Effect of nutrition education intervention to improve dietary diversity practice and nutritional status of the older people: A cluster randomized controlled trial

Abstract The growing aging population raises nutrition and health concerns, with malnutrition in the elderly linked to negative health outcomes. Our objective was to implement theory‐based nutritional education interventions to improve the nutritional status of the elderly through diversified dietary practices. A cluster randomized controlled trial was conducted from December 1, 2021, to May 30, 2021, among 782 older persons randomly selected from two urban and 12 semi‐urban areas in southwest Ethiopia. We used Social cognitive theory (SCT) in guiding the nutritional education intervention. Data were collected using an interviewer‐administered questionnaire. The Mini Nutritional Assessment (MNA) tool was used to assess nutritional status, and a qualitative 24‐h eating recall was used to evaluate dietary diversity. Difference‐in‐difference and generalized estimating equation models were used to assess the intervention effect. In total, 720 participants (361 in the intervention group and 359 in the control group) were included for analysis. The mean dietary diversity score differed significantly between the intervention group and the control group (p < .001). According to the multivariable generalized estimating equations model, the intervention group was 7.7 times (AOR = 7.746, 95% CI: 5.012, 11.973) more likely to consume a diverse diet than the control group. The nutrition status of the elderly in the intervention group improved significantly at the end of the intervention (p < .001). SCT‐based nutritional education interventions can effectively improve healthy eating and nutritional status. For older adults, with its convenient approach and low cost, SCT should be considered an effective and efficient nutritional education approach for behavior change.


| INTRODUC TI ON
The elderly population in developing countries is expected to increase over the next 25 years.This could lead to an increased demand for healthcare services and a strain on care resources.To meet these demands, governments will need to invest in healthcare infrastructure and develop new models of care.It represents over 25% of the total population (World Health Organization, 2015).
By 2030, about 75% of the elderly will live in less developed countries.Elderly people in less developed countries have difficult socioeconomic circumstances and their health status is much worse (Boraschi et al., 2010).Proper nutrition is a crucial predictor of effective aging.Poor eating habits in the elderly contribute to the advancement of noncommunicable and chronic illnesses such as type II diabetes, atherosclerosis, coronary heart disease, and malnutrition.
Loss of bone density is frequent in the elderly, increasing the risk of osteoporosis (Amarya et al., 2015).
Malnutrition is defined as an overall state of poor nutritional status, including under-nutrition and overnutrition of macronutrients.Malnutrition in the elderly further increases an individual's risk of developing general poor health or chronic diseases, such as hypertension, type 2 diabetes mellitus, sarcopenia, and cardiovascular disease (Corcoran et al., 2019;Wondiye et al., 2019;Wyka et al., 2012).
The prevalence of malnutrition among elderly people in different countries of the world varies between 10% and 60% (Abate et al., 2020;Konda et al., 2018;Tessfamichael et al., 2014).Malnutrition among older persons is common in Ethiopia, and it remains one of the country's significant public health issues (Abate et al., 2020;Ferede et al., 2022;Kushwaha et al., 2020;Tessfamichael et al., 2014;Wondiye et al., 2019).Studies recommended that further nutrition education intervention is the best strategy to prevent and manage malnutrition among the elderly so that they can achieve a better quality of life and health as they age (Bandura, 1986;Gezahegn et al., 2016;Paul et al., 2022;Poda et al., 2019).
A theoretical framework is the basic foundation upon which evidence-based interventions are built to achieve successful nutrition interventions (Bandura, 1986).Bandura's Social Cognitive Theory (SCT) (Bandura, 1997b(Bandura, , 2004) ) is the most commonly used theory in interventions to promote healthy eating behavior.It emphasizes the importance of personal, socio-environmental and behavioral factors and the interaction between these factors in influencing behavior.
SCT (Bandura, 1997a) provides a conceptual framework that simultaneously addresses psychological, social, and environmental factors related to physical activity.It includes four components: self-efficacy, outcome expectancies, self-regulatory behaviors, and barriers such as physical disability (Bandura, 1997b).SCT offers both predictors and principles on how to inform, enable, guide, and motivate people to adapt habits that promote health and reduce those that impair it.Core constructs of this framework include self-efficacy, outcome expectations, self-regulation, and perceived impediments and facilitators of behavior (Bandura, 1997b(Bandura, , 2004)).
Older adults who practice healthy eating behaviors may experience an increase in general wellness when socializing (Bandura, 1997a;Hamza et al., 2018).Outcome value precedes intention to make dietary changes, consistent with SCT (Bandura, 1986).
Long-term goals can serve as a general guide and short-term actions can inform current actions (Bandura, 2004).These factors are assumed to mediate the effect of self-efficacy on behavior (Hamza et al., 2018).
To the best of our knowledge, this SCT model is the first to mediate diversified dietary practices and healthy eating behaviors among the elderly in Ethiopia; this would make the current study unique.
We adapted the model from different sources (Brug et al., 2005;Hamza et al., 2018) to take the advantage of composite constructs of SCT.In Ethiopia, data regarding the effect of nutrition education interventions on the nutritional status of elderly people was scarce.
Within this context, the aim of our study was to assess the effectiveness of a health education intervention to improve dietary diversity practices and the nutritional status of older people through improving healthy eating practices using the SCT theoretical model (Bandura, 1997a;Brug et al., 2005) (Figure 1).

| Trial design
The effectiveness of a health education intervention to improve older people's dietary diversity practices and nutritional status is investigated in this cluster randomized controlled research (intervention and Control).Participants for a randomized controlled experiment were recruited from southwest Ethiopia's urban and semi-urban districts between December 2021 and May 2022.

| Eligibility criteria
Those eligible to participate in the study were individuals of both sexes, aged 60 years and beyond living in urban and semi-urban areas of Ilu Aba Bor zone southwest Ethiopia who agreed to participate in the research.The exclusion criteria comprised those who disagreed to give consent; geriatric individuals who were not willing to give an interview; seriously ill people and <60 years old people.

| Intervention
In this study, SCT was used to identify the psychosocial processes (behavior mediators) that lead to intervention outcomes.The intervention aimed to improve nutritional behaviors such as taste preferences, attitudes, beliefs, self-efficacy, health-related concerns, and body satisfaction (Allison & Keller, 2004;Anderson et al., 2007;Nematollahi & Eslami, 2021) following the completion of the baseline data collection, nutrition education was implemented among intervention groups.The nutrition education program was developed using nutrition education for older adults (Calfas et al., 1997) and other relevant materials devised for nutrition education and counseling guidelines were adopted (Calfas et al., 1997) and modified as a local situation for community-based nutrition intervention as well as using the findings of a baseline survey.It is easy, interactive, and user-friendly for all study participants and the research team (Table 1).
The Nutrition instructional Package (NTP) consists of five instructional modules and a range of supplemental and relevant educational (Contento, 2011;Nutrition Education for Older Adults, 1985).The communication mechanism was influenced by the context, cultural preferences, and common ways that people absorb and gather information.A number of educational techniques, including group discussions, interactive lectures, role-plays, active participation, demonstrations, and applicable pictorials, were used to teach the elderly about healthy eating habits.
Based on the best practice guidelines provided by the National Institutes of Health Behavioral Change Consortium, criteria were designed to evaluate the fidelity of the intervention (Bellg et al., 2004;Sheeshka et al., 1993).Checklists to evaluate the intervention design, counselor training, counseling process, acceptance of the intervention, and application of the skills learned from the intervention were part of the criteria (Borrelli, 2011;Sheeshka et al., 1993).

| Sample size determination
The sample size was calculated using the G power 3.1.9.4 program with a power of 80% for Fisher's exact test and a precision of 5%.
The prevalence of malnutrition among elderly people (P1) was 28.3% (Bandura, 1997b;Brug et al., 2005) and P2 was 13.3%, detecting a 15% difference after the intervention between P1 and P2 (Bandura, 1997a) and the calculated sample size was inflated by the Design effect (DE) assuming 12 clusters in each arm, an inter-class correlation of 0.007 (Abate et al., 2020;Konda et al., 2018) and cluster size 25 (Bandura, 2004;Rutterford et al., 2015) using the following formula: DE = 1+ (m−1) ICC = 1+ (25-1) 0.007 = 1.168.The total sample size, taking into account a 5% loss to follow-up, was 368 senior citizens in each arm.Since the elderly who met the criteria were enrolled and cluster randomization was used.Thus, 390 elderly participants in the control group and 392 elderly participants in the intervention group were included in this study.

| Randomization
The use of stratified sampling was made.Using a simple random selection procedure, the elderly have been chosen from a number F I G U R E 1 SCT theoretical model for this study.Source: (Bandura, 1997a;Brug et al., 2005).

Description
Teaching Methods

Teaching materials
Week of metropolitan and semi-urban locations.Two metropolitan areas and twelve semi-urban areas were chosen from a total of 36 areas.
Six clusters were chosen for the intervention based on their proximity, while the remaining six were chosen as the control groups.
Six arms were created, each with a buffer zone to prevent information manipulation, for each of the study's six clusters.The remaining six groups were then designated as control groups, and the remaining six clusters were then randomly assigned to an intervention group.The procedure used to determine the rough number of clusters in each arm has already been made public (Rutterford et al., 2015).

| Outcomes
The key finding of the study was the nutritional status of each group after the first 6 months, or when the experiment was formally concluded.The primary outcome was explained using, a measure developed by the Nestle Nutrition Institute exclusively for elderly persons (Guigoz, 2006;Kennedy et al., 2007Kennedy et al., , 2010;;Nematollahi & Eslami, 2018).The MNA's overall score allowed for the differentiation between senior individuals who were well-nourished (scoring 24-30), at risk for malnutrition (score 17-24), and malnourished (score 17).
The trial's secondary outcome required participants to recall every meal they had in the previous 24 h, both inside and outside the home.This was done to determine their dietary variety score.
On the day of the interview, unusual consumption (such as that connected with Ramadan and federal holidays) was prohibited.During the reference period, food items that were consumed were assigned a "1," whereas those that were not consumed were given a "0" (Guigoz, 2006; Swindale & Bilinsky, 2006).
With the most commonly used materials like cups, tablespoons, coconut spoons, food models, and other conventional household measurements, the respondent appraised the subject's consumption.After that, the foods were classified into ten classes using the following criteria: Examples of foods include cereals that are rich in carbohydrates, leafy greens, other fruits and vegetables that are vitamin-rich, organ meat, fish, eggs, beans, nuts, seeds, dairy products, and fats and oils (Kennedy et al., 2010;Swindale & Bilinsky, 2006).

| Data collection
The data were gathered using a pre-tested, validated, pretested, and structured interviewer-administered questionnaire that was modified from many sources of research (Contento, 2011;Kennedy et al., 2007Kennedy et al., , 2010)).The questionnaire was translated and administered in local language after translation from English language.The supervisors and data collectors received training on anthropometric techniques and data gathering.To determine the instrument's applicability, a pretest was conducted on 5% of the sample prior the actual data collection and the results were evaluated for any inconsistency and modified accordingly.Cranach's Alpha measure of reliability used and kappa above 0.7 was considered acceptable and all within the acceptable range.
The food security of households was assessed using a method that has been approved for use in other developing countries.
The Household Food Insecurity Access Scale (HFIAS), which runs from 0 to 27, was calculated for each participant.Families were categorized into food-secure, slightly, moderately, or seriously undernourished groups based on their level of food security (Guigoz, 2006).
Principal component analysis was used to build the wealth index.
The Household Wealth Index was developed utilizing a variety of variables, including household ownership of fixed assets, services, housing features, and other variables.Quintiles of wealth were then assigned to the latent factor reflecting the wealth index determined by principal component analysis (Arimond et al., 2010).within 1 cm of one another, the average is computed (Bagherniya et al., 2017;CDC, 2007).
The field supervisors provided on-site assistance to the data collectors every day.Each of their supervisors gathered the completed questionnaires and reviewed them overnight.The supervisors repeatedly conducted shady anthropometric tests and interviews.The study population was defined using summary statistics of means and percentages based on the findings, Sociodemographic characteristics, and other aspects.The data was entered using EpiData version 3.5.1 and exported to SPSS version 22 for analysis.Using Pearson correlation analysis, the relationship between the dietary diversity score, the nutritional state of the elderly, and the SCT components was looked at.

| Statistical analysis
The statistical analysis was carried out using SPSS version 22.0.
Categorical variables were summarized using numbers and percentages.Quantitative variables were examined for normality and expressed as mean standard deviations (SD).The Student's t-test for independent samples, Mann-Whitney, Chi-square, and Fisher's exact t-tests were used to compare the baseline characteristics of the intervention and the control groups.
Difference in Differences was used to evaluate the typical treatment effect on the treated by comparing the difference over time in the differences in outcome means in the control and treatment groups.Generalized estimating equations were used to determine how the intervention and control groups differed in their outcomes.
The Sociodemographic and economic factors, a few prevalent chronic diseases, food insecurity in households, and lifestyle decisions that can have an impact on older people's nutritional health were all covered by the generalized estimating equations.Each variable in the bivariable was fitted with a multivariable generalized estimating equation with a p-value of .25 and less.
According to the assertion, the adjusted odds ratio (AOR) with matching 95% confidence intervals demonstrated the strength of the relationship.Every analysis took into consideration the Intention to treat (ITT) idea.If a variable's p-value in the multivariable analysis was <.05, it was considered statistically significant.

| Ethical considerations
The study was conducted according to the principles of the Helsinki Declaration and the requirements of Good Clinical Practice (Hailemariam et al., 2016).The research protocol was approved by the Institutional Review Board of Jimma University (Ref.No: IHR-PGD/419/2019).Written informed consent was secured from each participant before starting the trial.

| Sociodemographic characteristics of the elderly
Almost 782 individuals (425 women and 357 men) constituted the sample (54.3% and 45.7%, respectively) in total who met the eligibility requirements were randomly assigned to the intervention or control group.Seven hundred twenty (92.1%) of the participants successfully completed the study, including 361 from the education group and 359 from the control group (Figure 2).If a variable's pvalue in the multivariable analysis was <.05, it was considered statistically significant.
At baseline, there was no discernible change in the Sociodemographic variables between the intervention and control groups (p > .05)(Table 2).
The post-intervention mean score change for the SCT components is shown in ( Similar to baseline, the sociocultural construct showed a substantial rise in the intervention group at the end of the study (t (718) = 18.497, p < .001).Except for self-efficacy vs. result expectations and sociocultural influences, the connection between SCT components, nutritional status, and dietary diversity score was statistically significant and in a positive direction (Table 4).
The results of this study showed that there was no statistically significant difference between the mean MNA at baseline between the intervention and control groups (20.99 vs. 21.29 points) (p = .403).However, the end-line data showed a significant difference between the intervention and control groups in terms of the mean MNA score (24.60 vs. 22.82 points; p < .001).Between the intervention and control groups, there was a net mean difference of 2.08 and 0.65 points difference in difference and standard error Table 5 shows that the differences were statistically significant (p < .001).
According to this study, the baseline mean dietary variety scores between the intervention and control groups were similar (t (718) = 0.104, p = .349).However, the end-line results reveal a significant difference between the intervention group and the control group in terms of mean dietary variety scores (t (718) = 14.059, p < .001).Following the nutrition education intervention, the independent t-test reveals a statistically significant difference in the mean DDS between the intervention and control group of older persons (DID = 2.20.16)(p < .000)(Table 5).
In the multivariable GEE (Table 6), the intervention groups were 7.7 times (AOR = 7.746, 95% CI: 5.012, 11.973) more likely to consume a varied diet than the control group after adjusting for potential confounding variables.A generalized estimating equation was utilized to identify independent parameters linked to the end-line baseline differences in the variances in mean MNA scores.
The factors in the model included sex, educational level, DDS, marital status, family size, occupation, home location, depression, alcohol use, smoking, social support, financial aid, and nutritional education.
Nutritional education, place of residence, sex, family size, social support, financial support, depression, age, occupational status, food security, DDS, and alcohol consumption were the only independent predictors of MNA among all the other factors.The elderly in the intervention group demonstrated a substantial increase in their nutritional status at the conclusion of the intervention (β = 1.119, p < .001)after accounting for potential confounders (Table 7).

| DISCUSS ION
The main purpose of this study was to assess the effectiveness of a health education intervention to improve dietary diversity practices and the nutritional status of older people through a SCT-guided framework.At baseline, the Sociodemographic parameters and nutritional status of the study participants were comparable.The scores on the SCT components of the intervention groups considerably increased as compared to the baseline score and control group.
The findings of this study indicated that even after a brief intervention, eating behavior modifications and nutrition intake status among female old women were positively impacted by the nutritional education program based on social cognitive theory (Bagherniya et al., 2017).Developing and implementing an educational program based on the social cognitive theory may improve pregnant women's patterns of fruit and vegetable consumption (Seo, 2016).

The results of an Iranian investigation, A Cluster Randomized
Controlled Trial (Maryam et al., 2019), which revealed that the SCT domains had altered over time and that there had been a substantial difference between the intervention and control groups, provide weight to this conclusion.In a research of mothers, it was discovered that there was a significant difference between the mean score of SCT structures in the experimental group and the control group before and after the intervention (Sebastian et al., 2021).
This figure shows the flow of the study participants through the trial according to the criteria recommended in the CONSORT guideline.
measurements were made both before and after the intervention.Weight was measured using a digital scale, light clothing, and no shoes(CDC., 2007).The person was measured for height while standing using a portable stadiometer.A calf circumference (CC) of less than 31 cm was given a score of 0(Smeeth & Woon, 2002; World Health Organization, 2011).The BMI was calculated as weight divided by the square of height (kg/m 2 ).Definition provided by the Centre for Disease Control and Prevention(World Health Organization, 2008).Adults with BMIs of 30 kg/m 2 , 25-30 kg/m 2 , and 18.5-25 kg/m 2 were categorized as obese, overweight, and normal weight, respectively.The waist-to-hip ratio was calculated as WC divided by HC.Utilizing a stretch-resistant tape that could endure 100 g of tension, we measured the hip and waist circumferences.If the two measurements differ by more than 1 cm, more measurements are conducted.If the two measurements are

Table 3
).After the intervention, the intervention group's score significantly surpassed the control group's score (t (718) = 25.633,p < .001).At baseline, the mean outcome value scores in the intervention and control groups were comparable.
Shows the average SCT construct score for the elderly in the Illu Aba Bora Zone of south west Ethiopia (n = 720) before and after the intervention.Shows the relationships between the SCT components and the DDS and MNA of senior people in the southwest Ethiopian region of Ilu Aba Bor (n = 720).Differences in mean DDS and Mini Nutritional Assessment of the group over the study period in Illu Aba Bor Zone, Southwest Ethiopia (n = 720).Significant difference between control and intervention (p < .001).Generalized estimating equation illustrating the impact of intervention on the elderly's practice of consuming a variety of food items in the Illu Aba Bor Zone, Southwest Ethiopia (n = 720).Notably, the model was modified to account for factors such as gender, marital status, drinking, smoking, family size, wealth index, social and financial support, occupation, residential area, education, age, household food security, depression, and DDS, Group * intervention, and interaction.Generalized estimating equation illustrating the impact of intervention on older people's nutritional status in Illu Aba Bor Zone, Southwest Ethiopia (n = 720).
Note: Values are presented as mean ± SD. *Significant difference between control and intervention (p = 0.05).**Significantdifference between control and intervention (p = .001).***Significant difference between control and intervention (p < .001).Abbreviation: SD, standard deviation.The dietary variety score of the structure and the elderly's nutritional state were significantly positively correlated in almost all of the SCT.The enhanced Behavioral mediator changes and their correlation with high DDS and nutritional status demonstrated the efficacy of the intervention in promoting wholesome attitudes, beliefs, values, and expectations of optimal nutrition as we age.This result is comparable to a study from India that found a significant positive link between type II diabetic patients' dietary diversity behavior and all of the Social Cognitive Theory's categoriesTA B L E 4Note: Values are presented as mean ± SD, (95% CI).**Significant difference between control and intervention (p = .001).***Significant difference between control and intervention (p < .001).Abbreviations: BL, baseline; EL, End line; SD, standard deviation.aTABLE 6Abbreviations: β, beta (coefficient); AOR, adjusted odds ratio; DDS, dietary diversity score; SE, standard error.TA B L E 7