A 16-week, single-center, parallel-group, randomized controlled trial compared a web-based CM intervention group (CMG) with a control group (CG) on their protein intake, cooking effectiveness, muscle strength, muscle mass, and physical activity. The study was conducted at Texas Tech University Nutrition and Metabolic Health Initiative (NMHI), Lubbock, Texas. Participants were permitted to remove themselves from the trial at any time.

A human study compliance review was submitted to the institutional review board at Texas Tech University, Lubbock, Texas. The study was expedited for review and received approval (IRB2021-693). Once participants were recruited and eligibility was determined, an initial appointment was set up at Texas Tech University NMHI. A research team member described the study in detail, and participants were asked to sign a consent form stating willingness to participate. The participants’ information collected for the study was deidentified, given a code number, and kept on the researchers’ computer at Texas Tech University NMHI. The research team offered the participants the vívofit 4 watch (Garmin) as compensation, which they used to complete the study.

Flyers, newsletters, and word of mouth were essential for recruitment. When participants agreed to enroll in the study, they filled out an initial screening questionnaire to help determine whether they met the eligibility criteria. The inclusion criteria involved individuals who are aged 65 years or older, able to cook for themselves, physically active (eg, no need for equipment for assistance), and able to use a computer and mobile device. The exclusion criteria included individuals aged <65 years; those with limited mobility (eg, need for equipment for assistance), cognitive dysfunction (eg, dementia), a heart pacemaker, or type 1 or type 2 diabetes with insulin use; current smokers; those with some form of amputation; those who unable to use a computer and mobile device or unable or unwilling to wear the vívofit 4 watch (Garmin) for the duration of the study; and those undergoing or had recently undergone a severe medical procedure or diagnosis.

Participants were recruited and enrolled from June 2022 to August 2022, with data collection completed in December 2022. If a participant dropped out of the study, a new participant would replace and be allotted to the same group as the participant they replaced. A total of 52 older adults, including both men and women, met the study’s eligibility criteria. Assessments were conducted at the prestudy, weekly, and poststudy time points.

The study was powered to identify pre- to poststudy changes between the groups. A similar study [60] was used to develop the necessary sample and effect size using the G*power software (version 3.1.9.6; Heinrich Heine University Düsseldorf). Calculations were made for a total sample of 52 participants (26 participants per group) to obtain a statistical difference in muscle strength and mass between the groups, assuming an α of 5%, effect size of 0.72, power of 80%, and 10% inflation for dropouts. Data were imported to SPSS (version 29; IBM Corp) for analysis. DXA measuring muscle and fat mass was the study’s primary outcome measure. Secondary outcomes included protein intake in grams, muscle strength in kilograms, average daily steps, frequency of physical activity in minutes per week, height in inches, and weight in kilograms.

Participants were randomized to the CMG or the CG by block randomization using 2 blocks with 26 codes. On the basis of the assigned participant’s study code, the primary researcher enrolled the participants into their group at the end of their initial visit. Therefore, the allocation was not concealed. The analysis assessed the effect of the intervention with the completers. Any missing data were replaced with the last observation carried forward before analyses of all measurements via single imputation. Participants were excluded from data analysis if they did not complete over 50% of their weekly questionnaires or, after enrollment, met an exclusion criterion.

Results are presented as mean (SD), mean (95% CI), ranges, or frequencies. P<.05 was considered statistically significant. Linear mixed models were used to assess the differences in protein intake between the groups at the end of the intervention. The model included the fixed effects of time, intervention, and time-by-intervention interaction. Participants were modeled as a random effect to account for the repeated measures design. When a significant main effect was observed, post hoc analyses were conducted and the Tukey-Kramer method was used to adjust for multiple comparisons. Within-group muscle mass and strength differences, as well as physical activity and diet quality differences, were estimated using an independent samples (1-tailed) t test for variables measured before and after the study.

In total, 64 participants expressed interest in the study. Of these, 8 (13%) were excluded during web-based or telephone screening due to failing to meet the inclusion criteria or losing contact. A total of 56 participants were eligible for inclusion and were randomized: 29 to the CMG and 27 to the CG. A total of 25 participants in the CMG, compared with 24 in the CG, completed the 16-week weekly questionnaires and both study visits. Of the eligible 56 participants, 7 (13%) withdrew or dropped out before the completion of the study. Of the 7 participants, 6 (86%) dropped out due to medical reasons unrelated to the study, and 1 participant (14%) dropped out due to family reasons. Of the 56 participants, 2 (4%) participants had to be excluded from the data analysis because 1 participant had bariatric surgery during the study and the other completed less than 50% of their weekly questionnaires. Therefore, a total of 49 participants were included for the data analysis (CMG: 24/29, 83%; CG: 23/27, 85%). See the CONSORT (Consolidated Standards of Reporting Trials) study flow diagram (Figure 1) for the study details.

The prestudy characteristics of the groups are presented in Table 1. The study included a greater proportion of female (38/47, 81%) and White (44/47, 94%) participants. The mean age, weight, and BMI of the participants in the CMG were 71.4 (SD 5.2) years, 76.6 (SD 17.4) kg, and 28.0 (SD 6.0) kg/m², respectively. In the CG, they were slightly older (mean 73.2, SD 5.8 years) but had lower weight (mean 69.4, SD 15.0 kg) and BMI (mean 26.1, SD 5.0 kg/m²). The CG was found to be more physically active than the CMG. Regarding diet, the CG consumed more fiber, calcium, dairy, vegetables, and fruit than the CMG. Meanwhile, the CMG consumed more daily added sugar than the CG. However, both groups consumed the same amount of daily whole grains.

There was no between-group difference in the muscle mass change from the prestudy visit (P=.88; Table 2). Using the EWGSOP sarcopenia diagnosis [61], 21% (5/24) of the CMG and 26% (6/23) of the CG had low muscle mass at the prestudy visit. At the poststudy visit, 21% (5/24) of the CMG and 22% (5/23) of the CG had low muscle mass.

Similar results were seen for muscle strength. There was no between-group difference in the muscle strength change from the prestudy visit (dominant: P=.92 and nondominant: P=.72). When comparing the classification of muscle strength for the dominant hand, the CMG was considered 29% (7/24) weak, 67% (16/24) normal, and 4% (1/24) strong at the prestudy visit. At the poststudy visit, the CMG was considered 33% (8/24) weak, 46% (11/24) normal, and 21% (5/24) strong. The CG was considered 13% (3/23) weak, 83% (19/23) normal, and 4% (1/23) strong at the prestudy visit. At the poststudy visit, the CG was considered 13% (3/23) weak, 74% (17/23) normal, and 13% (3/23) strong.

When comparing the classification of muscle strength for the nondominant hand, the CMG was considered 42% (10/24) weak, 54% (13/24) normal, and 4% (1/24) strong at the prestudy visit. At the poststudy visit, the CMG was considered 38% (9/24) weak, 50% (12/24) normal, and 13% (3/24) strong. On the other hand, the CG was considered 30% (7/23) weak, 65% (15/23) normal, and 4% (1/23) strong at the prestudy visit. At the poststudy visit, the CG was considered 30% (7/23) weak, 57% (13/23) normal, and 13% (3/23) strong.

Per the EWGSOP sarcopenia diagnosis [61], 38% (9/24) of the CMG and 30% (7/23) of the CG could be diagnosed with probable sarcopenia. In comparison, 8% (2/24) of the CMG and 9% (2/23) of the CG could be diagnosed with sarcopenia at the prestudy visit. At the poststudy visit, 33% (8/24) of the CMG and 17% (4/23) of the CG could be diagnosed with probable sarcopenia, whereas 8% (2/24) of the CMG and 9% (2/23) of the CG could be diagnosed with sarcopenia at the poststudy visit.

Figure 2 reveals the mean (SD) daily protein intake in grams for each week of the study for each group. A nonsignificant time-by-intervention interaction was observed for daily protein intake (Figure 2 and Table 3; P=.08). There was also no statistically significant difference in protein intake between the interventions (P=.50). However, a statistically significant time effect was observed (P≤.001). Post hoc testing showed that daily protein intake was significantly higher at weeks 1 to 16 versus week 0 (P<.05) in the cohort. At week 16, protein intake was 16.9 (95% CI 5.77-27.97) g higher than that at the prestudy visit.

Each group was evaluated to see how many participants met their protein needs (1.0-1.2 g/kg body mass per day). In the CG, 39% (9/23) participants did not meet their needs, 26% (6/23) did meet their needs, and 35% (8/23) exceeded their needs during the study. In the CMG, 58% (14/24) participants did not meet their needs, 8% (2/24) did meet their needs, and 33% (8/24) exceeded their needs during the study. Additionally, in all the completed protein questionnaires, the CMG and the CG had blank answers for 15.4% (63/408) and 12.5% (49/391) of their questions, respectively . When evaluating the daily intake for each dietary component from the DSQ (Table 3), the components stayed close to the same when comparing pre- with poststudy results.

For the CMG, participants reported watching 82.8% (318/384) of the intervention videos. The primary reason reported on why they did not watch the videos was “not interested in watching” (21/56, 38%). Additional reasons included personal reasons, traveling or vacation, or they did not receive the video. For the CG, participants reported that they read 94.8% (349/368) of the recipes sent to them. The primary reason why the participants did not read the recipe was “busy” (5/13, 39%). Additional reasons included personal and medical reasons, laziness, uninterest, not receiving the video, and having their spouse read it.

When examining whether both groups cooked the recipe learned through web-based videos or just by reading the recipe, the CMG cooked more recipes than the CG (64.8%, 249/384, vs 62.5%, 230/368). Based on the questionnaires with responses outside of “N/A,” the CMG and CG did not cook primarily because of “holiday, traveling, or vacation” (CMG: 20%, 25/125, and CG: 26.5%, 35/132). See Table 4 for the remaining reasons. Barriers or complications that were reported from both groups when either watching the videos or preparing the recipe included borrowing ingredients from a neighbor; recipe serving size being too big; confusion toward either the ingredients or methods; changing or not including ingredients to meet taste or diet preference; finding certain ingredients at the store; too much spice or ingredient in the recipe for their palette; standing for an extended period was challenging; difficulties in scheduling time and energy to shop, prepare, or cook; or taking more initiative to prepare recipe themselves.

At the end of the study, both groups were asked about the main challenges or barriers to maintaining their protein intake (Tables 5 and 6). Meanwhile, the CMG participants were asked how the CM videos specifically helped clarify managing their protein intake (Table 7) and what the most memorable thing they recalled after watching the video or what their favorite part of the CM videos was. All CMG participants were reported having no technical difficulties accessing and watching the videos.

To the authors’ knowledge, a study has yet to be performed with CM explicitly targeting the older adult population to enhance their protein intake. However, a statistically significant time effect was observed (P≤.001). Furthermore, post hoc testing showed that daily protein intake was significantly higher at weeks 1 to 16 versus week 0 (P<.05). At week 16, protein intake was 16.9 (95% CI 5.77-27.97) g higher than that at the prestudy visit. This result indicates that protein intake increased in the cohort with the information provided to both groups.

Nevertheless, there was no additive effect of the CMG over what the CG received because no between-group differences were observed for any primary or secondary outcomes. Insufficient consistent protein intake, low physical activity, adherence to the intervention, and accuracy of the questionnaires could explain the results. Also, participants’ ethnicity, average age, gender, and BMI were similar in both groups and affected the diversity of the study’s population; therefore, the outcomes were not tested against them because there was no vast difference to indicate a relationship. Given the limited representation of men in the cohort, the results cannot be generalized to men, Hispanic participants, and African American participants.

The accuracy of each group’s protein questionnaire could play a factor because they were self-administered. Self-administered questionnaires are more susceptible to item nonresponse [62]. The CMG and the CG had blank answers for 15.4% (63/408) and 12.5% (49/391) of their questions, respectively, suggesting that their intake could have been higher and explained how their muscle mass was overall maintained. Additionally, the participants were not asked to change their diet outside their protein intake. The DSQ reported that participants’ diets did remain the same.

Before the study started, both groups were recommended to consume 25 to 30 g of protein per meal in addition to daily physical activity. These recommendations are similar to Paddon-Jones and Rasmussen’s [63] findings, reporting that approximately 25 to 30 g of protein per meal is a valuable strategy for maintaining muscle mass in older adults. This strategy would mean that the participants would have to eat approximately 75 to 90 g of protein daily. The CMG met this range from weeks 6 to 15, but the CG met this range during weeks 4, 11, and 12. Specifically, 39% (9/23) of the CG and 58% (14/24) of the CMG did not meet their needs (1.0-1.2 g/kg body weight per day). The 2005-2014 NHANES [26] reported that 31% to 50% of older adults did not meet their protein recommendations. Our population was in this range. Therefore, these results could also explain why muscle mass did not significantly increase between the groups. However, the estimated average requirement for 51 to 70 years is 0.66 g/kg/d, and the recommended dietary intake is 0.8 g/kg/d for all adults over 18 years old, including older adults [64]. Therefore, in the context of adequate energy intake, muscle mass was maintained in this cohort if their protein intake was consistent with these levels.

Grip strength has been used in research to determine overall body strength [65,66]. However, there were no between-group differences in muscle strength change from the prestudy visit. Kim et al [67] found no association between the amount and change (increase or decrease) in daily total protein intake with the incidence or prevention of low muscle strength, which was similar to our results. Additionally, a longitudinal study [68] indicated that 25 to 30 g of protein per meal is associated with greater muscle strength in older adults. However, this recommended intake did not consistently happen in our study, and participants did not meet their calculated needs, which could affect their muscle strength. Physical activity also did not impact muscle strength. Similar results were seen with Ramsey et al [69], who also saw no association between the number of steps and handgrip strength.

When looking at their steps, current evidence suggests that healthy older adults should meet approximately 7000 to 10,000 steps per day [70]. However, our study showed that 67% (16/24) of the CMG and 57% (13/23) of the CG did not meet this range. Also, Park et al [34] reported that individuals who walked at least 7000 to 8000 steps daily likely have muscle mass above the sarcopenia threshold. Because only 33% of the CMG and 44% of the CG met this threshold, it is unsurprising that their lack of steps may have impacted our results.

Lastly, the dropout rates were similar, 14% (4/29) in the CMG and 11% (3/27) in the CG. This rate is lower than the reported average of 20% to 49%, which is commonly seen in dietary clinical trials [71]. In the CMG, 10% (3/29) of participants dropped out due to medical reasons, whereas 3% (1/29) dropped out due to family reasons. In the CG, all the participants dropped out due to medical reasons. These are all common reasons for dropouts in clinical trials [72]. The dropouts were not related to the study, and no adverse effects were reported throughout the study.

To the authors’ knowledge, this study is the first to examine the outcomes of CM in the form of web-based cooking demonstrations and nutrition education to enhance protein intake and muscle quality in older adults. The results reveal insufficient evidence because no between-group differences were observed for primary or secondary outcomes. However, most of the intervention group reported that the cooking demonstrations helped them prepare and cook recipes at home, providing more confidence in the kitchen, and its learning was feasible for them.

In the future, it would be valuable to further investigate the factors that could have affected this study. In developing and implementing this study, exercise training sessions and a dietary supplement could be included. Additionally, the research study design could include RDs, chefs, exercise physiologists, health coaches, or psychologists. The staff would be essential in creating the study protocol, kitchen equipment checklist, consent forms, scripts, and questions. During recruitment, it would be ideal to obtain a broad age range with an equal gender and ethnicity ratio to help reciprocate the general population. The recipes should consider the season, 1-person portion size, time, cost, and mild flavors. A protein food diary could help keep track of protein intake during the week and help answer the protein questionnaire accurately.

It could be interesting to incorporate muscle biopsy and biomarkers, such as vitamin B₁₂, folate, and creatinine, to evaluate muscle mass further and see if this intervention impacts or could explain why muscle mass outcomes were nonsignificant due to predispositions. However, there are challenges in successfully performing a muscle biopsy in older men and women who are frail or have low body mass [74], so that would be a concern to consider. For biomarkers, no specific recommendations, references, or cutoff values are available to assess muscle mass or quality. Therefore, the biomarkers could be used to notice any significant change within a short time duration. Overall, given the current concern of sarcopenia, these concepts could enhance this intervention further with the information gathered in this study to impact public health.