Perception of risk for hypertension and overweight / obesity in Cape Coast , Ghana

1Department of Community Medicine, School of Medical Sciences, College of Health and Allied Sciences, University of Cape Coast, Ghana 2Department of Surgery, School of Medical Sciences, College of Health and Allied Sciences, University of Cape Coast, Ghana 3Harvard Medical School, 250 Longwood Avenue, Boston, MA, USA 4Department of Surgery, Beth Israel Deaconess Medical Centre, 330 Brookline Avenue, Boston, MA, USA


INTRODUCTION
Obesity is no longer a problem only for Western or industrialised countries, but has now reached global proportions. 1With an estimated 500 million obese adults and another 1.5 billion of them overweight or obese 1 , projections are that nearly 60% (3.3 billion) of the world's population could be overweight or obese by 2030. 2 Available evidence reveals that between 2000 and 2004, 10% of adults in West Africa were obese with about half of the urban population and 60% of women being overweight or obese. 3,4 Ghana, the Demographic and Health Survey (DHS) reports that the proportion of overweight or obese women increased from 25% in 2003 to 40% in 2014. 5Obesity is a key risk factor for hypertension, which is an in-creasing public health challenge in both the developed and developing world. 6ecent trends indicate that the prevalence of hypertension in developing countries is already as high as is seen in the developed countries. 6,7The evidence in Ghana suggests an increasing prevalence of hypertension in both rural and urban populations, with high body mass index as a risk factor. 8,9 Ghana, the prevalence of hypertension has increased sharply over the last two decades.For instance, in public health facilities alone, the number of reported new cases of hypertension is estimated to have increased 10fold from about 49,087 in 1988 to 505,180 in 2007. 10he prevalence of hypertension in Ghana reported from various studies ranges between 25% and 48% with higher rates in urban populations than in rural populations. 11ealth behaviour models suggest that recognizing a significant health risk is an important step towards the adoption of protective health behaviours such as seat belt use or diabetes self-management. 12In that regard, perceiving a certain degree of risk for a weight-related chronic disease such as hypertension may strongly impact on weight-reduction behaviour and the consequent minimisation of such risks.Studies elsewhere have examined the association between obesity and other obesity-related chronic disease risk perception 13 , however, there remains a dearth of knowledge regarding the subject in the West African sub-region, particularly in Ghana.
The few studies 9,14,15 conducted in Ghana, have mostly assessed awareness of general health risks associated with overweight or obesity, without looking at perception of risks.Though knowledge of risks associated with obesity is important, an individual's perception of risk presents a stronger motivation for weight-reduction behaviour. 16,17This study examined the association between perception of risk for hypertension and overweight/obesity.Given the increasing prevalence of overweight/obesity and hypertension, this study highlights the importance of people's perceptions of weightrelated health risks and the promotion of healthy weight outcomes.

Sampling and data collection
The study was conducted at the Central Regional Hospital (now Cape Coast Teaching Hospital) in Cape Coast, Ghana, the main referral centre for both primary and secondary healthcare facilities within the Region.The 400-bed hospital serves the residents within the Cape Coast municipality, as well as the other districts in the region.On the average about 110,068 patients are seen at the outpatient department (OPD) annually.
Based on the annual average number of OPD cases, a minimum sample size of 383 was estimated to provide a representative sample assuming overweight/obesity prevalence of 50% in the population, absolute precision/confidence level of 5% with a design effect of 1.0 for a random sample.Although a minimum sample size of 383 was calculated, this was adjusted upwards by 10% to 421 to increase statistical power and cater for non-response or incompleteness of data.
A systematic random sampling procedure was employed in selecting participants who were at least 18 years old from the OPD between June and July 2012.The process began with a construction of a sampling frame based on patient folder numbers following which every n th (systematic) folder was selected.For instance, on the first day every 5 th folder was selected, while on the last day every 3 rd folder was selected.Individuals whose folder numbers were systematically selected were then identified for informed consent to be obtained and subsequent participation in the study.This process continued until the required sample size was attained.
To avoid response or measurement bias, pregnant women, lactating mothers, and patients who were too sick or had chronic diseases, particularly hypertension, diabetes mellitus, chronic kidney disease and heart failure were excluded from the study.Each eligible participant underwent a 15 to 20-minute survey interview conducted by the first and second authors.The survey was designed to elicit information covering relevant themes including socio-demographic details, mass media exposure, obesity-related chronic disease awareness, chronic disease risk perception, weight change among others.
In addition, each participant's weight and height were measured in kilograms and metres respectively, using standard procedures and equipment.The body mass index (BMI), expressed as kg/m 2 , was calculated and recorded.The BMI scores where then used to categorise participants into WHO standard BMI groupings -underweight, BMI < 18.5; normal weight, BMI 18.5-24.9;overweight, BMI 25.0-29.9;obese, BMI > 30.0. 18hical approval was obtained prior to the study by the Institutional Review Boards of Harvard Medical School, Boston, MA, USA and the University of Cape Coast Institutional Review Board, Cape Coast, Ghana.

Statistical Analysis
The data was entered into Microsoft Excel, 2007 and transferred to STATA version 11.0 for cleaning and further management.The current analysis is based on 400 (196 men and 204 women) participants with complete data for the purpose of this study.The analysis involved using proportions to describe the sample and factors associated with overweight/obesity.A BMI of 25 kg/m 2 increases one's risk for mortality and noncommunicable diseases 18,19 , thereby informing the construction of the dependent variable (overweight/obesity) for the study.Thus, participants with BMI scores of 25 kg/m 2 or above were classified as overweight/obese, while those with a BMI below 25 kg/m 2 were classified as not overweight/obese.This categorisation has widely been ap-plied in similar studies. 5,20,21and ensures that robust estimates are obtained in analysis. 20Perception of risk for hypertension was the main independent variable.The question on perception of risk for hypertension was "How likely do you think you are to develop HIGH BLOOD PRESSURE?"This was constructed as a dichotomous variable (not likely and likely).
In order to examine the robustness of association between overweight/obesity and perception of risk for hypertension, other covariates (knowledge factors, and socio-demographic factors) were controlled for.The knowledge factors considered were the participants' knowledge of the effect of overweight/obesity on health, and mass media exposure (radio and television).The assumption is that knowledge of how overweight/obesity affects one's health could influence his/her perception of risk for chronic diseases such as hypertension.
Similarly, perception of risk may be influenced by one's exposure to obesity-related health information through the mass media.The second group of mediating factors considered in this study were socio-demographic factors including age, sex, educational level and marital status.In that regard, three successive binary logistic models were estimated, starting with a model with perception of risk for hypertension (Model 1).This was followed with a second model (Model 2) which included knowledge factors and a final model (Model 3) with the addition of the socio-demographic factors (age, sex, educational level and marital status).This modelling technique aimed at assessing the influence of other factors on the initial relationship between perception of risk for hypertension and overweight/obesity.Consequently, the estimates from the composite model (Model 3) were relied on in assessing the overall association between perception of risk for hypertension and overweight/obesity.Post-hoc model fitness was obtained using the Hosmer-Lemeshow goodness-of-fit test.

RESULTS
Table 1 presents background characteristics of participants used for this study, as well as the prevalence of overweight/obesity by each characteristic.With the sample almost equally distributed between men (49%) and women (51%), the mean age was 33 years (SD=12.5)with more than half (64%) of participants younger than 35 years.Over three-quarters (80%) of the participants had obtained higher education, while 38% were married.
More than two-thirds (64%) of participants perceived themselves as not likely to develop hypertension.In terms of knowledge of the effects of obesity on health, a greater proportion (74%) of participants believed that obesity has a large effect on health.Mass media exposure was very high, with about 92% and 90% of participants having access to radio and television, respectively.39% of the participants were overweight/obese.More than half of the respondents in each age group were found to be overweight/obese, except for participants in the age groups below 25 years.Half of the women were overweight/obese, compared with about 28% of men.Half of those with primary education were overweight/obese, compared with their counterparts with higher education (39%).More married participants (55%) were overweight/obese.Overweight/obesity was higher among those who perceived themselves as likely to develop hypertension (46%) than those who perceived themselves as unlikely to develop the condition (33%).
Overweight/obesity was higher among those who acknowledged that obesity has a large effect (41%) on health, compared with those who indicated that obesity has no effect (26%) on health.From the results, overweight/obesity was more common among those with mass media exposure (40% for radio and 42% for television) than those without mass media exposure (36% for radio and 22% for television).
Table 2 presents the results of three logistic regression models fitted to examine the association between perception of risk for hypertension and overweight/obesity, starting with the bivariate association between perception of risk for hypertension and overweight/obesity (Model being overweight/obese were significantly higher (OR = 1.77, 95% CI = 1.09, 2.88) for those who perceived themselves to be likely to develop hypertension than those who did not perceive themselves as likely to develop hypertension (reference category).
With the inclusion of knowledge factors in Model 2 there was no change in the initial association found between perception of risk for hypertension and overweight/obesity.Those with exposure to television were found to be about 3.7 (95% CI = 1.17, 11.63) times more likely to be overweight/obese compared to those with no exposure.
When socio-demographic factors (age, sex, educational level and marital status) were fitted in Model 3, the association became stronger and more significant between overweight/obesity and those who perceived themselves to be likely to develop hypertension (OR = 2.2195% CI = 1.23, 3.96), and those with exposure to television (OR = 6.29, 95% CI = 1.57, 25.15).
Compared to those who were single, those in relationships (but not married) were significantly less likely (OR = 0.41, 95% CI = 0.200, 0.84) to be overweight/obese.The Hosmer-Lemeshow goodness of fit test was applied in Model 3 to assess the fitness of the data for the model.The model's chi-square test statistic was 3.44 (degrees of freedom = 8) and a p-value (0.9036) greater than 0.05, implying that the model fits the data.

DISCUSSION
This study found a notable disparity between men and women in terms of overweight/ obesity, with half the women categorized as overweight/obese about 28% of men.This confirms the findings of a previous study in Cape Coast 22 and the burgeoning problem of obesity in Ghana and other developing countries. 23his investigation determined that an independent association exists between perception of risk for hypertension and overweight/obesity, while controlling for the effect of other factors (socio-demographic and knowledge factors) using logistic regression models.Contrary to our expectation, that a positive risk perception for hypertension serves as a motivation to engage in healthy weight-seeking behaviours to prevent overweight/obesity, we found higher probabilities of overweight/obesity among those who perceived themselves as likely to develop hypertension.
Although this finding deviates from our theoretical expectations, we attempt First, once overweight/obesity has occurred, there is the possibility for perception of risk to be modified to reflect that weight status. 24Given the cross-sectional nature of the study, it is difficult to say whether perception of risk for hypertension is based on overweight/obesity status or otherwise.One way to disentangle this type of interdependence between perception of risk and overweight/obesity is through a longitudinal study. 25ain, our findings could also be linked with the fact that higher health literacy is often associated with more knowledge of chronic diseases which may influence perception of risk for chronic diseases such as hypertension. 13Majority of the study participants had high exposure to electronic media (92% radio and 91% television), and a greater proportion (74%) also believed that obesity had a large effect on health.This could be an indication of high knowledge of chronic diseases which could have influenced their perception of risk for hypertension based on their weight status.
Perhaps such participants are yet to or in the process of translating their knowledge of the impact of obesity on health and perception of hypertension risk into weight reducing behaviours towards attaining healthy weight.
That notwithstanding, the nature of association between perception of risk for hypertension and overweight/obesity found in the current study presents a window of opportunity for public health intervention, particularly the promotion of healthy weights.
The influence of television exposure on obesity is well documented.Evidence indicates that the more time people spend on watching television, the more they are likely to gain weight or become overweight/obese. 26Similarly, this current study found significant higher odds of overweight/obesity among participants with exposure to television than those without exposure, as well as those with exposure to radio.Television viewing promotes overweight/obesity by displacing the amount of time spent in physical activity, and through one's exposure to advertisement or marketing of high calorie foods which influence diet patterns. 27The high (91%) level of television exposure in the sample suggests a greater amount of "sit time" is spent in television viewing than engaging in physical activity.This also suggests a greater susceptibility to the influence of food advertising or marketing on their diet.However, these assumptions may be limited by the fact that this study did not measure the amounts of daily media exposure, as well as the content of the messages participants were exposed to, therefore, this finding should be interpreted with caution.
The amount and content of exposure to media messages is one research area that needs to be explored further.Except for those aged 45-54, the study found the probabilities of being overweight/obese to significantly increase with age.The association between increased age and overweight/obesity generally corroborates with prior studies. 28,29This could be explained by changes in metabolism, as well as changes in people's daily routines characterized by less physical activity as they grow older.Like in most urban environments, majority of people in Cape Coast are not engaged in strenuous physical labour to earn their living.Thus weight gain becomes more likely, particularly if there is no conscious effort to exercise or engage in other physical activity.
Gender disparity in obesity is another issue that has been well documented in the literature.Studies in developing countries have observed women to be more overweight or obese than men.In that regard, this study found overweight/obesity to be highly significant and more likely among women than men.Amegah et al. 22 similarly found overweight and obesity to be higher among women than men in Cape Coast.
Other studies in Ghana also indicate higher prevalence of overweight and obesity among women, plausibly as a result of the historic Ghanaian social desirability for a larger female body as ideal. 4,14 om the study, those who were in relationships but not married were less likely than their single and married counterparts to be overweight/obese.
Studies on obesity and marital status have generally found overweight and obesity to be more associated with being married that being unmarried. 30,31The predominantly young nature of our study participants could account for the negative association between those overweight/obesity and those in relationship but not married.With high levels of media exposure, young adults may be more conscious of their body image and as a result they may take dietary or physical activity measures to control their weight.

CONCLUSION
Overall, the study demonstrates a strong positive association between perception of risk for hypertension and overweight/obesity.The results highlight the need for the design of programmes to help individuals appreciate the reality of weight-related health risks, as well as the need to embrace lifestyles that promote healthy weight.In Cape Coast, overweight/obesity appears to cut across all adult age groups, and more so among women and those with exposure to television.Given the high levels of exposure, the mass media particularly television should be seen as a useful means to channel behavioural change communication messages.Beyond that, interventions aimed at promoting healthy weight outcomes should be instituted.

Table 1
Characteristics of respondents and overweight/obesity prevalence a Proportion in each category found to be overweight/obese

Table 2
Results of logistic regression on perception of risk for hypertension and overweight/obesity