Is gender a factor in socioeconomic disparities in undiagnosed, and untreated hypertension in Bangladesh?

Abstract

Our objectives were to ascertain the following: (1) the prevalence and socioeconomic distribution of hypertension (HTN), undiagnosed for HTN, and untreated cases of HTN‐diagnosed individuals; (2) the relationship between SES and the prevalence of HTN, undiagnosed for HTN, and untreated for HTN; and (3) whether sex moderate this association. Data from the 2017–18 Bangladesh Demographic Health Survey were used. 11,776 participants who were 18 years of age or older responded to our analysis. The age‐adjusted prevalence of HTN, undiagnosed for HTN, and untreated cases was 25.1%, 57.2%, and 12.3%. Compared to females, males were less likely to have HTN but more likely to have undiagnosed HTN. People in the rich SES groups had a higher odd of (adjusted odds ratio [aoR] 1.25; 95% confidence interval [CI] 1.08–3.45) of having HTN compared to those in the poor SES group. When compared to individuals in the poor SES group, those in the rich SES group had lower odds of undiagnosed (aoR 0.57; 95% CI 0.44–0.74) and untreated (aoR 0.56; 95% CI 0.31–0.98) for HTN. Sex moderated the association between SES and HTN prevalence, which showed that men from rich SES were more likely to suffer from HTN than men from poor SES. According to this study, the government and other pertinent stakeholders should concentrate more on developing suitable policy measures to reduce the risk of HTN, particularly for men in rich socioeconomic groups. They should also concentrate on screening and diagnosing HTN in socioeconomically disadvantaged populations, regardless of sex.

1. INTRODUCTION

A life‐threatening condition called hypertension (HTN) raises the risk of several ailments, including those that affect the kidneys, heart, and brain. ¹ In 2017, it caused more than 10.4 million deaths and 218 million disability‐adjusted life‐years globally. ² Another health risk is undiagnosed HTN, as many people do not know they have high blood pressure and do not take medicine to control it. A thorough review found that over half of HTN patients are ignorant of their hypertensive condition. ³ The substantial worldwide burden of undiagnosed and untreated HTN highlights the need for enhanced HTN screening and early detection to develop health promotion and disease prevention strategies.

Recent changes in epidemiology and demography have resulted in notable changes in lifestyle and behavior, including a rise in the prevalence of HTN in Bangladesh. According to a systematic review, ⁴ Bangladesh possesses a weighted pooled prevalence of HTN of 20%, with a prevalence range of 1.1%–75%. Going forward, among the primary non‐communicable diseases (NCDs) in Bangladesh and other South Asian nations including India, Nepal, Bhutan, and Sri Lanka, HTN continues to be the leading cause of disease burden. ⁵ In Bangladesh in particular, there is a dearth of systematic synthesizing data on undiagnosed and untreated HTN and its associated factors. Two small‐scale local studies carried out in rural Bangladesh found that among people 35 years of age and older, the prevalence of undiagnosed HTN was 45% and 82%, respectively. ⁶ , ⁷ A further nationwide survey carried out in 2011 revealed that 49.9% and 48.9% of Bangladeshis had undiagnosed and untreated HTN, respectively. ⁸ Nevertheless, as these studies only included participants who were 35 years of age or older, they were unable to accurately depict the current state of HTN, as data indicates that the disease has been concentrating more and more in younger people in recent years. ⁹

Socioeconomic status (SES) is one of the major independent factors among several sociodemographic, health, and behaviorally linked factors, and its effects on health and well‐being are widely recognized. ¹⁰ Surprisingly, SES disparities have not gotten much attention in attempts to prevent and manage HTN in LMICs, even though the vast bulk of research on SES disparities and HTN prevalence, diagnosis, and treatment has been undertaken in high‐income nations. ¹¹ , ¹² , ¹³ , ¹⁴ Research indicates that in high‐income nations, ¹¹ , ¹² , ¹³ , ¹⁴ HTN is more common among lower socioeconomic categories. On the other hand, the socioeconomic gradient of HTN in low‐ and middle‐income countries (LMICs) has shown the reverse pattern, ¹⁵ , ¹⁶ with socially disadvantaged groups having a lower prevalence of the disease, less access to HTN care, and a higher likelihood of experiencing complications from the disease. There is still a dearth of research in Bangladesh on the SES disparity in the prevalence of HTN, undiagnosed for HTN, and untreated for HTN. Bangladesh is experiencing a swift transition in its economy and epidemiology, leading to a rise in socioeconomic disparity because of urbanization and industrialization. ¹⁷ Therefore, developing health promotion and disease prevention strategies for HTN requires a focus on the link.

Furthermore, sex is a factor that may modify the association between SES and several health outcomes. ¹⁸ , ¹⁹ Wu and colleagues reported that the association between SES and diabetes was inverse for female participants and positive for male participants, respectively. ²⁰ This underscores the need for sex‐specific research on the discrepancy between SES and HTN diagnosis and treatment. Moreover, these kinds of studies are required in Bangladesh, a patriarchal country where women are subservient to men in all areas of life. ²¹ Contradictory results have been obtained from numerous research that looked at the relationship between sex and HTN in LMICs. Investigations have revealed that women are more prevalent than men to have HTN ²² , ²³ although other investigations have found the contrary. ²⁴ , ²⁵

Poor SES women with HTN may face unique challenges in managing their disease because of resource constraints that may restrict their access to healthcare, educational opportunities, lifestyle choices, and healthcare knowledge. These factors may also make it more difficult for them to become aware of HTN and to diagnose the condition. Therefore, a comparison analysis is required to determine whether sex can help mitigate the relationship between SES and the prevalence of HTN, undiagnosed for HTN, and untreated for HTN. Based on these considerations, the objectives of this research are to determine: (1) the prevalence and socio‐economic distribution of HTN, undiagnosed for HTN, and untreated for HTN; (2) the relationship between SES and the prevalence of HTN, undiagnosed for HTN, and untreated for HTN; and (3) if this relationship is moderated by sex.

2. METHODS

2.1. Data source and study population

The Bangladesh Demographic and Health Survey (BDHS) 2017–18 ²⁶ provided the study's data. The National Institute of Population Research and Training of Bangladesh's Ministry of Health and Family Welfare carried out the BDHS. It is a nationally representative probability sample of men and women drawn from households. The BDHS 2017–18 collected biomarker measures in addition to other relevant data including social and economic variables. The DHS Program received technical help from ICF International and financial support from the US Agency for International Development. The BDHS employed a stratified two‐stage sample of dwellings, with distinct strata for rural and urban areas. Initially, enumeration areas from the most recent 2011 Bangladesh census were used to determine primary sampling units (PSUs), which are groups of households that average 120.

A total of 675 PSUs were chosen at random from a total of 293,579 PSUs in the first stage. There were 672 PSUs in total (192 in urban areas and 480 in rural areas). Because of the floods, the other three PSUs were not sampled. Thirty households from each chosen PSU made up the second stage's sample of 20,160 households for data gathering. Of them, 19,457 families had interviews finished, yielding a 96.5% inclusion rate. Blood pressure (BP) readings were obtained from one‐fourth of the chosen homes (7–8 households per cluster) by the BDHS 2017–18, resulting in 4864 households. In those chosen households, 12,929 men and women who were at least 18 years old gave consent for BP measurement. Finally, the BP test was done on 11,776 respondents. The detailed flowchart for the selection of study participants is depicted in Figure 1.

FIGURE 1.

Selection of the sample.

2.2. Measures

2.2.1. Outcomes

The prevalence of hypertension—both diagnosed and undiagnosed—as well as its untreated states—piqued our attention. The survey utilized millimeters of mercury (mmHg) to express the systolic and diastolic blood pressure. Trained health technicians took three readings with the LIFE SOURCE® UA‐767 Plus blood pressure monitor at around ten‐minute intervals. The average of the second and third measurements was used to calculate each person's blood pressure. In accordance with the National recommendations for the Management of Hypertension in Bangladesh, ²⁷ HTN individuals were defined as those whose systolic blood pressure (SBP) was ≥140 mmHg, whose diastolic blood pressure (DBP) was ≥90 mmHg, and/or who were taking anti‐hypertensive drugs to lower or control their blood pressure or being previously diagnosed as hypertensive by any health professional.

If a participant had a diagnosis of HTN (SBP ≥140 mmHg or DBP ≥90 mmHg) at the time of the survey but had never taken any prescribed anti‐hypertensive medication to lower or control their blood pressure, or if they had never been informed by a medical professional that they had HTN prior to this study, they were deemed to be undiagnosed for the condition. If, at the time of the survey, a participant's blood pressure was diagnosed as hypertensive (SBP ≥140 mmHg or DBP ≥90 mmHg) and they had been informed by a healthcare provider that they had high blood pressure prior to this study, but they had not taken any prescribed anti‐hypertensive medication to lower or control their blood pressure, they were considered untreated for HTN.

2.2.2. Exposure

To assess socioeconomic disparities in HTN and its diagnosis, and treatment, we employed a wealth index as a stand‐in for SES. The proprietorship of durable goods (like TVs and bicycles) and housing (including a source of drinking water, sanitary facilities, and building supplies) were among the household assets that made up the BDHS wealth index. ²⁶ Based on principal component analysis (PCA), a weight (factor score) was assigned to each asset. Subsequently, a standard normal distribution with a mean of zero and a standard deviation of one was used to standardize the asset scores. Every asset was given a score for every household, and the scores were then added together. Following that, the sample was split up into population terciles, and each was given a rating of either zero (poor), one (middle), or two (rich). Based on the overall score of the household they reside in, people were ranked.

2.2.3. Moderator

In this study, sex is categorized as a dichotomous variable (either male or female).

2.2.4. Covariates

We embarked on a comprehensive two‐stage selection process to incorporate an extensive array of explanatory variables. Initially, we compiled a list of potential explanatory variables derived from an exhaustive review of existing literature. ¹ , ² , ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ²² , ²³ , ²⁴ , ²⁵ Subsequently, these identified variables were cross‐referenced with the survey data to verify their availability. The age groupings of the respondents were as follows: 18–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, and 65+ years. The respondents' educational backgrounds were categorized into four groups based on Bangladesh's official educational system: no education (0 years), primary education (1–5 years), secondary school (6–10 years), and higher education (11 years and above).

The area where people lived was categorized as either rural or urban. The current marital status and job status of the respondents were categorized as either yes or no. The weight in kilograms multiplied by the squared height in meters (kg/m²) yielded the body mass index (BMI). A BMI of less than 18.5 kg/m² was considered underweight for both men and women, 18.5–24.99 kg/m² was considered normal, and ≥25 kg/m² was considered overweight/obese, according to the 2017–2018 BDHS report. ²⁶ Terciles were used to classify the number of household members and a number of adult members in the household.

2.3. Statistical analysis

Through direct standardization, we calculated the age‐adjusted prevalence of HTN in individuals (diagnosed + undiagnosed). We used Bangladeshi citizens who were 18 years of age or older in the 2011 census to create an age‐specific reference population for the age‐adjusted prevalence estimates. As a reference group, we utilized the age distribution of study participants' HTN to compute the age‐adjusted percentage of undiagnosed and untreated HTN. For our nominal predictor variables, we used the Pearson chi‐squared test for independence to examine differences in the prevalence of HTN, undiagnosed for HTN, and untreated for HTN. For our ordinal predictors, we used the chi‐square for linear trend statistics.

To quantify socio‐economic disparities in the prevalence of HTN (diagnosed + undiagnosed), undiagnosed for HTN, and untreated for HTN, we used three regression‐based approaches: (1) adjusted odds ratio (aoR); (2) relative index of inequality (RII); and (3) slope index of inequality (SII). The aoR measures variations in HTN, undiagnosed for HTN, and untreated for HTN only in relative terms, while the RII and SII are regression‐based measures of inequality that consider the prevalence of HTN, undiagnosed for HTN, and untreated for HTN across the entire socioeconomic spectrum of the study population.

The SII was calculated following Mackenbach and Kunst's guidelines. ²⁸ The gradient was found using linear regression with the age‐adjusted prevalence serving as the outcome variable and the relative rank of the socioeconomic assessment component serving as the predictor variable. When the number is 0, it indicates that there is no relationship between SES and poor health. Positive SIIs indicate that the health indicator rises as SES rises, while negative SIIs imply that the unfavorable health indicator falls as SES rises.

To calculate the relative rank of social class, the cumulative percentage of the sample in each category was estimated. The chosen social class code was the midpoint. For example, the poor social class category, comprising 33.3% of the population, had a value of 0.167 (0.333/2); the middle social class category, comprising 33.4% of the population, had a value of 0.501 (0.333+ [0.334/2]); and the rich social class category, comprising 33.3% of the population, had a value of 0.834 (1– [0.333/2]). We computed RII using a modified Poisson's method as described by Zou, ²⁹ which results in more reliable estimates than the binary method. The impoverished are more prone than the wealthy to experience negative health outcomes when the RII is less than one. We created three fully adjusted logistic regression models, each of which contained SES, to evaluate each binary outcome variable.

To determine whether sex status affects the relationship between SES and the prevalence of HTN, undiagnosed for HTN, and untreated for HTN, we also conducted logistic regression analyses to examine the adjusted association between SES and the condition after stratification by sex. All the covariates were included simultaneously in the multiple regression models. Through the analysis of the variance inflation factors, which came out to be 2.0, multicollinearity was ruled out.

We used the contrast post‐estimation command to do a test for linear trend in the logistic model containing an ordered categorical independent variable, and the resulting p‐value for the trend was given. For every analysis, the significance threshold was fixed at p < .05. The results of the study were presented using the guidelines for improving the reporting of observational studies in epidemiology. Stata version 14.0 (Strata Corp. LP, College Station, USA) was used for all analyses to account for sample weights based on the complex survey design of the BDHS.

3. RESULTS

3.1. Descriptive statistics

The sociodemographic characteristics of the respondents are shown in Table 1. This study included a total of 11,776 respondents. A total of 73.5% of the respondents lived in rural areas and 43.8% of the respondents were between the ages of 18 and 34 years. About 62.1% of respondents reported having jobs, 80.5% of the respondents were currently married, and about 26% of respondents had no education.

TABLE 1.

Descriptive statistics according to the sage‐adjusted prevalence of HTN, undiagnosed for HTN, and untreated cases of HTN‐diagnosed individuals: 2017–2018, Bangladesh Demographic and Health Survey

		n (age‐adjusted prevalence in %)
Measures	n (%) a	HTN (n = 11,776) b	Undiagnosed for HTN (n = 3384) c	Untreated for HTN (n = 1479) d
Age, yrs
18–34	5128 (43.8)	683 (13.0)	492 (70.9)	24.4
35–39	1392 (11.8)	384 (27.6)	250 (68.0)	12.3
40–44	1036 (8.9)	329 (30.7)	184 (55.8)	15.6
45–49	1026 (8.5)	389 (37.2)	203 (53.9)	7.2
50–54	671 (5.7)	279 (41.2)	142 (53.5)	13.1
55–59	692 (5.9)	324 (45.1)	148 (46.7)	7.5
60–64	685 (5.8)	344 (49.8)	158 (45.7)	7.3
65+	1146 (9.6)	652 (55.4)	322 (51.8)	11.0
p‐value *		<0.001	<0.001	<0.001
Currently married
No	2411 (19.5)	741 (25.0)	418 (62.1)	11.5
Yes	9365 (80.5)	2643 (24.7)	1481 (56.9)	12.5
p‐value		0.015	0.077	0.687
Currently working
No	4517 (37.9)	1486 (28.6)	725 (50.2)	10.4
Yes	7259 (62.1)	1898 (23.4)	1174 (62.5)	14.2
p‐value		<0.001	<0.001	0.050
BMI
Underweight	2046 (17.6)	381 (13.3)	257 (70.6)	15.2
Normal	6873 (58.4)	1752 (22.2)	1027 (60.6)	11.7
Overweight/obese	2857 (24.0)	1251 (40.4)	615 (47.5)	12.4
p‐value *		<0.001	<0.001	0.675
Education
No education	2929 (26.0)	1079 (23.3)	589 (60.8)	65.4
Primary	3568 (30.1)	1002 (24.3)	543 (54.8)	61.6
Secondary	3353 (29.0)	822 (27.8)	473 (55.6)	60.0
Higher	1926 (14.9)	481 (29.6)	294 (55.6)	61.3
p‐value *		<0.001	0.011	0.209
Sex
Female	6588 (56.2)	1980 (27.8)	973 (50.1)	11.4
Male	5188 (43.8)	1404 (22.3)	926 (67.4)	14.1
p‐value		0.002	<0.001	0.202
Living in a patriarchal family
No	1345 (11.9)	407 (27.7)	224 (54.4)	10.6
Yes	10,431 (88.1)	2977 (24.8) 0.455	1675 (57.6) 0.327	12.5 0.482
No. of adult member
1–2	4150 (36.3)	1115 (25.6)	673 (60.0)	14.2
3	2927 (25.0)	915 (25.6)	502 (58.3)	9.4
4+	4699 (38.7)	1354 (24.6)	724 (54.5)	12.7
p‐value *		0.056	0.001	0.830
No. of household member
1–4	4987 (43.6)	1471 (25.8)	851 (59.1)	13.4
5	2282 (19.3)	663 (25.7)	368 (55.6)	8.7
6+	4507 (37.1)	1250 (24.1)	680 (55.6)	12.9
p‐value *		0.059	0.069	0.874
Place of residence
Rural	7561 (73.5)	2113 (24.3)	1237 (59.1)	13.4
Urban	4215 (26.5)	1271 (27.6)	662 (51.7)	9.7
p‐value		0.209	0.009	0.082
Division
Barisal	1237 (5.6)	405 (27.6)	219 (56.7)	12.1
Chittagong	1583 (17.2)	490 (28.5)	264 (52.6)	7.9
Dhaka	1475 (22.3)	363 (22.6)	196 (54.2)	7.7
Khulna	1642 (12.6)	504 (25.7)	274 (56.5)	15.6
Mymensingh	1332 (8.2)	321 (20.0)	182 (58.1)	8.1
Rajshahi	1565 (14.7)	438 (25.1)	265 (59.9)	21.3
Rangpur	1541 (12.8)	489 (27.4)	318 (66.9)	20.9
Sylhet	1401 (6.6)	374 (23.9)	181 (49.7)	5.3
p‐value		0.047	0.098	0.053
SES
Poor	3857 (33.3)	952 (21.4)	694 (66.7)	14.8
Middle	3798 (33.3)	1033 (24.2)	603 (55.4)	12.7
Rich	4121 (33.4)	1399 (29.8)	602 (48.9)	9.3
p‐value *		<0.001	<0.001	0.012
n (prevalence)	11,776	3384 (25.1)	1899 (57.2)	177 (12.3)

^a Unweighted numbers and weighted percentage were presented.

^b Comprise all individuals who were screened for blood pressure, including patients with normotension and hypertension.

^c Include HTN individuals were defined as those whose systolic blood pressure (SBP) was ≥140 mmHg, whose diastolic blood pressure (DBP) was ≥90 mmHg, and/or who were taking anti‐hypertensive drugs to lower or control their blood pressure or being previously diagnosed as hypertensive by any health professional.

^d Include participant who had a diagnosis of HTN (SBP ≥140 mmHg or DBP ≥90 mmHg) at the time of the survey and if they had been informed by a medical professional that they had HTN prior to this study.

^* p‐value for chi square for linear trend statistic were provided.

BMI calculations showed that 24% of respondents were overweight/obese, 58.4% were normal weight, and 17.6% were underweight. In addition, 88.1% of respondents were living in a patriarchal family and 33.3% were from the poor SES group. The age‐adjusted prevalence of HTN in our study sample was 25.1%. The age‐adjusted prevalence of undiagnosed cases of HTN and untreated cases of HTN‐diagnosed individuals in our sample study was 57.2% and 12.3%, respectively.

The sociodemographic differences in HTN, undiagnosed HTN, and untreated HTN are also shown in Table 1. The prevalence of HTN was higher in the rich and middle SES groups (29.8% and 24.2%, respectively) than in the poor SES group (21.4%). However, in the case of undiagnosed HTN, we saw the opposite situation: undiagnosed HTN was higher in the poor (66.7%) and middle SES groups (55.4%) than in the rich SES group (48.9%). The same thing happened with untreated HTN: the poor and middle SES groups had more untreated HTN (14.8% and 12.7%, respectively) than the rich SES group (9.3%).

3.2. Multivariable analyses

3.2.1. Association between HTN, undiagnosed for HTN, and untreated for HTN with SES and other covariates

The aORs of the relationship between SES and the prevalence of HTN, undiagnosed for HTN, and untreated for HTN in our study sample are shown in Table 2. When compared to people in the poor SES group, people in the rich SES group had higher odds of developing HTN (aoR 1.25; 95% CI 1.08–1.45). When compared to individuals in the poor SES group, those in the rich (aoR 0.57; 95% CI 0.44–0.74) and middle SES (aoR 0.68; 95% CI 0.53–0.86) groups had lower odds of undiagnosed for HTN. When compared to individuals in the poor SES group, those in the rich SES group (aoR 0.56; 95% CI 0.31–0.98) had lower odds of untreated HTN.

TABLE 2.

Adjusted odds ratio for the association between SES and other covariates with the prevalence of HTN, undiagnosed HTN, and untreated cases of HTN‐diagnosed individuals: 2017–2018, Bangladesh Demographic and Health Survey

	aOR (95% CI)
Measures	HTN 1 (n = 11,776)	Undiagnosed for HTN 2 (n = 3384)	Untreated for HTN 3 (n = 1479)
Age, yrs
18–34	1.00	1.00	1.00
35–39	2.53 (2.12–3.01) a	0.91 (0.64–1.27)	0.44 (0.23–0.88) c
40–44	3.08 (2.56–3.71) a	0.52 (0.37–0.71) a	0.60 (0.32–1.12)
45–49	4.20 (3.45–5.12) a	0.49 (0.36–0.68) a	0.25 (0.13–0.50) a
50–54	5.83 (4.71–7.22) a	0.37 (0.27–0.52) a	0.44 (0.23–0.85) c
55–59	6.87 (5.63–8.40) a	0.28(0.19–0.39) a	0.25 (0.12–0.52) a
60–64	8.91 (7.21–11.02) a	0.24 (0.17–0.32) a	0.25 (0.12–0.49) a
65+	11.83 (9.78–14.32) a	0.26 (0.19–0.36) a	0.36 (0.20–0.64) b
P‐trend of odds	<0.001	<0.001	<0.001
Currently married
No	1.00	1.00	1.00
Yes	0.78 (0.81–1.05)	0.74 (0.59–0.93) c	0.71 (0.43–1.18)
Currently working
No	1.00	1.00	1.00
Yes	0.93 (0.67–0.90) b	1.05 (0.87–1.28)	1.21 (0.79–1.85)
BMI
Underweight	1.00	1.00	1.00
Normal	1.93 (1.69–2.21) a	0.64 (0.49–0.83) b	0.85 (0.43–1.69)
Overweight/obese	4.98 (4.20–5.89) a	0.45 (0.34–0.61) a	0.89 (0.43–1.82)
P‐trend of odds	<0.001	<0.001	0.675
Education
No education	1.00	1.00	1.00
Primary	1.05 (0.92–1.21)	0.80 (0.65–0.99) c	1.57 (0.99–2.48)
Secondary	1.11 (0.95–1.29)	0.84 (0.65–1.08)	1.10 (0.63–1.91)
Higher	1.09 (0.91–1.32)	0.88 (0.64–1.21)	1.82 (0.94–3.52)
P‐trend of odds	<0.001	0.011	0.209
Sex
Female	1.00	1.00	1.00
Male	0.80 (0.71–0.91) a	2.37 (1.93–2.91) a	1.19 (0.77–1.86)
Living in a patriarchal family
No	1.00	1.00	1.00
Yes	1.06 (0.89–1.25)	0.97 (0.73–1.28)	1.35 (0.74–2.49)
No. of adult member
1–2	1.00	1.00	1.00
3	0.96 (0.82–1.11)	1.05 (0.81–1.32)	0.89 (0.53–1.48)
4+	0.89 (0.75–1.06)	0.99 (0.77–1.30)	1.44 (0.82–2.53)
P‐trend of odds	0.056	0.070	0.830
No. of a household member
1–4	1.00	1.00	1.00
5	0.99 (0.85–1.16)	0.88 (0.69–1.13)	0.57 (0.32–1.00)
6+	0.94 (0.80–1.10)	0.94 (0.73–1.20)	0.72 (0.43–1.22)
P‐trend of odds	0.059	0.069	0.874
Place of residence
Rural	1.00	1.00	1.00
Urban	1.06 (0.93–1.21)	0.93 (0.75–1.14)	0.79 (0.50–1.23)
Division
Barisal	1.00	1.00	1.00
Chittagong	0.88 (0.70–1.10)	1.03 (0.73–1.47)	0.55 (0.25–1.20)
Dhaka	0.64 (0.50–0.80) a	1.06 (0.75–1.49)	0.61 (0.27–1.29)
Khulna	0.81 (0.65–1.01)	1.12 (0.78–1.57)	1.40 (0.65–3.01)
Mymensingh	0.67 (0.54–0.84) b	1.05 (0.72–1.51)	0.59 (0.26–1.38)
Rajshahi	0.87 (0.68–1.10)	1.11 (0.79–1.56)	1.84 (0.86–3.92)
Rangpur	1.04 (0.84–1.30)	1.41 (1.01–1.97) c	2.02 (1.03–3.97) c
Sylhet	0.85 (0.67–1.07)	0.73 (0.50–1.06)	0.37 (0.15–0.91) c
SES
Poor	1.00	1.00	1.00
Middle	1.07 (0.93–1.23)	0.68 (0.53–0.86) b	0.71 (0.43–1.18)
Rich	1.25 (1.08–1.45) b	0.57 (0.44–0.74) a	0.56 (0.31–0.98) c
P‐trend of odds	<0.001	<0.001	0.012

Note: ^1,2,3 Models were adjusted for age, currently married, currently working, BMI, education, sex, living in a patriarchal family, no. of adult members, no. of household members, place of residence, division, and SES; Here a, b, and c indicate p < .001, p < .01, and p < .05.

Abbreviations: aOR, adjusted odds ratio; CI, confidence interval.

The aORs of the association between other sociodemographic factors and the prevalence of HTN, undiagnosed for HTN, and untreated for HTN are also shown in Table 2. Respondents aged 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, and 65+ years had significantly higher odds of getting HTN but significantly lower odds of undiagnosed and untreated (except age group 35–39 yrs) for HTN than those aged 18–34 years. Males had a greater likelihood of going undiagnosed for HTN, but they also had lower odds of getting HTN than females. Normal and overweight/obese respondents showed a lower likelihood of going undiagnosed for HTN, but higher odds of developing HTN compared to their counterparts. The current working population had lower odds (aoR 0.93; 95% CI 0.67–0.90) than the non‐working group to develop HTN. Married people had lower odds (aoR 0.74; 95%CI 0.59–‐0.93) of undiagnosed for HTN than non‐married people. Respondents living in Dhaka and Mymensingh divisions were less likely to develop HTN. In addition, respondents from the Rangpur division had higher odds of undiagnosed and untreated HTN (Table 2)

3.2.2. Summary measures of SES inequality

The summary measures of inequality are shown in Table 3. RII indicates that there is a 14% increase in HTN prevalence when moving from the bottom to the top of the SES distribution, and a corresponding 30%, and 56% decrease in the prevalence of undiagnosed and untreated for HTN when moving from the top to the bottom of the SES distribution. From SII, a one‐unit shift from the top to the bottom of the SES group is connected to a 0.04‐unit drop in HTN prevalence. On the other hand, a one‐unit shift from the top to the bottom of the SES group is connected to a 0.20 and 0.10‐unit drop in undiagnosed and untreated HTN.

TABLE 3.

Summary measure of SES inequality and the prevalence of HTN, undiagnosed HTN, and untreated cases of HTN‐diagnosed individuals: 2017–2018, Bangladesh Demographic and Health Survey

	Relative index of inequality (RII)	Slope index of inequality (SII)
Measure	RR (95% CI)	ß coefficient (95% CI)
HTN (n = 11,776)	1.14 (1.01, 1.30)c	0.04 (0.01, 0.08)c
Undiagnosed for HTN (n = 3384)	0.70 (0.60, 0.81)a	–0.20 (–0.29, –0.12)a
Untreated for HTN (n = 1479)	0.44 (0.22, 0.89)c	–0.10 (–0.19, –0.013)c

Note: Here a, b, and c indicate p < .001, p < .01, and p < .05.

Abbreviations: CI, confidence interval; RR, risk ratio.

3.2.3. Association between HTN, undiagnosed for HTN, and untreated for HTN with SES by sex

Table 4 shows the aORs for the association between HTN, undiagnosed for HTN, and untreated for HTN with SES by sex. Male people in the rich SES group have a higher odd of developing HTN than their poorer SES counterparts (aoR 1.28; 95% CI 1.01–1.62). Respondents in the middle or rich bands of wealth had lower odds of undiagnosed for HTN than those in the poor bands, regardless of male or female. Furthermore, compared to female respondents with a poor SES, those with a richer SES had a lower likelihood of going untreated for HTN.

TABLE 4.

Adjusted odds ratio for the association between HTN, undiagnosed HTN, and untreated cases of HTN‐diagnosed individuals with SES by sex: 2017–2018, Bangladesh Demographic and Health Survey.

	Male (aOR, 95% CI) 1			Female (aOR, 95% CI) 2
Measure	HTN (n = 5188)	Undiagnosed for HTN (n = 1405)	Untreated for HTN (n = 477)	HTN (n = 6588)	Undiagnosed for HTN (n = 1979)	Untreated for HTN (n = 1002)
SES
Poor	1.00	1.00	1.00	1.00	1.00	1.00
Middle	1.03	0.63	1.17	1.09	0.70	0.82
	(0.84–1.26)	(0.42–0.94)	(0.47–2.94)	(0.91–1.31)	(0.53–0.93)	(0.50–1.32)
	0.772*	0.025*	0.737*	0.340*	0.013*	0.409*
Rich	1.28	0.53	0.79	1.18	0.60	0.51
	(1.01–1.62)	(0.34–0.83)	(0.27–2.27)	(0.96–1.44)	(0.43–0.85)	(0.27–0.98)
	0.040*	0.005*	0.665*	0.112*	0.004*	0.042*

Note: ^1,2 Models were adjusted by age, currently married, currently working, BMI, education, sex, living in a patriarchal household, no. of adult members, no. of household members, place of residence, and division.

^* p‐value

4. DISCUSSION

4.1. Major findings

The findings of a nationwide representative large survey conducted in Bangladesh revealed a high prevalence of HTN (25.1%), along with a substantial proportion of undiagnosed cases of HTN (57.2%) and untreated cases (12.3%) of HTN‐diagnosed individuals. According to the study, SES has a significant influence on a person's likelihood of acquiring HTN, influencing the likelihood that a person will go undiagnosed for HTN, and influencing a patient's decision to take anti‐hypertensive medication. While SES had an independently detrimental effect on HTN prevalence, sex moderated the association; males from rich SES were more likely to suffer from HTN than males from poor SES.

4.2. Discussion of major findings

The obtained age‐adjusted HTN prevalence (25.1%) was greater than that reported systematic review and meta‐analysis conducted in Bangladesh ⁴ where the pooled prevalence of HTN was estimated to be 20.0%. The obtained prevalence was also higher than the age‐standardized prevalence of South Asian countries (20.1%), but lower than overall LMICs (31.5%), and high‐income countries [HICs] (28.5%). ³⁰ This alarmingly high prevalence of HTN in Bangladesh is regarded as a warning indication of the disease's rapid spread and suggests that Bangladesh contributes significantly to the burden of HTN in South‐East Asia.

This data indicates that the country has a poor diagnosis rate for HTN and a low treatment‐seeking behavior. Furthermore, our study's rate of undiagnosed and untreated HTN was comparable to other earlier studies carried out in areas with limited resources, including Ghana (65%, 22.2%), ³¹ India (42.3%, 6%), ³² Pakistan (58.1%, 25.2%), ³³ and Nepal (57.3%, 24.1%). ³⁴ These results demonstrate that more treatment and screening coverage is needed even though considerable progress has been made. Bangladesh must increase public knowledge of HTN considering these situations and offer suitable education and follow‐up for HTN patients.

According to our findings, women were more likely than men to develop HTN. The results varied according to sex and the prevalence of HTN, with geographic location showing up as a significant impact. In contrast to industrialized countries like Korea ³⁵ and the United States, ³⁶ where males were more likely to have HTN, it was discovered that females had a greater prevalence of the disease than males in several developing countries, including Bangladesh, ³⁷ India, ²² , ²³ and Nepal. ³⁸ It is believed that multifactorial biological and environmental factors, including genetic risk, epigenetic factors, poor dietary quality, insufficient physical activity, and a higher prevalence of overweight/obesity, influence the higher likelihood of HTN in females, ³⁹ though the precise mechanism underlying this finding is unknown. Furthermore, an additional analysis was carried out in our study to support this theory, and it was discovered that women were more likely than men to be overweight or obese.

This study found that males were more likely than females to have undiagnosed HTN. Nonetheless, conflicting findings of the sex gap in undiagnosed for HTN have been reported globally. For instance, in a cross‐sectional study that was conducted between 2007 and 2017 using the Canadian Health Measures Survey, ⁴⁰ which was nationally representative, women over a 10‐year period showed lower diagnoses of HTN, which was not the case for men. According to data from the China Hypertension Survey, women had greater diagnosis (51.9% vs. 42.5%) rates than men. ⁴¹ Furthermore, compared to men, women showed a greater prevalence and diagnosis of HTN in a cross‐sectional study conducted in Bangladesh. ³⁷ These variations do point to the potential for sex‐ and location‐specific health interventions for early detection to slow the rise of possibly avoidable heart attacks and strokes because of high blood pressure.

Interesting connections between SES and HTN prevalence were found in this investigation. The results of nationally representative data from Bangladesh showed that people with higher SES had higher rates of HTN than people who had poorer SES. Our findings are in line with those of earlier research carried out in LMICs. ¹⁵ , ¹⁶ In contrast, people with poor SES are more likely to have HTN in Western countries. ¹¹ , ¹² , ¹³ , ¹⁴ When interpreting this contrast, it is important to consider the food security and energy expenditure patterns of individuals in poor SES in South Asia. These patterns include food scarcity, lower consumption of refined foods, and high energy expenditure due to moderate to intense physical activity at work. ⁴² On the other hand, due to their higher rates of smoking, higher BMIs, and less activity than higher socioeconomic groups, poorer socioeconomic groups in western countries may have a higher prevalence of HTN, according to several research, including meta‐analyses. ⁴³ , ⁴⁴

Consistent with earlier systematic reviews conducted in LMICs, we discovered that patients with HTN who belonged to the rich socioeconomic strata had a higher likelihood of having their condition diagnosed and treated compared to those in lower socioeconomic strata. ⁴⁶ Higher living standards, easier access to HTN information, and easier access to medication, ⁴⁵ may be the cause of the higher diagnosis and treatment rates of HTN among individuals with higher SES.

The data also revealed that sex was a moderator in the association between having a rich SES and developing HTN and that this association was only evident for males, males who came from rich SES group were at a disadvantage in terms of their chance of developing HTN. One explanation could be that men in Bangladesh who belong to higher socioeconomic groups tend to consume more energy, work in white‐collar jobs, and lead sedentary lives, ⁴⁷ all of which raise the risk of HTN. The effect of sex on the relationship between SES and HTN should be carefully considered and requires further investigation, as there was not a strong enough association seen in men with high SES and HTN.

The findings also revealed that belonging to the rich SES group boosts the reduced risk of undiagnosed HTN in both males and females, implying that it is poor SES per se for both males and females that disadvantages individuals with undiagnosed HTN. The importance of this finding therefore needs to be underscored. Whether a patient is male or female, having poor SES has a negative impact on their ability to get a diagnosis for their HTN.

The results also showed that being in the rich SES group increases the likelihood that women with HTN diagnoses will not go untreated as comparted to HTN‐diagnosed poor SES group women. Due to resource limitations that may limit their access to healthcare, educational opportunities, lifestyle choices, and healthcare information, poor women with HTN may encounter difficulties in managing their illness. It might also be more challenging for them to learn about HTN and get treatment because of these issues.

Previous studies have established that age is a significant risk factor for the development of HTN, with the prevalence of HTN increasing with age. ⁶ , ¹⁶ , ⁴⁶ , ⁴⁸ Our results showed data that was consistent with the conclusions of previous studies. This may be explained by the fact that older people tend to be less active daily, have greater rates of insufficient physical activity, more time spent inactive, and lower levels of physical activity. ⁴⁹ In line with earlier research, ⁵⁰ , ⁵¹ we also found that the prevalence of undiagnosed and untreated HTN declined with aging. One theory is that older individuals are more likely to have HTN, have been aware of the condition, and have the resources to receive the necessary treatment.

Our analysis revealed an association between HTN and overweight/obesity, which is in line with international research. ⁵¹ , ⁵² Undiagnosed for HTN were also significantly higher in people with overweight/obese compared with underweight. ⁵¹ Numerous earlier research has found that the most significant factor influencing healthcare consumption is marriage. ⁵³ , ⁵⁴ Our findings were consistent with earlier research in showing a lower likelihood of undiagnosed HTN, ⁵⁵ , ⁵⁶ , ⁵⁷ , ⁵⁸ in married individuals. By causing changes in lifestyle and health/illness awareness as well as by enabling postponed/hidden illnesses due to societal stigmatization, marriage may increase the utilization of HTN health services.

In line with a prior Bangladeshi study, our research revealed that individuals with jobs had a lower likelihood of developing HTN than those without jobs. ⁵⁹ The likelihood of undiagnosed and untreated HTN was higher among residents of the Rangpur division. This discrepancy could be the result of unequal public infrastructure development, such as hospitals, poor access to urban areas, and living in resource‐poor environments, ⁵⁹ , ⁶⁰ which could make it more difficult to identify and treat adults from Rangpur who have elevated blood pressure early on.

4.3. Strengthens and limitations

The study has the following strengths: (1) our analysis is based on nationally representative data, therefore the findings may be generalized to the Bangladeshi population; (2) clinical factors, such as BP, body weight, and height, were measured with high‐quality methods; (3) PCA was used to create the household wealth index, which yielded a more accurate estimate of SES in Bangladesh than income or consumption expenditures. Furthermore, wealth is more resilient to shocks to the economy because of the high volatility of consumption, particularly in LMICs; and (4) rather than focusing solely on the lowest or highest SES groups, the RII and SII examines enable us to investigate the inequality in the burdens associated with HTN along the entire socioeconomic distribution.

The study's shortcomings are as follows: (1) clinical record information, including history of HTN, was not examined because the patients were numerous members of the community. Some of the individuals have never visited a hospital; (2) due to the lack of data collection in the BDHS 2017, we were unable to include several lifestyle‐related factors, such as smoking, eating, or exercise habits, nor could we include clinical records of other diseases, such as a history of diabetes, cardiovascular disease, or other blood biochemical criteria, as potential covariates. But considering the strong and well‐established associations between HTN and SES, it seems uncertain that including more lifestyle factors in the model will result in a non‐significant relationship when it comes to predicting the chance of developing HTN and SES; and (3) since this is a cross‐sectional design, we were only able to collect each participant's blood pressure data on a single day, making it unable to account for the “white coat effect.”

5. CONCLUSIONS

According to this study, Bangladesh encompasses a comparatively high prevalence of HTN, undiagnosed HTN, and untreated cases of HTN‐diagnosed individuals. Rich SES groups had higher rates of HTN than poor SES groups, but those with poor SES with HTN were also less likely to receive treatment and be aware that they had the condition. Our findings also suggested that although SES had an independently adverse effect on developing HTN sex moderated the likelihood of illness. Participants who were both poor and male were uniquely disadvantaged in terms of their developing HTN than participants who were rich and male. In addition, regardless of sex, respondents from poor‐SES backgrounds were more likely to have been undiagnosed for HTN. This study suggests that the government and other relevant stakeholders may focus more on creating appropriate policy measures to lower the risk of HTN, especially for males in rich‐socioeconomic groups, as well as focused efforts to screen and diagnose of HTN in socioeconomically disadvantaged populations. Future longitudinal research is nevertheless required to examine the impact of putative processes mediating the association between SES and the prevalence of HTN, undiagnosed for HTN, and untreated for HTN.

AUTHOR CONTRIBUTIONS

Tapan Kumar Roy and Mosiur Rahman originated the study and contributed to the study design, statistical analysis, and the writing of the article. Nityananda Halder, Md. Sohanur Rahman, and Md Mamunur Rashid contributed to the analysis and interpretation of data and revisions of the article. All authors read and approved the final manuscript

CONFLICT OF INTEREST STATEMENT

The authors state that the work was carried out in the absence of any commercial or financial relationships.

PATIENT CONSENT STATEMENT

Every responder gave their informed consent before the interview began. Following that, the interviewers gave an oral explanation in line with BDHS protocol. Prior to the studies, the respondents received instructions and information about the significance of BP measurement.

Kumar Roy T, Rahman M, Rahman MS, Halder N, Rashid MM. Is gender a factor in socioeconomic disparities in undiagnosed, and untreated hypertension in Bangladesh? J Clin Hypertens. 2024;26:964–976. 10.1111/jch.14858

DATA AVAILABILITY STATEMENT

The data are available in a public, open access repository at https://dhsprogram.com/data/Access‐Instructions.cfm