We aim to collaboratively develop (1) the forms of delivery of IMPACTFrail’s core components and (2) implementation strategies to support implementation. This collaboration will involve the research team, health care administrators (clinic leadership teams with decision-making authority), and frontline health care providers, including doctors, nurses, and allied health professionals.

Prior to this study, the research team conducted 6 months of informal stakeholder engagement (total of 22 interviews and focus group) with the clinics’ leadership, frontline health care providers, and older adults. Overall, we aimed to explore their awareness of frailty and IC, elicit anticipated barriers and facilitators of implementing or participating in the program, and understand the contextual operational and logistical constraints at each clinic. Our study builds on the information from the stakeholder engagement. To elaborate, we conducted 3 interviews with 3 older adults who were existing patients of the clinics to elicit expected barriers to participating in IMPACTFrail and to provide suggestions to encourage older people to attend IC screening. We conducted 7 interviews with 8 individuals who held mid-to-senior management positions in the clinics’ leadership team and 5 interviews and 5 focus groups with family medicine doctors (n=10), nurses (n=10), allied health professionals (n=2), care coordinators who assist doctors and nurses (n=4), and nonmedical support staff (n=6). Health care professionals were across different levels of seniority and whom we anticipated to be involved across different steps of IMPACTFrail’s patient journey. Sampling was guided by using the principle of sufficiency [51,52]. Additionally, we interviewed 2 private general practitioners to gain some insights from the private sector on implementing an IC detection and management program in a primary care setting.

The individual or paired interviews were for health care administrators and older adults to encourage open sharing, particularly in light of potential power dynamics and different language and cultural backgrounds. Focus groups with multidisciplinary teams were to gather different perspectives on the feasibility of the program. Interviews with older adults were conducted in their preferred language, either English or Chinese, by multilingual research team members. Engagement sessions with health care administrators and frontline health care providers were conducted in English. The engagement aimed to explore participants’ awareness of frailty and IC, elicit anticipated barriers and facilitators to implementing or participating in the program, and understand the operational and logistical context at each clinic. The study design and implementation were informed by insights through informal stakeholder engagement and guided by complex intervention research and implementation science frameworks.

The core functions of IMPACTFrail (the “what”) are informed by Singapore’s National Frailty Strategy, while the forms of delivery (the “who” and “how”) will be collaboratively developed with stakeholders in this study. Examples of specific output from this co-development stage may include: (1) IC screening questions adapted from the original WHO ICOPE screening tool (see Multimedia Appendix 2) tailored to each clinic’s clinical protocols and preferences; (2) modifications to electronic medical record systems to capture new data required by IMPACTFrail; (3) clear patient journey pathways; (4) identification of relevant clinic staff (doctors, nurses, allied health professionals, clinic coordinators, and support staff) and delegation of roles; and (5) capability building plans to prepare for the upcoming feasibility study.

We will select, specify, and develop implementation strategies to mitigate anticipated barriers and leverage anticipated facilitators to implementation. Implementation strategies are defined as “methods or techniques used to enhance the adoption, implementation, and sustainability of a clinical program or practice” [45]. As we will conduct a smaller scale feasibility study in the next stage, each clinic will only implement a tailored set of 2-5 implementation strategies. These implementation strategies could be from the list of the 73 ERIC strategies or created inductively [32]. Each clinic could select and develop a different set of implementation strategies to align with their specific contexts.

The anticipated barriers and facilitators identified during the informal stakeholder engagement will be mapped to the CFIR determinant constructs. Subsequently, we will use the CFIR-ERIC matching tool to match CFIR-identified determinants to ERIC strategies [53,54]. The research team, polyclinic leadership, and frontline health care providers will collaboratively review the identified barriers and facilitators, the output from the CFIR-ERIC matching tool, and local contextualized considerations (eg, clinic operations, manpower) to prioritize and select implementation strategies. Strategy selection will be guided by 5 literature-informed criteria: relevance, opportunity for improvement, feasibility, level of difficulty, and validity [55]. The selected implementation strategies will be specified using Proctor framework for reporting and specifying implementation strategies across 6 domains, including actor, action target, action, temporality, dose, and implementation outcomes [45], and subsequently developed for operationalization in the feasibility study [46,56].

The feasibility study is defined as a small-scale assessment aimed at determining whether IMPACTFrail can be implemented at this level, evaluating its potential for scalability, and identifying how it might be adapted for broader implementation [46,56]. In other words, we seek to evaluate the readiness of implementers delivering IMPACTFrail in real-world settings at a larger scale and the preparedness of older persons to participate in the program [30,47,48]. This protocol precedes a subsequent hybrid effectiveness-implementation study that will be conducted at a larger scale with more sites and older adult participants [57,58]. The study was registered on ClinicalTrials.gov on December 22, 2024 (Identifier: NCT06753643).

This will be a single-arm feasibility study conducted at five public clinics across all 3 regional health systems in Singapore. Specifically, 2 clinics were selected from each of 2 regional health systems, and 1 from the third System. Each clinic has contextualized barriers and facilitators to implementation, clinic workflows, and infrastructure. Thus, each clinic is conceptualized as a single site for feasibility testing. This study does not include a comparator arm or a pre-post design, as it does not aim to assess effectiveness. Instead, it focuses on a small-scale feasibility assessment to inform the design and implementation of a subsequent, statistically powered hybrid effectiveness-implementation. The phased approach increases the likelihood of successful real-world implementation by enabling early refinement and adaptation in practice settings.

Eligibility criteria include (1) registered patient of the clinic, (2) aged 60 years and older (3 of 5 clinics) or aged 65 years and older (2 of 5 clinics), (3) mildly frail, assigned by a trained health care provider as CFS score 4 or 5, and (4) able to provide informed consent independently or through a legally authorized representative. Exclusion criteria include (1) administered comprehensive geriatric assessment in the past 12 months or (2) at the time of this writing, being under the care of a geriatrician or patient of a geriatric specialist outpatient clinic.

Older adults who meet these eligibility criteria will be given information on the IMPACTFrail health care program and the research study verbally and via a hardcopy brochure by a paraprofessional care coordinator or nurse. A trained research assistant will obtain written informed consent from each older adult individually and address any queries from potential participants. Consent-taking to be recruited as a participant of the IMPACTFrail study will be conducted prior to intrinsic capacity screening. Recruitment will be conducted over approximately 9 months, continuing until each clinic reaches its target of 36 participants. All older adults will only need to go through this process once. Additionally, a subset of these older adult study participants of IMPACTFrail who participate in in-depth interviews of the qualitative study will undergo a separate informed consent process before their interviews. This informed consent-taking will be facilitated by trained researchers.

This set of consistent and unified inclusion and exclusion criteria will facilitate meaningful comparison of findings across sites and generalizability. Importantly, the broad objective is to translate the National Frailty Strategy to real-world practice. To facilitate this, the intervention will be structured around core functions that are common across all 5 participating clinics, including Clinical Frailty Scale score assignment, IC screening for mildly frail older adults, further clinical assessment, tailored referrals to clinics, specialists or community services, and monitoring and follow-up. While the form of implementation may vary across clinics, such as the precise age cut-offs for inclusion, further clinical assessment tools, and referral workflows, the underlying core functions remain consistent and intentional. This approach is aligned with the core functions and forms paradigm, recognizing that complex interventions must adapt to context while preserving their core purpose [59-61].

For the 2 of 5 clinics that will apply a slightly higher age cut-off of 65 years (instead of 60 years) for inclusion, it serves to facilitate integration of IMPACTFrail’s frailty screening with existing health programs and clinical protocols. They anticipated operational challenges and a lack of manpower to intentionally identify older adults 60-64 years old for the feasibility study.

Additionally, 2 clinics under the same regional health system have additional inclusion criteria. These criteria include (1) enrolment in Healthier SG [56], (2) absence of dementia diagnosis, and (3) active chronic disease appointments for any one of the following chronic conditions: Type 2 diabetes mellitus, chronic ischemic heart disease, congestive cardiac failure, osteoporosis, Stage 4 chronic kidney disease, and chronic obstructive pulmonary disease. These criteria are driven by contextual operational considerations within these clinics, including manpower capacity, logistics, and the need to align with ongoing preventive and management care programs for patients with chronic conditions. This more targeted approach was assessed to be more acceptable and pragmatic for small-scale feasibility testing in settings with constrained resources, while keeping the same core frailty screening, IC screening, and management functions.

Overall, this study protocol serves the purpose of disseminating knowledge on co-developing a health care intervention and implementation strategies, with built-in flexibility to accommodate system-level differences across clinics. In line with implementation research, we will subsequently assess how clinics could gradually align in eligibility criteria and protocols as IMPACTFrail scales up nationally.

No formal sample size calculation is performed, as this is a feasibility study not designed to be statistically powered for assessing effectiveness and health outcomes. There is no consensus in the literature on the required sample size for pilot and feasibility studies [62]. Recommendations of sample size vary depending on the primary objective of the study, the feasibility parameter being evaluated, or clinical scale evaluation [63-66]. Published reviews from the United Kingdom have reported median sample sizes of 30 and 36 participants per arm in feasibility studies [67,68]. In addition, clinics leads were also consulted on the practical projections of recruitment and uptake, taking into account the study timeframe and available resources. Hence, our study aims to recruit 36 older adults per clinic. This was computed based on targeting a minimum of 30 participants per site while accounting for an anticipated 20% loss to follow-up (eg, nonattendance at follow-up assessment post screening). Each clinic is conceptualized as an individual site due to contextualized barriers, workflows, and infrastructure.

Table 2 describes the four feasibility assessment criteria: (1) implementation, (2) acceptability, (3) practicality, and (4) adaptability, along with the corresponding study components mapped to each criterion. The criteria are informed by the literature on feasibility studies [47]. In summary, we prioritize the need to assess the ability to carry out the new screening and integrating the core functions with existing clinical protocols (implementability), how a satisfying or attractive a new program that focuses on domains beyond the conventional chronic conditions is to health care providers and older people (acceptability), whether the resources can support such a new program (practicality), and the degree it can be improved or changed if we progress beyond the pilot (adaptability) [47]. The criteria collectively inform us on the feasibility to progress from the small-scale pilot study to a main study that will be statistically powered to evaluate effectiveness and implementation at scale.

We will collect and document a comprehensive list of quantitative process indicators to measure the processes of care (Table 3). This addresses the research question on whether the program activities are delivered as planned and informs us on the follow-through of patients for the entire patient journey.

The sample includes all recruited older adult participants. Each of the 5 clinics will collect and provide data to the research team. Table 3 summarizes the list of process indicators.

We aim to assess the fidelity of the selected implementation strategies.

We will monitor and track the fidelity of operationalizing each selected ERIC implementation strategy across the specified domains (actor, action, action target, dose, or temporality), using a comprehensive set of indicators [45]. In other words, we will assess whether each strategy is delivered in the feasibility study as originally intended during the co-development phase.

Indicators will be collaboratively designed by the research team and clinic implementation leads. Data will be collected regularly by clinic implementation leads and frontline implementers. The research team will systematically document any deviations from the planned implementation and discuss them with clinic implementation leads. Adaptations to strategies, if any, will also be documented. For example, in the case of the implementation strategy involved educational sessions for capability building on IC screening, we will assess whether the planned elements, such as the trainers (actor), attendees (target), and the frequency and duration of training (dose), are delivered as intended. If discrepancies arise, we will explore whether fidelity can be improved or whether adaptations are needed. In addition, we will focus on sustainability and scalability, such as whether the frequency and dose of the strategies are sustainable within the clinic and adaptable for scaling up to other clinics.

Broadly, we aim to qualitatively assess:

Study participants will include older adult participants, clinic implementation leads who oversaw the planning and execution of IMPACTFrail, and frontline health care providers involved in the program’s core functions (doctors, nurses, care managers, care coordinators, allied health professionals, and support staff). We will conduct purposive sampling and sample sizes will be guided by the concept of sufficiency [51,52]. First, we will include older adults who participated in IMPACTFrail. We aim to interview a minimum of 3-4 older adults per clinic, with a total of at least 15 participants (approximately 10% of the targeted 36 recruited participants per clinic). We will sample the older adults using maximum variation for representation across demographic variables such as sex, ethnic groups, and age groups. Second, the qualitative study component will include clinic implementation leads who oversaw the planning and execution of IMPACTFrail and frontline health care providers involved in the program’s core functions (doctors, nurses, paraprofessional support staff). We anticipate only a small number of health care workers being involved in this small-scale pilot study. Hence, we will aim to sample all implementation leads at each clinic, estimated to range from 2 to 4 individuals per site. Additionally, we aim to sample frontline health care workers across every occupation who was involved across each central function of IMPACTFrail. Nurses and paraprofessional support staff in the qualitative study would need to have conducted IC screening for at least 5 older adults and doctors would need to conduct follow-up clinical assessments to assess intrinsic capacity deficits for at least 5 older adults. Interviews and focus groups with clinic implementation leads and frontline health care providers will be conducted in English, while interviews with older adults will be in either English or Chinese.

As mentioned, both data collection and data analysis will be guided by the 4 feasibility criteria, the updated CFIR and its Outcomes addendum, TDF, and ERIC taxonomy [31,32,49]. We will expand CFIR’s Individual Characteristics domain by mapping determinants of capability, opportunity, and motivation to TDF domains [69,70]. The ERIC taxonomy informs the refinement of existing implementation strategies and selection of new strategies [32]. The semistructured topic guides elicit data that inform the 4 feasibility criteria, CFIR-based and TDF-based barriers and facilitators, and suggestions on new ERIC strategies. For example, for the criteria on implementation, we will obtain perspectives from implementers on the barriers and facilitators to delivering IMPACTFrail’s core components and implementation strategies (eg, mobilizing program champions, attending training), and perspectives on the extent that the strategies enabled the implementation of IMPACTFrail. Older adults will share about the barriers and facilitators of participating in IMPACTFrail and receiving implementation strategies (eg, patient educational materials) and the extent to which the strategies have facilitated their participation.

We aim to conduct 2 rounds of 14 interviews with clinic implementation leads across all 5 clinics. The first round of interviews will be conducted within the first 2 months of IMPACTFrail’s roll out. This will inform early implementation barriers to refine its delivery and inform additional support required. The second round of interviews will be conducted toward the last 4-5 months of the study and after the first 20-25 older adults have been recruited. This is to elicit their perspectives on how implementation has progressed, new barriers and facilitators that arose, and insights on improving the program. Additionally, we will conduct interviews or focus groups with frontline health care providers involved in delivering IMPACTFrail’s core functions. In the first instance, we will aim to conduct focus groups, as opposed to interviews, among the health care providers who worked as a team, if scheduling allows.

Deductive analysis will be guided by the frameworks and concurrent inductive analysis will reveal emergent barriers, facilitators, and themes. Two trained research team members will code verbatim transcripts independently and subsequently reconcile discrepancies. Framework analysis will be conducted [71]. This allows timely charting of data into a framework matrix, generation of themes by linking conceptually related codes, and directly addressing the feasibility criteria (Table 2). We will conduct a thematic analysis using data collected across all 3 regional health systems, as we anticipate cross-cutting barriers, facilitators and learning points across study sites. We will also identify contextual insights specific to each regional health system. This analytic approach allows us to synthesize cross-site patterns with a larger amount of data, while accounting for differences in organizational context, and directly address the feasibility of implementation across settings. A pragmatic approach of concurrent data collection and analysis will be adopted to ensure timely reporting of findings to clinic leadership and stakeholders to support the preparation of the scaled-up phase [72-75]. Microsoft Excel and NVivo 12 (QSR International) will be used for data management.

We aim to estimate the cost of developing and delivering the program and implementation strategies, with the purpose of informing the amount of resources required to sustain and to expand the program beyond the feasibility study.

First, we will compute developmental costs, which comprise (1) developing IMPACTFrail’s core functions, and (2) developing each implementation strategy. The expenditures on manpower (eg, health care providers involved in the co-development phase) and other expenses (materials, printing, etc) will be aggregated. Second, program costs will be derived from the costs of (1) delivering IMPACTFrail’s core components, and (2) operationalizing each implementation strategy. The expenditures on manpower, equipment, and operating expenses will be aggregated.

A time-driven, activity-based micro-costing approach will be used to estimate costs related to manpower resourcing [76,77]. Specifically, this will be derived from the unit cost multiplied by the intensity of use and duration of use. The unit cost will be estimated based on financial records or the standardized workforce unit cost for implementers obtained from the clinics. The intensity and duration of use will be derived from self-reported time spent for a select list of activities by implementers to deliver the program [78]. We will only use financial statements or costs self-reported by implementation leads to obtain unit costs on equipment and other operating expenses and their intensity of use.

Within each site, we will present the cost associated with (1) developing IMPACTFrail’s core functions, (2) developing each implementation strategy, (3) delivering IMPACTFrail’s core functions, (4) delivering each implementation strategy, (5) total developmental cost, (6) total program cost, and (7) total costs.