Health-Seeking Behavior and Its Determinants Among Hypertensive Patients in Selected Public Hospitals, Addis Ababa Ethiopia

Abstract

Background

Health-seeking behavior refers to the actions or inactions individuals take when they perceive themselves to have a health issue or illness, seeking an appropriate remedy. The national prevalence of hypertension among adults aged 15-69 years is 19.2%. Limited context specific evidence exists on behavioral, sociocultural, and health system factors influence poor health seeking-behavior.

Understanding Health-seeking behavior is crucial for identifying the health needs and priorities of a population, which in turn addresses these needs and prevent the worsening of health conditions.

Objective

This study aimed to assess health-seeking behavior and associated factors among hypertensive patients at selected government hospitals in Addis Ababa, Ethiopia.

Method

An institution-based cross-sectional study was conducted from March 1 to April 30, 2024, across four public hospitals in Addis Ababa. A total of 214 participants were selected using convenience sampling. A structured data collection tool adapted from previous studies was used to collect data from participants. Data were analyzed using SPSS version 26.0, and descriptive statistics and logistic regression identified factors associated with health-seeking behavior.

Results

The overall prevalence of good health-seeking behavior of participants was 47.7%. The findings indicated that educational status, presence of comorbidities, age, and income significantly influenced health-seeking behavior, specifically, participants with a degree or higher education were significantly ore likely to seek health care (AOR=6.25, 95% CI 1.38-28.29), history of coronary artery disease (AOR=2.70, 95% CI:2.28-27.98). Age 20-40 years (AOR=1.14, 95%CI: 0.34-3.88), and income (AOR=1.13, 95% CI: 0.26-4.94).

Conclusion

Health-seeking behavior among hypertensive patients in Addis Ababa was suboptimal, indicating that a considerable proportion of participants demonstrated suboptimal engagement with recommended care. Education, and coronary artery disease significantly influenced health-seeking behavior, emphasizing the need for interventions targeting awareness, and access to affordable healthcare are recommended.

Introduction

Background

Hypertension is a major global public health challenge and a leading contributor to cardiovascular morbidity and mortality worldwide (Holman, 2020). The national prevalence of hypertension among adults aged 15-69 years is 19.2% (Koye et al., 2022). Effective prevention and management of hypertension depend on appropriate health-seeking behavior (HSB), defined as any action or inaction taken by individuals who perceive a health problem and a remedy (Dave et al., 2025). Ideally, this involves early consultation with qualified healthcare providers in recognized medical facilities (Fatma & Ramamohan, 2023). However, many individuals with hypertension remain unaware of their condition due to its often-asymptomatic nature, and even those who recognize symptoms frequently delay or avoid seeking care (Chekol et al., 2025). Poor HSB is associated with adverse health outcomes, increased disease complications, and higher mortality rates (Latunji & Akinyemi, 2018).

In low-and middle-income countries (LMICs), health-seeking behavior is shaped by complex sociocultural, economic, and health system factors. Cultural beliefs, socioeconomic status, accessibility, cost of care, disease perceptions and trust in healthcare services significantly influence healthcare utilization (Mou et al., 2022). Across Africa, barriers such as long travel distances financial constraints, reliance on traditional medicine, and weak health system responsiveness continue to hinder timely and appropriate health-seeking (Lyimo et al., 2024). Global literature further categorizes determinants of HSB into cognitive factors such as perceives illness severity and treatment effectiveness and non-cognitive factors shaped by social and economic contexts (Ojangiba et al., 2023). As a result, delayed care-seeking often leads patients to present only after experiencing severe complications, including stroke, heart failure, or kidney damage (Musinguzi et al., 2018; Oberoi et al., 2016).

In Ethiopia, hypertension is an increasing public health concern, with a growing number of newly diagnosed cases (Mills et al., 2020). Despite expanding healthcare services, HSB among hypertensive patients remains suboptimal. Studies indicate low awareness and treatment uptake; for example, in Nekemte, only 52.7% of hypertensive patients were aware of their condition, and just 22.4% were receiving treatment (Geleta et al., 2019). Higher educational attainment and the presence of comorbidities have been linked to better care-seeking and adherence (Amare et al., 2020; Sahile & Nurhussien, 2023). Qualitative evidences further highlight systemic barriers such as long waiting times, limited health literacy, and inadequate access to health information (Endrias et al., 2024). Although health extension workers play a key role in rural screening and referral, they face training and resource limitations (Feysal et al., 2022). At the same time, facilitators such as family support, awareness of illness severity, and trust in local healthcare providers offer promising opportunity for intervention.

Despite increasing evidence on hypertension prevalence and management, there remains limited context-specific research in Ethiopia examining the behavioral and psychosocial determinants of health-seeking behavior among hypertensive patients. This gap restricts the development of targeted evidence-based interventions tailored to the country’s healthcare context. Understanding HSB among hypertension patients particularly those attending outpatient departments in Addis Ababa’s public hospitals is therefore essential to improving care uptake, strengthening prevention strategies, and reducing hypertension-related complications (Solomon et al., 2023). Accordingly, this study aimed to assess health-seeking behavior patterns and associated factors among hypertensive patients receiving care in selected public hospitals in Addis Ababa, Ethiopia.

Understanding the HSB of hypertension patients in Ethiopia, particularly those attending outpatient departments in Addis Ababa’s public hospitals, is crucial. This knowledge can inform targeted interventions aimed at improving health outcomes and reducing hypertension related complications (Solomon et al., 2023). Unfortunately, evidence on this topic remains limited, leaving significant gaps in the development of evidence-based prevention and management strategies tailored to Ethiopia’s unique health care context (Musinguzi et al., 2018). This study aimed to examine HSB patterns and associated factors among hypertension patients seeking care in selected public hospitals in Addis Ababa, Ethiopia.

Literature Review

Health-seeking behavior (HSB) represents the actions individuals take when addressing perceived health issues, ideally involving consultation with a licensed healthcare provider (Fatma & Ramamohan, 2023). In low- and middle-income countries, HSB is shaped by a complex interplay of factors, including cultural beliefs, socio-demographic, economic conditions, and the accessibility of services (Mou et al., 2022). For hypertension patients in Ethiopia and across sub Saharan Africa, HSB remains particularly inadequate, with studies revealing alarmingly low awareness and poor treatment uptake. For instance, in Nekemte, Ethiopia, only 52.7% of hypertensive patients were aware of their condition, and a mere 22.4% were on treatment (Geleta et al., 2019). These challenges are compounded by structural barriers such as long travel distance, high treatment costs, and low health literacy (Endrias et al., 2024; Lyimo et al., 2024; van Rensburg, 2014). While higher education levels and comorbidities like diabetes have been linked to better care-seeking, significant gaps persist in understanding the specific behavioral and psychosocial determinants of HSB among hypertensive patients in Ethiopia’s unique urban settings (Amare et al., 2020; Sahile & Nurhussien, 2023).

The consequences of poor HSB for hypertension are severe, as many individuals delay medical care until they experience life-threatening complications such as stroke, heart attack, or kidney failure (Musinguzi et al., 2018; Oberoi et al., 2016). Despite the growing prevalence of hypertension in Ethiopia (Mills et al., 2020), research has identified important facilitators for appropriate care, including family support, awareness of illness severity, and trust in local health care providers (Endrias et al., 2024). However, there is a critical lack of context-specific research examining HSB patterns among hypertension patients in Addis Ababa’s public hospitals (Solomon et al., 2023; Musinguzi et al., 2028). This evidence gap limits the development of targeted evidence-based prevention and management strategies, underscoring the urgent need for focused research to improve health outcomes and reduce the burden of hypertension-related complications in the country (Feyisa et al., 2023).

Hypothesis Testing

Null Hypothesis (H_o).

There is no significant association between educational status, comorbidity, or income level and health-seeking behavior among hypertension patients.

Alternative Hypothesis (H₁).

There is a significant association between educational status, comorbidity, or income level and health-seeking behavior among hypertension patients.

Materials and Methods

Study Settings

The current study was conducted in four public hospitals in Addis Ababa: St Paul’s hospital Millennium medical college, St. Peter specialized hospital, Zewditu Memorial hospital, and Yekatit 12 hospital. Addis Ababa, the capital city, is Ethiopia’s political, cultural, and economic hub, home to landmarks like the National Museum, the Ethnological Museum, and the African Union headquarters (Wubneh, 2013). The city was chosen due to its higher patient inflow driven by the availability of advanced healthcare services.

Addis Ababa has a total of 21 hospitals. Among these, only seven hospitals: Tikur Anbessa Specialized hospital, St. Paul’s Millennium medical College, St. Pitter’s Specialized Hospital, Zewditu Memorial Hospital, Yekatit 12 Hospital, Minellik II Specialized Hospital, and Tirunesh Beijing General Hospital provide hypertensive follow-up services. For this study, four hospitals were selected using simple random sampling method due to a limited time and budget. St. Paul’s Hospital Millennium Medical College, established in 1968 by Emperor Haile Selassie and expanded in 2010, offers a wide range of medical services, with more than 700 inpatient beds and an average of 1200 outpatient visits daily. It is a leader in Ethiopia’s medical education, training over 2800 staff and providing care to over 400,000 patients annually.

St. Peter Specialized Hospital, founded in 1963, specializes in tuberculosis and HIV care, and also offers services in internal medicine, dermatology, ophthalmology, pediatrics, and mental health. The hospital includes a cardiac care unit and collaborates with St. Paul’s Hospital Millennium Medical College for advanced cardiac services (Abate et al., 2012; Muluneh et al., 2021). Zewditu Memorial Hospital, named after Empress Zewditu, is a key healthcare provider offering specialized care in internal medicine, surgery, pediatrics, and obstetrics. It also serves as a teaching hospital, contributing to medical education in Ethiopia (Solomon et al., 2023).

Yekatit 12 Hospital, one of the oldest and largest hospitals in Ethiopia, provides services in internal medicine, surgery, pediatrics, and obstetrics, as well as diagnostic services like radiology and laboratory testing. It plays a vital role in public health, particularly for underserved populations (Abebe et al., 2021; Assefa et al., 2016). Study participants were selected using convenience sampling from these hospital clinics.

Design and Period

An institution-based cross-sectional study was conducted from March 01 to April 30, 2024. This design was selected because it is cost effective, time efficient, and practical, when resources are limited and when the research objective does not require longitudinal tracking.

Population

All adult hypertensive patients who visited the chronic disease outpatient departments in the four randomly selected public hospitals during the data collection period were eligible for inclusion, and those study participants available during this period were included using convenience sampling method.

Eligibility Criteria

All hypertension patients aged 18 years and older who visited the chronic care outpatient departments of St. Paul’s Hospital Millennium Medical College, St. Peter Specialized Hospital, Zewditu Memorial Hospital, and Yekatit 12 Hospital were considered for inclusion. Patients who were unwilling to participate, those with a history of stroke (communication difficulties), critically ill individuals, and patients with severe mental illness were excluded from the study.

Health care seeking behavior in this study was assessed using six dimensions: healthcare utilization, physical activity (measured in MET-minutes/week according to the established physical activity assessment guidelines), dietary practices, preventive screening, substance use avoidance, and health information seeking practice (World Health Organization, 2012). MET (Metabolic Equivalent of Task) criteria can be appropriately used as part of measuring HSB because they provide a standardized objective way to quantify physical activity, which is a key behavioral component of the protective health practices. This classification produced categorical outcome (Low, Moderate, High), rather than a continuous outcome. It was classified as: High when a person engaged in any combination of walking, moderate-intensity, or vigorous-intensity activities for 7 or more days, achieving at least 3,000 MET minutes per week; Moderate when they engaged in such activities for 5 or more days, achieving at least 600 MET minutes per week; and low when they did not meet either of these criteria, including those engaging in less than one day of activity per week. Categorizing health-seeking behavior based on physical activity levels is a valid and meaningful approach that aligns with a global health standard, reflects individuals’ engagement in preventive health practices, and serves as a predictive indicator for health behaviors and outcomes (Kifle, 2013).

Sample Size Determination

The sample size for the study was determined by using the single population proportion formula. According to a community study done in Jimma (Kifle, 2013), Ethiopia where the health-seeking behavior of respondents was 17.5, accordingly, the sample size was calculated to be 221.9

$$\mathrm{n}=\frac{\left(\mathrm{Z}{\mathrm{a}}_{/2}\right){}^{2\hspace{0.17em}}\hspace{0.17em}\mathrm{p}\left(1-\mathrm{p}\right)}{{\mathrm{d}}^2}$$

$$\mathrm{n}=\frac{\left(1.96\right){}^{2\hspace{0.17em}}\hspace{0.17em}0.175\left(1-0.175\right)}{{\left(0.05\right)}^2}$$

$$\mathrm{n}=\frac{3.8416*\hspace{0.5em}0.175\left(0.825\right)}{0.0025}=221.0$$

This formula is appropriate for estimating a single proportion in a cross-sectional study. In this case, the expected proportion (p) of health-seeking behavior was set at 17.5% (0.175), based on findings from a prior study. A 95% confidence level was used, corresponding to Z_a/ ₂ = 1.96, and a margin of error (d) typically set at 5% (0.05).

Sampling Techniques

Public hospitals were selected using a simple random sampling method to ensure representativeness across institutions. However, participants within these hospitals were recruited through convenience sampling method, due to the absence of a complete sampling frame, the continuous and unpredictable flow of hypertensive patients, and time and resource constraints.

Data Collection Tool, and Process

A structured data collection instrument was used to gather information from participants. The instrument was adapted and modified from previously conducted studies in the literature (Fetensa, 2018; Li, Zhao et al.,. 2025) It comprised four sections; the first captured socio-demographic variables, the second assessed health seeking behavior, the third explored the factors influencing health-seeking behavior, and the final addressed behavioral risk factors. The original instrument was developed in English and subsequently translated into the local language, “Amharic,” by language experts. To ensure semantic and conceptual consistency, the tool was then, back translated into the English. To assess the reliability of the instrument, Cronbach’s alpha was calculated using 60 participants. This statistical measure evaluates internal consistency, or how closely related a set of items are as a group. A Cronbach’s alpha score of 0.79 was obtained indicating an acceptable level of internal consistency for the items measuring health-seeking behavior. This suggests that the items within the instrument reliably measured the intended constructs. The instrument was also reviewed by experts in hypertension care to ensure face validity, confirming that the items appeared appropriate and relevant for measuring health-seeking behaviors. Additionally, A pretest was conducted on 5% of the total sample size at Minilk-I hospital, a facility not included among the main study sites, to assess clarity, relevance, and effectiveness of the instrument in the actual data collection setting. Health-seeking behavior (HSB) was operationally defined as the extent to which participants engaged in action intended to maintain, promote or improve their health through preventive, promotive, and curative practices. In this study, HSB was assessed across six key dimensions: healthcare utilization, physical activity, dietary practice, preventive screening, substance-use avoidance, and health information-seeking practices. The physical activity dimension was measured using MET-minutes per week (MET-minutes/week), a standardized measure of energy expenditure calculated according to established physical activity assessment guidelines.

Participants achieving the recommended threshold of ≥ 600 MET-minutes/week were scored as having adequate physical activity, while those below this threshold were categorized as inadequate.

The remaining five HSB items were coded dichotomously (1=healthy/appropriate behavior, 0 unhealthy/inappropriate behavior). The scores from all six items were summed to generate a composite HSB score for each participant. Thus, the total composite score reflected the overall level of engagement in positive health-seeking behaviors, with higher scores indicating better HSB.

To classify participants, the mean composite HSB score of the study population was calculated. Participants scoring greater than or equal to the mean were categorized as having “good health-seeking behavior” whereas those scoring below the mean were categorized as having “poor health-seeking behavior” (Babbie, 2010).

We assumed that the total study population in the study area was small, and that the available time for data collection was limited. Two data collector nurses and one supervisor were involved in the data collection process. Prior to the starting the actual data collection, the data collectors received a one-day training to enhance their understanding of the data collection tool. Participants were provided with a detailed explanation of the study’s objectives and their rights before data collection began. Primary data was used for extraction, and the data collection team ensured the confidentiality of participant information. The data was reviewed for completeness at the end of each data collection day.

Possible Source of Biases

Possible sources of bias in this study include selection bias, as participants attending health facilities may differ from those who do not seek care; recall bias, since respondents were asked to remember past health practices; and social desirability bias, where participants may overreport positive health behaviors.

Data Quality Assurance

To ensure the quality and reliability of date, several measures were implemented throughout the data collection process. First, the data collectors underwent a one-day training session to familiarize themselves with the data collection tool, ensuring consistent understanding and application. A pretest was conducted on 5% of the sample at Minilik II hospital, to assess the instrument’s clarity before the actual data collection began. The principal investigator (PI) closely supervised the data collection process, providing oversight and ensuring that the interviewers were conducted according to the established protocol. To maintain data accuracy, completeness checks were performed daily, and any missing and incomplete data was promptly addressed and replaced.

Statistical Analysis

Data was entered into Ep-data version 4.2.0.0 and then exported to SPSS version 26.0 for analysis Descriptive and inferential statistics were used to analyze the data, descriptive statistics such as percentages, frequencies, means, and standard deviations were calculated. Logistic regression which include bivariate and multivariate analysis was performed to examine the associations between dependent and independent variables. The model was chosen to identify factors influencing HSB, as it allows for assessing the relationship between multiple predictors and a binary outcome. Parameters were estimated using Maximum Likelihood Estimation (MLE), providing robust and interpretable results as odds ratio. The Hosmer-Lemeshow goodness-of-fit test was used to assess the calibration of the logistic regression model, and a p value >0.05 was considered indicative of adequate model fit. Level of significance was determined at p-value < 0.05. The results were presented using frequency tables, figures, and charts.

Result

Socio-Demographic Characteristics of Hypertensive Patients in Selected Public Hospitals

Although the calculated sample size for the study was 222 participants, only 214 eligible participants were available during the data collection period. Therefore, 214 participants were included in the final analysis, resulting in a 96.4% achievement.

The demographic profile of respondents reveals a balanced gender distribution, a predominance of middle-aged adults, and a majority who are married with basic literacy levels. The gender distribution shows a near-equal split, with females representing 113 (52.8%) and males 101 (47.2%). The most represented age group is 41-50 years, comprising 73 (34.3%) of respondents, followed by the over-61 age group at 51 (23.8%). In terms of marital status, married individuals make up the majority at 150 (70.1%), while widowed 23 (10.7%), single 22 (10.3%), and divorced 19 (8.9%) individuals form smaller segments. Regarding education, 84 (39.3%) of respondents possess basic literacy skills, and 27 (12.6%) have attained a degree or higher education. Notably, the proportions of illiterate respondents 28 (13.6%) and those with preparatory education 20 (9.3%) are similar (Table 1).

Table 1.

Socio-Demographic Characteristics of Hypertensive Patient in Selected Public Hospitals in Addis Ababa (n=214)

Variables	Category	f (n)	%
Gender	Male	101	47.2
	Female	113	52.8
Age	18-30	10	4.7
	31-40	31	14.5
	41-50	73	34.1
	51-60	49	22.9
	≥ 61	51	23.8
Monthly Income in Eth. Birr	< 3000	42	19.63
	3001-6000	70	32.71
	6001-9000	52	24.30
	9001-12000	18	8.41
	>12000	32	14.95
Marital status	Divorced	19	8.9
	Married	150	70.1
	Single	22	10.3
	Widowed	23	10.7
Educational status	Able to read and write	84	39.3
	Degree and more	27	12.6
	Diploma	21	9.8
	Illiterate	29	13.6
	Preparatory (10 &11)	20	9.3
	Primary school (from 1 -8)	17	7.9
	Secondary school (9 & 10)	16	7.5

Health-Seeking Behavior Levels Among People Living With Hypertension

The cut-off point used to categorize HSB was the mean score of the total HSB scale. Participants scoring greater than or equal to the mean were classified as having good health-seeking behavior, while those scoring below the mean were categorized as having poor health-seeking behavior.

The overall health-seeking behavior in this study showed that 112 (52.3%) participants had poor health-seeing behavior, while the proportion of participants with good health-seeking behavior was 102 (47.7%) (95% CI: 41-0.54.4) (Figure 1).

Figure 1.

Health-seeking behavior of patients who were first diagnosed with hypertension at the outpatient departments of selected government hospitals in Addis Ababa, Ethiopia

Clinical/Comorbid Characteristics of People Living With Hypertension

Among respondents, 56 (26.2%) reported having diabetes mellitus, 33 (15.4%) acknowledged a history of cardiovascular disease, Stroke occurrence was less common, with only 18 (8.4%) of respondents having experienced a stroke. Regarding participants were awareness of being having hypertension: only 112 (52.3%) of respondents were familiar with the condition, leaving 102 (47.7%) unaware (Figure 2).

Figure 2.

Clinical characteristics of patients who were first diagnosed with hypertension at the outpatient departments of selected government hospitals in Addis Ababa, Ethiopia

Health-Seeking Behaviors

The table provides an overview of health-related behaviors and experiences within the surveyed population, highlighting key trends in healthcare utilization and practices. A majority of the respondents 110 (51.4%) reported using physician-prescribed medication as their primary health-protective measure. However, 35 (16.4%) admitted to taking random drugs without a prescription, while 69 (32.2%) chose to ignore symptoms of illness. Hypertension emerged as the leading self-reported illness prompting hospital visits in the past two months 141 (65.9%), followed by atrial fibrillation 3 (1.4%) and chronic heart failure 4 (1.9%). Additionally, 54 (24.3%) cited other illnesses. In terms of hospital visit frequency, 122 (57.0%) visited once when ill, 63 (29.4%) visited twice, 13 (6.1%) visited three times, and 16 (7.5%) reported no visits. The primary reason for hospital visits was sickness 139 (65.0%), though 48 (22.4%) visited for routine check-ups. Monitoring of weight and blood pressure was predominantly reactive, with 89 (41.6%) and 137 (64.0%), respectively, reporting they only checked these metrics when ill. Regarding preferred treatment locations, most respondents 148 (69.2%) favored hospitals, followed by pharmacies 41 (19.2%), while smaller numbers turned to traditional remedies such as herbs or prayer houses (Table 2).

Table 2.

Health Seeking Behavior and Health Protective Practice of Participants in Selected Public Hospitals in Addis Ababa (n=214)

Variables	Category	f (n)	%
Health-protective measures	Ignore any feeling of sickness	69	32.2
	Take random drugs not prescribed by physician	35	16.4
	Take regular medication prescribed by physician	110	51.4
Preferred place of treatment	Use herbs and traditional alternatives	12	5.6
	Go to a prayer house	12	5.6
	Buy drugs from a chemist/Pharmacy	41	19.2
	Go to the hospitals	148	69.2
	Hypertension	141	65.9
	Obesity and diabetes	3	1.4
	Others	53	24.8
Number of hospital visits in the past 2 months	Once only when sick	122	57.0
	Twice	63	29.4
	Thrice	13	6.1
	None	16	7.5
Reasons for hospital visits	I was sick	139	65.0
	Went for check-up	48	22.4
	Advised to do so by family/friends	22	10.3
	Other	5	2.3
Weight monitoring frequency	Monthly	34	15.9
	Only when I am sick in the hospital	91	42.5
	I do not check	89	41.6
Blood pressure monitoring frequency	Daily	1	.5
	Weekly	5	2.3
	Monthly	46	21.5
	Only when I am sick in the hospital	137	64.0
	I do not check	25	11.7

Factors Affecting the Health Seeking Behavior

Bivariate and multivariate analyses were conducted to examine predictors of health seeking behavior. Variables with a p-value less than 0.25 in the bivariate analysis were candidates for the multivariate model. A p-value threshold of 0.25 is often used in bivariate analysis to elect candidate variables for multivariate logistic regression because variables that are not significant at the conventional 0.05 level may still be important predictors or confounders. This relaxed threshold helps ensure that potentially relevant variables are not excluded permanently (Hosmer et al., 2013; Bursac et al., 2008).

This included age, educational status, presence of coronary artery disease, and monthly income. Before running the multivariate analysis, multicollinearity was assessed using variance inflation factor (VIF). All VIF values were found to be below 2.5, indicating no significant multicollinearity among the independent variables (table 3). In the final model, educational status and comorbidity remained significant predictors. Individuals with a degree or higher education had 6.25 times higher Odds of engaging in good health-seeking behavior compared to those without formal education controlling for other factors in the model. This represents a strong positive association, Participants with a degree or higher education were 6.25 times more likely to exhibit health-seeking behavior (AOR=6.25, 95% CI:1.38-28,29, p=0.017), while those with preparatory education (grade 10-11) had an even 7.65 times stronger association (AOR = 7.65, 95% CI: 1.57–37.32). Conversely, illiteracy was linked to poorer health-seeking behavior (AOR: 0.041, 95% CI: 0.003–0.564). The presence of coronary artery disease was another significant predictor (AOR: 2.70, 95% CI: 1.28–27.98). Additionally, age and monthly income appeared to be associated with HSB (AOR=0.041; 95% CI: 0.003-0.564; p=0.017), and (AOR= 1.13; 95% CI: 0.26-4.94). This was consistent with the study conducted in Philippines (Laurito et al., 2025), however, interaction terms between key predictors (e.g. age and education) were tested but did not show statistically significant interaction effect. Model diagnostics were also performed: the Hosmer-Lemeshow goodness-of-fit test yielded a p-value > 0.05, (X²=5.87, df=8, p=0.66) indicating that the model fit the data adequately. The overall predictive power of the model was assessed using the area under the ROC curve (AUC), which showed acceptable dimension (Table 4).

Table 3.

Factors Affecting the Health Seeking Behavior of Study Participants, Results From the Multivariate Binary Logistic Regression Analysis

Variables	Categories	Health seeking behavior		COR 95% CI	AOR of 95% CI	p-value
		Good	Poor
Age (in Years)	20-40	24	17	2.33 (1.01-5.36)	1.14 (0.34-3.88)	0.047*
	41-60	58	62	1.54 (0.80-2.99)	0.85 (0.35-2.10)	0.123
	61-90	20	33	1	1	​
Educational status	Illiterate	2	27	.1 (0.02-0.54)	.041 (.003-.564)	0.017*
	Read and write	33	51	.83 (0.28-2.45)	1.180 (.36-3.88)	.786
	Primary school (from 1 -8)	8	9	1.14 (0.29-4.51)	1.18 (.267-5.28)	.821
	Degree or above	21	6	4.50 (1.18-17.2)	6.254 (1.38-28.29)	0.017*
	Diploma	15	6	3.21 (0.82-12.63)	3.881 (.89-16.94)	.071
	Secondary school (9 & 10)	7	9	1	1	​
	Preparatory (11&12)	16	4	5.14 (1.18-22.48)	7.646 (1.57-37.32)	0.012*
	Diploma	15	6	3.21 (0.82-12.63)	3.881 (.89-16.94)	.071
	Degree or above	21	6	4.50 (1.18-17.2)	6.254 (1.38-28.29)	0.017*
Presence of coronary artery disease	Yes	10	2	5.97 (1.28-27.79)	2.70 (1.28-27.98)	0.008*
	No	92	110	1	1	​
Monthly Income (in ETB)	(1000 – 4999)	48	61	0.30 (0.10-0.91)	1.13 (0.26-4.94)	0.025
	(5000 - 9999)	41	46	0.34 (0.11-1.04)	0.82 (0.19-4.94)	.074
	(≥10000)	13	5	1	1	​

*ETB; Ethiopian Birr.

Table 4.

Variance Inflation Factor (VIF) for Independent Variables in Multivariable Logistic Regression

Variable	VIF	Interpretation
Age in years	1.22	No multicollinearity
Sex	1.08	No multicollinearity
Educational status	1.14	No multicollinearity
Marital status	1.03	No multicollinearity
Monthly income	1.35	No multicollinearity
Physical activity level	1.28	No multicollinearity
Comorbidity (CAD, DM, past history of Stroke)	1.37	No multicollinearity
Health protective measures	1.29	No multicollinearity
Number of visits	1.83	No multicollinearity
Blood pressure monitoring frequency	1.77	No multicollinearity
Weight monitoring frequency	1.22	No multicollinearity

Note. All VIF values are <5, indicating that multicollinearity is not a concern in this model. Independent variables indicated in the multivariable model were selected based on bivariate analysis (p≤ 0.25).

Discussion

This study assessed the health seeking behavior (HSB) of people living with hypertension attending selected public hospitals outpatient departments in Addis Ababa, Ethiopia, and found that 47.6% of participants demonstrated appropriate HSB. This proportion is highly lower than reports from Southern Ethiopia (60%) and Gondar (55%) (Fetensa, 2018; Wolde et al., 2022). Such variation may reflect differences in healthcare access, infrastructure, and educational status across regions, as documented in previous Ethiopian studies (Anota & Nedi, 2022; Wubneh, 2013).

The relatively moderate level of HSB observed in Addis Ababa may be attributed to its urban context, where access to healthcare services, exposure to healthcare information, and participation in healthcare promotion activities are generally better. Compared to findings from Nekemte, where awareness was (52%) but treatment uptake only (22.4%) (Geleta et al., 2019).

Differences in reported prevalence across studies may also be explained by methodological variations. Unlike earlier studies that focused mainly on healthcare utilization or medication adherence, this study employed a broader multidimensional definition of HSB by incorporating physical activity measured in MET-minutes per week. This more comprehensive and behaviorally specific approach may have influenced the observed estimates.

Clinically, (26.2%) of respondents had diabetes mellitus, indicating a substantial burden of comorbidity, consistent with global patterns of hypertension clustering with metabolic disorders (Abebe et al., 2021). However, only (52.3%) of participants were aware of their hypertensive status, underscoring gaps in screening and health literacy. Although over half (51.4%) adhered to prescribed medications, self-medication (16.4%) remains a concern due to potential safety risks (Kovesdy, 2022; Solomon et al., 2021). Additionally, most healthcare visits (65%) were illness driven, suggesting a predominantly reactive rather than preventive care-seeking pattern.

Educational status and comorbid conditions, particularly coronary artery disease, were significant predictors of HSB. Individuals with higher education were more likely to seek care, likely due to better health literacy and awareness, whereas, illiterate participants had reduced engagement, probably due to communication barriers and reliance on informal care. Similarly, patients with comorbidities may perceive greater disease severity, prompting increased healthcare utilization. This finding aligns with previous studies in Ethiopia (Amare et al., 2020; Sahile & Nurhussien, 2023).

Although younger age (20-30 years) and lower-to-middle income categories initially appeared to be associated with better HSB at the crude analysis level, these relationships did not persist after controlling for potential confounding variables in the multivariate model. Specifically, participant aged 20-40 years showed higher odds of good SHB in the unadjusted analysis; however, this association lost statistical significance after adjustment, indicating that the observed effect was likely influenced by other factors such as educational status and comorbid conditions.

Similarly, while certain income groups (particularly those earning 1000-5999 ETB) demonstrated a significant association with HSB in the crude analysis, this effect was attenuated and become non-significant in the adjusted model. This suggests that income alone may not independently predict HSB when other socio-demographic and clinical variables are taken into account.

Overall, the findings emphasize that both individual-level factors (such as education and awareness) and clinical conditions (such as comorbidities) play critical role in shaping HSB. Addressing barriers such as low health literacy and healthcare navigation challenges is essential to improve early and proactive care among people living with hypertension.

Strengths and Limitations

This study possesses several methodological strengths, including the random selection of hospitals, a well justified sample size, and a rigorously developed data collection tool that was translated, pretested, and demonstrated good reliability. The implementation of robust data quality assurance measures and comprehensive statistical analysis, including model diagnostics, further enhances the credibility of the findings. However, the research is significantly constrained by its most critical limitation: the unconventional and problematic operationalization of “Health-Seeking Behavior” (HBV) based on physical activity levels (MET-minutes) rather than direct measures of healthcare engagement. This fundamental issue makes the study’s primary outcome difficult to interpret and compare with existing literature. The weakness is compounded by the use of convenient sampling for participant recruitment, which induces selection bias and limits the generalizability of the results to the specific population of patients in Addis Ababa’s public hospitals. Additionally, the cross-sectional design prevents causal inferences, and the reliance on self-reported data raises concerns about potential recall and social desirability biases.

Implications for Nursing Practice

The study highlights a critical need for nurses to assume a more proactive role in hypertension management. Given that low educational status is a key predictor of poor health-seeking behavior, nurses must become primary patient educators, developing tailored communication strategies for varying literacy levels to improve understanding of the disease and the importance of adherence. Furthermore, since most patients seek care only when sick, nurses are positioned to champion preventive care by encouraging regular screening and follow-ups. By also acting as patient advocates to help navigate systemic barriers like long waiting times, nurses can significantly improve patient engagement, treatment adherence, and ultimately, health outcomes for hypertensive patients in Ethiopia.

Conclusions and Recommendations

Based on the findings, this study concludes that health-seeking behavior among hypertensive patients in Addis Ababa remains suboptimal with only 47.6%, of participants demonstrating appropriate health-seeking practices, and is significantly influenced by educational level and presence of comorbidities like coronary artery disease. Despite relatively better healthcare access in the urban setting, challenges such as long waiting times, high test costs, and low health literacy continue to hinder proactive care-seeking, with over half of participants unaware of their hypertensive status. The inclusion of physical activity as a behavior metric offered more comprehensive understanding of care-seeking patterns, helping to explain prevalence differences from previous regional studies. The observed reliance on illness triggered visits and instances of self-medication reflect gaps in preventive care and chronic disease management. Therefore, it is recommended that public health strategies prioritize targeted education campaigns to improve awareness, enhance screening programs, and streamline access to care, particularly focusing on individuals with low literacy and high comorbidity burdens, while also addressing systemic barriers within healthcare facilities. Implementing a qualitative component is essential to gain deeper insights into participants’ experiences, perceptions, and contextual factors that cannot be fully captured through quantitative data alone.

ORCID iDs

Tigistu Gebreyohannis Gebretensaye https://orcid.org/0000-0002-2292-7865

Boka Dugassa Tolera https://orcid.org/0000-0003-2905-6844

Ethical Considerations

Ethical approval for the study (Ref: CHS/NSG/008/29/24) was obtained from the Ethical Review Committee of the College of Health Sciences, Addis Ababa University. The study adhered to the principles outlined in the Declaration of Helsinki. A support letter was submitted to the study hospitals, and permission to conduct the study was granted. Participants were fully informed about the study objectives, and written informed consent was obtained from those who volunteered. Participants who were unable to read and write gave their consent to participate by providing fingerprints. To ensure confidentiality, participants’ names were anonymized, and their information was securely stored.

Data Availability Statement

The datasets used in the current study are available from the corresponding author on reasonable request.*