Hypertension is the force of blood pushing up against the walls of blood vessels and it is defined as a systolic blood pressure ≥ 140 mm Hg and/or diastolic blood pressure ≥ 90 mm Hg. The risk of dying from high blood  pressure  in low- and middle-income countries is more than double that in high-income countries. In the high income countries, only 7% of deaths caused by high blood pressure occur under age 60; in the African Region, this has increased to 25% (WHO, 2010). Non-Communicable Diseases (including hypertension) are becoming an increasingly important public health problem as epidemiological transition is progressing in Ethiopia (Awoke et al, 2014).World Health Organization reported that the prevalence of hypertension in Ethiopia was 35.2% for the crude adjusted estimates of both sexes (WHO, 2010).

Hypertension is the growing public health problem in many developing countries including Ethiopia. However, its prevention, early diagnose and control have not received greater attention. Hypertension does not have its own signs and symptoms like other chronic disease and its treatments are not given on time. It causes many deaths before it is treated, and it is usually known as silent killer disease. Hypertension increases the risk of ischemic heart disease, stroke, peripheral vascular disease and other cardiovascular diseases, and is also a risk factor for cognitive impairment and dementia and chronic kidney disease (Assefa et al., 2014). In this study, by classifying the severity of hypertension as severe, moderate and mild ordinal logistic regression model was applied to assess the effect of multiple covariates in different levels of hypertension severity.

This study used hypertension data collected from hypertension diseased patients linked to Gondar University Hospital chronic illness medication and follow up clinic. Four hundred twenty hypertension diseased patients were included. From the total hypertension diseased patients’ data 44.5% (187) were severe, 35.2% (148) were moderate and 20.2% (85) were mild (Table 1). Table 2 shows the distribution of socio-demographic related variables with the severity of hypertension. 44.23, 35.10 and 20.67% men had severe, moderate and mild hypertension as compared to 44.81, 35.38 and 19.81% women, respectively. Other variables also can be interpreted in a similar way.

Table 3 shows the distribution of risk factor variables related with the severity of hypertension.  42.6% of individuals who have hypertension in their family had severe hypertension as compared to 46% of individuals who do not have hypertension in their family. 35.5% of individuals who have hypertension in their family had moderate hypertension as compared to 35% of individuals who do not have hypertension in their family. Similarly, 21.89% of individuals who have hypertension in their family had mild hypertension as compared to 19% of individuals who do not have hypertension in their family. Similarly, other variables can be interpreted like this.

The result of Akaike’s Information Criteria (AIC) suggests that the main effect model is reasonable since a model with low AIC is preferred model. Therefore, a cumulative consideration of the above tests confirmed that the main effect model is appropriate. The models can be also compared  by  using  BIC  value.  From  the  SAS output the -2log likelihood values for main effect and interaction effect model are 780.199 and 639.970, respectively. This implies that the BIC values of the two models are 851.0254 and 904.9132, respectively. Since the main effect model has smaller values, it is concluded that the main effect model is appropriate (Tables 4 to 6). The hypothesis that the model adequately fits the data can be examined by the Pearson and Deviance tests.  

The result reveals that the goodness of fit test is fulfilled, which is confirmed by both the deviance and Pearson with P-values of 0.2124 and 0.6325, respectively.

Consequently, the observed and expected cell counts are similar and the value of each statistic is small (Table 7). As shown in Table 8, the tolerance of all the variables greater than 0.1 implies there is no a problem of collinearity between the dependent and independent variables. Moreover, the VIF values are greater than 10, which indicates there is no problem of mulitcollinearity. As shown in Table 9, the tolerance of all the variables greater than 0.1 implies there is no a problem of multicollinearity between age and other independent variables.Moreover, the VIF values are greater than 10, which indicates there is no a problem of mulitcollinearity between age and other independ variables.Similarily, we can continoue for other independent variables.

The study shows that there is significant association between age of patients and the severity of hypertension. Older age group patients (that is, 60 and above), who develop hypertension belong to “severe hypertension” category and they are also liable to high prevalence of hypertension. The result is consistent with Mulugeta (2015), Assefa et al. (2014), Kassawmar et al. (2015), Mungati et al. (2014), Erhun et al. (2005), Marwiro (2012) and Kumar et al. (2002) that age [being aged] has significant effect on the occurrence of hypertension. This study  also  agrees  with  national  academy  on an aging society.

There is a significant difference in developing hypertension between urban and rural residences. This shows that people living in urban areas have a high risk of developing hypertension than people living in ruralareas. The study result is consistent with Iwelunmor et al. (2014). The study shows that there is significant association between alcohol drinking habit of patients and the severity of hypertension. Current alcohol drinker individuals are more likely to develop severe hypertension than never drinker individuals. This implies most current alcohol drinker individuals who develop hypertension belong to “severe HTN” category and they are also liable to high prevalence of hypertension. This result of the study agrees with that of Mulugeta (2015), Erhun et al. (2005), Birhanu (2014) and Kumar et al (2002); alcohol drinking was a significant factor of hypertension. There is high significant difference between individuals with excessive salty food user and non-users in terms of developing hypertension. That is, individuals who are salty food users are more likely to develop “severe hypertension” than those non-users. In addition, the result reveals that the overall prevalence of hypertension is high with those individuals who are salty food users. This result of the study agrees with that of Seid (2014), Mulugeta (2015) and Mizumoto (2004); use of excessive salty foods was the significant factor of hypertension.

There is high significant difference in developing hypertension between vegetable/ fruit users and non-users. That is, individuals who are vegetable or fruit users are less likely to develop “severe hypertension” than non-users. In addition, the result reveals that the overall prevalence of hypertension is high with those individuals who are non-users.The study result is consistent with the result of Mulugeta (2015) that use of fruit or vegetable was significantly associated with hypertension. Consequently, vegetable or fruit users are less susceptible to develop severe hypertension.

There is a high significant difference in developing hypertension between individuals who are physically active and not physically active individuals. Among all hypertension developed in physically inactive individuals most of them belong to “severe hypertension” category and it is also responsible for high incidence of hypertension. This shows people who are physically inactive are at a high risk of hypertension than physically active individuals. The study result is consistent with Seid (2014), Mulugeta (2015), Akilew et al. (2012) and Kumar et al. (2002) that not doing physical activity is significantly associated with hypertension.

The study shows that there is significant association between body mass index of patients and the severity of hypertension. Obese individuals are more likely to develop severe hypertension than normal/underweight body mass index individuals. This implies most “obese body mass  index”  individuals who develop hypertension belong to “severe hypertension” category and they are also liable to high incidence of hypertension. The study result is consistent with Seid (2014), .Mulugeta (2015), Henok (2014), Assefa et al. (2014), Kassawmar et al. (2015), Jenei et al.  (2002), Erhun  et al. (2005), Akilew  et al. (2012), Marwiro (2012), Mizumoto  (2004), Birhanu  (2014), Jasmine  et al. (2013), Kumar  et al. (2002) that body mass index was a significant factor of hypertension.

The study examined the socio-demographic and risk factor determinants of hypertension in GoUTH. Results of proportional odds model show that socio-demographic and risk factor variables are very important for determinant of hypertension outcome. The finding of the study shows that different factors such as patients’ age, Residence, Alcohol use, using salty foods, habit of using vegetable/fruit, Doing physical exercise and BMI have statistically significant effect on the outcome of hypertension.

Based on the findings of this study, the following issues should be considered and promoted for improving the prevention of hypertensive condition in community people.  The government and concerned bodies should work on the preventive strategies of hypertension including promoting and advocating healthy lifestyles, by improvement of dietary habits, eradication of alcohol consumption, decreasing the consumption of salty foods, increasing physical activity and encouraging vegetable or fruit users.

One of the limitations of this study is all patients who were proposed to be included in the study did not come to the clinic during the three months of data collection time and hence the proposed sample size decreased. The other major limitation of this study is related to the problem of the availability of complete secondary data. Patients did not volunteer to give response during primary data collection. Some variables were not included in the study due to lack of equipment and limitation of resource or budget (like waist-to-hip ratio).

The authors have not declared any conflict of interests.