Patterns, socioeconomic inequalities and determinants of healthy eating in Kenya: results from a national cross-sectional survey

Abstract

Objective

The burden of non-communicable diseases is rising in low-and-middle-income countries, with diet being a key risk factor. This study aimed to assess the patterns, socioeconomic inequalities and determinants of eating healthy in Kenya. The study is the first in Kenya to use a healthy diet index to assess dietary patterns.

Design and methods

We analysed cross-sectional data from the 2015/16 Kenya Integrated Household Budget Survey. The study’s outcome variable was a continuous healthy diet index (HDI) constructed using principal component analysis from nine WHO/Food and Agriculture Organization (FAO) healthy diet recommendations. The HDI score and WHO/FAO healthy diet recommendations met were summarised for Kenyan households. Using the concentration index, we examined the socioeconomic disparities in healthy eating. In addition, multivariable linear regression was used to determine factors that influence healthy eating in Kenya.

Results

A total of 21 512 households in Kenya were included, of which 60% were rural and about two-thirds headed by males. The HDI score ranged between −1.13 and 1.70, with a higher value indicating healthier eating. Overall, the average HDI score was 0.24 (95% CI: 0.24 to 0.25), interpreted as moderate. We identified key determinants including socioeconomic status and urban–rural residency differences. Healthy eating was concentrated among higher socioeconomic households, regardless of gender or location. Higher socioeconomic status (β=0.28, 95% CI 0.26 to 0.30), rural residence (β=0.18, 95% CI 0.15 to 0.20), household head being in union (β=0.04, 95% CI 0.02 to 0.06) or employed (β=0.05, 95% CI 0.02 to 0.08) were significantly associated with increased HDI scores, whereas male-headed households and lack of education were associated with significant decreases in HDI scores on average.

Conclusions

Most Kenyan households do not meet all the healthy dietary recommendations, and socioeconomic inequalities exist in eating healthy. Targeted interventions that promote healthy eating based on key determinants in Kenya are required.

Correspondence to Elvis Omondi Achach Wambiya; [email protected]

STRENGTHS AND LIMITATIONS OF THIS STUDY

• This study used a nationally representative dataset making the results generalisable to the Kenyan population.

• The generation and analysis of the healthy diet index as a continuous score reduced the potential bias of information loss in the outcome variable, which increases the validity of our results.

• Some mixed foods were excluded from the analysis because of the absence of their nutritional component values in the Kenya Food Composition Tables.

• The study being cross-sectional, we cannot infer causality on the determinants of healthy eating in Kenya.

• Despite this being the latest available survey, the fact that this survey was conducted between 2015 and 2016 means that dietary habits may have changed since and therefore may not accurately reflect the current patterns in Kenya.

Introduction

Non-communicable diseases (NCDs) are becoming more prevalent in low-and-middle-income countries (LMICs).^1–5 NCDs have been linked to unhealthy lifestyle practices such as poor dietary habits, sedentary behaviour, cigarette use and harmful alcohol use.^{6 7} In LMICs, a nutrition transition has been documented, with shifts from traditional diets to more processed and refined foods high in saturated fats, trans fats, sugar and salt.^6–8 Obesity, diabetes and cardiovascular disease are among the diet-related NCDs that have been linked to this.^{9 10} For instance, global rates of overweight and obesity have almost tripled since 1975, with LMICs accounting for about 70% of overweight or obese people.¹¹ Furthermore, cardiovascular diseases and diabetes are major causes of morbidity and mortality globally and have been increasing in LMICs.^{12 13}

Healthy eating habits are essential for avoiding diet-related chronic diseases. The WHO established a Global NCD Action Plan in 2013 to combat the rising global burden of NCDs.¹⁴ As a result, norms and criteria for eating a healthy diet have been established. The Eatwell Guide, the NOVA (not an abbreviation) classification and the WHO’s and the Food and Agriculture Organization (FAO)’s healthy diet recommendations are among them.^15–18 Evidence-based interventions that promote healthy diets include salt intake reduction, replacement of trans-fat with polyunsaturated fat and public awareness on eating healthy. Despite the available standards and guidelines, increasing evidence shows poor dietary practices in LMICs.¹⁹ The public health implications of NCDs in these countries may become unmanageable without effective policies, initiatives and interventions.²⁰

The rising burden of NCDs in developing countries has led to a growing body of research on dietary patterns and their drivers.^{1 21} Studies conducted in high-income countries (HICs) have often used a composite healthy diet index (HDI) based on WHO healthy diet guidelines to assess dietary patterns in their population.^22–24 In contrast, most studies in LMICs have assessed dietary behaviour based on individual food components such as salt, sugar and fruit and vegetable consumption rather than using a composite score.^25–27 Evidence from studies in LMICs has identified food costs and socioeconomic status as key determinants of dietary behaviour.^28–31 Urban–rural disparities and gender differences in dietary choices have also been reported.^{25 32–34} However, context-specific factors influencing these dietary patterns remain inadequately explored.

Despite growing research on diet and NCD prevention, few studies in Kenya have comprehensively examined healthy dietary patterns using a composite indicator. Most existing studies have focused on single dietary components, such as fruit and vegetable consumption, salt or sugar intake rather than an overall measure of healthy eating.^{26 35 36} To inform evidence-based interventions and policies aimed at promoting healthy eating, a deeper understanding of the patterns and determinants of healthy eating in Kenya is required. Socioeconomic status has been identified as a key determinant of dietary behaviour. However, socioeconomic inequalities in dietary behaviours remain underexplored. According to time preference theory, individuals with higher socioeconomic status are more likely to invest in long-term health benefits, including healthier food choices, due to lower discount rates and greater awareness of the health consequences of poor nutrition.^37–40 Conversely, individuals from lower socioeconomic backgrounds may face short-term financial constraints that lead to prioritisation of cheaper, energy-dense but less nutritious foods over healthier alternatives.

To address these knowledge gaps, our study examined the patterns, socioeconomic inequalities and determinants of healthy eating in Kenya using a composite HDI. By generating this empirical evidence, this study aims to inform policy interventions and targeted programmes to improve dietary habits and reduce the burden of diet-related NCDs in Kenya.

Methods

Study design and participants

This study used cross-sectional data from the 2015/16 Kenya Integrated Household Budget Survey (KIHBS). The KIHBS is a nationally representative household survey that provides integrated household-level data on a wide range of indicators in order to assess the progress made in improving the living standards of the population.⁴¹ Data from three main tools of the KIHBS survey were used for our analysis: (1) household members’ information questionnaire, (2) household level information questionnaire and (3) household consumption expenditure information questionnaire. The household members’ questionnaire collected information on the gender of the household head, education status of the household head, age of household head and household members and marital status of the household head. The household-level questionnaire collected information on household size, residence of the household (urban/rural) and household-level expenditure on non-food items. The consumption expenditure questionnaire collected information on the types and quantities of foods consumed by Kenyan households from purchases, own production, own stock and gifts over a 7-day recall period.

Sampling strategy

A two-stage cluster sampling design was employed to select a nationally representative random sample for the KIHBS survey. In the first stage, 2400 clusters were randomly selected comprising 988 in urban and 1412 in rural areas. In the second stage, 16 households were randomly selected per cluster, with a subsample of 10 households designated for the main KIHBS. This resulted in a final sample of 24 000 households, including 14 120 rural and 9880 urban households. Sampling weights were applied based on selection probabilities of enumeration areas, clusters and households derived from the fifth National Sample Survey and Evaluation Programme master sample.

For this study, analysis was restricted to households that reported consuming foods and dietary quantities available in the Kenya Food Composition Tables 2018 (KFCT). Households that reported consuming restaurant mixed foods, canteen foods, beer, wine, spirits, tobacco or stimulants, or narcotics, whose dietary component quantities could not be obtained, were excluded. After excluding 2488 households, the final analytical sample comprised 21 512 households (online supplemental figure 1).

Measures

Outcome variable

The outcome variable for this study was a continuous HDI developed using the 2003 WHO/FAO expert recommendations on diet, nutrition and prevention of chronic diseases¹⁸ and the 2018 updated WHO healthy diet fact sheet.⁴² Nine dietary components and their WHO/FAO cut-off values were used to construct the composite HDI index as shown in online supplemental table 1. Nutrient composition and energy information for foods consumed in the KIHBS survey were obtained from KFCT. The KFCT lists the energy quantities (kcal) of the nutritional components per 100 g edible portion on fresh weight for commonly consumed foods in Kenya.⁴³ Principal component analysis was used to generate the composite HDI from the nine dietary components.

Explanatory variables

Predictor variables included in the study were the household head’s gender, age, education, marital status, occupation, residence, socioeconomic status, household size and the number of members in the household in different age groups. Age was categorised into four groups: below 30 years, 30–44 years, 45–59 years and 60 years and above. The education level of the household head comprised four categories: no education, completed primary, secondary and above and other education. Marital status had two categories: in union (this included those married and cohabiting) and not in union (this included those separated/divorced, widowed and never married). Occupation of the household head was categorised into three categories: employed, self-employed and unemployed. Residence comprises rural or urban areas.

Socioeconomic status was measured using the total aggregated consumption expenditure per adult equivalent in the household. This was an aggregate measure of food and non-food consumption expenditures of the households following the best-practice guidelines provided by Deaton and Zaidi.⁴⁴ Adult equivalents at the household level were calculated using the steps described by Smith and Subandoro.⁴⁵ The food consumption component included expenditures of food consumed from purchases, own production, own stock and gifts over a 7-day recall period. The non-food expenditure components included household expenditure on house rent, water, electricity, gas, other cooking fuels and healthcare over the last 1 month; expenditure on clothing and footwear over the last 3 months; expenditure on education, household goods, furniture and fittings, communication, recreation and culture, insurance, financial, new/secondhand motor vehicles and accessories and miscellaneous over the last 12 months. The aggregate consumption expenditure per adult equivalent was categorised into five quintiles, that is, poorest, poor, middle, rich and richest.

Data analysis

The sociodemographic characteristics of the study sample were described using frequencies and proportions. Counts and proportions were presented to show the distribution of categorical variables, while means and SD summarised the continuous variables. Means and 95% CI summarised the proportions of households meeting each of the healthy diet recommendations for each of the nine dietary components. The difference in proportion tests was used to assess the differences in the proportions meeting recommendations by gender of the household head, residence and socioeconomic status. Two sample t-tests were used to test differences in HDI scores by gender and residence. One-way analysis of variance (ANOVA) and Bonferroni tests were used to test differences in HDI by the different quintiles of socioeconomic status. Means, 95% CI, minimum and maximum were used to summarise the HDI index overall by gender and residence. The summary of HDI by county was mapped on the Kenyan map using ArcGIS software.

The concentration index (CI) was used to examine whether healthy eating is evenly distributed across poorer or richer households, with respect to gender and place of residence. To compute the CI, households were ranked by wealth quintiles beginning with the poorest in the population. A concentration curve (L(s)) was plotted that presents the cumulative percentage of the population ranked by wealth quintiles against their cumulative percentage of healthy and unhealthy eating.^46–51 The CI is computed between −1 and 1, with a negative value signifying that healthy eating was concentrated among the poorest households, while a positive index value implied that healthy eating was concentrated among the richer households. When there is no inequality, the CI value will be zero. The CI is two times the area between the concentration curve (L(s)) and the diagonal, and it is given by the following formula:

(1)CI=2μcov(FCi,ri)

where μ is the mean of real food expenditure use; ${\displaystyle r_i}$ is the fractional rank of the i^th individual; ${\displaystyle F{C}_i}$ is the household consumption and cov() is the covariance. Socioeconomic inequalities nationwide, by gender and residence, were presented as the CI value, SE and p value. The index was considered significant if the p value was <0.05.

A multivariable linear regression model was fitted to assess factors associated with healthy diet consumption in Kenya. Crude and adjusted marginal effects and 95% CI were presented for each determinant. Variables were considered significant determinants of eating healthy if p values were <0.05. For all the analyses, survey weights were used to account for survey design and clustering. Data analysis was performed using STATA statistical software V.15.0 (StataCorp, College Station, Texas, USA).

Patient and public involvement

Patients or community members were not involved in this study, as it was an analysis of secondary data. However, the larger study within which this analysis was embedded disseminated the results of this analysis to members of the community residing in selected counties in Kenya, namely Nairobi, Uasin Gishu, Kisumu, Isiolo and Mombasa.

Results

Sociodemographic characteristics of the study sample

A total of 21 512 households in Kenya were included in the sample, of which 60% were from rural areas and about two-thirds were headed by males. The average household size was 4.0 (SD 2.4) members, and the average age of household heads was 43 years (SD 15.7); male 42 years (SD 14.9) and female 46 years (SD 17.0), with about two-thirds of the study sample falling between 30 and 59 years of age. The average monthly per adult equivalent consumption expenditure was US$76.5 (SD 75.6) with 57% in the rich and richest category and 25% falling in the poor and poorest categories. About three-quarters of household heads were in union (71%) while slightly more than two-fifths had attained primary level education (44%). About half of the household heads were employed, while 41% and 11% were self-employed and unemployed, respectively (table 1).

Table 1

Sociodemographic characteristics of the study sample

	N	%
Monthly per adult equivalent total consumption expenditure, mean (SD) in US$	76.5 (75.6)
Age group
Below 30 years	3890	18.1
30–44 years	8234	38.3
45–59 years	5308	24.7
60 years and above	4080	19.0
Residence
Urban	8556	39.8
Rural	12 956	60.2
Gender of household head
Female	7266	33.8
Male	14 246	66.2
Education of household head
No education	4446	20.7
Primary	9540	44.4
Secondary and above	7387	34.3
Other*	139	0.7
Employment status†
Unemployed	2262	10.5
Employed	10 431	48.5
Self-employed	8819	41.0
Marital status of household head
Not in union	6229	29.0
In union	15 283	71.0
Total	21 512	100.0

*Other education category comprised informal education, that is, madrassa/duksi.

†Employed comprised those in salaried employment at the public and private sector, self-employed included those with private business, and the unemployed included those with no employment. Sampling weights and clustering are used to account for the sampling design of the survey.

Patterns of healthy diet consumption in Kenya

Number of dietary recommendations met by Kenyan households

Online supplemental figure 2 depicts the number of healthy diet recommendations followed by Kenyan families in general and by location. In Kenya, no household met seven or more of the healthy eating guidelines. Only 3% of Kenyan households followed six of the nine healthy eating guidelines. The majority of Kenyans (84% cumulative proportion) followed four or fewer of the healthy eating recommendations, with no household following all seven. The same was observed with regard to residence, with 82% and 86% of rural and urban households meeting four or less of the healthy diet criteria, respectively.

Proportion of households meeting WHO/FAO recommendations for HDI components

Online supplemental table 2 displays the percentage of Kenyan households who followed the WHO/FAO recommendations for each HDI component. The required fruit and vegetable intake was met by 45% of households, with more female-headed households (50%) and urban households (52%) meeting the recommendations than their male-headed and rural counterparts, which was statistically significant. When it came to the recommended total fat consumption, the majority of households (87%) followed the guidelines, with more female-headed (88%) and rural homes (88%) following the guidelines than male-headed (86%) and urban households (86%). Only 25% of households met the recommended total carbohydrate intake, with female-headed households (30%) and rural households (29%) being more likely to do so. Total protein and dietary fibre requirements followed a similar pattern. Overall, about a third of the households met the recommended saturated fat intake, with more urban households meeting the recommendations compared with their rural counterparts. Only 5% of households met the recommended polyunsaturated fats intake level, with more urban households meeting the recommendations compared with their rural counterparts. Overall, only 3% of households met the recommended total trans fat energy requirements, with more male-headed households meeting the recommendations. For total carbohydrates and total proteins, more than 90% of the households were above the recommended daily intake, while for polyunsaturated fats, about two-thirds were above the healthy diet recommendations (online supplemental tables 3 and 4).

Distribution of HDI scores by gender, residence and socioeconomic status

The average HDI scores in the study sample are summarised in table 2 by gender, residence and socioeconomic level. The mean HDI index ranges from −1.13 to 1.70, with a higher number suggesting healthier eating habits in accordance with WHO/FAO dietary guidelines. Kenya’s average HDI score was 0.24, with urban residents scoring higher (0.25) than rural residents (0.23). There was no significant difference observed in the HDI score by gender of the household head, as observed in the overlapping confidence intervals. With regard to socioeconomic status, the findings indicate an increasing trend in the HDI score with increasing socioeconomic status, meaning that households with higher socioeconomic status were eating healthier.

Table 2

Summary of HDI score overall, by gender, residence and socioeconomic status

	Mean	95% CI
Overall	0.241	0.235 to 0.247
Age group
Below 30 years	0.278	0.263 to 0.292
30–44 years	0.248	0.238 to 0.257
45–59 years	0.209	0.197 to 0.221
60 years and above	0.225	0.212 to 0.239
Residence
Urban	0.254	0.245 to 0.264
Rural	0.231	0.223 to 0.239
Gender of household head
Female	0.247	0.236 to 0.258
Male	0.239	0.231 to 0.246
Education of household head
No education	0.016	0.002 to 0.030
Primary	0.240	0.231 to 0.249
Secondary and above	0.322	0.312 to 0.332
Other*	−0.046	−0.122 to 0.030
Employment status†
Unemployed	0.140	0.120 to 0.160
Employed	0.236	0.228 to 0.245
Self-employed	0.266	0.256 to 0.275
Marital status of household head
Not in union	0.257	0.246 to 0.268
In union	0.234	0.227 to 0.242
Socioeconomic status
Poorest	−0.021	−0.032 to −0.010
Poor	0.170	0.157 to 0.182
Middle	0.253	0.240 to 0.266
Rich	0.340	0.327 to 0.354
Richest	0.465	0.449 to 0.480

*Other education category comprised informal education, that is, madrassa/duksi.

†Employed comprised those in salaried employment at the public and private sector, self-employed included those with private business, and the unemployed included those with no employment. Sampling weights and clustering are used to account for the sampling design of the survey.

HDI, healthy diet index.

Distribution of HDI score by county

The distribution of the HDI scores by the 47 counties in Kenya is presented in figure 1. The results showed that Western counties had higher HDI values as compared with counties in arid and semiarid lands (ASAL) areas, which had the lowest HDI scores. It was also evident that counties that were neighbouring higher HDI counties had moderate HDI scores. Online supplemental table 5 shows the average HDI scores for each of the 47 counties.

Figure 1. The distribution of HDI scores in Kenya by county. HDI, healthy diet index.

Socioeconomic inequalities of eating healthy in Kenya

Online supplemental table 6 shows the results of socioeconomic inequality analysis of HDI overall, by gender and residence, based on the CI. The results indicated that eating healthy foods in Kenya was concentrated among the richest households (CI=0.40, p<0.01). This result is confirmed by the finding in figure 2A, which shows that the concentration curve was below the line of perfect equality. Similar results (figure 2B,C) were observed with respect to gender (female CI=0.46, p<0.01; male CI=0.37, p<0.01) and residence (rural CI=0.50, p<0.01; urban CI=0.41, p<0.01). The results suggest that pro-rich inequality in eating healthy food was greater among female-headed households (diff=0.09, p<0.01) and rural households (diff=0.07, p<0.01).

Figure 2. (A) Concentration curve of HDI in Kenya. (B) Concentration curve of HDI in Kenya by gender. (C) Concentration curve of HDI in Kenya by residence. HDI, healthy diet index.

Determinants of healthy food consumption in Kenya

Table 3 presents the marginal effects from the adjusted multivariable linear regression model assessing the determinants of eating healthy in Kenya. Socioeconomic status was a strong positive predictor of healthy eating, with a one-unit increase in the log of per adult equivalent household expenditure associated with a 0.280 increase in HDI (95% CI: 0.262 to 0.297; p<0.001). The gender of the household head was significant, with male-headed households having lower HDI scores than female-headed households (β=−0.066, 95% CI: −0.087 to –0.045; p<0.001). Compared with households where the head had secondary education or higher, those with no formal education had significantly lower HDI scores (β=−0.149, 95% CI: −0.181 to –0.118; p<0.001). Similarly, households in the ‘other’ education category (including informal education) had significantly lower HDI scores (β=−0.257, 95% CI: −0.357 to –0.157; p<0.001). Compared with unemployed household heads, those who were self-employed had a moderately higher HDI (β=0.039, 95% CI: 0.009 to 0.069; p<0.01), while those employed in formal sectors had the highest HDI scores (β=0.050, 95% CI: 0.017 to 0.084; p<0.001). There were significant rural–urban differences in dietary behaviour, with rural households having higher HDI scores compared with urban households (β=0.175, 95% CI: 0.150 to 0.200; p<0.001). Finally, marital status was significantly associated with dietary behaviour. Households headed by individuals in a union had higher HDI scores compared with those headed by individuals not in a union (β=0.039, 95% CI: 0.018 to 0.061; p<0.001).

Table 3

Determinants of eating healthy in Kenya

	Unadjusted model		Adjusted model
	Marginal effects	95% CI	Marginal effects	95% CI
Socioeconomic status	0.233†	0.217 to 0.250	0.280†	0.262 to 0.297
Household size	−0.031†	−0.036 to −0.027	−0.003	−0.007 to 0.000
Gender of household head
Female (Ref)
Male	−0.009	−0.029 to 0.012	−0.066†	−0.087 to −0.045
Log of household head age (years)	−0.057†	−0.087 to −0.028	0.091†	0.064 to 0.117
Education status of household head
Secondary and above (Ref)
No education	−0.262†	−0.289 to −0.235	−0.149†	−0.181 to −0.118
Primary	−0.002	−0.023 to 0.018	0.012	−0.009 to 0.032
Other	−0.288†	−0.406 to −0.170	−0.257†	−0.357 to −0.157
Occupation of household head
Unemployed (Ref)
Self-employed	−0.009	−0.029 to 0.012	0.039*	0.009 to 0.069
Employed	0.045†	0.024 to 0.067	0.050†	0.017 to 0.084
Residence
Urban (Ref)
Rural	−0.023	−0.053 to 0.007	0.175†	0.150 to 0.200
Marital status of household head
Not in union (Ref)
In union	−0.023*	−0.044 to −0.002	0.039†	0.018 to 0.061

Socioeconomic status = log of monthly per adult equivalent total consumption expenditure. Survey weights were used to account for the survey design and clustering.

**p<0.05, ***p<0.01

*p<0.05

†p<0.01

Ref, reference category.

Discussion

This study investigated the patterns, socioeconomic inequalities and determinants of healthy eating in Kenya. The findings reveal that only 3% of Kenyan households met six of the nine healthy eating recommendations. Furthermore, more than 80% of the population met four or fewer of the recommendations, suggesting that a significant percentage of Kenyan families are not following the full set of WHO/FAO healthy eating guidelines. Kenyan households scored moderately on the HDI, with urban households scoring higher than rural ones. We also discovered that eating healthy was associated with socioeconomic position, residence and the household head’s gender, age, occupation, marital and educational status. In the same vein, we found that healthy eating was concentrated among households with higher socioeconomic status. The findings of this study could be used to inform programmes that help Kenyan families make healthy food choices.

This study is the first in Kenya to use a healthy diet index to assess patterns of eating healthy. Previous studies conducted in HICs have used a healthy diet index.^{22–24 52} The findings of the present study show that the majority of households were not meeting the healthy diet recommendations. This is in line with previous studies conducted in LMICs that show evidence of poor dietary behaviour.^{7 53–55} While Kenya’s vegetable and fruit intake was low, it was much higher than what was found in South Africa (32%)⁵⁶ and what was found in an analysis that involved 52 LMICs (22%).⁵⁴ A study in South Africa also reported that fruits were considered luxuries that were only bought if money was left over after the purchase of staple food.⁵⁷ The low fruit and vegetable consumption observed in these countries may be linked to the low supply of fruits and vegetables that has been reported in SSA.⁵⁸

The findings suggest substantial regional variations in eating healthy in Kenya. The Western and Central counties had the highest HDI values indicating healthier eating compared with counties in ASAL. These variations are somewhat expected because of differences in climatic conditions and social and economic factors, among other factors in different regions of Kenya.⁵⁹

The study found that socioeconomic inequalities exist with healthy eating more concentrated among the wealthy households. This was later substantiated by the multivariate analysis, which also showed a positive and significant association between eating healthy and higher socioeconomic status. Recent global evidence indicates that healthier diets are more costly in LMICs as compared with HICs.^{60 61} Furthermore, the affordability of healthy diets in HICs including the USA and Europe is better than LMICs including parts of Asia and Africa.^{30 62} However, in both settings, higher socioeconomic status is associated with healthier dietary patterns.^{29 30} Our findings are supported by studies in other parts of Africa, including South Africa and Ghana, where socioeconomic status is an important determinant of eating healthy.^{29 31} A study conducted in Kenya showed that the foods the urban poor could afford were not sufficient for them to meet FAO dietary recommendations.²⁸

Gender differences exist in choices made regarding the types of foods consumed in a household.³² Our study found that households headed by a female were more likely to eat healthy. This is consistent with other studies that showed that women generally make healthier food choices by eating more fruits and fibres, avoiding foods high in fats and limiting their salt intake.^{32 34} A study by Sedibe et al⁵⁷ reported that female caregivers were the main promoters of healthy eating practices.

Our study also found that households headed by an educated individual were more likely to eat healthy, which corroborates findings from studies conducted in rural and urban South Africa.^{63 64} Another study in the same setting demonstrated that low education was associated with inadequate fruit and vegetable intake.⁵⁶ Lastly, urban–rural differences have been reported in healthy diet consumption.⁵⁴ It also emerged from our study that rural households were more likely to eat healthy compared with their urban counterparts. A systematic review and meta-analysis of salt intake in SSA found a higher consumption of salt in urban areas compared with rural areas.²⁵ A study conducted in Soweto, South Africa, also demonstrated that urbanisation has led to increased consumption of diets higher in energy and containing more salt, saturated fat and sugar.⁶⁵

The findings of this study highlight critical policy implications to improve access to and consumption of healthy foods in Kenya. Given the observed socioeconomic inequalities in healthy eating, policies should aim to address affordability barriers, particularly for low-income households. One potential strategy is to implement subsidies for fruits and vegetables, making them more accessible to households with lower socioeconomic status. This has been shown to be effective in changing population dietary behaviour.⁶⁶ Additionally, taxation on unhealthy, ultra-processed foods such as sugar-sweetened beverages and bad fats could be explored in Kenya as it has proven effective in shifting consumer behaviour towards healthier options in other settings.^66–68

The regional disparities in healthy eating patterns further emphasise the need for localised interventions. Counties in arid and semiarid regions may benefit from targeted agricultural investments, improved market infrastructure and supply chain enhancements to increase availability from other food-producing regions like the central and western region counties. Furthermore, the study’s findings on gender and education highlight the importance of integrating nutrition education into broader public health campaigns. Community-based programmes that empower male and female household heads with knowledge on nutrition and meal planning could enhance household dietary practices. Similarly, school-based initiatives that promote healthy eating from an early age may help establish long-term behavioural changes. Addressing these multifaceted challenges requires a combination of economic, agricultural and public health policies to ensure that all Kenyans, regardless of socioeconomic status or geographic location, have equitable access and consume healthy diets.

Study strengths and limitations

A major strength of this study is the use of a nationally representative dataset, which makes our results generalisable to the Kenyan population. Second, this study provides novel evidence on patterns, determinants and socioeconomic inequalities of healthy eating in Kenya. Third, the use of the HDI as a continuous score reduced the potential bias of information loss in the outcome variable, which increased the validity of the results. However, there are some limitations to note. Some mixed foods were excluded because their nutritional values were not found in the KFCT, which may have limited the accuracy of nutrient composition data to be used in our analysis. In addition, there may have been potential for recall bias in self-reporting of dietary behaviour that could introduce inaccuracies. Because the study was cross-sectional, we cannot infer causality on the determinants of healthy eating in Kenya. The fact that the secondary data used in this study were collected between 2015 and 2016 means that dietary habits may have changed since, and the results may not accurately reflect the current consumption patterns in Kenya. However, the 2015/2016 survey is the only currently available national survey with food consumption data. The Kenya National Bureau of Statistics has piloted survey instruments for data collection for the updated survey to be conducted in the period 2024–2025.⁶⁹ As soon as the latest data is out, we will update our analysis to capture current evidence. Despite these limitations, this research contributes to our understanding of Kenya’s healthy eating behaviours and associated determinants that can inform health policy interventions and programmes.

Conclusions

In conclusion, the majority of Kenyan households do not meet all the nine healthy dietary recommendations, and socioeconomic inequalities exist with eating healthy concentrated among households with higher socioeconomic status. Furthermore, eating healthy is associated with higher socioeconomic status, living in a rural area and the household head being older, having education, being in employment and in union. The findings from this study could be used to inform policies and tailored interventions that promote healthy eating and the prevention of diet-related NCDs among the Kenyan population. Future research should explore the impact of specific policy interventions, such as food subsidies, taxation on unhealthy foods and nutrition education programmes, on improving dietary patterns and reducing diet-related diseases, particularly among low-income and urban households. This will benefit from updated representative surveys on food consumption to enable longitudinal tracking of changes in dietary patterns and outcomes over time.

We acknowledge the support provided by Kenya National Bureau of Statistics from the time of data request to approval. We also acknowledge the valuable support of Mr. Francis Onyambu of the African Population and Health Research Center in mapping the HDI scores in Kenya.

Data availability statement

Data are available in a public, open access repository. This study used publicly available data from the Kenya National Bureau of Statistics. The data described in the manuscript can be accessed via the following link: https://statistics.knbs.or.ke/nada/index.php/catalog/13.

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

This study used anonymised data from the KIHBS, which are publicly available from the Kenya National Bureau of Statistics (KNBS), thus this study did not require institutional review board approval. The KNBS sought ethical approval and study participants provided informed consent before participating in the original survey.

Footnote

X @shukrifmohamed

Contributors SFM: Funding acquisition, conceptualisation, methodology, writing—review and editing and supervision. HPPD: Funding acquisition, conceptualisation, methodology, formal analysis, supervision, writing—review and editing and designed the research. EOAW: Conceptualisation, methodology, formal analysis, data curation and writing—original draft. LK: Writing—review and editing. JO: Writing—review and editing. IK: Writing—review and editing. JOO: Formal analysis and writing—review and editing. PGI: Formal analysis and writing—review and editing. SK: Formal analysis and writing–review and editing. All authors have read and approved the final manuscript. EOAW is the guarantor for this manuscript.

Funding This work was supported by the International Development Research Centre (grant number 108984-001) and managed by the African Population and Health Research Center. The corresponding author (EW) acknowledges funding support from the Wellcome Trust as part of a doctoral training grant [218462/Z/19/Z] at the University of Sheffield.

Disclaimer The findings, interpretations and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank to its affiliated organisations, or to members of its Board of Executive Directors or the countries they represent.

Map disclaimer The depiction of boundaries on this map does not imply the expression of any opinion whatsoever on the part of BMJ (or any member of its group) concerning the legal status of any country, territory, jurisdiction or area or of its authorities. This map is provided without any warranty of any kind, either express or implied.

Competing interests None declared.

Patient and public involvement Patients and/or the public were not involved in the design, conduct, reporting or dissemination plans of this research.

Provenance and peer review Not commissioned; externally peer reviewed.

Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

References

¹ Alleyne G, Binagwaho A, Haines A, et al. Embedding non-communicable diseases in the post-2015 development agenda. Lancet 2013; 381: 566–74. doi:10.1016/S0140-6736(12)61806-6

² Gaziano TA, Bitton A, Anand S, et al. Growing epidemic of coronary heart disease in low- and middle-income countries. Curr Probl Cardiol 2010; 35: 72–115. doi:10.1016/j.cpcardiol.2009.10.002

³ Gaziano TA. Cardiovascular disease in the developing world and its cost-effective management. Circulation 2005; 112: 3547–53. doi:10.1161/CIRCULATIONAHA.105.591792

⁴ Yusuf S, Reddy S, Ounpuu S, et al. Global burden of cardiovascular diseases: part I: general considerations, the epidemiologic transition, risk factors, and impact of urbanization. Circulation 2001; 104: 2746–53. doi:10.1161/hc4601.099487

⁵ GBD 2021 Diseases and Injuries Collaborators. Global incidence, prevalence, years lived with disability (YLDs), disability-adjusted life-years (DALYs), and healthy life expectancy (HALE) for 371 diseases and injuries in 204 countries and territories and 811 subnational locations, 1990-2021: a systematic analysis for the Global Burden of Disease Study 2021. Lancet 2024; 403: 2133–61. doi:10.1016/S0140-6736(24)00757-8

⁶ Bowry ADK, Lewey J, Dugani SB, et al. The Burden of Cardiovascular Disease in Low- and Middle-Income Countries: Epidemiology and Management. Can J Cardiol 2015; 31: 1151–9. doi:10.1016/j.cjca.2015.06.028

⁷ Popkin BM, Adair LS, Ng SW. Global nutrition transition and the pandemic of obesity in developing countries. Nutr Rev 2012; 70: 3–21. doi:10.1111/j.1753-4887.2011.00456.x

⁸ Popkin BM. The nutrition transition and the global shift towards obesity. Diabetes Voice 2004; 49: 38–41.

⁹ Malik VS, Popkin BM, Bray GA, et al. Sugar-sweetened beverages, obesity, type 2 diabetes mellitus, and cardiovascular disease risk. Circulation 2010; 121: 1356–64. doi:10.1161/CIRCULATIONAHA.109.876185

¹⁰ Malik VS, Schulze MB, Hu FB. Intake of sugar-sweetened beverages and weight gain: a systematic review. Am J Clin Nutr 2006; 84: 274–88. doi:10.1093/ajcn/84.1.274

¹¹ Shekar M, Popkin B. Obesity: Health and economic consequences of an impending global challenge. Washington, DC: World Bank Publications, 2020.

¹² Miranda JJ, Barrientos-Gutiérrez T, Corvalan C, et al. Understanding the rise of cardiometabolic diseases in low- and middle-income countries. Nat Med 2019; 25: 1667–79. doi:10.1038/s41591-019-0644-7

¹³ World Health Organization. Diabetes fact sheet. 2024. Available: https://www.who.int/europe/news-room/fact-sheets/item/diabetes

¹⁴ World Health Organization. Global action plan for the prevention and control of noncommunicable diseases 2013-2020. 2013.

¹⁵ Jones NRV, Conklin AI, Suhrcke M, et al. The growing price gap between more and less healthy foods: analysis of a novel longitudinal UK dataset. PLoS One 2014; 9: e109343. doi:10.1371/journal.pone.0109343

¹⁶ Slavin JL, Lloyd B. Health benefits of fruits and vegetables. Adv Nutr 2012; 3: 506–16. doi:10.3945/an.112.002154

¹⁷ Monteiro CA, Cannon G, Moubarac J-C, et al. The UN Decade of Nutrition, the NOVA food classification and the trouble with ultra-processing. Public Health Nutr 2018; 21: 5–17. doi:10.1017/S1368980017000234

¹⁸ Nishida C, Uauy R, Kumanyika S, et al. The joint WHO/FAO expert consultation on diet, nutrition and the prevention of chronic diseases: process, product and policy implications. Public Health Nutr 2004; 7: 245–50. doi:10.1079/phn2003592

¹⁹ Ronto R, Wu JHY, Singh GM. The global nutrition transition: trends, disease burdens and policy interventions. Public Health Nutr 2018; 21: 2267–70. doi:10.1017/S1368980018000423

²⁰ Abegunde DO, Mathers CD, Adam T, et al. The burden and costs of chronic diseases in low-income and middle-income countries. Lancet 2007; 370: 1929–38. doi:10.1016/S0140-6736(07)61696-1

²¹ Naghavi M, Forouzanfar MH. Burden of non-communicable diseases in sub-Saharan Africa in 1990 and 2010: Global Burden of Diseases, Injuries, and Risk Factors Study 2010. Lancet 2013; 381: S95. doi:10.1016/S0140-6736(13)61349-5

²² Berentzen NE, Beulens JW, Hoevenaar-Blom MP, et al. Adherence to the WHO’s healthy diet indicator and overall cancer risk in the EPIC-NL cohort. PLoS One 2013; 8: e70535. doi:10.1371/journal.pone.0070535

²³ Kanauchi M, Kanauchi K. The World Health Organization’s Healthy Diet Indicator and its associated factors: A cross-sectional study in central Kinki, Japan. Prev Med Rep 2018; 12: 198–202. doi:10.1016/j.pmedr.2018.09.011

²⁴ Cade J, Upmeier H, Calvert C, et al. Costs of a healthy diet: analysis from the UK Women’s Cohort Study. Public Health Nutr 1999; 2: 505–12. doi:10.1017/s1368980099000683

²⁵ Oyebode O, Oti S, Chen Y-F, et al. Salt intakes in sub-Saharan Africa: a systematic review and meta-regression. Popul Health Metr 2016; 14: 1. doi:10.1186/s12963-015-0068-7

²⁶ Mwenda V, Mwangi M, Nyanjau L, et al. Dietary risk factors for non-communicable diseases in Kenya: findings of the STEPS survey, 2015. BMC Public Health 2018; 18: 1218. doi:10.1186/s12889-018-6060-y

²⁷ Lara-Castor L, Micha R, Cudhea F, et al. Intake of sugar sweetened beverages among children and adolescents in 185 countries between 1990 and 2018: population based study. BMJ 2024; 386: e079234. doi:10.1136/bmj-2024-079234

²⁸ Olielo T, Maseno University, Kenya. Food security problems in various income groups of Kenya. AJFAND 2013; 13: 1–13. doi:10.18697/ajfand.59.SC011

²⁹ Temple NJ, Steyn NP, Fourie J, et al. Price and availability of healthy food: a study in rural South Africa. Nutrition 2011; 27: 55–8. doi:10.1016/j.nut.2009.12.004

³⁰ McDermott AJ, Stephens MB. Cost of eating: whole foods versus convenience foods in a low-income model. Fam Med 2010; 42: 280–4.

³¹ Osei-Kwasi HA, Laar A, Zotor F, et al. The African urban food environment framework for creating healthy nutrition policy and interventions in urban Africa. PLoS One 2021; 16: e0249621. doi:10.1371/journal.pone.0249621

³² Shiferaw B, Verrill L, Booth H, et al. Sex-based differences in food consumption: Foodborne Diseases Active Surveillance Network (FoodNet) Population Survey, 2006-2007. Clin Infect Dis 2012; 54 Suppl 5: S453–7. doi:10.1093/cid/cis247

³³ Sedibe M, Feeley A, Voorend C, et al. Narratives of urban female adolescents in South Africa: dietary and physical activity practices in an obesogenic environment. South Afr J Clin Nutr 2014; 27: 114–9. doi:10.1080/16070658.2014.11734499

³⁴ Wardle J, Haase AM, Steptoe A, et al. Gender differences in food choice: the contribution of health beliefs and dieting. Ann Behav Med 2004; 27: 107–16. doi:10.1207/s15324796abm2702_5

³⁵ Okemwa S, Nyakundi C, Ngesa RW, et al. Prevalence and factors associated with sugar-sweetened beverage consumption among adolescents and women aged 10-49 years in Kenya: a cross-sectional. 2024. Available: https://www.researchsquare.com/article/rs-5004584/v1

³⁶ Vila-Real CP de M, Pimenta-Martins AS, Kunyanga CN, et al. Nutritional intake and food sources in an adult urban Kenyan population. Nutr Bull 2022; 47: 423–37. doi:10.1111/nbu.12582

³⁷ Lawless L, Drichoutis AC, Nayga RM. Time preferences and health behaviour: a review. Agric Econ 2013; 1: 17. doi:10.1186/2193-7532-1-17

³⁸ Zhang L, Rashad I. Obesity and time preference: the health consequences of discounting the future. J Biosoc Sci 2008; 40: 97–113. doi:10.1017/S0021932007002039

³⁹ Andersen S, Harrison GW, Lau MI, et al. Eliciting risk and time preferences. Econometrica 2008; 76: 583–618. doi:10.1111/j.1468-0262.2008.00848.x

⁴⁰ van der Pol M, Cairns J. A comparison of the discounted utility model and hyperbolic discounting models in the case of social and private intertemporal preferences for health. J Econ Behav Organ 2002; 49: 79–96. doi:10.1016/S0167-2681(02)00059-8

⁴¹ Kenya National Bureau of Statistics (KNBS). Kenya integrated household budget survey. Nairobi: Kenya National Bureau of Statistics, 2018. Available: https://sun-connect-news.org/fileadmin/DATEIEN/Dateien/New/KNBS_-_Basic_Report.pdf

⁴² WHO. Healthy diet [fact sheet]. 2018. Available: https://www.who.int/nutrition/publications/nutrientrequirements/healthydiet_factsheet/en/

⁴³ FAO, Ministry of Health, Ministry of Agriculture and Irrigation. Kenya food composition tables 2018. 2018.

⁴⁴ Deaton A, Zaidi S. Guidelines for constructing consumption aggregates for welfare analysis. Washington, D.C: World Bank Publications, 2002.

⁴⁵ Smith LC, Subandoro A. Measuring food security using household expenditure surveys. International Food Policy Research Institute, 2007.

⁴⁶ Allen L, Williams J, Townsend N, et al. Socioeconomic status and non-communicable disease behavioural risk factors in low-income and lower-middle-income countries: a systematic review. Lancet Glob Health 2017; 5: e277–89. doi:10.1016/S2214-109X(17)30058-X

⁴⁷ Moradi G, Mohammad K, Majdzadeh R, et al. Socioeconomic Inequality of Non-Communicable Risk Factors among People Living in Kurdistan Province, Islamic Republic of Iran. Int J Prev Med 2013; 4: 671–83.

⁴⁸ O’donnell O, Doorslaer E, Wagstaff A, et al. Analyzing health equity using household survey data: a guide to techniques and their implementation. Washington, D.C: The World Bank, 2007.

⁴⁹ Kakwani N, Wagstaff A, van Doorslaer E. Socioeconomic inequalities in health: measurement, computation, and statistical inference. J Econom 1997; 77: 87–103. doi:10.1016/S0304-4076(96)01807-6

⁵⁰ Wagstaff A, Paci P, van Doorslaer E. On the measurement of inequalities in health. Soc Sci Med 1991; 33: 545–57. doi:10.1016/0277-9536(91)90212-u

⁵¹ Wagstaff A, Watanabe N. On decomposing the causes of health sector inequalities with an application to malnutrition inequalities in Vietnam. Washington, D.C: The World Bank, 2001.

⁵² Stefler D, Pikhart H, Jankovic N, et al. Healthy diet indicator and mortality in Eastern European populations: prospective evidence from the HAPIEE cohort. Eur J Clin Nutr 2014; 68: 1346–52. doi:10.1038/ejcn.2014.134

⁵³ Afshin A, Sur PJ, Fay KA. Health effects of dietary risks in 195 countries, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet 2019; 393: 1958–72. doi:10.1016/S0140-6736(19)30041-8

⁵⁴ Hall JN, Moore S, Harper SB, et al. Global variability in fruit and vegetable consumption. Am J Prev Med 2009; 36: 402–9. doi:10.1016/j.amepre.2009.01.029

⁵⁵ Mazzocchi M, Brasili C, Sandri E. Trends in dietary patterns and compliance with World Health Organization recommendations: a cross-country analysis. Public Health Nutr 2008; 11: 535–40. doi:10.1017/S1368980007000900

⁵⁶ Peltzer K, Phaswana-Mafuya N. Fruit and vegetable intake and associated factors in older adults in South Africa. Glob Health Action 2012; 5: 1–8. doi:10.3402/gha.v5i0.18668

⁵⁷ Sedibe HM, Kahn K, Edin K, et al. Qualitative study exploring healthy eating practices and physical activity among adolescent girls in rural South Africa. BMC Pediatr 2014; 14: 211. doi:10.1186/1471-2431-14-211

⁵⁸ Gebremedhin S, Bekele T. Evaluating the African food supply against the nutrient intake goals set for preventing diet-related non-communicable diseases: 1990 to 2017 trend analysis. PLoS One 2021; 16: e0245241. doi:10.1371/journal.pone.0245241

⁵⁹ Cgiar, Ccafs. Kenya county climate risk profiles. 2018. Available: https://ccafs.cgiar.org/resources/publications/kenya-county-climate-risk-profiles

⁶⁰ Springmann M, Clark MA, Rayner M, et al. The global and regional costs of healthy and sustainable dietary patterns: a modelling study. Lancet Planet Health 2021; 5: e797–807. doi:10.1016/S2542-5196(21)00251-5

⁶¹ Rao M, Afshin A, Singh G, et al. Do healthier foods and diet patterns cost more than less healthy options? A systematic review and meta-analysis. BMJ Open 2013; 3: e004277. doi:10.1136/bmjopen-2013-004277

⁶² Moncayo JR, Hamadeh N, Rissanen M, et al. Over 3.1 billion people could not afford a healthy diet in 2021 - an increase of 134 million since the start of COVID-19. 2023. Available: https://blogs.worldbank.org/, https://blogs.worldbank.org/en/opendata/over-31-billion-people-could-not-afford-healthy-diet-2021-increase-134-million-start-covid [Accessed 25 Jan 2023 ].

⁶³ Taruvinga A, Muchenje V, Mushunje A. Determinants of rural household dietary diversity: the case of Amatole and Nyandeni districts, South Africa. Int J Dev Sustain 2013; 2: 2233–47.

⁶⁴ Muzigaba M. Nutrition education and the cost of healthy food-do they collide? Lessons learned in a predominantly black urban township in South Africa. Int Public Health J 2013; 5: 279.

⁶⁵ Pretorius S, Sliwa K. Perspectives and perceptions on the consumption of a healthy diet in Soweto, an urban African community in South Africa. SAHJ 2011; 8: 178–83. doi:10.24170/8-3-1897

⁶⁶ Niebylski ML, Redburn KA, Duhaney T, et al. Healthy food subsidies and unhealthy food taxation: A systematic review of the evidence. Nutrition 2015; 31: 787–95. doi:10.1016/j.nut.2014.12.010

⁶⁷ Boysen O, Boysen-Urban K, Bradford H, et al. Taxing highly processed foods: What could be the impacts on obesity and underweight in sub-Saharan Africa? World Dev 2019; 119: 55–67. doi:10.1016/j.worlddev.2019.03.006

⁶⁸ Cabrera Escobar MA, Veerman JL, Tollman SM, et al. Evidence that a tax on sugar sweetened beverages reduces the obesity rate: a meta-analysis. BMC Public Health 2013; 13: 1072. doi:10.1186/1471-2458-13-1072

⁶⁹ Kenya National Bureau of Statistics. Pilot survey report: 2024/25 Kenya integrated household budget survey. KNBS, 2024.