Prevalence and Associated Factors of Hypertension among Women in Southern Ghana: Evidence from 2014 GDHS

Background Hypertension, coupled with prehypertension and other hazards such as high blood pressure, is responsible for 8·5 million deaths from stroke, ischaemic heart disease, other vascular diseases, and renal disease worldwide. Hypertension is the fifth commonest cause of outpatient morbidity in Ghana. Some evidence have illustrated geographical variation in hypertension and it seems to have a heavy toll on women in southern Ghana compared to the north. This study seeks to determine the prevalence and associatedfactors of hypertension among women in southern Ghana using the most recent demographic and health survey (DHS) data set. Materials and Methods This study used data of 5,662 women from the current DHS data from Ghana that was conducted in 2014. Data were extracted from the women's file of the 2014 Ghana DHS. The outcome variable of this current study was hypertension and it was measured by blood pressure, according to guidelines of the Joint National Committee Seven (JNC7). Multivariable binary logistic regression analyses were performed to establish the factors associated with hypertension at the individual and community levels. Results Prevalence of hypertension among women in southern Ghana was 16%. Women aged 40–44 years (aOR = 8.04, CI = 4.88–13.25) and 45–49 years (aOR = 13.20, CI = 7.96–21.89] had the highest odds of hypertension relative to women aged 15–19 years. Women with two births (aOR = 1.45, CI = 1.01–2.07) and those with three births (aOR = 1.47, CI = 1.01–2.15) had a higher likelihood of being hypertensive. Greater Accra women had higher odds (aOR = 1.35, CI = 1.02–1.79) of being hypertensive relative to the reference category, women from the Western region. Women of Guan ethnicity had a lesser likelihood (aOR = 0.54, CI = 0.29–0.98) of being hypertensive. Women who engaged in agriculture had the least likelihood (aOR = 0.72, CI = 0.52–0.99) of being classified hypertensive compared to unemployed women. Conclusion This study has revealed the prevalence of hypertension among women in southern Ghana. The associated factors include age, parity, region, and occupation. As a result, existing interventions need to be appraised in the light of these factors. Of essence is the need for Ghana Health Service to implement wide-embracing health promotion initiatives that accommodate the nutritional, exercise, and lifestyle needs of women in southern Ghana. Having more children is associated with higher propensity of hypertension and consequently, women need to limit childbearing to reduce their chances of being hypertensive. It will also be advisable for women in the Greater Accra region to have frequent hypertension screening, as women in the region exhibited higher hypertension prospects.


Background
Hypertension, coupled with prehypertension and other hazards, such as high blood pressure, is responsible for 8·5 million deaths from stroke, ischaemic heart disease, other vascular diseases, and renal disease worldwide [1]. In high-income countries, hypertension prevalence has declined while health systems have achieved treatment rates of about 80% and control rates of 60% [1]. Hypertension prevalence has generally increased while obesity, which is a risk factor for hypertension, decreased between 1990 and 2019 [1].
Hypertension reduction is a primary goal of the World Health Organization (WHO) global monitoring framework [2]. Besides, the expansion of universal health coverage and primary care in places with low rates of diagnosis, especially sub-Saharan Africa and South Asia, provide an opportunity for improving hypertension care [1].
Hypertension has been a key public health challenge in Ghana. It is known to be the fifth commonest cause of outpatient (OPD) morbidity in Ghana whilst at the same time, 63% of women with hypertension symptoms are oblivious of thier hypertension status [3]. Geographical variations in hypertension have been reported from some countries, including China and Gambia [4,5].
In the case of Ghana, there is a seemingly wide difference in the Southern-Northern prevalence of hypertension with about half of the adults living in some parts of southern Ghana (including Eastern, Volta, Western, and Brong Ahafo regions) having the condition while about one quarter are either overweight or obese [6,7]. e relatively lower prevalence in the north has been associated with the traditional way of living characterized by manual farming, housekeeping, and greater physical activity which people mostly take a walk or travel by bicycle [6][7][8]. e Southern-Northern differences in the prevalence of hypertension warrant empirical investigation since higher prevalence has persisted over the last four decades [6,7].
It is based on this background that the current study aims to examine the prevalence and factors associated with hypertension among women in southern Ghana using the most recent Demographic and Health Survey (DHS) data set. Results from the study will contribute significantly towards the development of local and national level interventions and policies that will contribute to the prevention of hypertension across Ghana.

Study Setting.
e study was implemented in southern Ghana. Our study was based on 2014 Ghana Demographic and Health Survey (GDHS); and it is worth noting that prior to 2018 Ghana had ten administrative regions [9]. As stated by the Ghana Statistical Service, six out of the ten regions constitute southern Ghana and these are Western, Central, Greater Accra, Ashanti, Eastern, and Volta region as illustrated in green in Figure 1. One key distinction between southern and northern Ghana manifest in skeweness in development towards the south [10].

Sources of Data.
is study used current DHS data of Ghana which was conducted in 2014 [11]. Specifically, data was pulled from the women's files of the DHS data set. DHS are national surveys in respective countries carried out approximately every five years in over 90 low-and middleincome countries in the world [12]. DHS concentrates on maternal and child health issues, including physical activity, non-communicable diseases, sexually transmitted infections, fertility, tobacco use, health insurance, and alcohol consumption.
e survey categorically provides data to monitor the demographic and health profiles of included countries [12]. e sample for the present study consisted of women whose blood pressure were taken (aged 15-49 years) and had complete cases on all variables of interest (N = 5, 662). e DHS program permitted us access to the dataset after the evaluation of our concept note. e datasetis freely available and accessible to the public at https://www. measuredhs.com/.

Outcome Variable.
e outcome variable of this study was hypertension, measured by blood pressure. For 2014 DHS, blood pressure was monitored and measured thrice and these followed the UA-767F/FAC (A & D Medical) blood pressure computation with a minimum of 10 minutes interval [10]. Hypertension status was determined by computing the average of the second and third measurements. It was in line with the calibration by similar empirical studies on hypertension that relied on DHS datasets [13]. Guided by the guidelines of the Joint National Committee Seven (JNC7), hypertension was computed as the average systolic blood pressure of ≥140 mmHg and/or an average diastolic blood pressure of ≥90 mmHg. Consequently, a hypertensive woman was identified as 1, and non-hypertensive women were identified otherwise "0."

Explanatory Variables.
Eleven explanatory variables were examined in the study. e selection of these variables was based on conceptual relevance and their significant association with the outcome variable based on findings from previous studies [14][15][16][17]. A number of studies have followed this approach [18,19]. All variables were grouped into personal and community level variables based on the orderly nature of the dataset. e variables were determined based on their availability in the dataset, practical significance, and theoretical relevance for hypertension.

Individual Variables.
e individual level explanatory variables were wealth status, age, marital status, education, parity, occupation, and the consumption of salted fish in the last 24 hours. Age was recorded as 15-19, 20-24, 25-29, 30-34, 35-39, 40-44, and 45-49 years. Wealth status was categorized into poorest, poorer, middle, richer, and richest. Education was classified into four categories: no education, primary education, secondary education, and higher education. Occupation was identified unemployed, professional/clerical, sales/services, agricultural worker, and manual worker. Consumption of salted fish in the last 24 hours was coded as Yes and No.

Community Variables.
Four variables were selected at the community level. ese are region, place of residence, ethnicity, and religion. e region was coded as Eastern, Western, Greater Accra, Central Volta, and Ashanti, since the study was limited to the Southern part of Ghana. Ethnicity was coded Ga/Dangme, Akan, Ewe, Guan, Mole-Dagbani, Gurma, Grusi, Mande, and others. Religion was identified as Islam, Christianity, and others. Place of residence was identified as urban and rural.

Data
Analysis. Data were extracted, cleaned, and analyzed using Stata software version 13.0. Percentage was used to summarize the prevalence of hypertension among respondents. Cross-tabulation was adopted to examine the distribution of hypertension across explanatory variables. Results of cross-tabulation were displayed using percentages with their corresponding confidence intervals. Subsequently, multivariable binary logistic regression analysis was used to determine the factors related with hypertension. ree (3) models were built to examine the factors associated with hypertension. e first model (Model I) consisted of only individual-level variables. Model II was built to contain community-level variables, whilst in Model III, all explanatory variables were combined to examine their association with hypertension. Results of the study were presented using adjusted odds ratio (aOR) with their respective 95% confidence intervals. e women's sample weights (v005/ 1,000,000) were utilized to achieve unbiased estimates, and the Stata survey command "svy" was used to correct for the data's complex sampling structure in the chi-square test and regression analyses, as recommended by DHS.

Model Fit and Specifications.
We evaluated the fitness of all models with Akaike's information criterion (AIC) and Bayesian information criterion (BIC). e presence of multicollinearity between independent variables was checked before fitting these models. e variance inflation factor (VIF) test revealed the absence of high multicollinearity between variables (Mean VIF � 3.01).

Ethical Approval.
is study included participation of human subjects; however, the authors of this manuscript were not directly involved in data collection processes. According to the final report of 2014 GDHS, the survey protocol, including biomarker collection, was reviewed and approved by the Ghana Health Service' Ethical Review Committee and the Institutional Review Board of ICF International [10]. For every research participant, either written or verbal consent was obtained.

Hypertension Prevalence by Socio-Demographic
Characteristics.
e prevalence of hypertension was 16%. Four in ten women aged 45-49 were hypertensive. A higher proportion of women of richest wealth index (20%) were hypertensive. Twenty-one percent of women with no education (21%) were seen to be hypertensive (see Table 2). A higher proportion of women from urban residences (18%) were also hypertensive (see Table 2).  Table 3).

Multivariate Logistic
Women with two births (aOR = 1.45, CI = 1.01-2.07) and those with three births (aOR = 1.47, CI = 1.01-2.15) had higher likelihood of being hypertensive compared to those with zero birth. Greater Accra women had higher odd (aOR = 1.35, CI = 1.02-1.79) of being hypertensive compared to the reference category, thus women from the Western region (see Table 3). Women of Guan ethnicity had lesser likelihood (aOR = 0.54, CI = 0.29-0.98) of being hypertensive compared to women with the Akan ethnic group. Women who engagedin agricultural work had lesser likelihood (OR = 0.72, CI = 0.52-0.99) of being hypertensive compared to unemployed women (see Table 3).     e objective of the study was to investigate the magnitude of hypertension in women across southern Ghana and underlying socio-demographic factors, as previous studies have not prioritized this inquiry. Age, parity, region, ethnicity, and occupation had significant association with hypertension among women in southern Ghana. e total prevalence was 16% which is relatively lower than the reported prevalence for women in the middle belt of Ghana, which was 28.1% [20]. Relatedly, a systematic review has revealed that urban residents in Ghana tend to have a higher prevalence. Considering the high urbanization rate of southern Ghana relative to other parts of the country, coupled with the variations in physical and dietary patterns as well as migration induced psychosocial distress, the hypertension situation in southern Ghana may rise if effective public health interventions are not implemented [21][22][23].
e finding is suggestive that southern Ghana may require well-tailored and effective anti-hypertension interventions to overcome the condition. Besides, health promotion interventions can be rolled out through various media platforms across southern Ghana.
Women who were older had a higher probability of hypertension relative to younger women. High inclination of older women to hypertension is extensively reported in the literature from Ghana [24,25], as well as other low and middle-income countries (LMICs) [26,27] and advanced countries like USA [28]. It is worth noting that in the global west (specifically USA and Europe), between one quarter and one-third of adult populations are hypertensive [29,30]. Our finding is consistent with the existing literature and aging-induced arterial stiffening has been identified to increase the likelihood of hypertension among persons who are advanced in age [31,32].
Women with two or three births had the highest probability of being hypertensive relative to those without birth. Contrary to this, in Ethiopia, a case-control study revealed that women at parity zero had about seven times higher likelihood of hypertension compared with those with one or higher parity or birth [33]. Several factors may account for the discrepancy in findings. For instance, whilst the present study employed a cross-sectional design, the Ethiopian based study was a facility-based retrospective unmatched case-control which was organized in a single south-western based referral hospital. Further, considering the strong correlation between diet, lifestyle, and hypertension, it is unlikely that different groups in distinct environmental contexts will have equal propensities of hypertension in their lifetime [34,35]. Our finding here accentuates the need for maternity healthcare providers to consider hypertension education as a critical component of reproductive health services, with emphasis on antenatal care (ANC) and postnatal care (PNC) in southern parts of Ghana. Greater Accra women experienced increased odds of being hypertensive relative to the women in the Western region. Guan women had a lesser probability of being hypertensive relative to women of Akan ethnicity. ese plausibly imply that some ethnic-and region-based peculiar characteristics dictate women's chances of hypertension. For instance, production and distribution of fruits are uneven across the various regions of Ghana and these could disadvantage residents of some regions whilst making it easy for others to get access to fruits, hence having less likelihood of hypertension [36,37].
Lastly, women who engaged in agricultural work had lesser probability of hypertension relative to unemployed women. Considering extensive documentation on the inverse relationship between vigorous/intensive work and hypertension, this finding is anticipated [37,38]. Subsequently, unemployed women require behavioral change and continuous engagement in at least moderate exercises in order to subside their chances of hypertension.

Strengths and Limitations.
It quite interesting to note that this is the first study to investigate the prevalence of hypertension and underlying conditions among women in southern Ghana. Besides, appropriate statistical techniques have been applied. One limitation of the study is that it was limited to females. We focused on femelaes because consistent evidence have revealed that women bear the greatest risk of hypertension relative to men [39,40]. Second, a crosssectional study design was employed by the study. Due to this, causal inference cannot be drawn.

Conclusion
is study aimed at expanding the frontiers of knowledge on hypertension by investigating the prevalence and underlying factors among women in southern Ghana. From the findings, it will be useful for the Ghana Health Service to implement wide-embracing health promotion initiatives that accommodate the nutritional, exercise, and lifestyle needs of women in southern Ghana. e study has revealed a number of associates of hypertension including age, parity, region, ethnicity, and occupation; hence, existing interventions need to be appraised in the light of these identified factors. Finally, fruit consumption habits and maintenance of the normal body weight ought to be encouraged among women in southern Ghana.

aOR:
Adjusted odds ratio CI: Confidence interval CrI: Credible interval DIC: Deviance information criterion EA: Enumeration areas GDHS: Ghana demographic and health survey GSS: Ghana statistical service IRB: Institutional review board MCMC: Markov chain monte carlo MOR: Median odds ratio NHIA: National health insurance authority NHIS: National health insurance scheme OR: Odds ratio PHC: Population and housing census SSA: Sub-Saharan Africa VIF: Variance inflation factor.

Data Availability
Data used for the study are freely available to the public at https://dhsprogram.com/data/available-datasets.cfm.

Conflicts of Interest
e authors declare that they have no conflicts of interest.

Authors' Contributions
CID conceptualized the entire study. CID, KSD, EKA, and KSA contributed to developing the manuscript. All authors proofread and gave consent for final version for the submission.