Hidden crisis of undiagnosed hypertension among women of reproductive age with elevated blood pressure: a regional disparities and determinants using Tanzania nation survey

Abstract

Background

Hypertension, the leading risk factor for cardiovascular diseases, remains largely undiagnosed in low-resource settings. This study focuses on the hidden crisis of undiagnosed hypertension among women of reproductive age with elevated blood pressure (BP) in Tanzania, where up to 90% remain unaware of their condition.

Methods

Using data from the 2022 Tanzania Demographic and Health Survey and Malaria Indicator Survey (TDHS-MIS), this cross-sectional study analyzed a nationally representative sample of 853 women aged 15–49 years with elevated BP. Hypertension was defined as an average systolic blood pressure (SBP) ≥ 140 mmHg, diastolic blood pressure (DBP) ≥ 90 mmHg, or current antihypertensive medication use. Undiagnosed hypertension was categorized based on self-reported lack of prior diagnosis. Multilevel logistic regression was employed to identify determinants at individual and community levels, while forest plots assessed regional disparities.

Results

The prevalence of undiagnosed hypertension was 77%. Regional disparities were evident, with the Western (90%) and Central (89%) zones showing the highest prevalence, these zonal estimates are elevated because key regions within them report very high proportions, Tabora (97%) and Dodoma (90.5%) respectively. In contrast, Zanzibar had the lowest prevalence at 48%. With each additional year of age, the odds of being undiagnosed decreased by 8.0% (AOR: 0.92, 95% CI: 0.89–0.95), while married women (AOR: 0.21, 95% CI: 0.08–0.59)women with internet access (AOR: 0.32, 95% CI: 0.15–0.69) and women from wealthier households all had significantly lower odds of undiagnosed hypertension ((middle (AOR: 0.32, 95% CI: 0.16–0.67)) and ((higher socioeconomic classes (AOR: 0.37, 95% CI: 0.17–0.80)). Community-level variation accounted for 25–32% of disparities, with a Median Odds Ratio (MOR) of up to 3.30, indicating substantial contextual influence.

Conclusions

The high prevalence of undiagnosed hypertension among Tanzanian women highlights the urgent need for targeted interventions. Improving health awareness, integrating screening into primary care, and addressing regional inequities can significantly reduce this burden and improve maternal and cardiovascular health outcomes.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12872-025-05473-2.

Background

Noncommunicable diseases (NCDs) caused an estimated 43 million deaths in 2021, representing 75% of all non-pandemic-related deaths globally. Of these, 73% occurred in low- and middle-income countries. Cardiovascular diseases (CVDs) were the leading cause of NCD-related deaths, accounting for 19 million fatalities, followed by cancers (10 million), chronic respiratory diseases (4 million), and diabetes, including diabetes-related kidney disease, which collectively caused over 2 million deaths [1]. Arterial hypertension defined as persistently elevated blood pressure in the systemic arteries, is recognized as the leading risk factor for CVDs and a major contributor to global mortality [2]. For clarity, we refer to this condition as hypertension throughout the remainder of the manuscript. Over the past 30 years, the number of individuals living with hypertension has doubled, reaching an estimated 1.28 billion globally [3]. Each year, hypertension causes approximately 10 million deaths, making it the single largest cause of mortality worldwide and surpassing the combined deaths from all infectious diseases. Alarmingly, nearly 60% of adults with hypertension remain undiagnosed, with the highest prevalence in low- and middle-income countries, including sub-Saharan Africa (SSA) [3, 4] This highest prevalence in SSA is aggravated by limited health system capacities, inadequate screening, and poor health literacy in resource-constrained settings [5].

SSA is undergoing an epidemiological transition marked by a growing burden of hypertension with the prevalence reaching 48% (CI. 42–54%) in women and 34% (CI. 29–39%) in men in 2019 [4]. However, awareness and treatment levels in this region are dismally low, with only one in four individuals aware of their condition [6–9]. Gender disparities further compound the problem, with women, particularly those of reproductive age, facing unique challenges related to undiagnosed hypertension. These challenges include physiological changes, pregnancy-related risks, and sociocultural barriers that limit their access to healthcare [6, 9, 10].

In Tanzania, hypertension has emerged as a public health priority, with prevalence rates rising alongside urbanization, changing dietary habits, and increasing life expectancy. Recent estimates suggest that 65% up to 90% of hypertensive individuals in the country are unaware of their condition [11, 12], due to fewer opportunities for routine blood pressure screening. Therefore, undiagnosed hypertension among women of reproductive age is a critical issue due to its implications for maternal health and long-term cardiovascular outcomes. Early detection and intervention are essential to mitigating these risks. However, despite the severity of this problem, evidence on the prevalence, determinants, and regional/zonal disparities of undiagnosed hypertension in Tanzania remains limited.

Hence, this study aims to address these gaps by quantifying the prevalence of undiagnosed hypertension among women of reproductive age with elevated BP in Tanzania, identifying individual- and community-level determinants, and highlighting regional disparities using data from the 2022 Tanzania Demographic and Health Survey and Malaria Indicator Survey (2022 TDHS-MIS). The findings will provide evidence to inform targeted interventions and strengthen the health system to improve hypertension screening and management in Tanzania.

Methods

Study design and data source

This study analyzed data from the 2022 TDHS-MIS, conducted by the National Bureau of Statistics (NBS) in collaboration with the Office of the Chief Government Statistician Zanzibar (OCGS). The survey was implemented with support from the Ministries of Health in Tanzania Mainland and Zanzibar and the Tanzania Food and Nutrition Centre (TFNC) for biomarker data collection. Technical guidance was provided by ICF International through the DHS program. Data collection was conducted between February and July 2022. The survey, part of the ongoing DHS program conducted every four years in Tanzania since 1990, provided nationally representative data on demographic, health, and socio-economic factors, including blood pressure measurements. This dataset enabled a detailed exploration of the hidden crisis of undiagnosed hypertension among women with elevated BP, highlighting regional disparities and key determinants in a limited-resource setting.

Sample size and sampling procedure

The 2022 TDHS-MIS employed a stratified two-stage cluster sampling method to generate a nationally representative sample for Tanzania Mainland (urban and rural areas) and Zanzibar. In the first stage, enumeration areas (EAs) from the 2012 Tanzania Population and Housing Census were used to select sampling points (clusters) using probability proportional to size within each stratum. A total of 629 clusters were chosen. In the second stage, 26 households were systematically selected from each cluster, resulting in an initial target of 16,354 households. However, due to security issues, one EA was excluded, and 16,312 households were ultimately selected. Among these, 15,907 were occupied, and 15,705 were successfully interviewed, achieving a 99% response rate [13].

The TDHS-MIS primarily targets women of reproductive age (15–49 years). A total of 15,699 eligible women were identified from the interviewed households, and 15,254 (97.2%) completed individual interviews. Although men are also interviewed in selected households for specific health-related modules, they were not included in this analysis. Importantly, blood pressure (BP) measurements were conducted in a randomly selected subsample comprising half of the surveyed households, consistent with DHS biomarker collection procedures aimed at reducing participant burden and fieldwork costs while ensuring national representativeness. Among women of reproductive age residing in these households, 7,627 were eligible and consented for blood pressure measurement, of whom 7,610 (99.8%) agreed and were successfully measured. For this analysis, women were classified as hypertensive if they had an average SBP of 140 mmHg or higher, an average DBP of 90 mmHg or higher, or if they were currently taking antihypertensive medication. Based on these criteria, 853 women with hypertension were included in the final analysis. Detailed information on the sampling procedure is presented in Fig. 1.

Fig. 1.

Selection of study respondents included in this analysis

Undiagnosed hypertension

In the 2022 TDHS-MIS, women were considered hypertensive if they were currently taking antihypertensive medication or if their average blood pressure measurements met the diagnostic criteria. BP was measured using a validated automatic digital device, the Life Source blood pressure monitor (UA-767 F) which is commonly used in large-scale population surveys and meets international standards for accuracy. Before blood pressure measurements, respondents were seated and rested for at least five minutes. Three readings taken at intervals of at least 10 min, and the average of the second and third readings was used for analysis. BP categories followed World Health Organization/International Society of Hypertension (WHO/ISH) aligned thresholds: Acceptable (< 120/<80 mmHg); At the high end of acceptable (120–139 and/or 80–89 mmHg); Above acceptable/Grade 1 (140–159 and/or 90–99 mmHg); Moderately high/Grade 2 (160–179 and/or 100–109 mmHg); High/Grade 3 (≥ 180 and/or ≥ 110 mmHg); and Very high/Malignant (≥ 180 and/or ≥ 120 mmHg with acute target-organ damage). For prevalence estimates, hypertension was defined as SBP ≥ 140 mmHg and/or DBP ≥ 90 mmHg based on the averaged measure. Undiagnosed hypertension was categorized as “YES” for women with elevated BP who reported never being informed by a doctor or healthcare provider that they had high BP, and “NO” for those who had been informed of their condition.

Statistical analysis

Forest plots were constructed to compare the proportions of individuals with undiagnosed hypertension across zones and regions, along with their 95% confidence intervals (CI). The prevalence of undiagnosed hypertension was estimated as the weighted proportion of hypertensive women who were unaware of their condition, using the total number of hypertensive women as the denominator for each region and zone. The Wilson score method was used to compute the CIs. Regions were sorted in descending order of their proportions, with point estimates represented as markers and horizontal error bars indicating the 95% CI for each region. A vertical red dashed line represented the overall mean proportion, serving as a reference to identify regions with higher or lower proportions. Regions with point estimates and CIs entirely above or below the mean indicated significantly higher or lower proportions of undiagnosed hypertension, respectively, while overlapping intervals suggested no significant differences. This visual approach enabled the identification of regions with significantly higher or lower proportions of undiagnosed hypertension and regions with overlapping confidence intervals, indicating potential similarities in their estimates. The forest plot was generated using Python’s Matplotlib library for clear and accurate visualization.

The fixed effects component of the multilevel logistic regression models included both individual- and community-level variables. Individual-level variables comprised age (continuous), education level, marital status, employment status, health insurance coverage, internet use in the last 12 months, media exposure, and alcohol consumption. Community-level variables were modeled as fixed contextual effects and included household wealth status, perceived distance to health facility, geographical zone, and place of residence (urban/rural). This specification allowed us to evaluate how both individual characteristics and broader socioeconomic context are associated with undiagnosed hypertension and follows standard guidance for multilevel logistic models [14].

To account for the hierarchical structure of the 2022 TDHS-MIS data, we further specified a two-level random intercept model, where individual women (level 1) were nested within communities (level 2). Communities were defined as enumeration areas (EAs)/clusters, which served as the primary sampling units in the DHS design. These clusters were treated as the random grouping unit to capture unobserved community-level heterogeneity; between-cluster variability was summarized using the random-intercept variance, intraclass correlation coefficient (ICC), and median odds ratio (MOR). Broader contextual identifiers (geographical zone; urban/rural) were included as fixed effects rather than additional random levels, aligning with DHS methodological guidance on multilevel weighting/structure and common analytical practice in DHS applications [15, 16]. Notably, recent TDHS-MIS 2022 analyses in Tanzania have used the same approach of multilevel mixed-effects logistic regression with individuals nested in clusters and contextual covariates (e.g., zone, residence, wealth status, distance to health facilities) entered as fixed effects [17–19].

We fitted four sequential models to examine the contribution of variables at each level:

• Model 1 (Null model): Contained only the random intercept, with no explanatory variables, to estimate baseline community-level variance.

• Model 2 (Individual-level model): Included only individual-level covariates.

• Model 3 (Community-level model): Included only community-level covariates.

• Model 4 (Full model): Combined both individual- and community-level covariates.

Comparison of model fit was assessed using the Akaike’s Information Criterion (AIC) and the log pseudolikelihood, with the lowest AIC and highest log likelihood indicating the best-fitting model. All analyses incorporated DHS sampling weights, strata, and primary sampling units (clusters) to ensure representativeness of the national population.

Results

Respondents characteristics

Table 1 summarizes the background characteristics of 853 women who classified as having hypertension in this study. The median age was 38 years (IQR: 30–45), with the largest proportion (42.68%) aged 40–49 years. Most of women with hypertension were married or living with a partner (69.68%), and a majority had primary education (59.04%), with only 0.89% attaining tertiary education. Over half (52.32%) were employed for cash, while 26% were unemployed. A significant majority lacked health insurance (91.85%) and internet access in the past year (81.25%). Media exposure was nearly evenly distributed (51.01% exposed). Most of women with hypertension abstained from alcohol (87.06%) and belonged to wealthier households (57.94% rich). Rural residents (58.32%) were more prevalent than urban residents, with women with hypertension distributed across geographical zones, predominantly in the Coastal zone (30.66%). Based on WHO (1999) cut-off points for blood pressure classification, more than nine in ten women (96%) had levels above the optimal range, with the majority (63.74%) classified as Grade 1 hypertension.

Table 1.

Percent distribution of women of reproductive age with elevated BP by selected background characteristics, 2022 TDHS-MIS (n = 853)

Variable	n (%) weighted
Age (median (IQR) = 38 (30–45))
15–29	223 (26.13)
30–34	117 (13.68)
35–39	149 (17.50)
40–49	364 (42.68)
Marital status
Never married	98 (11.51)
Married/living together	594 (69.68)
Separated/widow/divorced	161 (18.81)
Education level
None	136 (15.94)
Primary	504 (59.04)
Secondary	205 (24.12)
Tertiary	8 (0.89)
Employed in last 12 months
Not employed	222 (26.00)
Employed for cash	446 (52.32)
Employed but paid-in-kind	185 (21.68)
Covered by health insurance
No	784 (91.85)
Yes	69 (8.15)
Media exposure
No	418 [48.99]
Yes	435 [51.01]
Use of internet in last 12 months
No	693 (81.25)
Yes	160 (18.75)
Alcohol consumption
No	743 (87.06)
Yes	110 (12.94)
Household wealth status
Poor	204 (23.86)
Middle	155 (18.19)
Rich	494 (57.94)
Geographical zone
Central	47 (5.48)
Coastal	262 (30.66)
Lake	194 (22.80)
Northern	72 (8.46)
Southern	77 (9.09)
Southern Highlands	84 (9.82)
Western	79 (9.25)
Zanzibar	38 (4.43)
Residence
Rural	497 (58.32)
Urban	356 (41.68)
Blood pressure category
Acceptable range	23 (2.70)
At the high end of acceptable	10 (1.22)
Above acceptable (Grade 1)	544 (63.74)
Moderately high (Grade 2)	179 (20.96)
High (Grade 3)	65 (7.64)
Very high (Malignant)	32 (3.73)

Acceptable = (< 120/<80 mmHg); At the high end of acceptable = (120–139 and/or 80–89 mmHg); Above acceptable/Grade 1 = (140–159 and/or 90–99 mmHg); Moderately high/Grade 2 = (160–179 and/or 100–109 mmHg); High/Grade 3 = (≥ 180 and/or ≥ 110 mmHg); and Very high/Malignant = (≥ 180 and/or ≥ 120 mmHg with acute target-organ damage)

Prevalence of undiagnosed hypertension

The study reveals that a substantial proportion (77%) of women with elevated BP remain undiagnosed, while only a smaller percentage (23%) have been diagnosed. This indicates a significant gap in hypertension awareness and detection among women with elevated BP.

Zonal and regional disparities in undiagnosed hypertension

Figures 2 and 3 present forest plots showing the proportions of undiagnosed hypertension across zones and regions among women classified as having hypertension in Tanzania. The plots reveal significant disparities, with the Western zone exhibiting the highest rate at 90%, primarily driven by high proportions in Tabora (97%) and Kigoma (91%) regions. Similarly, the Central zone has a high rate of 89%, heavily influenced by the Dodoma (90.5%) region. Both these zones and their respective regions are significantly above the national average, as indicated by the red dashed line. Although other zones do not show rates significantly above the national average, some of their regions exhibit disproportionately high rates of undiagnosed hypertension. These include Geita (89%) and Simiyu (92%) from the Lake zone, Manyara (90%) from the Northern zone, Rukwa (92%) from the Southern zone, and Ruvuma (89%) from the Southern Highlands. In contrast, regions such as Dar es Salaam (61%) from the Coastal zone, Mwanza (61%) from the Lake zone, and Mjini Magharibi (48%) from Zanzibar have significantly lower proportions of undiagnosed hypertension, falling below the national average of 75%. The details are provided in the Supplementary Tables S1 and S2.

Fig. 2.

Forest plot for proportions of undiagnosed hypertension among women with elevated BP by region, 2022 TDHS-MIS (n = 853)

Fig. 3.

Forest plot for proportions of undiagnosed hypertension among women with elevated BP by zones, 2022 TDHS-MIS (n = 853)

Factors associated with undiagnosed hypertension

Table 2 presents the results of a stepwise multilevel (mixed-effects) logistic regression analysis examining factors associated with undiagnosed hypertension among women with elevated BP. The analysis incorporates both fixed and random effects, with the final model (Model 4) combining individual- and community-level variables along with a random intercept to account for clustering at the community level.

Table 2.

Multilevel logit models for the factors associated undiagnosed hypertension among women of reproductive age with elevated BP, 2022 TDHS-MIS (n = 853)

Variable	Model 1 AOR [95% CI]	Model 2 AOR [95% CI]	Model 3 AOR [95% CI]	Model 4 AOR [95% CI]
Fixed Effects Component
Individual-level variables
Age
As continuous		0.92 [0.89–0.95]		0.92 [0.89–0.95]
Education level (ref: None)
Primary		0.39 [0.17–0.87]		0.50 [0.22–1.13]
Secondary		0.25 [0.09–0.65]		0.40 [0.15–1.09]
Tertiary		0.14 [0.01–1.61]		0.21 [0.02–2.29]
Marital status (ref: single)
Married/living with partner		0.22 [0.08–0.61]		0.21 [0.08–0.59]
Separated/widow/divorced		0.37 [0.11–1.21]		0.34 [0.10–1.16]
Employed in last 12 months (ref: Not employed)
Employed for cash		1.58 [0.89–2.80]		1.64 [0.93–2.90]
Employed but paid-in-kind		1.77 [0.83–3.76]		1.54 [0.68–3.51]
Health insurance coverage (ref: No)
Yes		1.37 [0.55–3.41]		1.36 [0.54–3.47]
Use of internet in last 12 months (ref: No)
Yes		0.28 [0.13–0.60]		0.32 [0.15–0.69]
Media exposure (ref. No)
Yes		0.68 [0.40–1.14]		0.85 [0.50–1.45]
Alcohol consumption (ref: No)
Yes		2.28 [0.90–5.80]		1.94 [0.76–4.97]
Community-level variables
Household wealth status (ref: lower class)
Middle class			0.40 [0.20–0.79]	0.32 [0.16–0.67]
Higher class			0.36 [0.18–0.73]	0.37 [0.17–0.80]
Distance to health facility (ref: not barrier)
Barrier			1.72 [0.83–3.56]	1.37 [0.65–2.90]
Geographical zone (ref: Central)
Coastal			0.32 [0.08–1.31]	0.32 [0.07–1.55]
Lake			0.44 [0.10–1.90]	0.45 [0.09–2.32]
Northern			0.61 [0.13–2.77]	0.70 [0.13–3.91]
Southern			0.51 [0.11–2.30]	0.41 [0.07–2.21]
Southern highlands			0.23 [0.06–0.98]	0.25 [0.05–1.27]
Western			1.13 [0.22–5.73]	1.47 [0.25–8.68]
Zanzibar			0.12 [0.03–0.50]	0.15 [0.03–0.71]
Residence (ref: Rural)
Urban			1.63 [0.93–2.86]	1.23 [0.66–2.31]
Random Effects Component
Community-level variance	1.35 [0.71–2.56]	1.57 [0.80–3.08]	1.09 [0.50–2.37]	1.41 [0.66–3.02]
Intraclass correlation	0.29 [0.18–0.44]	0.32 [0.20–0.48]	0.25 [0.13–0.42]	0.30 [0.17–0.48]
Median Odds Ratio (MOR)	3.03 [1.94–4.14]	3.30 [1.97–4.69]	2.71 [1.60–3.77]	3.10 [1.73–4.50]
Akaike’s information criterion	916.10	835.86	873.85	819.85
Log pseudolikelihood	-456.05	-403.93	-423.92	-384.92

The random effect model

The random effects model revealed substantial community-level variability in undiagnosed hypertension, with the variance ranging from 1.09 to 1.57 and an intraclass correlation coefficient (ICC) of 0.25 to 0.32, indicating that 25–32% of the variability in undiagnosed hypertension was attributable to differences between communities while the remaining 68–75% are due to individual variances within communities. The Median Odds Ratio (MOR) ranged from 2.71 to 3.30, highlighting that the odds of undiagnosed hypertension differed by a factor of up to 3.30 between randomly selected communities, signifying a consistent increase in odds of undiagnosed hypertension for women in higher-risk communities (i.e., a cluster with a larger positive EA-level random intercept) than lower-risk ones. The Akaike Information Criterion (AIC) and log pseudolikelihood values showed that the fourth model (AIC: 819.85, log pseudolikelihood: -384.92) indicate a better-fitting model and is used for the interpretation of fixed effect component.

Fixed effect model

The factors associated with undiagnosed hypertension are based on findings from the fixed-effects component of Model 4. This model is the most comprehensive model incorporating both individual- and community-level variables. The study identified significant factors associated with undiagnosed hypertension, reflecting individuals unaware of their hypertension status. Older age was protective, with younger individuals more likely to have undiagnosed hypertension (AOR: 0.92; 95% CI: 0.89–0.95). Being married or living with a partner significantly reduced the likelihood of undiagnosed hypertension (AOR: 0.21; 95% CI: 0.08–0.59), indicating the role of spousal or partner support in health awareness. Regular internet use was also protective, with users 68% less likely to have undiagnosed hypertension (AOR: 0.32; 95% CI: 0.15–0.69), emphasizing the role of digital access in improving health literacy. At the community level, individuals in the middle (AOR: 0.32; 95% CI: 0.16–0.67) and higher socioeconomic classes (AOR: 0.37; 95% CI: 0.17–0.80) were significantly less likely to have undiagnosed hypertension compared to those in the lower class. Regionally, living in Zanzibar was associated with a significantly lower likelihood of undiagnosed hypertension (AOR: 0.15; 95% CI: 0.03–0.71).

Discussion

This study aims to address significant gaps in hypertension diagnosis by quantifying the prevalence of undiagnosed hypertension among women of reproductive age with elevated BP in Tanzania. It also identifies individual- and community-level determinants and highlights regional disparities using a nationally representative data. The evidence generated is intended to inform targeted interventions and strengthen the health system strategies for hypertension screening and management in resource-limited settings.

Among women of reproductive age with elevated BP, 96% had blood pressure values above the optimal range with the majority falling into Grade 1 hypertension. This pattern is consistent with findings from other Sub-Saharan African countries (Kenya, Nigeria, Central African Republic), where a large burden of mild-to-moderate but often undiagnosed hypertension has been documented. This finding aligns with previous national and regional studies, including earlier rounds of the TDHS, which consistently report a large burden of undiagnosed hypertension among women [11, 12, 20]. While direct comparisons with male counterparts were beyond the scope of this study, the 2015–16 Tanzania DHS reported a slightly higher prevalence of hypertension among men, but similar challenges in diagnosis and treatment for both sexes [21]. Moreover, similar patterns have been observed in other SSA countries [22–24], where the prevalence of undiagnosed hypertension remains high across both sexes but disproportionately affects women in rural and low-income settings [25, 26]. This trend likely reflects systemic challenges within many SSA health systems, which often face constraints in providing routine NCD screening, especially at the primary care level [27, 28]. In light of these findings, there is a compelling need to scale up hypertension screening through integrated service delivery platforms. Incorporating blood pressure checks into existing health programs such as HIV clinics, antenatal care, and other chronic care services could improve early detection, particularly among women who are already accessing these services.

This study also assessed zonal and regional disparities in undiagnosed hypertension among women with elevated BP. The results reveal that despite their proximity to major cities with tertiary hospitals, women in regions like Dodoma and Arusha exhibit high proportions of undiagnosed hypertension. This likely due to persistent barriers in healthcare access and preventive service delivery in these regions. In addition, while tertiary hospitals focus primarily on curative care, the lack of systematic screening and hypertension management within primary healthcare systems may contributes to undiagnosed cases in surrounding communities [22]. Moreover, inequities in healthcare utilization mean that tertiary hospitals primarily serve urban populations, leaving peri-urban and rural communities underserved [29]. Additionally, rapid urbanization in these regions has also led to lifestyle changes, including increased sedentary behavior and poor dietary habits, exacerbating the hypertension burden while preventive measures remain insufficient. Conversely, regions with lower rates of undiagnosed hypertension likely benefit from stronger healthcare systems, well-organized community health initiatives, and targeted health education programs. Evidence from previous research suggests that systematic screening and integration of hypertension management into primary care are crucial for reducing the burden of undiagnosed hypertension [24].

The random effects analysis revealed substantial community-level variation in undiagnosed hypertension among women, with a notable proportion of the variability attributed to community-level differences. This finding aligns with previous studies in low- and middle-income countries, which have consistently highlighted the significant influence of community-level factors, such as socioeconomic disparities and healthcare access, on hypertension outcomes [30, 31]. The MOR further underscores the disparities across communities, indicating that women in higher-risk communities face considerably greater odds of undiagnosed hypertension compared to their counterparts in lower-risk settings. This finding aligns with previous research highlighting the influence of high-risk communities on hypertension outcomes among women [32, 33].

The fixed-effects analysis of Model 4 highlights several important factors associated with undiagnosed hypertension among women with elevated BP, both at the individual and community levels. Notably, older age was found to be protective against undiagnosed hypertension, with younger individuals being more likely to remain unaware of their hypertension status. This finding aligns with previous studies that have shown younger populations often have lower rates of hypertension awareness, potentially due to less frequent health checks and less health-seeking behavior [11, 34]. Also, younger individuals may remain undiagnosed due to lower perceived susceptibility to hypertension. They might not prioritize regular health screenings, contributing to lower awareness levels. However, this finding contrast with other previous studies which indicated that older ages are at risk of remained undiagnosed compared to younger ones. This might be due to different factors, such as being asymptomatic, attributing symptoms to aging, or having less access to healthcare facilities [35, 36]. The contrasting findings regarding age and undiagnosed hypertension across studies may be attributed to differences in population demographics, healthcare access, and study methodologies. Therefore, these findings highlight the need for context-specific strategies to enhance hypertension awareness across different age groups.

The study found that being married or living with a partner was associated with a significantly lower likelihood of undiagnosed hypertension. This highlights the importance of spousal or partner support in promoting health-seeking behaviors and enhancing health awareness. Research indicates that individuals in stable relationships often share health-related information, which can lead to greater awareness, increased participation in health screenings, and improved management of chronic conditions like hypertension, ultimately contributing to lower rates of undiagnosed hypertension [37, 38]. Previous studies have consistently demonstrated the protective role of being married or living with a partner in reducing the likelihood of undiagnosed hypertension [39, 40].

Use of internet and digital health has been used to empower individuals to access healthcare services at the point of care or remotely [41]. The current study found that regular internet use was a protective factor, with internet users being 68% less likely to have undiagnosed hypertension. This finding underscores the role of digital platforms in improving health literacy, particularly in rural or underserved areas. Similar studies have shown that access to online health information can significantly influence individuals’ health knowledge and prompt them to seek medical advice [42]. A study in China found that internet access was positively associated with greater health awareness, particularly among women [43]. This indicate availability of health information through online resources, social media, and mobile health applications could be key drivers in reducing the burden of undiagnosed hypertension in the digital era [44, 45].

At the community level, this study found that individuals in the middle and higher socioeconomic classes were significantly less likely to have undiagnosed hypertension compared to those in the lower class. This is consistent with existing literature that shows individuals from wealthier households are more likely to have access to healthcare services, including hypertension screening and management [46, 47]. For instance, research in Bangladesh has demonstrated that poor socioeconomic status is significantly associated with untreated hypertension and undiagnosed cases of the condition [48]. This observed finding from previous and current study may be attributed to limited access to hypertension-related information, inadequate availability of healthcare facilities, and low health insurance coverage.

Regionally, the study identified Zanzibar as an area with a significantly lower likelihood of undiagnosed hypertension. Zanzibar’s lower levels of undiagnosed hypertension can be attributed to several possible factors, including effective local health policies, targeted public health interventions, and a relatively efficient healthcare system. The region may benefit from better accessibility to healthcare services, regular screening programs, and community-based health awareness campaigns that emphasize hypertension detection and management. Additionally, Zanzibar’s smaller population and unique governance structure could allow for more tailored and responsive health initiatives compared to mainland Tanzania.

This study has several strengths. It is the first nationally representative analysis to quantify the prevalence and determinants of undiagnosed hypertension among women of reproductive age with elevated BP in Tanzania, providing robust and generalizable findings. The use of data from the 2022 TDHS-MIS ensures high-quality, population-based evidence with standardized methodologies. Additionally, the multilevel analytical approach captures both individual and community-level factors, offering comprehensive insights into the drivers of undiagnosed hypertension. Furthermore, the identification of regional disparities highlights critical areas for targeted interventions and resource allocation. However, the study also has some limitations. Being a cross-sectional analysis, it cannot establish causal relationships between the identified determinants and undiagnosed hypertension. The reliance on self-reported data for some variables, such as prior diagnosis of hypertension, introduces potential recall and reporting biases. As this study is based on a secondary analysis of the 2022 TDHS-MIS data, whose sampling framework was not specifically powered to detect regional or zonal differences in hypertension prevalence. As such, the number of hypertensive women varied across geographical zones, and some subgroup estimates may have wide confidence intervals due to small sample sizes. This limits the precision and generalizability of regional comparisons. The forest plots presented are descriptive in nature and reflect variability after accounting for community (cluster) random effects, but not higher-level random effects such as regions or zones, which were not modeled due to the TDHS-MIS sampling design. Therefore, findings related to regional inequalities in hypertension diagnosis should be interpreted with caution. Despite these limitations, the findings provide valuable evidence to inform public health strategies and policy development. Additionally, some regions had relatively small sample sizes, and although sampling weights were applied, regional estimates may remain unstable and should be interpreted with caution. Last, the blood pressure distribution observed in this study reflects methodological limitations of DHS sampling and measurement procedures and should not be interpreted as representative of true community-level hypertension prevalence or severity.

Conclusions

Overall, the findings from this study emphasize the multifaceted nature of undiagnosed hypertension, highlighting the importance of both individual and community-level factors. These results reinforce the need for targeted interventions that focus on enhancing health awareness, improving access to healthcare services, and utilizing digital health platforms, particularly for younger populations and those from lower socioeconomic backgrounds who were disproportionately affected in this study. Moreover, the regional differences observed in this study point to the potential impact of localized health policies and regional healthcare infrastructures on reducing burden of undiagnosed hypertension disparities.

Abbreviations

AOR

Adjusted Odds Ratio

CI

Confidence Interval

CVDs

Cardiovascular diseases

DHS

Demographic and Health Survey

ICF

Inner City Fund (referring to ICF International, which provides technical support)

IQR

Interquartile Range

MOR

Median Odds Ratio

NCDs

Non-communicable Diseases

NBS

National Bureau of Statistics

OCGS

Office of the Chief Government Statistician Zanzibar

SSA

Sub-Saharan Africa

TDHS

MIS-Tanzania Demographic and Health Survey and Malaria Indicator Survey

TFNC

Tanzania Food and Nutrition Centre

WHO

World Health Organization

Authors’ contributions

D.B. conceptualized the study, designed the methodology, led the primary data analysis and drafted manuscript. C.C.K, E.S, and E.K contributed to data interpretation, and drafted the manuscript. Both authors critically revised the manuscript for intellectual content and approved the final version for submission.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

The data supporting this study’s findings were obtained from the 2022 Tanzania Demographic and Health Survey (TDHS-MIS), accessible through the DHS Program. The dataset is publicly available for research purposes upon request from the DHS Program website (https://dhsprogram.com/data/).

Declarations

Ethics approval and consent to participate

This study is a secondary analysis of data from the 2022 TDHS-MIS, which was conducted by the Tanzania National Bureau of Statistics (NBS) in collaboration with the Office of the Chief Government Statistician (OCGS) Zanzibar, with technical support from ICF International. Ethical approval for the original survey was obtained from the Tanzania National Institute for Medical Research (NIMR), the Zanzibar Health Research Institute (ZAHRI), and the ICF Institutional Review Board (IRB). However, the specific Ethics Approval Number for the TDHS-MIS is not publicly available in the dataset documentation. The TDHS-MIS adhered to strict ethical guidelines to ensure participant confidentiality and privacy. Informed consent was obtained from all participants before data collection, and no personally identifiable information was included in the publicly available dataset. The dataset used in this study was accessed through the DHS Program after obtaining formal permission.

Since this study involved the secondary analysis of anonymized data, no additional ethical approval was required from the principal investigator’s institution. The study was conducted in accordance with the principles of the Declaration of Helsinki, ensuring the ethical protection of participants’ rights and data integrity.

Consent for publication

This study uses de-identified, publicly available data from the 2022 TDHS-MIS, obtained with permission from the DHS Program. Therefore, no additional consent was required, as all results are presented in aggregate to ensure confidentiality.

Competing interests

The authors declare no competing interests.