Overweight and obesity prevalence among public servants in Nadowli district, Ghana, and associated risk factors: a cross-sectional study

Study area

This study was conducted in the Nadowli District of Ghana. Farming accounts for about 85% of the labour force with maize, millet, groundnut and beans as some of the staples and relatively available all year round.

Study design

This study was cross-sectional in design, and was carried out among public servants in the Nadowli District. Public servants in this study refer to the totality of human resources employed by state or government departments established in accordance with the constitution of Ghana as public service, and those who are on government payroll within the Nadowli District. Women who were public servants but were pregnant at the time of the study were not included. At the time of this study, a total of 775 public servants were employed in a total of 12 public service departments in the district.

We focused on public servants in this study for a number of reasons. First, previous researchers have observed that public servants are an important group of interest when it comes to NCDs because, the lay public generally expect them to be leaders when it comes to living and practicing healthy lifestyles because of their relatively better access to information [14]. At the same time, this group is also at risk of leading sedentary and unhealthy lifestyles partly due to long periods of sitting in work places, and partly because of their relatively better socio-economic status which makes it easier for them to access less physically active modes of transportation such as cars. Second, we focus on this group because less attention has been paid to them in research across Africa and in Ghana, despite the recognition that they could be at heightened risk of NCDs.

Sample size determination

We determined the appropriate sample size for the study using the Public Service of Creative System survey software (http://www.surveysystem.com). Using the population size of 775, with a confidence level of 95%, a sample size of 257 was obtained. However, 271 questionnaires were administered to cater for data entry errors and non-responses.

Sampling methods

In terms of sampling, all the 12 public service departments within the district were listed and each department coded. The codes were folded and put in a basket and picked at random until half of them were picked. The subjects were selected through a proportionate random sampling. This sampling technique ensured the inclusion of greater number of respondents from bigger departments. This ensured that our sample was representative of all the different categories of public servants in the district. In order to select a respondent, any person who was present at the time of entering a particular department was interviewed and was asked to call the next available person. This was done in all selected departments until the total sample size was obtained.

Data collection instruments and procedures

We used structured questionnaires (see Additional file 1) to gather data on respondents’ socio-demographic characteristics such as age, sex, religion, occupation, marital status, health risk behaviours (smoking and alcohol consumption), nutrition knowledge/health and physical activity. Anthropometrics were measured for the calculation of their Body Mass Index. Respondents were weighed wearing light cloths and without shoes. We also measured height of respondents with a suspended microtoise tape to the nearest 1 m without shoes and in minimum clothing standing erect with hands hanging loosely by their sides.

Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters. Overweight/obesity was defined as BMI of ≥ 25 kg/m2; and obesity as BMI of ≥ 30 kg/m2.

Pretesting of data collection instruments

The questionnaire was pretested on selected number of public servants outside the sample area and necessary corrections and adjustments made. The pretest enabled all ambiguities in questionnaire items and responses to be identified and corrected. The pretest also helped to determine how much time was required to complete the questionnaire.

Data management and statistical analysis

The SPSS statistical software (version 20) was used to enter data, clean the data, and as well edit the data for inconsistencies. All questionnaires that were incomplete or did not meet other quality control tests were not included for analysis. Descriptive statistical analysis such as frequency and percentage distributions were first used to summarize the data as well as describe important characteristics of study respondents. Bivariate and regression analyses were then performed to examine the possible relationship between a number of independent variables such as age, sex, marital status, level of education and leisure time activity, and dependent variables such as BMI and Hypertension and Diabetes status. Statistical significance was held at 95% confidence level and at a p-value of < 0.05.

Operational definitions

Socio-demographical status and obesity

Both overweight and obesity were detected among respondents of different socio-demographic backgrounds. Overall, overweight and obesity prevalence were 29.9 and 4.8% respectively. The prevalence of obesity in our study population is higher than the 4.6% prevalence reported among the general adult population in Ghana by [12]. Together with previous findings from other African urban settings such as Nigeria [10, 20], the findings in this study suggests that public servants are indeed one sub-population at particular risk of overweight and obesity, and related consequences of hypertension and diabetes. This result is however counter-intuitive because public servants generally are educated and also have greater opportunities for obtaining healthy behavioural change information. Therefore, one would expect better outcomes and lifestyle indicators for public servants than for the general public [14]. However, our results that both overweight and obesity levels among respondents are higher than the general population in Ghana clearly do not support the assumption above. Rather, the finding lends support to previous research which suggested statistically significant association between obesity and high socioeconomic status (SES) [8]. Indeed, although the link between SES and overweigh/obesity was not directly investigated in our study, it is plausible to think that increased SES among public servants is a factor contributing to the relatively high levels of overweight and obesity prevalence among our study’s respondents. As noted earlier, salaries of public servants over the past 5 years in Ghana have marginally improved. This has ultimately improved their economic status above the general population. In addition to the risks associated with sedentary work and general lifestyle among public servants, consumption of processed foods, unhealthy snacking, consumption of high fat diets away from home are all behaviours that are associated with increasing purchasing power particularly in low-income settings [20]. Coupled with the fact that several of the public servants surveyed in this study were not regularly engaged in any vigorous physical activity or rode cars or public buses to work, our findings suggest the need for promotive health interventions (e.g. healthy eating) targeted at public servants not only in our study context of Ghana but also in other low-income settings such as Nigeria where similar high levels of overweight and obesity have been reported among public servants.

Hypertension/diabetes associated with obesity

The prevalence of 20% in hypertension/diabetes obtained from this study compares favourably with the results of one Nigerian study which reported 27.1% [10]. Overweight and obesity are known risk factors for hypertension and diabetes. This is partly because increases in body weight can lead to increases in insulin resistance as well as defects in insulin secretion. Also, the risk of developing Type 2 diabetes and/or hypertension is increased with increases in overweight and/or obesity [21]. As indicated, respondents who had either diabetes or hypertension were more likely to be overweight and obese compared to those who were not hypertensive or diabetic and was significantly related with BMI (P <0.012). Indeed, hypertension and obesity have been known to be positively related [14]. Findings indicated abdominal obesity in particular played a greater role in elevated blood pressure. In this study however, no association was found between these behavioural factors and selected NCDs. This could partly be explained by the cross- sectional nature of the study. It could also be due to the fact that the behavioural factors examined in this study were self-reported by respondents. Therefore, the possibility of exaggeration of socially acceptable practices by respondents may have affected the relationship.

Smoking and alcohol consumption

This study also examined the link between smoking, alcohol intake and weight status. No association was found between smoking and weight status (p > 0.730). Previous studies have however found alcohol consumption to be positively correlated with BMI in men or in both sexes [22, 23]. Other recent studies have however indicated that the association between alcohol consumption and BMI could be explained by differences in consumption patterns [24–26]. For example, among adults, the amount or intensity of drinking per drinking occasion has been found to be positively correlated with BMI [24, 25]. At the same time, the frequency of drinking has been found to be negatively correlated with BMI [26]. This latter finding would suggest that frequent light drinking might offer a protective effect or benefits. Also no statistically significant association was found between alcohol consumption and BMI (p > 0.109). The lack of association could partly be due to the relatively small proportion of alcohol drinkers and smokers in our study. Again current smokers have lower BMI than former smokers and this study involved only current smokers.

Leisure time and physical activity

Increasingly, a number of researchers are recognizing that physical inactivity arising from sedentary forms of work and lifestyles, changing modes of transportation, and increasing urbanization, is a major factor contributing to obesity [2, 4]. Among the respondents in this study, leisure time activity, which was apparently dominated by sedentary behaviours such as long sitting hours watching television and drinking with friends, was positively and significantly associated with BMI (P < 0.000). This finding supports earlier results by [15] which suggested that sedentary behaviours were associated positively with obesity. The result is also consistent with another study among subjects from two urban and one rural community, which found high levels of overweight and obesity among subjects who were sedentary and engaged in light activity [10]. The lack of physical activity among the public servants surveyed in this study could partly be one of the factors responsible for the high overweight prevalence. In terms of policy response, our findings here suggest the need for promotive health interventions targeted at public servants such as more physical activities such as playing sports and regular exercising. This recommendation is indeed in line with the recent WHO’s recommendations related to the prevention of NCDs in the workplace through diet and physical activity’ [27].

Study limitation

This study was a cross- sectional study which could lead to limited study conclusion between risk factors and overweight. Again BMI alone is not always an accurate predictor of body fat or distribution of adiposity, particularly as it may overestimate body fat in muscular individuals and underestimate in older subjects because of muscle mass loss.