A sequential mixed methods study design was chosen that used quantitative and then qualitative research techniques given the difficulty of evaluating a complex intervention and to maximize the breadth and depth of the research findings. The protocol was previously published [22], and the trial was registered prior to recruitment (NCT05943366). This study was reported in accordance with the COREQ (Consolidated Criteria for Reporting Qualitative Research) guidelines [25] to ensure comprehensive and transparent reporting of the qualitative methods and findings (Multimedia Appendix 1).

Ethics approval was obtained from the University of Manchester Research Ethics Committee on August 18, 2023 (reference 2023-17372-30569), and this study complied with the Declaration of Helsinki. Written informed consent was obtained from all participants prior to commencement of the research.

Participants were included in the study if they were aged 65 years and older, living independently in the community in Greater Manchester, and willing to use an iPad or tablet (their own or one provided by the research team) for the duration of the study. Participants were excluded if they were unable to communicate in English or had a known cognitive impairment.

Men and women aged 65 years and older living independently in the community were recruited on a rolling basis through purposive sampling via local Age UK groups, assisted living facilities, and university networks. Initially, a phone call was arranged to determine eligibility. The researcher verbally explained the aims of the study, how data were collected, and what was required from participants and encouraged participants to ask any questions or raise any concerns.

Recruitment to the qualitative study was done purposefully toward the end of the intervention period to recruit a demographically representative group of participants, including variation in sex, age, and treatment group. Researchers continued to invite participants until data saturation was reached, defined a priori as the point at which no new themes emerged from successive interviews during data coding and charting.

Upon meeting the eligibility criteria and providing written informed consent, participants were randomized into either the intervention or control group by an independent member of the research team using sealedenvelope [26]. Participants were informed of their group allocation after baseline data had been collected.

KOKU-Nut [6] is a digital platform that can be downloaded on tablet or iPad devices. The platform includes progressive and personalized strength and balance exercises alongside a suite of educational games, including a nutrition matching pair card game to educate and nudge users to support their dietary intake (see Multimedia Appendix 1 for screenshots). In particular, the educational content focuses on (1) increasing intake of high-quality protein items such as milk, yogurt, and fish; (2) increasing fluid intake to 6 to 8 glasses a day; and (3) educating users on the signs and risks of malnutrition.

Participants in the intervention group were asked to engage with KOKU-Nut in their own time for approximately 30 minutes 3 times a week for 12 weeks. At the baseline visit, participants were supported with downloading the digital platform onto their iPad or tablet. In cases in which participants did not have the necessary device to use the intervention, an iPad (Apple Inc) with KOKU-Nut installed was provided to the participants for the duration of the intervention. The researcher explained and demonstrated the features of the digital platform and assisted with any technical queries. A crib sheet and contact details for the researcher were provided to participants for additional support and to help with technical issues. A family member or friend was invited to attend the baseline visit if the participant wanted additional support. Participants were advised to continue their usual care and treatment.

Participants assigned to the control group received a leaflet developed by Age UK that contained information about the importance of a healthy lifestyle, including the benefits of staying active and eating healthy [27]. Participants in the control group were also advised to continue with usual treatments and care from their health care professionals.

Written informed consent was collected from all participants prior to collecting the baseline case report file or conducting any assessments at the participants’ homes. The baseline case report file collected data on the participants’ sociodemographic characteristics, as well as the following measures. Health-related quality of life was assessed using the EQ-5D-5L, which considers mobility, self-care, usual activities, pain and discomfort, and anxiety and depression [28]. Each domain is rated across 5 levels of severity, which are combined and converted into a single summary value [29]. The instrument also includes a visual analog scale, where participants rate their overall health from 0 (worst health imaginable) to 100 (best health imaginable). Risk of malnutrition was assessed using the validated Malnutrition Universal Screening Tool [30], and overall risk was assessed as a binary measure to compare those with low vs medium or high risk of malnutrition. Mood was assessed using the short version of the Geriatric Depression Scale [31]. Items were scored dichotomously and summed; overall risk was then assessed as a binary variable so that participants scoring 2 or higher were at risk of depression. These assessments were chosen due to their ease of administration, relevance to the research question, and widespread use across community and clinical settings in the United Kingdom.

During the baseline and 12-week assessments, the researcher also measured grip strength (in kilograms), the time taken to complete 5 sit-to-stand exercises (seconds), and frailty status based on the Clinical Frailty Scale. Further details on these procedures are provided in the study protocol [22].

Participants were asked to complete a 24-hour dietary recall using the online questionnaire Intake24 (version 2) at baseline and 12 weeks [32,33]. Participants were encouraged to follow usual food and drink consumption and then report all food and drink consumed in the online software as accurately as possible. If participants were unable to access Intake24 or experienced technical difficulties, the researcher input this information based on a paper-based version of the participants’ food diary.

The researcher arranged the follow-up assessment 12 weeks after the baseline visits. Baseline measurements were repeated, and participants in the intervention group also reported outcome data related to the usability and acceptability of the intervention, assessed using the technology acceptance model (TAM) [34] and System Usability Scale (SUS) [35].

A subsample of participants completed one-to-one semistructured interviews at the end of the intervention period with a female researcher (CF) experienced in qualitative methodology. Interviews were conducted online or face-to-face based on the preference of the participant, audio recorded, and expected to last no more than 60 minutes. Prior to completing the interview, participants provided additional consent regarding the recording, storage, and use of their interview data. A topic guide was informed by the research aims and similar studies and was used to explore barriers to recruitment, the appropriateness of the outcome measures, and the feasibility of implementing the KOKU-Nut intervention.

The main outcome was the feasibility, acceptability, and usability of the KOKU-Nut digital platform (Multimedia Appendix 1). Additional outcomes were the potential benefits of engaging with KOKU-Nut, specifically considering protein intake, risk of malnutrition, physical function, and health-related quality of life, as previously described [22].

Feasibility outcomes and baseline characteristics of the participants were summarized using descriptive statistics, including proportions or means with SDs. Between-group differences in health-related measures at 12 weeks were explored using analysis of covariance adjusting for baseline values of the outcome and participants’ age at recruitment. These analyses were conducted on an exploratory basis to provide preliminary estimates of potential intervention effects. Given the feasibility nature of the study, the sample size was not powered to detect statistically significant differences.

Interviews were transcribed verbatim and then checked by a researcher to ensure accuracy. Transcripts were then managed using NVivo (version 12; Lumivero) [36] and analyzed inductively using the 5 stages of framework analysis [37,38]. The initial stage involved repeatedly reading the transcripts, followed by the development of codes and the identification of an initial thematic framework (CF). Subsequently, data were systematically indexed according to this framework. The fourth stage involved organizing and summarizing the data into thematic charts to facilitate comparison and synthesis. Finally, patterns within the data were interpreted and discussed with all coauthors to generate overarching insights. To enhance analytical rigor, all stages of the process were reviewed and discussed by a minimum of 2 researchers.

Table 1 shows the baseline characteristics of all 31 participants enrolled in the study. The mean age was 81.3 (SD 8.3) years, and most participants were female (n=24, 77.4%) and White (n=27, 87.1%). Almost all (n=28, 90.3%) participants were retired, and just under half (n=14, 45.2%) were widowed. Overall, most participants were at a low risk of depression based on the 5-item Geriatric Depression Scale. In total, 12.5% (2/16) of the participants in the intervention group were at risk of depression at baseline, which may have affected motivation to use and engage with KOKU-Nut.

More than half (16/31, 51.6%) of participants were taking 5 or more medications, and there were more participants with polypharmacy in the intervention group (10/16, 62.5%) compared to the control group (6/15, 40%). More than half (16/31, 51.6%) of all participants self-reported one or more falls in the previous 12 months, and fall history was similar between the 2 groups. At baseline, 16.1% (5/31) of the participants were at risk of malnutrition based on a score of 1 or more on the Malnutrition Universal Screening Tool, and there was no difference in risk between the groups. The median frailty score was 4 (IQR 3-5), indicating that the sample was living with very mild frailty. This has been described as “while not dependent on others for daily help, often symptoms limit activities.” A common concern is being “slowed up” and/or being “tired during the day” [39].

Participants undertook assessments of functional ability, dietary intake, and quality of life at baseline and the 12-week follow-up. An exploratory intention-to-treat analysis was conducted, but as this was a feasibility study, it was not sufficiently powered to detect significant differences between groups.

There was no difference in grip strength or time taken to complete 5 sit-to-stand exercises at the follow-up assessment between groups after adjusting for age and baseline values.

EQ-5D-5L scores were similar between the groups at baseline and reflect the norms for this age group [42]. After the 12-week intervention, there was a difference in EQ-5D-5L scores using analysis of covariance (F_1,27=7.26; P=.01), again adjusting for age and baseline values. Participants in the intervention group had higher mean EQ-5D-5L scores than participants in the control group (at 12 weeks; 0.72, 95% CI 0.67-0.78 vs 0.62, 95% CI 0.57-0.78 in the control group; P=.01).

Table 3 shows that baseline mean protein intake was above the reference nutrient intake of 0.75 grams per kilogram of body weight per day [43,44]. There was no difference in energy (kilocalories per day) or protein (grams per kilogram of body weight per day) intake at the follow-up assessment between participants in the intervention and control groups when adjusting for baseline values (Table 3).

This 12-week intervention informed by the Medical Research Council guidelines for the development and evaluation of a complex intervention [45] in older adults living in the community in Greater Manchester was feasible and acceptable. Participants reported high usability of KOKU-Nut and found the digital platform acceptable and easy to use, as demonstrated by the SUS [35] and TAM [34]. The current study had retention rates above 90%, and framework analysis of the interviews demonstrated that the study design, assessments used, and allocated treatment groups were acceptable to participants.

This study managed to recruit an at-risk population with low attrition at 12 weeks (3/31, 9.7%). Participants had a mean age of 81.3 (SD 8.3) years, ranging from 70 to 96 years. At baseline, more than half (10/16, 51.6%) of the sample had fallen at least once in the previous 12 months, 16.1% (5/31) were at risk of malnutrition, and most of the participants (median score 4, IQR:3,5) were considered vulnerable according to the Clinical Frailty Scale.

The KOKU-Nut intervention had no impact on energy or protein intake and physical function outcomes at 12 weeks; however, this was a feasibility study with a small sample size and, thus, not powered to detect statistical differences. In addition, interpretation of these findings is further limited by the baseline characteristics of the sample, who were well nourished with adequate protein intake of approximately 1 gram per kilogram of body weight per day at baseline [46]. This introduces a potential ceiling effect whereby participants were already meeting dietary requirements and limits the scope for measurable improvement. Future trials should consider additional eligibility criteria to recruit individuals with suboptimal dietary intake at baseline to enhance sensitivity to change and increase the likelihood of detecting a meaningful intervention effect. Increasing protein intake can support the anabolic effects of strength training and minimize age-related muscle loss, and this has previously been demonstrated in older adults [47], including those with frailty [48] and sarcopenia [49]. However, these dietary interventions did not use a digital platform, and very few apps simultaneously improve dietary intake and support strength training in this population [50]. The study by van den Helder et al [51] demonstrates that combining a home-based exercise program delivered via a tablet-based app with dietary protein counseling delivered via face-to-face dietitian sessions can effectively increase protein intake in older adults. Additional work should focus on the development of a digital tool to support dietary intake among older adults. Additional educational games, high-protein recipes, and goal-setting features are likely to support users in changing their behavior, promote protein intake, and increase the effectiveness of KOKU-Nut.

Most participants in the intervention group (10/14, 71.4%) self-reported using KOKU-Nut 3 or more times a week, demonstrating a satisfactory level of adherence to the intervention. This indicates that older adults can regularly and independently engage with a digital platform such as KOKU-Nut even without prior experience of using an iPad or tablet. A feasibility study with 25 older adults found that 80% of participants regularly used a tablet-based self-monitoring nutrition app (Appetitus) [52]; this complements findings from the current study demonstrating the potential to engage older adults with digital nutrition tools, but further research is required on long-term use and effectiveness. Despite high levels of engagement and usability in the current study, future digital interventions may benefit from increased involvement of caregivers or family members, particularly for those with lower digital literacy, to support confidence, engagement, and uptake.

Unlike the interactive KOKU-Nut intervention, the leaflet-based control condition provided a relatively passive form of intervention delivery. Leaflets are often provided to this population for health guidance and information and, thus, reflect treatment as usual; however, an active control condition may provide greater insight into the effectiveness of the KOKU-Nut intervention. Moreover, qualitative findings indicated perceived behavior change in participants across both trial arms, suggesting that study participation itself may have influenced diet and lifestyle behaviors. This is likely influenced by self-monitoring effects associated with completing assessments and heightened motivation due to involvement in a research study.

This study had many strengths. The approach aligns with the Medical Research Council framework for the development and evaluation of complex interventions [45]. The mixed methods design provides a more comprehensive and nuanced understanding of the feasibility and delivery of the study compared to either methodology alone. Attrition bias was low given that this study had a 90.3% (28/31) retention rate, with only 3 participants dropping out during the 12 weeks. This demonstrates that it was possible to recruit and retain a group of older adults who are at a high risk of falls and often not included in research. Moreover, KOKU-Nut is a minimally invasive digital intervention, so it would be suitable for wide-scale implementation, and a small effect in a large population has the potential to have a big overall effect with low cost or risk.

There are also several limitations that must be acknowledged. Participants were unaware of their allocation until after completing baseline assessments; however, randomization had already been performed prior to the assessment. The lack of allocation concealment may introduce selection bias. This study was not blinded due to resource constraints that, again, may introduce detection bias. This study could have compared the use of KOKU-Nut with that of a generic health app that was not designed for this population to blind the intervention. However, due to limited access to iPads, the control group received a leaflet explaining the benefits of healthy lifestyle habits, and this is likely to better represent standard practice or usual care. The nutrition game was embedded into the Keep-on-Keep-up app, and therefore, it was not possible to specifically look at engagement with the nutrition component or identify whether self-reported use and improvements in quality of life were related to the nutrition game or a different part of the digital platform. Engagement with the digital platform was self-reported, so future studies should aim to develop a dashboard or use Google Analytics data to capture use time objectively. Methodologies to inform the best practice to conduct a gold-standard clinical trial for a digital intervention should be prioritized. Although many assessments were objective, dietary recall is often subject to recall and response bias from underreporting and poor memory. Dietary assessment is prone to misreporting of portion sizes and foods for social approval [53-55]. Moreover, the qualitative work highlighted challenges with completing the 24-hour recall, including burden, difficulties entering complex or homemade foods, and a tendency to simplify intake to facilitate recording. These issues suggest that dietary assessment tools may require greater flexibility and user support to reduce respondent burden and improve data completeness.

Overall, this feasibility randomized controlled trial demonstrates that the study design was appropriate and acceptable to older adults, as indicated by the recruitment and retention rates as well as the themes that emerged from the interviews. This is promising and demonstrates the potential for conducting a larger study in the future that will be powered to detect effectiveness.

KOKU-Nut had good usability, acceptability, and engagement according to the SUS, TAM, and self-report use, indicating that the intervention was designed appropriately to engage this population. However, the exploratory outcomes exhibited no difference in physical function or dietary intake. This demonstrates that further developments should be made to KOKU-Nut to support dietary change before conducting a larger study to see whether KOKU-Nut can lead to clinical and meaningful impact.