Sub-Saharan Africa (SSA) is currently the world’s fastest urbanizing region with a growth in urban population of 3.5% per annum [1]. Urbanization on the continent has resulted in the nutrition transition [2], characterized by the emergence of Western eating patterns such as the preference of convenient and non-home prepared foods [3]. An epidemiological transition characterized by a rise in cardio-metabolic diseases and cancers is also apparent in SSA [4]. A pooled analysis reported a high prevalence of overweight/obesity, hypertension and low-density lipoprotein cholesterol in Uganda [5]. The 2016 Non-communicable disease (NCD) collaboration report indicated that 30.9% and 8.6% of women and 13.7% and 1.8% of men were overweight and obese, respectively [6]. Obesity is an established risk factor of cardiovascular disease, type 2 diabetes and certain types of cancer [3, 5].

The epidemiological transition has been linked to the nutrition transition. Evidence suggests that a shift from a traditional diet high in carbohydrates and fiber and low in fat and sugar to a Western diet high in energy, saturated fat, sodium, sugar, and low in fiber, increases the risk of obesity and other NCDs [7]. A multi-country study in SSA countries, including Uganda, reported a positive association between processed food dietary patterns and an increased risk of overweight and obesity [8]. To identify public health efforts for tackling the non-communicable disease epidemic, it is necessary to understand the determinants of the changing dietary patterns in a particular population.

Away-from-home (AFH) eating is one of the eating habits that is being adopted in Uganda and could be an important factor in understanding the context of the evolving epidemiological patterns in Uganda. Due to structural consequences of urbanization that do not favor home-food preparation, urban consumers have become more reliant on highly processed, non-home prepared foods [9, 10]. In a 2014 nationwide survey, close to 40% of Ugandans reported eating at least a meal away from home in a week [11]. AFH eating is linked to excessive consumption of energy dense, high fat and sugary foods[10]; and low fruit and vegetable intake [12, 13]. AFH eating also increases the risk of obesity [12, 13], and has been linked to a low-quality diet in developed nations [14, 15].

However, the influence of AFH eating on diet quality in low-income countries is largely not studied. Previous research on this topic has either been conducted in Western countries [14, 15, 16, 17, 18, 19, 20], focused on specific foods and nutrients [10, 21, 22, 23, 24], or region-specific diet indices such as the Mediterranean diet scores (MDS), Health Eating index (HEI) [14, 25] or Dietary Approaches to Stop Hypertension (DASH) [17]. A few studies that were conducted in Uganda only targeted street food consumers and did not analyze the overall diet quality [26, 27]. We found only two studies that investigated empirical dietary patterns in relation to AFH, but none of these studies was conducted in Uganda [19, 20]. Empirically derived dietary patterns provide insights into dietary patterns of countries without established diet indices. Since dietary patterns vary by region and culture, evaluating food intake in form of empirical dietary patterns captures the local overall dietary intake and can be used to offer public health advice focused on locally available foods.

Therefore, this study examined the influence of AFH on empirically derived dietary patterns of Ugandan adults aged 18 ~ 65 years.

Socio-demographic, lifestyle and clinical characteristics of participants according to frequency of AFH eating are displayed in Table 1. About 75% of participants reported eating at least a meal away from home per week. The distribution of age, gender, place of residence and perceived food insecurity significantly differed across AFH categories. Frequent AFH eaters were more likely to be less than 30 years old, men, urban residents and food insecure compared to rare eaters.

Table 1
Characteristics of participants according to frequency of away-from-home eating (n = 375)

• Download Table

The 29 food groups and corresponding food items used to derive dietary patterns are shown in Table 2. Table 3 shows the factor loadings of different food items on the 3 principal components. Three components that explained 33.2% of the total variance in food groups were generated. The components were labelled based on the food items with a factor loading of ≥ 0.4. The first component highly correlated with meats, fish, insects, eggs, SSB, alcoholic and coffee beverages and hence was labelled “animal-based, beverage pattern”. The second component had high factor loadings on bread and buns, fats, milk, chips, sweets and confectionary and was thus labelled “high-fat, sweet pattern”. The third component was highly correlated with traditional cereals, legumes, nuts and seeds, starchy roots, plantain, and fresh juices and was thus labelled “traditional, plant-based pattern”.

Table 2
Food groups used to generate dietary patterns of participants

• Download Table

Table 3
Factor loadings of food groups explaining dietary patterns of participants

• Download Table

Results of the bivariable association between AFH eating and tertiles of the three dietary patterns are presented in Table 4. The distribution of participants across tertiles of the animal-based, beverage pattern significantly differed across frequency of AFH. The proportion of participants in the extreme tertile of the animal-based, beverage pattern was highest among frequent AFH eaters compared to rare eaters. However, except the traditional, plant-based pattern, both the animal-based beverage and high fat, sweet patterns were significantly associated with AFH after adjusting for potential covariates (Table 5). Participants who ate AFH ≥ 5 times/week were 2.85 times more likely to be in the second tertile and 5.64 times more likely to be in the third tertile of the animal-based, beverage pattern compared to rare eaters (OR = 2.85, 95% CI:1.35-6.06 for T2 vs T1; and OR = 5.64, 95% CI:2.50-12.73 for T3 vs T1). In addition, those who ate AFH 3 ~ 4 times/week were 2.70 times and 5.04 times more likely to be in the second and third tertile of the animal-based, beverage pattern respectively, compared to rare AFH eaters (OR = 2.70, 95% CI: 1.15-6.32 for T2 vs T1; and OR = 5.04, 95% CI: 2.02-12.58 for T3 vs T1). On the other hand, the odds of being in the second tertile of the high fat, sweet pattern were significantly higher for frequent AFH eaters compared to rare eaters (OR = 2.61, 95% CI:1.23-5.52).

Table 4
Bivariable association between tertiles of dietary pattern scores and frequency of away-from-home eating (n = 375)

• Download Table

Table 5
Multivariate association between tertiles of dietary pattern scores and frequency of away-from-home eating (n = 375)

• Download Table

For the first time, we sought to understand the influence of AFH eating on the overall diet quality of Ugandan adults. We observed that three quarters of adult Ugandans who participated in this survey ate at least a meal away from home in a week. We identified three dietary patterns: the animal-based beverage pattern, the high-fat sweet pattern and the traditional, plant-based pattern. Frequent AFH eating was associated with increased adherence to the animal-based, beverage pattern and the high fat, sweet pattern.

The prevalence of AFH reported in this study is comparable to that in Singapore [31] but higher than that in the UK [32] and the US Hispanic community study [14]. In nationwide study in Uganda, AFH eating was estimated at about 40% [11]. The sampling method and overall survey methodology used in our study make it difficult to compare our results with those of previous studies. However, our findings suggest that AFH eating behavior may have increased and exceeded that reported in some developed nations. Notably, fast-food use was reported at 70% by previous studies in Uganda [33] while street food users were estimated at 63.5% in Kampala [26]. The rapid socio-economic transformation and urbanization evidenced by an increase in Gross-Domestic Product, a decrease in youth unemployment rate, and an increase in the working and urban population are probable drivers of the increasing popularity of AFH eating in Uganda [34, 35]. In line with the previous association between AFH eating and obesity [12, 13, 14, 17, 25, 36], nutritional and public healthy interventions aiming at reducing AFH eating or encouraging healthier choices during dining out are urgently needed.

This study examined the overall diet quality specific to the Ugandan population, rather than specific nutrients or regional specific diet indices. Furthermore, we used principal components analysis to derive dietary patterns. This approach has the advantage of capturing variations in eating patterns that exist in different populations depending on socio-economic status, ethnicity and cultural background [37]. The influence of AFH eating on empirically derived dietary patterns has not been widely reported which makes comparisons with other studies difficult. Nevertheless, our dietary patterns are comparable to previously derived dietary patterns and reflect the healthiness of the diet in this population [29]. Our findings show that AFH eating is associated with a poor diet quality, consistent with several studies. Full-service and fast-food restaurant use was associated with non-adherence to the Alternate Healthy Eating Index 2010 (AHEI-2010) [14], while fast-food use was associated with non-adherence to the HEI and MDS [25]. Consumption of high energy from fast foods was associated with non-adherence to the DASH diet. Moreover, abstaining from AFH eating appeared to favor consumption of a prudent dietary pattern [19]. Frequent AFH eating was associated with consumption of an ultra-processed food pattern [20], low intake of vegetables [13], high intake of energy , fat, and low intake of micronutrients [23], and consumption of less healthy foods [18]. The two dietary patterns associated with AFH in this study resemble a Western dietary pattern. The animal-based, beverage pattern was correlated with intake of meat, including red meat, eggs, sugar-sweetened beverages, alcohol and coffee. The high-fat, sweet pattern was rich in refined bread, processed fats, deep fried foods and sugary confectionary. Such food choices have been linked to obesity and cardiovascular diseases [7, 8, 12].

Our results highlight a shift from a traditional Ugandan dietary pattern among frequent AFH consumers. Substitution of a traditional dietary pattern rich in carbohydrates and fiber, fruits, vegetables, legumes, nuts and seeds, and low in fat and sugar, with a Western diet high in energy, saturated fat, sodium, sugar and low in fiber increases the risk of obesity and other NCDs [7]. A traditional Ugandan diet comprises a starchy staple, which could be plantain or grains like millet or sorghum; and sauce, which could comprise a mixture of vegetables, legumes or nuts, and meat or fish eaten occasionally [38].

There are probable reasons for the association between AFH eating and unhealthy dietary patterns. It is possible that AFH food establishments offer food items which are characteristic components of unhealthy dietary patterns. However, our study cannot confirm the menu items provided by AFH food establishments since we did not assess the specific foods consumed away from home. Studies from other countries reported that majority of the menu developers considered increasing sales and profits as primary considerations while offering menus. Only 21% considered health and nutrition as important aspects in their menus. In addition, healthy menu items have a low shelf life, increased preparation time, low sales and high labor demands [39]. Notably, street-foods offer foods made from cheap ingredients such as refined flour, maize and hydrogenated oils [2]. This type of AFH eating is very popular in low-income countries [26, 36, 40] suggesting that consumers in low-income countries may be at a high risk eating low quality AFH food. Nevertheless, consumer attitudes might also influence the choice of foods bought AFH. Only a few consumers considered health as an important determinant of the choice of AFH foods [41, 42]. Thus, interconnected consumer and supplier-related factors seem to mediate the relationship between AFH and poor diet quality and these factors will require thorough investigation if AFH eating is to be targeted for health promotion.

To unravel the motivators to AFH eating and design effective solutions that target AFH eating in Uganda, socio-demographic characteristics associated with AFH also deserve special attention. Frequent AFH eating was dominant among men [26, 43], the young [26, 31, 33], urban dwellers [44] and the food insecure. Thus, interventions such as nutritional education aimed at fostering health AFH food choices should target the young, men, the food insecure and urban dwellers.

Several limitations should be considered while interpreting these findings. First, the sampling technique used in this study and the small sample size make it difficult to generalize these findings to the entire Ugandan urban/peri-urban population. However, our sample was relatively close to the composition of semi-urban and urban Uganda, although there are no recent census reports for comparison with national statistics. The urban Ugandan population is 53% women [45], which is similar to the female proportion in our study. We sampled 54% of the employed, slightly above the proportion of urban dwellers in paid employment reported in 2016 (51.3%) and below the employment to population ratio of 58.2% [46]. But it is also important to note that our sample comprised a significant number of students, who could possibly occupy some form of employment but were not counted as employed. Second, the use of an online-based survey might have led to selection bias where young people who own computers and smart phones were more likely to participate. Nevertheless, the median age in our study was 25 years, consistent with the general urban population which is dominated by the 20-39-year-old as of 2019. In addition, those with at least an undergraduate degree were more likely to participate in the study since they have social media access and adequate internet navigation skills. Thus, the educational level in our study might not represent that of the general urban population. Census data estimated the proportion of the employed population in urban Uganda who attained post-secondary education or above to be at only 25.1% in 2016 [46]. Consequently, our findings are not generalizable to the less educated urban dwellers. Third, this was a cross-sectional study which precludes causal inference. Fourth, dietary data was collected using a food frequency questionnaire (FFQ), which relies on memory. Therefore, the study may not have captured dietary intake accurately. Moreover, the qualitative FFQ could not capture the portion sizes of food consumed by participants. This might have affected the derived dietary patterns in our study. Lastly, AFH eating was assessed using a self-reported questionnaire that was not validated for this population. This could have led to inaccurate reporting of dietary intake. In addition, there are different definitions of AFH eating between studies. We did not consider the place where food was prepared making comparison with other studies difficult. The place of food preparation rather than the place of consumption appears to influence the nutritional quality of AFH foods [36].

Nevertheless, this is a novel study, the first to examine the determinants of AFH eating and its influence on overall dietary quality defined by empirical dietary patterns in Uganda. Dietary patterns differ according to culture and ethnicity. This underscores the need to study population specific dietary patterns. In addition, the dietary pattern approach offers a broad picture of eating habits beyond single nutrients, and offers a practical way of giving dietary advice to the population [37]. Thus, we were able to capture a population specific overall eating pattern. We also adjusted for potential covariates to capture the independent relationship between AFH eating and dietary patterns.

Understanding the influence of away-from-home eating on the overall diet quality in Uganda is necessary for early interventions since the influence of this behavior on obesity and other NCDs is known in developed countries. The prevalence of AFH is appalling, and frequent AFH eating is associated with unhealthy dietary patterns. Future studies should investigate the motivating factors of AFH to inform nutritional interventions such as nutritional education aimed to foster healthy AFH choices, especially among the young, men, urban dwellers and the food insecure.