) expressed confidence to carry out medical SM behaviours and described performing a number of SM behaviours (described in Table 3 Another participant kept his medications inside an artificial plastic egg in his pants pocket, which served the dual purpose of avoiding theft and helping him to remember to take his pills. Other participants employed similarly adaptive SM behaviours to save money, reduce risks of infection, and minimize disease symptoms like fatigue, muscle pain and headache (See Table 3).

| Emotional SM is most challenging
Despite participants attributing importance to SM and expressing confidence to perform and modify their SM behaviours, managing negative emotions pertaining to the difficulties of chronic disease while homeless emerged as the aspect of SM that was most challenging (see Figure 2, SM needs). Many participants (61.1%)

| Emotional SM behaviours tied to presence of interpersonal relationships
When comparing transcripts, differences in emotional SM emerged based on the presence or absence of interpersonal relationships (see

| SM needs high for participants with severe physical symptoms and co-morbid mental illness
Two other subgroups of participants emerged as having high SM needs (see Figure 2, SM needs). First, participants with 'severe' physical disease symptoms (44.4%) described using almost twice as many medical SM behaviours as those with 'non-severe' symptoms. Physical symptoms were coded as 'severe', if participants

| Preferences for consistent, supportive HCP
Participants expressed a preference for consistent HCPs and peer-support SM (see Figure 2, preferences Seeking medical support (eg seeing on-site nurse or going to emergency room) Advocating for better care (eg asserting right to information about medical procedure) Sharing information with provider (eg carrying wireless cardiac rhythm device)

Making decisions
Monitoring symptoms (eg monitoring for signs of recurrent inflammation) Analysing symptoms in light of acquired knowledge (eg linking asthmatic chest pain to increase in smoking) Weighing pros and cons (eg living with joint pain vs. post-operative recovery) Deciding on action (eg going to emergency room when symptoms understood as critical) Taking medications (eg over-the-counter, prescribed and illicit medications) Linking mental state to physical well-being (eg managing anxiety to reduce physical pain)

Context adapted SM behaviours
Minimizing symptom impacts Improving sleeping conditions (eg seeking out a particular mattress) Reducing migraine pain (eg using earplugs to block out sound) and negative past experiences with the healthcare system (38.9%).
Perhaps the most deterring negative experience described by 22.2% of participants was the perception of prejudice on the part of healthcare workers, which undermined their confidence in the healthcare system and HCPs in general. As one participant described, They [doctors] size you up and they're like … this guy is a street guy … And if he's a street guy than he's here for drugs … they basically blow off everything you say (participant 9).
About a third of participants (33.3%) described utilizing the services of the shelter's on-site nurse. All of these participants expressed satisfaction with this service and described the nurse as their first line of medical assistance in lieu of attempting to see a physician.

| Peer-support interventions
Half-way through the study, one participant reflected on his positive experiences with a peer-support therapy programme. He described the peer-support programme's many benefits: …it's good to have people to talk to … even if you knew people were just around … it just made you feel, sort of, better (participant 8).
Eight of the ten subsequent participants agreed that peer-support SM interventions would be helpful in managing the emotional challenges of chronic disease and homelessness because, unlike HCPs, peers have 'been through all of it (participant 12)'. Some participants reflected on the utility of sharing their experiences and strategies that they learned through many years of managing their chronic illness in the context of homelessness. Furthermore, peer-based SM interventions have been associated with increases in self-efficacy, reduced pain, fewer emergency room visits, 52 improvements in physical activity, greater smoking cessation, 53 increased satisfaction with medical care, increased social support, better mood and increased sense of belonging. 54 When taken with existing evidence of the importance of peer-support and social relationships on psychological and physical well-being [44][45][46]53 peer-support and social network-building SM interventions emerge as crucial strategies in developing SM behaviours among men experiencing homelessness.

| Study limitations
As with much research on SM, self-selection bias may leave out the most marginalized segments of this already hard-to-reach population. 55 There is also a risk of social desirability bias. 56 The study additionally failed to elicit sociodemographic data pertaining to individual's racialized identities, which may have implications for the experiences of prejudice described by some participants.

| Practice implications
Barriers to developing chronic disease SM behaviours and SM support must be addressed at various levels of the healthcare system.
At the governmental-level, coordination and consistent funding could improve access to SM resources, such as on-site health-care and social services. Among service organizations, eliciting input from service users on policy development may reduce barriers, like access to personal storage, meal options or exercise spaces.
Providing on-site health-care services outside of work hours may diminish competing demands between SM and employment.
Increasing awareness and support of nursing services may foster linkages between this population and the health-care system. At the level of HCPs and researchers, eradicating stigma and prejudice is paramount. The most vulnerable segments of this population require SM support interventions that address and support their complex medical and emotional needs, while simultaneously drawing upon their strengths.

| Conclusion
The importance placed on chronic disease SM by study participants, their confidence to SM and their resourceful adaptation of SM behaviours are reflective of the many strengths and resiliencies upon which SM support interventions may be built. However, the intersecting challenges of chronic disease, homelessness, social isolation, severe physical symptoms and co-morbid mental illness render some subgroup of this population particularly vulnerable to poor health and psychological outcomes. Whereas there is no substitute for SM support provided by a consistent and trusted HCP, participant's enthusiasm for peer-support SM interventions provides a promising start.

ACK N OWLED G EM ENTS
The authors would foremost like to thank the staff and board of Welcome Hall Mission for their continued support and generous assistance with this research study, in particular Jean-Michel Paquette and James Jean-Denis. Many thanks to Lisa Merry for her on-going support throughout the research process.

CO N FLI C T O F I NTE R E S T
The authors whose names are listed immediately below certify that

DATA AVA I L A B I L I T Y S TAT E M E N T
Due to the sensitive nature of the qualitative data collected in this manuscript, research data are not shared.