A multidisciplinary professionals’ focus group was conducted at F2UH with participants from internal medicine (n = 2), endocrinology (n = 1), geriatrics (n = 1), nursing (n = 1), pharmacy (n = 1), psychology (n = 1), nutrition (n = 3), biology (n = 1), digital health/biomedical engineering (n = 2), and architecture (n = 2). Professionals were recruited via purposive sampling (role/seniority and direct involvement in the care of older adults with multimorbidity at F2UH); the sample size was determined pragmatically to ensure multidisciplinary coverage and was deemed adequate by thematic sufficiency across four sessions. The FG met four times (Sept 2023–Mar 2024) and followed a semi-structured guide covering: (i) unmet needs relevant to multimorbidity, (ii) candidate digital functions and data/interoperability requirements, and (iii) barriers/facilitators and acceptability. Each session lasted approximately 90 min, was audio-recorded, and a rapporteur kept structured minutes on a standardized electronic template; outputs were member-checked by participants before consolidation. The verbatim prompts used with professionals and the stimulus technologies are provided in Supplementary material S1. Any disagreements on item wording or categorization were noted in the minutes and finalized in a subsequent consensus meeting; unresolved items were adjudicated by a third senior investigator.

To analyze the focus-group discussions, audio recordings were transcribed verbatim and examined using reflexive thematic analysis following Braun & Clarke’s phases: (i) familiarization; (ii) generation of initial codes; (iii) construction of candidate themes; (iv) review of themes against coded extracts and the full dataset; (v) theme definition and naming; and (vi) reporting. Two researchers independently coded all materials using an inductive, primarily semantic approach (allowing latent insights when warranted). A shared codebook was iteratively developed and refined; discrepancies were resolved by consensus, with escalation to a third senior investigator when needed. To support trustworthiness, we maintained an audit trail (audio files, structured minutes, versioned matrices), sought negative/contradictory cases, and conducted a member-check of key takeaways at the end of sessions. In line with a reflexive approach to thematic analysis, we did not compute inter-rater reliability coefficients (e.g., Cohen’s κ); rigor was ensured through double coding, consensus meetings, and transparent documentation. The codebook (labels and definitions) and a textual theme map are provided in Supplementary material S2; anonymized exemplar extracts can be supplied to editors upon reasonable request.

In this manuscript, we analysed SUNFRAIL+ screening data collected at the Federico II University Hospital (F2UH) site. Patients with multiple chronic conditions are more likely to receive multiple drug treatments (36); accordingly, many studies use two or more conditions (multimorbidity) and five or more medications (polypharmacy) as consistent endpoints (37). The Focus Group (FG) reviewed preliminary data from the multicentre SUNFRAIL+ study promoted by the PROMIS network (38), with contributions from eight Italian regions and coordinated by F2UH (39). The study aimed to validate an innovative model of frailty screening in community-dwelling older adults, supported by information technology (IT). The screening model adopts the SUNFRAIL tool, consisting of nine questions investigating the physical, psychological, and socio-economic domains (Table 1) (40). After written informed consent, screening data were collected by clinicians during outpatient visits across five F2UH departments (Adapted Physical Activity Prescription, Geriatrics, Neurology, Diabetology, Maxillofacial Surgery). A total of n = 113 older adults were enrolled at the F2UH site of the multicentre SUNFRAIL+ study (see Ethical approval, Section 2.4). Analyses focused on the polypharmacy subgroup (n = 54) and were strictly descriptive. Given this descriptive aim, no a priori power calculation was performed. To profile common characteristics, the FG selected older adults who answered “yes” to the question “Do you regularly take five or more medications per day?” and identified the frailty risk factors they presented. Only counts and percentages are reported; no inferential statistics (e.g., p-values, confidence intervals) were computed.

The European Commission’s Blueprint methodology is a patient-centred approach for the design of IT-supported interventions that highlight patients’ unmet needs to design key use scenarios (41). The health, social and technological needs of the target population are represented through a “Persona” use case, a single, specific, hypothetical patient, with a realistic name and a brief description of their needs, goals, hopes, dreams and attitudes (42). In addition, the use case reports the typical behavioral characteristics of the target population, based on the experience of the experts contributing to the definition. Based on the analysis of the risk factors resulting from SUNFRAIL+ questionnaire, a theoretical elaboration of the Blueprint persona was developed by the FG. The professionals involved in the FG answered iteratively to the questions in the Blueprint Persona Development tool (Supplementary material S3) and agreed on the final response. Persona development followed the European Commission’s Blueprint methodology. The process combined SUNFRAIL+ data with qualitative insights from the FG to build a realistic profile, including socio-demographic characteristics, health status, behavioural traits, and contextual factors. The persona (‘Gennaro’) was iteratively refined until consensus was reached among FG members. This persona represents an expert-informed synthesis step intended to structure design hypotheses. End-user co-design and validation were not undertaken at this stage and are planned as a subsequent phase within the broader UCD (User-Centered Design) pipeline.

Based on the Persona use case, the FG identified digital solutions that can help meet the needs of patients in healthcare and other domains, such as social, psychological, environmental, etc., to design a digitally supported intervention, integrated into the current service provision, from organizational and technological perspectives. Finally, the FG identified the dataset to be collected through digitally supported intervention. In particular, the FG focused on the type of data, the tools adopted for data collection, the setup and interoperability requirements. Such a minimum dataset helps to collect and manage data from multiple sources (EHR, medical devices, wearables, etc.), identifying the interoperability requirements that digitally supported interventions must provide.

The studies involving humans were approved by the local Ethics Committee of University “Federico II”—Azienda Ospedaliera di Rilievo Nazionale “A. Cardarelli” (N. 284/22). The research protocol was registered at ClinicalTrials.gov on 9 December 2022 (registration number: NCT05646472). The studies were conducted in accordance with local legislation and institutional requirements. Written informed consent for participation was obtained from the participants’ legal guardians or next of kin.”

The FG extracted and analyzed preliminary data of the SUNFRAIL+ study collected at the F2UH, the answers provided by n = 54 patients who declared they took 5 or more drugs. These data were not intended to provide inferential statistics, but rather to generate a descriptive profile of multimorbid patients with polypharmacy, which served to contextualize and enrich the identification of unmet needs in the subsequent qualitative phase. The mean age of this subpopulation is 74.68 (±6.01) and an age range between 66 and 95. More than half of the participants (55.5%) are aged 74 years or younger. n = 33 are women. Half of participants (50%; n = 27) have a low level of education. Only n = 5 (9.2%) had attained a university degree (Table 2). The most recurrent risk factor in this sub-population is the perception of memory loss (n = 24; 44.4%). This finding corresponds to the cognitive domain of the SUNFRAIL+ tool, highlighting early vulnerability that may interact with pharmacological complexity. This symptom could be linked both to the age-related loss of cognitive function and to the side effects of polypharmacy. In 40.7% of cases, participants reported a decrease in physical activity and the adoption of a more sedentary lifestyle, linked to health conditions and loss of independence. This result is aligned with the physical activity domain of SUNFRAIL+, reinforcing the link between reduced mobility, frailty progression, and loss of independence. The feeling of loneliness found in n = 16 participants (29.6%) highlight a vulnerable psychological state and a weakening of social relationships due to the loss of independence. This indicator belongs to the social domain of SUNFRAIL+, underlining how psychological vulnerability and social isolation interact with health status. 25.9% (n = 14) of participants declared that they had not been visited by their General Practitioner (GP) in the last year. This could be explained by difficult communication with GP, or a problem with the accessibility of doctor’s outpatient clinic. Finally, it could be a question of low trust in GP and inappropriate use of specialist services (Table 3). Overall, these descriptive results provided an empirical foundation across the main SUNFRAIL+ domains (cognitive, physical, social, and healthcare access). They guided the Focus Group in selecting the most relevant unmet needs for persona development and digital health solution mapping.”

Using the Blueprint method and drawing on the SUNFRAIL+ questionnaire results of the sub-cohort, the Focus Group (FG) constructed a fictitious but realistic profile, named ‘Gennaro’. This persona was inspired by common characteristics and unmet needs observed in patients with multimorbidity and polypharmacy, and was iteratively refined through multidisciplinary discussion. Gennaro is a 78-year-old man with obesity, hypertension, and poorly controlled type 2 diabetes, complicated by cardiovascular disease and a previous stroke, which reduced his autonomy and forced early retirement. Despite the support of his wife, he reports depressive symptoms and feelings of loneliness. Living in a suburban area, in a third-floor apartment without a lift and with poor transport connections, he struggles to access healthcare services and maintain social contacts. This persona highlights unmet needs in pharmacological management, physical activity, sociability, disease self-management, and psychological well-being. These needs were systematically linked to potential mHealth and digital solutions identified in the subsequent analysis (Figure 1).

Based on the persona profile and SUNFRAIL+ data, the FG systematically mapped digital health solutions to the identified unmet needs. The following categories summarize the main solutions, which were organized by domain and technology to ensure a clear link between needs and interventions.

The FG identified requirements for a digital intervention to support integration of the proposed solutions. Specifically, participants outlined the types of data, the tools for data collection, the implementation contexts, and the interoperability standards needed for a shared database capable of managing inputs from multiple sources (e.g., clinical settings, self-monitoring devices, and wearables). These elements are summarized in Table 5.

While the persona approach provided a useful synthesis of quantitative and qualitative findings, it remains a theoretical construct. In our study, patient co-design was not directly included, which may limit ecological validity. Future studies should integrate patients’ perspectives to refine and validate personas against real-life experiences. Despite this limitation, the efforts of the interdisciplinary focus group in the present study allowed us to identify the IT solutions and the dataset pivotal to implement a digitally supported intervention addressing the identified needs. This approach aims both to enhance cost-effective healthcare services (45) and facilitate multidisciplinary integration in the management of non-communicable chronic diseases (46).

According to FG insights, shared care plans and telemonitoring systems were identified as potentially supporting patients’ day-to-day self-management and potentially assisting healthcare professionals in treatment optimization, though this requires prospective evaluation. Adherence to multiple medications therapy in multimorbid NCD patients is challenging (47). In our FG, medication adherence emerged as a critical issue for multimorbid patients. Among the solutions identified, smart pillboxes and mobile apps were considered relevant to simplify complex regimens. This aligns with previous evidence showing their role in reducing adverse events and improving compliance (48). The FG also emphasized the importance of supporting clinicians in managing polypharmacy. Clinical decision support systems were identified as potential enablers, in line with prior studies showing their value in improving prescribing and deprescribing practices (49, 50). Diabetes represents one of the prevalent chronic conditions within multimorbidity scenarios (51). To effectively manage diabetes, non-invasive blood glucose monitoring may help empower patients to make informed decisions regarding dietary choices, physical activity levels, and medication usage (52). Maintaining glucose levels within the target range contributes to mitigating additional health complications associated with diabetes (45).

Proper nutrition is essential for managing chronic diseases and improving overall health. Mobile apps can provide tailored nutritional advice and reminders and may support adherence to dietary prescriptions. These tools facilitate regular monitoring and adjustment of dietary plans based on the patient’s health status and taste preferences (53). Sedentary behavior is linked to an increased risk of multimorbidity and falls among older adults, highlighting the importance of public policies promoting physical activity (54). Adapted physical activity programs, administered at home through a web-based platform interface, coupled with smartwatches or other wearable monitoring devices, may encourage physical exercise and may enhance the overall health status of patients in some contexts, subject to validation (55). Self-care is a dynamic and complex process that requires psychological adjustment of NCD patients to their medical conditions (56). Motivational processes and emotional self-regulation are relevant both to adapt to disease conditions and to manage and maintain self-care behaviors (57). Personal motivation is a key element for patients to be ready for changes in disease self-management. Motivational Interviewing (MI) may constitute a very appropriate intervention in those cases in which the behavioural changes (i.e., dietary changes, physical activity, etc.) are hindered by ambivalence, resistance and denial (58). MI enhance patient engagement and self-care by providing personalized tasks and goals, fostering a sense of autonomy and competence. MI contribute to address psychological barriers to lifestyle changes, such as fear of losing independence or depressive symptoms. Technology-delivered motivational interviewing interventions (TAMIs) offer considerable potential to reduce costs, minimize therapist and training burden, and expand the range of clients that may benefit from adaptations of MI (59).

Beyond psychological support, the FG also emphasized environmental and technological solutions that directly sustain autonomy in daily life. Our FG highlighted the relevance of Ambient Assisted Living (AAL) systems to promote autonomy and safety in older adults. This finding is consistent with studies on indoor monitoring and data integration (60, 61) and on outdoor navigation support and fall prevention (62, 63). Nonetheless, the integration of heterogeneous data sources remains a challenge, as underlined by El Murabet et al. (64). Last but not least, the built environment significantly influences the health and well-being of older adults, both in outdoor and indoor settings (65, 66). Outdoor activities, such as walking, shopping, socialising, and engaging in physical exercise—are essential to prevent functional decline in older adults. Providing an adequate functional mix within neighbourhoods, together with accessible parks and green spaces, creates opportunities for movement and spatial orientation in open areas, while alternative, comfortable, and safe mobility systems support high levels of walkability. In addition, the smart-home modification process for ageing in place involves not only structural refurbishment or the rearrangement of the housing layout to increase usability and safety, but also the installation of smart technology devices (67, 68). Therefore, qualitative and quantitative assessments of neighbourhood and home features are instrumental in identifying modifications that support ageing in place. Technologies such as fall-detection systems and geolocation apps may help enhance safety and support an active lifestyle by reducing fall and wandering risks, pending prospective validation. Next, steps include multi-site, cross-context studies to co-design and prospectively validate the mapped digital options, evaluating feasibility, usability, acceptability, and contextual fit across different healthcare systems.

To move beyond isolated pilots, digitally supported home care should be embedded within governance and reimbursement frameworks; otherwise adoption risks remaining fragmented and time-limited (69). Integration through platforms interoperable with the Electronic Health Record (EHR) would allow clinicians to prescribe digital interventions within existing care pathways and reduce siloed use. Although this study did not assess social prescribing of digital tools, such an approach could be promising for prevention and health promotion; however, organisational and cost evaluations are required before scale-up (70). Future work should also broaden clinical perspectives by involving additional specialties (e.g., vestibular, somatosensory, visual) to capture needs that influence independence and well-being in older adults (71–73). In low-resource contexts, staged adoption can prioritise low-cost, high-impact functions (e.g., medication reminders, basic vitals logging) delivered via SMS/IVR or USSD and offline-first workflows. A minimal dataset (only safety-critical fields) and interoperability through open standards reduce technical burden and vendor lock-in. Task-shifting to community health workers, short micro-trainings for patients/caregivers, and clear governance/reimbursement micro-paths (e.g., bundles or capitated add-ons) can improve feasibility. A lean monitoring plan (simple indicators for uptake, adherence, flags) supports iterative scale-up. As a single-centre Italian study, the present findings are context-dependent. The methodological approach (persona-guided, expert-informed mapping) is portable, but content and priorities will require local adaptation to reflect cultural norms, service organisation, and reimbursement rules.

This study has several limitations. First, the persona and digital options were developed through expert insight rather than direct patient co-design; co-design and end-user validation are necessary next steps. Second, qualitative data reflect a single-center, professionals-only focus group in Italy, which may limit transferability to other contexts. Transferability may also be constrained by cultural factors in aging and care-seeking, which were not explored in this study. Third, the quantitative component is descriptive and restricted to a polypharmacy subgroup (n = 54), and therefore does not support inferential claims. Fourth, although multidisciplinary, the focus-group composition may have under-represented some clinical domains (e.g., vestibular, somatosensory, visual), and priority-setting reflected resource constraints. Fifth, we adopted reflexive thematic analysis and therefore did not compute inter-rater reliability coefficients; rigor was supported via double coding, consensus meetings (with third-investigator adjudication when needed), an audit trail, member-checks, and a search for negative cases. Sixth, the study did not assess social prescribing or the economic/organizational implications (e.g., cost-effectiveness, reimbursement models) of implementation. Despite these safeguards, residual interpretive bias may persist; future work will include patient/caregiver co-design, multi-site validation, and holistic feasibility assessment.