Association of cash transfer and nutrition message intervention with nutrition status of children aged 6-59 months in the universal child benefit pilot programme In Nyando Sub-County, Kenya

¹Department of Public and Community Health, School of Health Sciences, Jaramogi Oginga Odinga University of Science and Technology, Bondo, Kenya, ²Department of Biomedical Sciences, School of Health Sciences, Jaramogi Oginga Odinga University of Science and Technology, Bondo, Kenya, ³Department of Human Nutrition, Faculty of Health Sciences, Egerton University, Njoro, Kenya

^&Corresponding author
Phinhas Ochieng Atieno, Department of Public and Community Health, School of Health Sciences, Jaramogi Oginga Odinga University of Science and Technology, Bondo, Kenya

Introduction    

Children under 5 years are the most endangered society members to social problems, such as poverty and malnutrition. As of June 2020, over 385 million children were living below USD 1.9 per day, twice the adult population, as highlighted by Overseas Development Institute/ United Nations Children's Fund (ODI/UNICEF) [1]. Gudu et al. [2] confirmed that universally, approximately three million children between six and fifty-nine months of age die every year due to malnutrition. In sub-Saharan Africa, the double burden of malnutrition continues with high levels of undernutrition and growing levels of overnutrition coupled with increased burden of diet-related non-communicable diseases [3]. In Kenya, stunting currently stands at 18%, which is a sign of chronic malnutrition; wasting at 5%, depicting acute malnutrition; 3% being overweight, depicting signs of overnutrition; and finally 10% of children being underweight as per the Kenya National Bureau of Statistics [4].

The 2019 Cost of Hunger in Africa report in Kenya disclosed that 19.2 of every 100-child deaths are stimulated by malnutrition (hunger). This was due to household food insecurity rendering household members unable to meet even the minimum dietary diversity due to low purchasing power, hence the need for a cash transfer to help cushion the households against the negative coping strategies that impact negatively on the health and nutrition status of children under 5 years.

The term 'cash plus' is gaining considerable momentum due to its potential to complement cash transfer with additional elements that, in combination, may be more effective in realizing the intended outcome compared to cash alone [5]. The theory of change around 'cash plus' programming is built on the belief that, while cash transfer can impact beyond the poverty reduction objective, the utilization of the cash can be influenced by behavioral factors [6]. For example, a study by Knowles et al.found that cash transfer programs alone may not yield adequate behavior change to impact in areas such as nutrition and health [7]. To fill this gap, Universal Child Benefit (UCB) has been globally adopted by governments and international organizations to primarily provide social protection to children irrespective of their background. Universal Child Benefit, also known as Universal Childcare Benefit (UCCB), is a welfare cash transfer in which a monetary benefit is disbursed to guardians or caregivers.

In the same regard, in February 2022, the Kenyan government, through the Ministry of Public Service, Gender, Senior Citizen Affairs, and Special Programs, State Department of Social Protection, Senior Citizens Affairs, and Special Programs, in partnership with UNICEF, Save the Children, and the World Food Programme (WFP), initiated a pilot cash plus intervention program. A baseline assessment was conducted to unravel the standard estimates of given indicators before the imposition of UCB intervention programs in Mbeere North, Kajiado Central, and Nyando Sub-Counties.

In Nyando Sub-County, UCB was implemented in Ahero Ward. In this intervention, an unconditional cash transfer of eight hundred Kenya shillings was provided to families with minors aged three years and below for twelve months and on a bimonthly basis through M-Pesa transfer. Additionally, complementary messages were sent to beneficiary households in Ahero Ward to provide and educate them on nutrition, especially on maternal, infant, and young child nutrition.

Internationally, the major cash transfer programs include the popular Cash Transfer for Orphans and Vulnerable Children (CT-OVC). The social protection program was launched by the United States Agency for International Development through the United States President's Emergency Plan for AIDS Relief (PEPFAR). The chief intent of CT-OVC since its inception has been offering financial cushioning to children against various unpropitious situations, e.g., children living with HIV, children depending on parents, guardians, or caregivers who are HIV-positive, and juveniles who are at high risk of contracting the virus.

In Africa, an example of cash plus intervention program includes the Ethiopian Integrated Nutrition and Social Cash Transfer (IN-SCT) Pilot program in Southern Nations, Nationalities and Peoples (SNNP) and Oromia Regions. It is funded by Irish Aid partnering with UNICEF and the Ethiopian Ministry of Labor and Social Affairs (MoLSA). IN-SCT complements household social protection cash transfer with multisectoral nutrition services to help alleviate poverty-related shortcomings while strengthening the nutritional wellness of the benefiting households [8]. In Kenya, the two main cash transfer programs include Cash Transfers for Persons with Severe Disabilities (PWSD-CT) and Old Persons' Cash Transfer (OPCT), both of which are funded by the Kenyan government. PWSD-CT was activated in June 2011 to improve the financial protection of impoverished individuals providing care services to children and adults with severe disabilities, besides enhancing the livelihood of disabled adults and children in low-income households.

Although the benefits of cash transfer are well documented, it remains unclear if it has a direct impact on the nutritional status of children under 5 years. Even though considerable progress has been made to improve child nutrition outcomes, stunting rates in Kisumu County remain at 9%, as shown in the Kenya Demographic and Health Survey recent data. Stunting in Nyando Sub-County at baseline stood at 40% higher than the county overall, and underweight at 27.8%, with wasting being 5.6%, as shown in the UCB baseline report.

The household level of utilization influences the nutrition outcome and is yet to be clearly established. The efficacy of the nutrition messages sent and shared with beneficiaries also remains to be established, as the key role is to yield a behavior change that would ensure the children at the household are cared for to achieve optimal nutrition status. This study aimed to establish if there is an association of cash transfer and message intervention with diet quality and nutrition status of children aged 6-59 months in Nyando Sub-County.

In Nyando Sub-County, UCB was implemented in Ahero Ward. Here, the indicators included Infant and Young Child Feeding (IYCF) practices, hurdles to adopting the prototypical IYCF and positive parenting, and water, sanitation, and hygiene (WASH) practices. In this intervention, an unconditional cash transfer of eight hundred Kenya shillings was provided to families with minors aged three years and below for twelve consecutive months and on a bimonthly basis through M-Pesa transfer. Additionally, complementary nutrition messages were sent to beneficiary households in Ahero Ward to provide and educate them on positive parenting and nutrition. Complementary nutrition messages were sent to the caregivers in the form of automated phone calls, which, when received by the caregivers, provided them with information on good nutrition practices in the Swahili language. Community health volunteers were also used to reach out to caregivers with complementary nutrition messages during household visits. For complementary nutrition messages, child protection volunteers, community health volunteers, and lay volunteers were involved in capacity building through training, and in turn, they reached the community during household visits and through mother-to-mother support groups to pass the messages to the caregivers with the intention that the initiative would result in improved nutrition status of children under 5 years.

Conceptual framework of the study: this study was guided by a conceptual framework adapted from the UNICEF framework on the determinants of child malnutrition. The framework illustrates the relationship between cash transfer interventions, complementary nutrition messaging, and child nutrition outcomes among children aged 6-59 months in Nyando Sub-County. In this study, the independent variables were the cash transfer provided through the Universal Child Benefit (UCB) pilot project and the complementary nutrition messages delivered to beneficiary households. These interventions were expected to influence how households utilize financial resources and nutrition information to improve feeding practices and childcare. The intermediate variables included the level of utilization of the cash transfer, application of nutrition messages in household feeding practices, and diet quality of children, measured through dietary diversity and food consumption patterns. These intermediate factors influence the nutrition status of children aged 6-59 months, which constituted the dependent variable (outcome) in this study. The outcome variable was measured using anthropometric indicators, specifically stunting (height-for-age), wasting (weight-for-height), and underweight (weight-for-age). Additionally, the relationship between the interventions and child nutrition outcomes may be influenced by several household and caregiver characteristics, including household size, household headship, the individual receiving the cash transfer on behalf of the household, caregiver education level, and caregiver age. These sociodemographic characteristics may affect both the utilization of the cash transfer and the adoption of recommended nutrition practices, thereby influencing child nutrition outcomes as shown in Figure 1.

Methods     

Study area: the study was conducted in Nyando Sub-County of Kisumu County, consisting of those living in town and the rural population extending to Lake Victoria in August 2023. Nyando Sub-County consists of five wards, namely Kobura, Ahero, East Kano/Wawidhi, Awasi/Onjiko and Kabonyo/Kanyagwal Ward. This study was done in Ahero ward, Kakola and Kochogo Locations where the UCB pilot was implemented.

Sample size determination: the sample size was calculated using Cochran, William Gemmel. (1977): Cochran formula, stating sample size (N):

Where: P= proportion of the population with the desired characteristics=35% (Ministry of Public Service, Gender, Senior Citizen Affairs, and Special Program, Baseline Survey Report, 2022); Z = 1.96 the standard normal deviate (here at CI=95%); Q= 1-P = 65%; e = margin of error = 0.05. Population of interest in Ahero Ward (under 5 years population) = 6245 (KHIS, 2022); population enrolled for UCB cash transfer = 2000; n₀= 350.

Since the population of interest is under 10,000, the sample size was modified using the formula below from Cochran, W. G. (1977):

Three hundred and thirty two (332) children will be sampled in Ahero Ward. Attrition rate of 10%= 0.1(332) = 34. Total sample size = 332+34. With an attrition rate of 10%, the total sample size (n) was 366 children from 366 households, assuming at least one child per household. The study, however, reached a population of 448 children from 400 households who met the criteria. Five key informant interviews were purposively selected to collect qualitative data. The key informants were the sub-county nutrition coordinator, sub-county child protection officer, sub-county community health strategy focal person, Community health assistant, and health facility nutritionist. The key informants were physically interviewed in their offices in Nyando Sub-County while a focus group discussion was held in Ahero Ward within the community. One focused group discussion consisting of ten caregivers each was also conducted to collect additional information.

Study design, population, and sampling: this was a cross-sectional mixed-methods study that adopted both qualitative and quantitative approaches. Caregivers with children aged 6-59 months benefiting from the UCB cash transfer were the respondents for the study, and anthropometric measurements of children aged 6-59 months were taken. Purposive sampling was used, targeting households with children aged 6-59 months who were universal child benefit pilot program beneficiaries, reaching a total of 448 children from 400 randomly selected households in Ahero Ward where the UCB cash transfer was implemented.

Data collection tools

Household semi-structured questionnaire: a semi-structured household questionnaire was used to collect data on sociodemographic information such as household composition, socioeconomic and complementary nutrition messages sent to beneficiaries from the primary caregivers. A 24-hour recall questionnaire was used to collect data on food groups for assessment of dietary diversity. The child questionnaire was used to collect quantitative data on anthropometric measurements following standard protocol for taking body measurements, including mid-upper arm circumference (MUAC), weight, and height/length of children aged 6-59 months to assess their nutritional status. Weight measurements for children were taken using a standardized weighing scale (SECA; MD001; 1998; Germany) placed on a flat surface and zeroed. The caregivers were asked to remove their children's shoes and bulky clothing, and stand with both feet at the center of the platform on the scale for those who could stand, and indirect weight measurement was taken by weighing the caregivers and their children. Weight measurements were then recorded to the nearest 0.1kg (Centers for Disease Control and Prevention, 2007). Height measurements of the children were taken using a standardized stadiometer (SECA; MD001; 1998; Germany) placed on a firm, level ground. The caregivers were asked to remove their children's shoes and stand straight on the stadiometer, with heels together and the upper part of the back and the head in contact with the vertical part of the stadiometer. Then the headpiece was firmly lowered down to meet the vertex of the head. The height measurements were recorded to the nearest 0.1cm (Centers for Disease Control and Prevention, 2007). For children who could not stand, their lengths were measured while lying down. Mid upper arm circumference was measured by asking the caregivers to remove clothing, which allowed for the midpoint of the left upper arm to be identified.

The MUAC measurement was then taken using UNICEF-approved MUAC tape at the midpoint of the left arm and recorded to the nearest 0.1cm. Qualitative data was collected through focused group discussions using the developed guide and through key informant interviews with a KII guide. The data collection was done by the researcher with the help of enumerators arranged in five teams of three through the questionnaire uploaded in KoboCollect and accessible using mobile phones via face-to-face interviews at the household. Each team had one supervisor. Completed questionnaires were submitted to the server, which was monitored throughout data collection to flag off any discrepancies in the data through plausibility checks, especially anthropometric measurements in situations where a deviation from standard measurement protocols was noted.

Qualitative data was collected through focused group discussions using the developed guide and key informant interviews with a key informant guide. Two focused group discussions with caregivers from households who were not part of the sample and five key informant interviews were conducted to collect qualitative data. The key informants were the sub-county nutrition coordinator, sub-county child protection officer, sub-county community health strategy focal person, community health assistant, and health facility nutritionist. The key informants were physically interviewed in their offices in Nyando Sub-County while a focus group discussion was held in Ahero Ward within the community.

Inclusion criteria: households with children aged 6-59 months who were beneficiaries of the UCB program were included in the study.

Exclusion criteria: households with children aged 6-59 months who were not beneficiaries of the UCB program were excluded from the study, as well as UCB-benefiting households with children under 6 months and over the age of 59 months.

Recruitment of respondents for the household questionnaire: respondents for the household questionnaire were recruited from households that had participated in the universal child benefit (UCB) pilot programme in Nyando Sub-County and had eligible children aged 6-59 months at the time of the study. The UCB programme beneficiary list served as the sampling frame for identifying eligible households. From this list, households were selected using the study's sampling procedure. Upon reaching a selected household, the researcher or trained research assistant identified the primary caregiver of the eligible child and explained the purpose of the study, what participation involved, confidentiality of the information provided, and the voluntary nature of participation.

Data analyses: quantitative data were cleaned for analysis using Statistical Package for Social Sciences (SPSS) version 23 software. Descriptive statistics were used to summarize the participant characteristics, household sociodemographic characteristics, universal child benefit cash utilization, complementary nutrition messages, knowledge of respondents on infant and young child nutrition (IYCN), practice of IYCN, nutrition and health status of children enrolled in the study, infant and young child feeding practices, and household dietary diversity. Analysis of numerical data was done using measures of central tendency and measures of dispersion, while analysis of categorical data was carried out by calculating frequencies and proportions. Measures of association were used to explain the association between independent variables (cash transfer, nutrition message, dietary quality) and dependent variables (nutritional status). This was done using Chi-square and Fisher's Exact test when more than 20% of the cells had expected frequencies of less than five. All the analyses were done at a 5% significance level, two tailed at a 95% confidence interval. The results of this study were presented in text, tables, and graphs.

Qualitative data were analyzed using MAXQDA version 2020. Themes and sub-themes were generated based on the study objectives and the focus group discussion and Key informant interview transcripts uploaded into MAXQDA. Data were coded, and analysis was done on common themes.

Ethical considerations: ethical approval for this study was obtained from the Egerton University Institutional Scientific and Ethics Review Committee with reference number EUISERC/APP/253/2023, and a research permit with reference number 361191 was obtained from the National Commission for Science, Technology, and Innovation. Relevant permission was sought from the Department of Health-Kisumu County and Nyando Sub-County where the study was conducted. Consent of the respondent was also sought before collecting the data. The respondents were assured of confidentiality of the data collected.

Results     

Respondents and household sociodemographic characteristics: the respondents and sociodemographic characteristics are shown in Table 1. Over half, 219 (54.7%) of the respondents resided in Kakola Location. The average age of the respondents was 31.3 (SD = 9.17) years, with the majority, 377 (94.2%) of the respondents being females. More than three-quarters, 316 (79.0%) of the respondents were married to 1 spouse, and some 130 (32.5%) of them had completed primary education. The average household size was 5.5 (SD = 1.79) with the majority, 325 (81.2%), of the household heads being males. The mean age of the heads of the households was 38.0 (SD = 10.35) years. Most, 137 (34.3%), of the household heads completed primary education. The highest number of households, 163 (40.8%), had their source of income as business, and most, 278 (69.5%) of the households' monthly income was less than 5,000 Kenya shillings.

Utilization levels of cash transfer: from quantitative data collected, shown in Figure 2, 389 (97.3%) of the respondents had used the money on household food items. Another 106 (26.5%) spent the money on hospital bills, 11 (2.8%) on transport/fare, 138 (34.5%) on household assets, 59 (14.8%) on school fees, 93 (23.3%) on other uses such as child clothing, boosting businesses and paying rent, and one (0.3%) sent to a relative. A similar trend was also displayed from FGDs and KIIs conducted, with a larger proportion of the cash used for buying food at 44%, clothes (33%), taking care of children (17%), and meeting medical expenses for their children (6%).

From the focus group discussion, a larger proportion of the cash was used for buying food (44%), clothes (33%), taking care of children (17%), and meeting medical expenses for their children (6%). From the qualitative data, the bulk of the money was spent on household food items, with carbohydrates food like cereals forming the bulk of it. Priority was given to the foods consumed by the entire household rather than the children who were targeted by the project.

Dietary patterns: on dietary diversity for the 6-23 months old children, 22.2% attained minimum dietary diversity. The mean number of food groups from which the children ate was 3.5 (SD = 1.75). The proportion of 6-23 months old children who achieved minimum meal frequency in this study was 58.7%. The daily average number of meals for the children was 3.7 (SD = 1.25). The proportion of children 6-23 months old who achieved minimum acceptable diet (i.e., those who achieved both minimum dietary diversity and minimum meal frequency) was only 15.9% (Figure 3).

Regarding specific food groups that the child consumed in the last 24 hours, 77.8% of the children had taken grains, roots, and tubers; 42.9% consumed legumes, nuts, and seeds; 82.5% fed on milk and milk products. Another 38.1% ate flesh products (meat, fish, poultry); 30.2% ate eggs; 41.3% ate vitamin A-rich fruits and vegetables, and 36.5% ate other fruits and vegetables (Figure 4).

Effectiveness of complementary nutrition messages: most respondents, 347 (86.8%), had received training or information on how to improve infant and young child feeding (IYCF) practices. The majority, 310 (77.5%), of the respondents received information from community health volunteers (CHVs). Additionally, 152 (38.0%) received from health facility, 90 (22.5%), phone short message services (SMS), 27 (7.0%), mobile phone audios, 155 (38.8%) mother-to-mother support groups, 12 (3.0%) neighbors/friends with 40 (10.0%) receiving from other sources such as save the children team, UNICEF, and seminars. The key informant interviews also yielded diverse approaches in delivering nutrition messages to universal child benefit (UCB) beneficiaries like community health volunteers, social protection lay volunteers, and automated SMS. Additionally, utilizing SMS and mother-to-mother support groups was also key in disseminating the messages.

The study participants received training on various IYCF practices. Most of the respondents, 332 (83.0%), were trained on exclusive breastfeeding. Besides this, over three-quarters, 309 (77.3%), were trained on complementary feeding, over half, 248 (62.0%), on food hygiene, some 256 (64.0%) on proper food storage and cooking, some 162 (40.5%) on handwashing, and nearly one-third, 129 (32.3%), on appropriate nutrition care for sick children. These trainings and information were key in improving knowledge and practices of the respondents.

In this study, a participant who had a sum of at least six correct responses out of the nine knowledge-related variables on infant and young child nutrition in Table 2 (at least 60% score) was considered to have good knowledge on infant and young child nutrition (IYCN). A sum of five and below correct responses out of the nine variables (less than 60%) was considered a poor knowledge level. The majority (385, 96.3%) of the study participants had good knowledge of IYCN in this study.

To determine the proportion of caregivers whose practice of IYCN was good, this study analyzed the following fifteen IYCN practices. They included handwashing facility use, toilet use, household water treatment, handwashing after visiting the toilet, handwashing before/after eating, handwashing after handling a sick person, and handwashing after handling garbage. The other behaviors included were handwashing after changing a baby's diaper, handwashing before/during food preparation, soap use for handwashing, adequacy of antenatal care visits, iodized salt use, action taken when a baby rejects food, health-seeking behavior for a sick child, and prioritization of persons when serving food. A participant who had a sum of at least nine correct practices out of the fifteen IYCN practices (at least 60% score) was considered to have good practice of IYCN. A sum of eight and below correct practices out of the fifteen IYCN practices (less than 60%) was considered to have poor IYCN practice. This study found that about three-quarters (312, 78.0%) of the respondents had good practice of infant and young child nutrition.

The high knowledge and practice showed that the complementary messages shared with caregivers were effective in improving child nutrition outcomes. Table 2 summarizes the comparison of malnutrition among the children in this study at baseline and after nutritional interventions (cash transfer and messaging). The prevalence of underweight was significantly reduced by 20.2% (95% CI: 15.6%, 24.8%; p < 0.001). Similarly, the prevalence of stunting among children enrolled in this study was remarkably reduced by 24.6% (95% CI: 19.2%, 30.0%; p < 0.001). The prevalence of wasting was reduced by 0.9% (95% CI: -1.9%, 3.7%; p = 0.529). Early initiation of breastfeeding in the study area significantly increased by 13.5% (95% CI: 8.6%, 18.4%; p < 0.001). Bottle-feeding was reduced significantly by 80.6% (95% CI: 77.0%, 84.2%; p < 0.001). In addition, minimum dietary diversity (MDD) for the children reduced notably at the time of the study compared to the pre-study period by 14.9% (95% CI: 3.8%, 26.0%; p = 0.009).

Association of cash transfer and nutrition message intervention with diet quality and nutrition status of children aged 6-59 months in Nyando Sub-County: this study found out that cash transfer, nutritional messages, and breastfeeding were not statistically associated with either underweight, stunting, or wasting among children aged 6-23 months (Table 3). However, among children aged 24-59 months, receipt of nutritional messages through phone audios was significantly associated with stunting (χ² = 19.415, p < 0.001), nutritional messaging delivered at health facilities was significantly associated with wasting (χ² = 4.272, p = 0.039) while breastfeeding initiation also had a statistically significant association with underweight (χ² = 18.093, p = 0.001) and stunting (χ² = 12.874, p = 0.012).

Association between bimonthly receipt of cash transfer and minimum meal frequency, minimum dietary diversity, minimum acceptable diet: as shown in Table 4, there was no association (P>0.05) between bimonthly cash transfer and diet quality indicators of minimum dietary diversity, minimum meal frequency and minimum acceptable diet.

Bivariate binary logistic regression of cash transfer, message intervention, breastfeeding, and malnutrition: bivariate binary logistic regression was performed between the dependent and independent variables, which had a significant association with Pearson Chi-square analysis. This was done to understand the crude odds ratio (COR) and confidence intervals (CI) of the association. The results indicated that 24- 59-month-old children whose caregivers received nutritional messages via phone audios were approximately 5.9 times more likely to be stunted than those whose caregivers did not report phone audios as a source of nutritional messages (COR = 5.874, 95% CI: 2.457 - 14.046). Further surprisingly, odds of wasting increased among 24- 59-month-old children whose caregivers received nutritional messages at a health facility (COR = 2.648, 95% CI: 1.019 - 6.883). Breastfeeding initiation after 24 hours was significantly associated with higher odds of being underweight and stunting (COR = 16.500, 95% CI: 1.088 - 250.176) compared to breastfeeding initiation of less than 1 hour (Table 5).

Discussion     

Cash transfer programs have the potential to improve nutrition outcomes if correctly used. In this study, the majority of the caregivers used the money on household food items. This is because their income level was not adequate to support household food items that could eventually benefit the children targeted by the project, with almost three-quarters of the caregivers having reported an income of less than five thousand shillings. The caregivers also spent the cash received on other uses such as hospital bills, purchasing household assets, transportation fees, payment of school fees, buying clothing for children, boosting business, and rent payment. Similar studies like Manley et al. [9] have reported that a significant proportion of cash transfer beneficiaries, 50%, used their cash transfer in acquiring food. The high spending on food items was, however, not reflected in dietary diversity, especially among children aged 6-23 months thereby not significantly associated with the children's nutrition status. This could be due to household food insecurity because of low purchasing power that leads to spending any income received through cash transfer on majorly carbohydrate diets to benefit the entire household members as opposed to children 6 to 23 months of age. This leaves the children vulnerable to malnutrition from poor feeding and dietary practices. This is evident in this study where more than three-quarters of the caregivers showed good knowledge of infant and young child nutrition (IYCN), with less than those with knowledge displaying good practices, further demonstrating that having knowledge does not translate to good IYCN practices when the purchasing power is low. This might have been used as a coping strategy for household food insecurity, where food was purchased to benefit the entire household and not the targeted children in the program. The diversion of the money to other uses might have also contributed to low dietary diversity among children aged 6-23 months.

The minimum dietary diversity of children aged 6-23 months in this study was low, a decline from the baseline despite the majority of the respondents using cash transfer on food items. This contradicted the findings in a similar study by Save The Children International [10] which reported higher minimum dietary diversity among households benefiting from cash transfers compared with the control group, suggesting a positive relationship between cash transfers and minimum dietary diversity. In this study, cash transfer was not associated with dietary diversity of children again because the caregivers considered all household members when purchasing food items and not the children who were the target in the UCB project. This explains why the minimum dietary diversity of children was lower than the overall household dietary diversity. The targeting of all household members in food purchase resulted in slightly over half of the children attaining minimum meal frequency, but the minimum acceptable diet, which is a composite of minimum meal frequency and minimum dietary diversity, remained low. This could have been affected by low dietary diversity since the caregivers did not prioritize the targeted children in the project when purchasing food. This was clearly shown with the high carbohydrate consumption, which agrees with the children's diets in Kenya as shown by the Kenya National Bureau of Statistics [11] depicting children's diets in Kenya being majorly starchy staples and dairy products with little of high biological value proteins such as eggs and flesh foods. The food consumption patterns of children in the present study were also as that of Southern Benin, where starchy staples were widely consumed while only 2.1% of the children consumed eggs [12]. The same consumption pattern was also observed in Western Kenya by Waswa et al. [13].

Nutrition messages on infant and young child nutrition are key to improving the knowledge and practices of caregivers on child feeding practices, and they complement other interventions. In this study, most caregivers reported receiving training or information on how to improve infant and young child feeding (IYCF) practices, with the majority of the caregivers having received the information from community health volunteers (CHVs). There were multiple platforms for sharing nutrition messages, such as through community health volunteers, from health facilities, phone short message services, automated phone audios, mothers to other support groups, and through neighbors and friends. Community health volunteers reached many people because they conducted regular household visits; hence was the preferred method of nutrition message sharing with the caregivers. This was also supported by the fact that most community health volunteers were community members and people known to caregivers; hence, there was already established trust between them and the caregivers.

Most messages received were drawn from the Baby Friendly Community Initiative manual on infant and young child nutrition practices. The messages were aimed at improving knowledge and practices of the caregivers on infant and young child nutrition. The majority of the caregivers were knowledgeable about IYCF, with about three-quarters of the caregivers having good practice of infant and young child nutrition. Good knowledge and practice were determined by a score of 60% and above, with a score of less than 60% displaying poor knowledge and practice [14,15]. The findings agreed with those by the Food and Agriculture Organization (FAO) [16] in a similar study that established that message interventions in an integrated cash and message programs have the potential of benefiting the families with nutrition knowledge and training. This knowledge could result in improved practices that eventually help in improving children's nutrition status by enabling the households to prepare nutrient-sufficient meals. However, this was not the case in this study as the minimum dietary diversity and minimum acceptable diet for children aged 6 to 23 months remained low even after the intervention despite good knowledge. This could be attributed to caregivers prioritizing the entire household over the children targeted by the project as a means of addressing household food insecurity to cushion all household members at the expense of vulnerable children. Another reason could have been competing needs, where the cash transfer was used in other ways, such as transportation and household assets, thereby reducing the cash available for healthy foods for children. This agreed with the findings from key informants and focus group discussions that showed the same result, with the majority of the caregivers who were women using the money to pay for the respective village merry-go-round.

Infant and young child nutrition practices such as early initiation of breastfeeding and bottle feeding showed a significant increase compared to what was reported in the baseline data [17]. Early initiation of breastfeeding increased, while bottle-feeding significantly reduced from further showing that the complementary messages were effective in promoting good practices that contribute to good nutrition outcomes. Breastfeeding initiation after 24 hours was significantly associated with higher odds of being underweight and stunting compared to breastfeeding initiation of less than 1 hour. This emphasized the importance of early initiation of breastfeeding as a lifesaving intervention and was consistent with established evidence on optimal infant feeding practices. Bivariate binary logistic regression was conducted to estimate crude odds ratios for variables that showed significant associations in Chi-square analysis. The findings showed that children aged 24-59 months whose caregivers received nutrition messages through phone audios had higher odds of stunting, while receipt of nutrition messages at health facilities was associated with higher odds of wasting. These findings may reflect program targeting caregivers with already vulnerable children rather than a causal effect of the interventions. This is because caregivers tend to visit health facilities when a child is already sick or malnourished, so messaging is reactive and not causal.

The effectiveness of complementary messages was supported by qualitative data, with all the key informants reporting that the complementary messages were more effective compared to cash transfer. This can be seen in the excerpt below from one of the key informants: "It helped in strengthening multi-sectoral nutrition coordination, capacity building on Baby Friendly Community Initiative (BFCI) for community health assistants, community health volunteers, and health workers that in turn was disseminated to the beneficiaries. It is the BFCI messages that were effective" (key informant, August 2023).

An excerpt from a focus group discussion also showed a similar trend: R3: "Those trainings really helped me a lot. I left taking my child to the clinic when the child was nine months, but now after getting the training, I took my child to the clinic until the child turned two years and six months am still taking him to the clinic. I did not know the preferred quantities of porridge the child should be given after six months, so we were given a certain cup which we can use to feed the child, and we were taught on continued breastfeeding to two years and beyond" (FGD participant named R3, August 2023).

Cumulatively, the study found that the combination of cash transfers and message intervention had positive impacts on the nutrition status of children in the study area. Prevalence of stunting, underweight, and wasting reduced compared to baseline, which could imply the intervention was effective. This agreed with the findings from Awojobi et el. [18] that found that enhancing dietary access through cash transfer and healthcare information through complementary messages improves the nutritional and medical outcome of benefitting household since message interventions equips caregivers with the knowledge and skills of infant and young child feeding practices while the cash transfer supplements income to purchase nutritious food. This was clearly demonstrated through this study with a significant reduction in stunting and underweight. Wasting also reduced but was not a significant reduction. It also agreed with another study in Bangladesh by Hypher et al. [19] that found that the nutritional effectiveness of cash transfers was reflected when dietary-based messages were shared with the benefiting households. Like Manley et al. [20], the study recorded a significant drop in stunting rates among benefiting households, but this did not largely affect the weight outcome of the children, even though there was a slight drop in the cases of wasting.

The findings from the key informants also indicated an improvement in other indicators such as vitamin A coverage, further confirming the effectiveness of the cash plus interventions. However, cash transfer was not significantly associated with minimum dietary diversity, minimum meal frequency, and minimum acceptable diet, and no association was established between minimum dietary diversity, minimum meal frequency, and minimum acceptable diet with the nutrition status of children in the study population. This can be attributed to diversions of cash transfer to other uses such as buying clothes and payment of merry-go-round by the caregivers as reported by key informants and focus group discussions. The conflict witnessed in some of the households regarding the utilization of the cash transfer received, as reported in key informants and group discussions, could have resulted in diversion of the cash as the parents fought over the cash received.

Limitation of the study: the cross-sectional nature of the study made it impossible to determine the direction of the associations and therefore to establish causality.

Conclusion     

The UCB cash transfer received by the caregivers in Ahero ward, Nyando Sub-County was mainly used in purchasing household food items that benefited the entire household as opposed to children who were the main target for the cash transfer. The complementary nutrition messages were key in improving the nutrition status of children, as seen by an increase in early initiation of breastfeeding and a drop in bottle feeding, which are lifesaving interventions. Cash transfer, nutritional messages, and breastfeeding were not statistically associated with either underweight, stunting, or wasting among children aged 6-23 months. However, among children aged 24-59 months, receipt of nutritional messages through phone audios was significantly associated with stunting, with nutritional messaging delivered at health facilities being significantly associated with wasting and breastfeeding being significantly associated with underweight and stunting. Cumulatively, the study found that the combination of cash transfer and message intervention might have had positive impacts on the nutrition status of children in the study area when compared to the baseline. However, this can only be clearly established in a longitudinal study.

Competing interests     

The authors declare no competing interests.

Authors' contributions     

Phinhas Ochieng Atieno conceptualized and designed the study, conducted the fieldwork, developed the methodology, analyzed the data, and drafted the manuscript; Daniel Onguru and Maureen Cheserek provided academic guidance, supervised the study design and methodology, supported data interpretation, and critically reviewed the work from concept development through to manuscript preparation. All the authors read and approved the final version of this manuscript.

Acknowledgments     

The authors are grateful to Jaramogi Oginga Odinga University of Science and Technology for the technical and academic support provided throughout the study. We also sincerely thank the Kisumu County Department of Public Health for granting approval and facilitating the implementation of the study in Nyando Sub-County. Special appreciation goes to the technical teams on the ground, including the sub-county nutrition coordinator, sub-county public health coordinator, and community health volunteers, whose guidance and collaboration were invaluable. We further acknowledge the contribution of the research assistants for their dedication during data collection. Finally, we extend our heartfelt gratitude to the caregivers and parents who participated in the study, without whom this research would not have been possible.

Tables and figures     

Table 1: social demographic characteristics of caregivers (n=400) in Ahero Ward, Nyando Sub-County in August 2023

Table 2: effectiveness of the nutritional interventions in Ahero Ward (n=448), Nyando Sub-County in August 2023 (comparing study and baseline)

Table 3: association of cash transfer, message intervention, breastfeeding and malnutrition among 6-23 months and 24-59 months old children (n=448) in Ahero Ward, Nyando Sub-County in August 2023

Table 4: association between cash transfer and minimum meal frequency, minimum dietary diversity, minimum acceptable diet and nutritional status of children aged 6-59 months (n=448) in Ahero Ward, Nyando Sub-County in August 2023

Table 5: bivariate binary logistic regression of cash transfer, message intervention, breastfeeding, and nutrition status of children aged 6-59 months (n=448) in Ahero Ward, Nyando Sub-County in August 2023

Figure 1: conceptual framework (adapted from UNICEF conceptual framework of malnutrition)

Figure 2: levels of utilization of cash transfer money

Figure 3: proportion of children meeting minimum dietary diversity, minimum meal frequency and minimum acceptable diet in Ahero Ward, Nyando Sub-County in August 2023

Figure 4: food groups consumed by children aged 6-23 months in the previous 24 hours in Ahero Ward, Nyando Sub-County in August 2023

References