This study not only assesses the prevalence of overweight and hypertension but also describes the modifiable socio-behavioural factors associated with these in a low-income setting. The finding that 18% of participants were overweight shows a high burden of overweight among people aged 35–60 years. Compared with findings from a rural cohort in southern Uganda where 11% of the general population were overweight 
, these findings imply that the likelihood of being overweight may be significantly higher in people aged 35–60 years than in the general population. However, these rates are lower than those reported in rural South Africa 
, and in higher-income countries 
. Prevalence of hypertension in this population was 21%, consistent with other studies in sub-Saharan Africa 
. In addition, the finding that 46% of participants were pre-hypertensive demonstrates the relevance of targeted prevention through life-style measures. The high burden of overweight and hypertension suggest an increased risk for non-communicable diseases like type-2 diabetes and CVD, and hence the need for these to be addressed by the health system.
The non-modifiable factors found to be associated with being overweight included sex, age and family history of diabetes. Sex was the most significant of these factors, implying a possible connection with gender related factors. This is consistent with findings from other settings in sub-Saharan Africa (SSA) 
. In contrast to our findings, estimates for most higher-income countries show that obesity is higher in men 
. Studies have shown that there are positive socio-cultural attitudes towards being overweight among women in SSA 
. However, the WHO recommended cut-offs for classifying abdominal obesity may not be appropriate for African women 
. The finding that increasing age is associated with being overweight may be related to increased sedentariness with age, which is a modifiable factor. Similar findings were demonstrated in a study in Morocco 
. The importance of these non-modifiable factors lies in defining whom to target for prevention of overweight at primary care levels in low income countries.
This study affirms that age is an important non-modifiable factor associated with hypertension, similar to what was found in other contexts in sub-Saharan Africa 
. The increased likelihood of hypertension from 45–49 years and older age-groups has important implications for targeted screening. Decisions about the target age-group for primary care level screening for hypertension could be guided by this finding. Similar to a study in South Africa 
, sex was not significantly associated with hypertension. This contrasted with findings from a study in Rukungiri in rural western Uganda 
. However, the Rukungiri study targeted a wider age-range (18 years and above). Socio-economic status was not found to be associated with hypertension, a finding that has been demonstrated in other studies 
The modifiable risk factors associated with being overweight included location of residence, socio-economic status, physical activity and dietary diversity. The observed increase in likelihood of being overweight with socio-economic status and the rural-urban divide may be attributed to ability to afford the more expensive energy-dense foods 
. Our study population could be in the early phases of the nutritional transition, where obesity is more frequent in wealthier individuals 
. The link between physical activity and being overweight has been demonstrated in urban settings in sub-Saharan Africa 
. Our study also shows that 84% of participants meet the minimum recommended physical activity level. STEPS surveys in 22 African countries show that the majority of the population meet the minimum physical activity requirement, but with variability between countries, ranging from 46.8% in Mali to 96.0% in Mozambique 
. The finding that moderate diversity diets were associated with a lower likelihood of being overweight could mean that diverse diets are more likely to include vegetables and fruits 
. In urban Benin, dietary quality did not show significant association with obesity. This was attributed to their population being in the early phases of the nutritional transition 
. Modifiable factors associated with hypertension included location of residence and being overweight. The finding that overweight people in rural areas were significantly more likely to be hypertensive than those in peri-urban areas has implications for targeted screening.
Our study finds that those more knowledgeable about lifestyle diseases were more likely to be overweight and hypertensive. This paradox may be due to hypertensive people knowing more about lifestyle diseases than non-hypertensives. However, it may suggest that knowledge alone may not be sufficient to change behaviour. The finding that harmful alcohol taking and tobacco use were not associated with being overweight or hypertensive could be due to the low volumes used in this population, or that the sample sizes for these sub-populations were not sufficient.
We recognize a number of methodological limitations in this study, including using knowledge about diabetes as a proxy for knowledge about lifestyle diseases, categorisation of variables that were initially measured as numerical variables, and use of BMI to test for associations, despite its known shortfalls. However, BMI categorizations have been calibrated and recommended by the WHO as a measure of CVD risk, and are widely used in NCD risk assessments, including the STEPs. All categorisations (blood pressure, physical activity and BMI) were based on the standard criteria recommended by the WHO. Limitations arising from using self-reports to assess lifestyles are also noted, but were mitigated by using validated tools. In addition, dietary assessment was limited to dietary diversity scores that do not take into account the quantity of nutrients eaten. A full assessment of nutritional factors was outside the scope of this study. The study was conducted in a HDSS setting, where populations know that they are under observation. However, there are no on-going interventions on NCDs.