2.1. Recruitment and sampling

The study was conducted in an Irish primary care clinic for PEH. Over a five-month period between 13^th August 2025 until 17^th January 2025. All adults presenting to the clinic were screened for eligibility for the feasibility trial by two of the study authors (TC and JT), both general practitioners working at the site.

Inclusion criteria were:

• Aged ≥18 years at baseline
• Clinical Frailty Scale (CFS) score between 3 and 6 (pre-frail to moderately frail) [37]
• Current living circumstances classified under the “roofless,” “houseless,” or “insecure” categories according to the European Typology of Homelessness and Housing Exclusion (ETHOS) classification system for homelessness and housing exclusion [38]
• Capacity and willingness to provide informed consent

Exclusion criteria included:

• Need for emergency medical care
• CFS < 3 or >6
• Concurrent malignancy
• Chronic kidney disease (CKD) stage 3 or 4
• Coded diagnosis of severe dementia
• Inability to provide written informed consent

Participants were informed that they would be invited to a post-intervention interview as part of a process evaluation, and consent for this was obtained at the outset. Those who were followed up in person after completing the trial were eligible for interview and invited to participate, representing a convenience sampling approach.

2.2. Data collection

Data were collected through one-to-one semi-structured interviews, conducted in a private space at the clinic by a single interviewer (TC), who is a GP and the trial’s principal investigator. TC had an existing clinical relationship with participants, which was acknowledged during the interview process and addressed through reflexive practice (see S2 Table).

Interviews were conducted face-to-face and audio-recorded with consent. Field notes were taken where appropriate. At the end of each interview, participants were asked to confirm whether they consented to their data being used in this secondary analysis. Interviews typically lasted around 10 minutes. This shorter duration reflected the constraints of the clinical environment and needs of the patients, where participants were attending for healthcare support and time was limited. Repeat interviews were not conducted.

The interview topic guide (S3 Table) included questions related to participant experiences of the intervention, perceptions of frailty and resilience, the impact of homelessness on frailty, and views on how frailty might best be managed. The guide was not formally pilot-tested but was reviewed and refined by the research team to ensure clarity and relevance. Semi-structured interviews were chosen to allow flexibility in phrasing and sequencing of questions, facilitating deeper engagement with individual participants [39].

All interviews were transcribed verbatim by the principal investigator. Transcripts were pseudo-anonymised using unique study IDs and stored securely on password-protected institutional servers. Transcripts were not returned to participants for review, as this was considered impractical given the study context.

2.3. Data analysis

Participant characteristics, including age bracket, gender, baseline Clinical Frailty Scale (CFS) score, opioid agonist treatment (OAT) status, recent illicit drug use (within the past six months), and ETHOS homelessness category, were recorded.

Interview transcripts were analysed using thematic analysis, following Braun and Clarke’s six-step framework [40]. NVivo software supported the organisation and coding of the data. Codes were developed by TC and grouped into related categories. This was reviewed and refined through regular research meetings with co-author JT. Categories were refined into three overarching themes and associated sub-themes. This coding structure was developed iteratively during analysis to reflect the recurring patterns and relationships in the data. Themes were iteratively applied to subsequent transcripts and modified as needed.

A process of constant comparison, reflection, and engagement with raw data was used to ensure that the themes accurately captured participant views. As this was a secondary analysis of a process evaluation, the interviews were not originally designed to reach thematic saturation relating to participants’ views toward frailty and resilience. However, themes related to these concepts became consistently evident across the dataset. As such, although the sample was not expanded until saturation was achieved, no new frailty- or resilience-related codes were observed to emerge in the later interviews, suggesting adequate coverage of the topic within this secondary analysis. Participants did not provide feedback on findings, as this was not feasible due to the secondary nature of the analysis.

3.1. Frailty and homelessness

The experience of homelessness, as expressed by participants, was seen to both contribute towards and magnify the impact of frailty. In the interviews, participants outlined how their physical and mental health deteriorated from living without stable housing and resulted in challenges for them in achieving better health and well-being. This illustrated the ongoing cycle of vulnerability that interviewees experienced, whilst also highlighting both the visible and invisible struggles they face due to their homeless status.

3.2. Survival

Despite the overwhelming challenges of frailty and homelessness, participants frequently demonstrated resilience in their ability to adapt and survive. This was often a necessity in a response to their unstable living situation, whilst survival was also borne out of early life experiences such as childhood trauma. The support of others was seen as key to surviving, amongst interviewees, and psychological supports were frequently felt to be integral in combating frailty for PEH.

3.3. Systemic neglect

A frequent sense of neglect was a recurrent theme in the accounts of interviewees. This included deficiencies in available emergency accommodation services, inadequate mental health supports and policies that many felt left them behind. These factors were identified as contributing to frailty and diminishing resilience and were highlighted as areas for improvement to mitigate the effects of frailty.

4.1. Principal findings

This study is the first to explore the views and experiences of PEH who are pre-frail and frail towards frailty and resilience. Participants described homelessness as both contributing to and compounding frailty, through pathways such as exposure to violence, malnutrition, poor sleep, substance misuse, and limited access to healthcare. These accounts suggest that frailty among PEH may be influenced by cumulative life course exposures, including prolonged housing instability, adverse childhood events, repeated trauma, and disrupted access to institutional support, rather than chronological ageing alone.

Mental health was often central to participants’ understanding of frailty, which could represent a consequence of homelessness and as well as a routes to homelessness. In this way, frailty was not framed purely as a physical state but one shaped by complex psychosocial challenges. Participants also expressed how difficult living environments, such as hostels or street settings, could accelerate physical and psychological decline. These narratives suggested possible biographical pathways into frailty, where early-life adversity, inadequately treated mental illness, and prolonged substance use may interact with homelessness to accelerate vulnerability and functional decline.

Despite these compounding risks, participants demonstrated notable resilience, often drawing on internal coping mechanisms and support from peers, keyworkers, and addiction services. Yet this resilience frequently occurred in spite of, rather than because of, systemic supports. Barriers to accessing appropriate services, including mental health care, secure housing, and tailored health interventions were frequently cited and contributed to feelings of neglect and reduced the capacity to overcome frailty.

4.2. Comparison to existing literature

Previous research has reported that older people often regard frailty negatively, viewed as an inevitable and irreversible state of decline [28,29,31,41], which may undermine engagement with interventions aimed at delaying or reversing frailty [42]. In our study, frailty was similarly described by participants in adverse terms, closely associated with weakness, vulnerability and accelerated ageing. Conversely, resilience was perceived positively, consistent with prior research describing it as a desirable attribute linked to physical strength and wellbeing [24,30].

Qualitative research with PEH has consistently identified barriers to healthcare access [8,10]. Our findings reinforce this, particularly in relation to mental healthcare. Participants frequently interlinked homelessness, frailty and mental illness. This reflects wider evidence showing elevated frailty in those with serious mental illness [43,44], and the role of social isolation in the development of frailty [45].

Consistent with prior research [46,47] participants described emergency accommodation as frequently inappropriate, particularly for individuals with complex needs. Inadequate support, exposure to substance use, and lack of privacy or safety were identified as barriers to recovery, aligning with studies documenting how hostel environments can undermine addiction treatment efforts [48,49].

Oral health challenges were also prominently reported. This supports research showing that PEH experience disproportionately poor dental outcomes [50], often due to limited access to care [51], and the impact of substance use, such as methadone-induced xerostomia, on oral health [52]. Correspondingly, participants in this research reported issues with dentition and illustrated how this could be linked to frailty.

4.3. Strengths and limitations

All interviews were conducted by one interviewer which aided in consistency. Having prior knowledge, clinical experience and understanding of frailty helped inform the interviewer in interpreting and responding appropriately to participants’ perspectives. However, the interviewer’s role as a doctor was apparent to all interviewees and may have influenced the responses due to perceived power dynamics [53].

Participants were recruited from a feasibility trial of an intervention designed to reverse frailty and build resilience, which may have introduced selection bias in the sample. Notably, a number of individuals enrolled in the trial were lost to follow-up or declined to take part in the qualitative interviews. However, the final sample included a range of participants across age, gender, frailty status, and substance use backgrounds. As recruitment occurred within a healthcare-based setting, the sample is likely to over-represent individuals who are already engaged with healthcare services, which may limit generalisability to the wider homeless population. Nevertheless, recruitment was done during routine healthcare appointments, which means the sample reflects patients in this real-world setting.

Additionally, participants’ involvement in the trial itself may have shaped their views on frailty and resilience, compared to what might have been expressed in a pre-intervention context. These factors limit the generalisability of the findings. Nonetheless, the study addresses a significant evidence gap regarding frailty in PEH, a population often underrepresented in research and provides new insight into the topic.

Lastly, transcripts were not returned to participants for comment or correction, and member checking was not conducted, representing a limitation in confirming participant interpretations. However, this decision was influenced by the transient nature of the study population, the secondary analysis of the data, and the constraints of the clinical environment in which data collection took place.

4.4. Implications for policy, practice and future research

This research demonstrates the structural barriers PEH face in attempting to overcome frailty. Living in homeless hostels or sleeping rough was frequently associated with disrupted sleep, contributing to a cycle of worsening mental health and substance misuse [48]. Moreover, such precarious living situations often prevent access to cooking facilities to produce nutritious protein-rich foods to delay or reverse frailty, whilst the risk of violence may impact upon home exercise programmes, both proven treatments in the management of frailty [54–56]. Interventions that fail to consider lack of access to institutional support, exposure to violence, mental illness, substance use and duration of homelessness in this group may have limited effectiveness. This illustrates the foundational importance of safe and secure housing to both protect health and enable engagement in evidence-based frailty interventions.

This study also identifies potentially under considered contextual factors for addressing frailty among PEH. Poor dentition and reduced access to cooking facilities, described by several participants, may significantly limit the ability to consume appropriate foods and should be considered when determining nutritional strategies. In certain cases, this may justify the use of ONS, despite concerns about misuse in populations with substance use histories [57]. Additionally, improving access to dental care for this population should be prioritised in future health policy.

Given participants’ strong emphasis on the interplay between mental health and frailty, improving access to psychological and psychiatric services is vital. In particular, there is an urgent need to develop dual diagnosis services, which address co-occurring mental illness and substance use frequently experience by PEH [58]. Existing reports note that “there has been subdued response by policy-makers and services to providing effective health care to people with dual diagnosis in Ireland” [59], and access remains limited and inconsistent across regions [60].

While dominant frailty models largely focus on physical deficits, our findings underscore the importance of social and psychological contributors to frailty. Future research should consider the development of frailty assessment tools tailored specifically to the context and experiences of PEH, which may better capture the domains most relevant to this group, enabling more targeted frailty treatments. The Tilburg Frailty Indicator [61], for example, incorporates physical, psychological, and social domains and may represent a more holistic framework that could be adapted or applied more widely in this population. Such approaches may offer more suitable means of detecting frailty compared to existing instruments primarily designed for older adults [17].