Prolonged breastfeeding and hypertension risk among Kenyan women: exploring the roles of obesity and alcohol consumption

Abstract

Background

Breastfeeding is known to have positive effects on child health outcomes. However, little is known about its effects on physical health outcomes for mothers. Given that hypertension remains a significant public health concern in Kenya, it is imperative to understand whether breastfeeding has some protective effects for mothers. This study investigates the association between breastfeeding duration and hypertension risk while considering the moderating effects of overweight/obesity and alcohol consumption.

Methods

We relied on 2022 Kenya demographic and health survey data. A total sample of 5314 was analyzed. Prevalence rates of hypertension were assessed across different breastfeeding durations, and both unadjusted and adjusted odds ratios were calculated to evaluate the risk of hypertension. Interaction effects between breastfeeding duration, overweight/obesity status, and alcohol consumption were also analyzed with predictive margins.

Results

Overall, the prevalence of hypertension in the sample was 8.7% (95% CI: 7.7–9.9). After adjusting for confounders, women who breastfed for 12–24 months (AOR = 0.60, 95%CI: 0.43–0.84) and for more than 24 months (AOR = 0.60, 95%CI: 0.45–0.80) were at lower odds of being hypertensive. Despite the increased risk of hypertension among overweight or obese women and those who consumed alcohol, the predictive margins did not show a significant moderating effect of these factors on the association between breastfeeding duration and hypertension in this study.

Conclusion

Our study suggests that prolonged breastfeeding may have a protective effect against hypertension among parous women. This protective effect of longer breastfeeding durations remains relatively stable across different obesity and alcohol consumption statuses. It is essential for the Kenyan Ministry of Health to intensify its breastfeeding awareness. Health providers at antenatal and postnatal care centers must emphasize the protective effects of prolonged breastfeeding on physical health outcomes.

Background

Hypertension is a leading cause of cardiovascular diseases and renal failure worldwide [1, 2]. Data from a 2019 Global Burden of Disease study [1] projects that mortalities attributable to hypertensive heart disease will increase from 1.16 million in 2019 to 1.57 million by 2034. Despite hypertension being a growing global public health concern, low-and-middle-income countries (LMICs) such as Kenya bear two-thirds of the world’s prevalence of high blood pressure [3]. Available evidence shows that more than a quarter of the adult population in Kenya (25.8%) live with hypertension [4]. This high burden of hypertension in Kenya is attributable to the point that only 15.9% are aware of their status [5].

The existing body of literature on the epidemiology of hypertension has identified socio-economic, demographic, contextual, lifestyle and dietary factors that exacerbate one’s risk. Lifestyle and dietary factors such as low physical activity, alcohol consumption, tobacco use, low fruit and vegetable consumption, high screen time/internet use, and sedentary occupations are some of the known risk factors of hypertension [6–8]. Also, demographic factors such as level of education, ageing, and marital status have been reported to be significantly associated with hypertension risk [7]. Furthermore, wealth status and exposure to intimate partner violence constitute some socioeconomic factors that influence hypertension status [1, 9]. Contextual factors like urban residency and use of unclean cooking fuel used have also been identified as high-risk factors for hypertension [10, 11]. This suggests that hypertension could be prevented by capitalizing on protective factors.

In recent times, the role of breastfeeding in shaping long-term cardiovascular health has garnered increasing attention [12–15]. For instance, a meta-analysis of previous literature showed that “the pooled hazard ratio of hypertension was 1.34 (95%CI: 1.17–1.52) for women who did not breastfeed compared with women who breastfed for more than 12 months for their first child” [12]. Another study conducted among 3,119 postmenopausal women in Korea revealed that a longer duration of breastfeeding significantly reduces the risk of hypertension [13]. Additionally, Kirkegaard et al. [16] also report that breastfeeding for four months or more significantly reduced the risk of hypertension by 20–30%.

Despite the prior evidence [12–16] establishing the association between breastfeeding duration and hypertension risk, none of these studies were conducted in sub-Saharan Africa (SSA) which suffers a higher burden of hypertension and a high rate of undiagnosed hypertension [3, 4]. Furthermore, it is well-documented that alcohol consumption and obesity are strong risk factors for hypertension [17–20]. The accumulation of excess body fat, particularly central obesity, leads to increased sympathetic nervous system activity, insulin resistance, and inflammatory responses which accelerates hypertension risk [21, 22]. Similarly, alcohol consumption is known to exacerbate oxidative stress which is a well-known factor that increases likelihood of high-blood pressure [23, 24]. Given that alcohol consumption, overweight/obesity and hypertension share common metabolic pathways, we hypothesize that these factors might modify or attenuate the protective effect of breastfeeding on hypertension. Currently, there exists a dearth of knowledge regarding these nuanced associations in Kenya or other SSA countries. As such, this study investigates the association between breastfeeding duration and hypertension risk while considering the moderating effects of overweight/obesity and alcohol consumption.

Methods

Design and data source

Data for this study was obtained from the individual recode file of the 2022 Kenyan Demographic and Health Survey (KDHS), which marks the seventh iteration of a nationally representative survey conducted in the country since 1989. Conducted as part of the DHS-8 series, the 2022 KDHS not only extended existing inquiries but also introduced new modules focusing on early childhood development and chronic diseases [8, 9]. The survey was carried out under the auspices of the Kenya National Bureau of Statistics (KNBS) in collaboration with the Ministry of Health (MoH) and several other stakeholders. The sample for the 2022 KDHS was drawn from the Kenya Household Master Sample Frame (K-HMSF), which KNBS utilizes for conducting household-based sample surveys across the country [9].

The 2022 KDHS utilized a two-stage stratified sampling design. In the first stage, 1,692 clusters were selected from the Kenya Household Master Sample Frame (K-HMSF) using the Equal Probability Selection Method (EPSEM), with clusters chosen independently within each sampling stratum [25]. Following this, household listings were conducted in all selected clusters to create a comprehensive list of households for the second stage of selection, where 25 households were to be randomly chosen from each cluster. However, it was found that some clusters had fewer than 25 households; as a result, all households in those clusters were included in the sample [8, 25]. This adjustment led to a final total of 42,022 households being sampled for the 2022 KDHS. For more details on the design and sampling methodology, please refer to the following link: https://www.dhsprogram.com/pubs/pdf/FR380/FR380bis.pdf. In this study, we analyzed data on a weighted sample of 5,314 parous women.

Measures

Outcome

The outcome variable was self-reported hypertension status which was derived from the question, ‘Have ever been told by a health provider that you have high blood pressure?’. The binary response was either yes or no. All who responded ‘yes’ were considered to be living with hypertension.

Exposure

Breastfeeding duration constituted the exposure variable, which was measured in months. Respondents who did not breastfeed were excluded from the analysis to avoid unnecessary noise in our estimation. We classified the breastfeeding duration into four categories: breastfeeding < 6 months, between 6 and 12 months, between 12 and 24 months, and > 24 months.

Confounders

We selected factors that have been be reported in previous literature to independently predict duration of breastfeeding and/or hypertension risk as confounders. These factors included age, place of residence, educational level, wealth index, exposure to media, alcohol consumption and overweight/obesity status. Overweight/obesity status was computed using the body mass index (BMI). We computed the BMI for women by dividing their weight (kg) with the square of their height in meters. A BMI of < 24.99 was coded ‘not overweight/obese’ while ≥ 25 was coded as ‘overweight/obese’.

Statistical analysis

Descriptive statistics were utilized to summarize the characteristics of the study sample. This included frequencies and percentages for categorical variables. To assess the association between breastfeeding duration and hypertension risk, both unadjusted and adjusted odds ratios were computed using binary logistic regression models. The unadjusted models provided initial insights into the association between breastfeeding duration and hypertension, while the adjusted models accounted for potential confounders (i.e., age, place of residence, educational level, wealth index, exposure to media, alcohol consumption and overweight/obesity status). A backward stepwise approach was employed to iteratively exclude statistically non-significant variables, ensuring a parsimonious final model. The backward stepwise procedure was employed as a variable reduction strategy to improve model parsimony rather than as a causal selection approach. Consequently, estimates from the final model should be interpreted as associative rather than causal.

The interaction effects between breastfeeding duration and both obesity and alcohol consumption were also examined (Fig. 1). Predictive margins were calculated to illustrate the predicted probabilities of hypertension across varying breastfeeding durations while considering obesity and alcohol consumption statuses. This approach provided a clearer interpretation of the potential impact of breastfeeding duration on hypertension risk within different subgroups. It must be noted that all estimates were weighted using the appropriate sample weight (v005), as well as considering the complex design of the survey (accounting for the primary sample unit and the stratifications). All analyses were performed in STATA version 18 (StataCorp, College Station, TX, USA). We also checked for multicollinearity using the variance inflation factor.

Fig. 1.

Diagram illustrates the associations and pathways between breastfeeding duration, hypertension, confounders, and interaction terms

Results

Characteristics of study participants

The study included a total weighted sample of 5,314 participants. The majority of participants (39.6%) breastfed for more than 24 months, while 21.3% breastfed for 12–24 months, and 18.2% breastfed for less than 6 months. Most participants were under 25 years (33.4%), resided in rural areas (63.5%), had secondary education (36.7%), were exposed to media (75.6%), and belonged to the richer wealth index (21.7%). Only 2.5% of the sampled population consumed alcohol. In terms of obesity, 36.4% of participants were classified as overweight or obese (Table 1).

Table 1.

Prevalence of self-reported hypertension by participants’ characteristics

Characteristics	Weighted sample n (%)	Proportion hypertensive n (% [95%CI])	p-values
Duration of breastfeeding			0.047
< 6 months	966 (18.2)	96 (9.9 [7.8–12.4])
6–12 months	1109 (20.9)	111 (10.0 [7.9–12.6])
12–24 months	1136 (21.3)	68 (6.0 [4.6–7.9])
> 24 months	2107 (39.6)	189 (9.0 [7.1–11.3])
Age			< 0.001
< 25 years	1773 (33.4)	127 (7.1 [5.6–9.1])
25–29 years	1573 (29.6)	104 (6.6 [4.9–8.9])
30–34 years	1057 (19.9)	113 (10.7 [8.1–14.0])
35–39 years	658 (12.4)	76 (11.5 [9.0–14.6.0.6])
40–44 years	203 (3.8)	41 (20.1 [13.4–28.9])
45–49 years	49 (0.9)	5 (9.6 [3.7–22.7])
Residence			0.014
Urban	1938 (36.5)	206 (10.6 [8.5–13.2])
Rural	3376 (63.5)	259 (7.7 [6.7–8.8])
Education			0.019
No formal education	494 (9.3)	22 (4.5 [2.9–6.9])
Primary	1908 (35.9)	185 (9.7 [7.9–11.8])
Secondary	1951 (36.7)	155 (7.9 [6.4–9.7])
Higher	961 (18.1)	102 (10.7 [7.9–14.3])
Wealth Index			< 0.001
Poorest	1129 (21.2)	65 (5.7 [4.3–7.6])
Poorer	971 (18.3)	66 (6.8 [5.2–8.9])
Middle	952 (17.9)	85 (9.0 [7.0–11.4.0.4])
Richer	1155 (21.7)	114 (9.9 [7.6–12.9])
Richest	1106 (20.8)	134 (12.1 [9.3–15.6])
Exposed to media			< 0.001
No	1296 (24.4)	72 (5.6 [4.2–7.3])
Yes	4018 (75.6)	392 (9.8 [8.5–11.2])
Consumes alcohol			0.216
No	5180 (97.5)	448 (8.7 [7.6–9.8])
Yes	132 (2.5)	17 (12.4 [7.1–20.8])
Obesity status			< 0.001
Not overweight/obese	3380 (63.6)	210 (6.2 [532 − 7.3])
Overweight/obese	1934 (36.4)	255 (13.2 [11.0–15.8.0.8])
Overall	5314 (100.0)	465 (8.7 [7.7–9.9])

Prevalence of self-reported hypertension among parous women in Kenya

Overall, the prevalence of self-reported hypertension in the sample was 8.7% (95% CI: 7.7–9.9) (Table 1). This prevalence was high among women who breastfed for less than 6 months (9.9% [95% CI: 7.8–12.4]). Also, the prevalence of self-reported hypertension was significantly high among those aged 40–44 years (20.1%), urban residents (10.6%), those with higher education (10.7%), and women in the richest wealth index (12.1%). Women exposed to media had a higher prevalence of self-reported hypertension (9.8%) compared to those without media exposure. There was no significant difference between alcohol consumption and self-reported hypertension prevalence in this study, although the proportion was high among those who consumed alcohol (12.4%); however, overweight and obese participants had a significantly higher prevalence (13.2%) compared to those with normal weight (6.2%).

Association between duration of breastfeeding and self-reported hypertension risk

Table 2 shows the association between duration of breastfeeding and self-reported hypertension risk. In the unadjusted model, breastfeeding for 12–24 months and more than 24 months was associated with a significantly reduced risk of self-reported hypertension, with crude odds ratios of 0.70 (95%CI: 0.50–0.96, p < 0.05) and 0.71 (95%CI: 0.54–0.94, p < 0.05), respectively, compared to breastfeeding for less than 6 months. After adjusting for confounding variables, the risk reduction became more pronounced. Women who breastfed for 12–24 months (AOR = 0.60, 95%CI: 0.43–0.84) and for more than 24 months (AOR = 0.60, 95%CI: 0.45–0.80) remained at lower odds of being hypertensive. Additional factors influencing self-reported hypertension risk included age with participants aged 40–44 years having more than three times the odds of self-reported hypertension (AOR = 3.38, 95%CI: 2.19–5.24) compared to those under 25 years. Higher odds of self-reported hypertension were found among women with higher educational level (AOR = 2.49, 95%CI: 1.54–4.02), those exposed to media (AOR = 1.41, 95%CI: 1.06–1.88), those who consumed alcohol (AOR = 1.85, 95%CI: 1.05–3.27), and those overweight/obese (AOR = 1.78, 95%CI: 1.42–2.23). The final model showed improved model fitness with a lower AIC (2659.995) and BIC (2758.667), indicating a better fit compared to the unadjusted model.

Table 2.

Association between duration of breastfeeding and self-reported hypertension risk

Variables	Unadjusted model (Crude odds ratio; 95%CI)	Adjusted model (Adjusted odds ratio; 95%CI)
Duration of breastfeeding
< 6 months	Ref.	Ref.
6–12 months	0.94 [0.69–1.27]	0.90 [0.66–1.22]
12–24 months	0.70 [0.50–0.96]*	0.60 [0.43–0.84]**
> 24 months	0.71 [0.54–0.94]*	0.60 [0.45–0.80]***
Age
< 25 years		Ref.
25–29 years		1.04 [0.77–1.41]
30–34 years		1.75 [1.28–2.38]***
35–39 years		2.19 [1.57–3.05]***
40–44 years		3.38 [2.19–5.24]***
45–49 years		2.80 [1.06–7.39]*
Education
No formal education		Ref.
Primary		2.54 [1.68–3.83]***
Secondary		2.45 [1.58 [3.81]***
Higher		2.49 [1.54–4.02]***
Exposed to media
No		Ref.
Yes		1.41 [1.06–1.88]*
Consumes alcohol
No		Ref.
Yes		1.85 [1.05–3.27]*
Obesity status
Not overweight/obese		Ref.
Overweight/obese		1.78 [1.42–2.23]***
Model fitness
AIC	2792.114	2659.995
BIC	2818.427	2758.667
Prob > chi2	0.029	< 0.001

***p < 0.001, **p < 0.01, *p < 0.05; Ref: reference category; NB: Backward stepwise approach was used to iteratively exclude statistically non-significant variables

Interaction effect of obesity status

The analysis explored the interaction effects between breastfeeding duration and both obesity and alcohol consumption on the risk of hypertension. Although being overweight or obese significantly increased the risk of hypertension (AOR = 2.25, 95% CI: 1.42–3.55), the interaction terms between breastfeeding duration and obesity status were not statistically significant (Table 3). For example, the interaction effect between breastfeeding for 6–12 months and obesity yielded an odds ratio of 0.69 (95% CI: 0.37–1.29), while the interaction for breastfeeding over 24 months had an odds ratio of 0.73 (95% CI: 0.41–1.29). Alcohol consumption, although associated with a higher risk of hypertension, did not show a statistically significant interaction with breastfeeding duration.

Table 3.

Interaction effect of overweight/obesity and alcohol consumption on the association between breastfeeding duration and hypertension

Variable	Adjusted Odds ratio (95%CI)	p-values
Interaction effect of overweight/obesity a
Breastfeeding Duration
< 6 months	Ref.
6–12 months	1.07 [0.70–1.64]	0.761
12–24 months	0.66 [0.41–1.05]	0.076
> 24 months	0.70 [0.47–1.05]	0.082
Obesity status
Not overweight/obese	Ref.
Overweight/obese	2.25 [1.42–3.55]	0.001
Obesity status
Interaction Terms (Breastfeeding Duration#Obesity status)
< 6 months	Ref.
6–12 months # Overweight/obese	0.69 [0.37–1.29]	0.246
12–24 months # Overweight/obese	0.84 [0.43–1.64]	0.613
> 24 months # Overweight/obese	0.73 [0.41–1.29]	0.281
Interaction effect of alcohol consumption b
Breastfeeding Duration
< 6 months	Ref.
6–12 months	0.90 [0.65–1.23]	0.487
12–24 months	0.62 [0.44–0.86]	0.005
> 24 months	0.62 [0.46–0.83]	0.001
Alcohol Consumption
No	Ref.
Yes	3.52 [0.87–14.29]	0.079
Interaction Terms (Breastfeeding Duration#Alcohol Consumption)
< 6 months	Ref.
6–12 months # Consume Alcohol	0.90 [0.15–5.36]	0.910
12–24 months # Consume Alcohol	0.45 [0.06–3.57]	0.448
> 24 months # Consume Alcohol	0.35 [0.07–1.85]	0.219

^aAdjusted for: age, education, media exposure, alcohol consumption, wealth, residence

^bAdjusted for: age, education, media exposure, overweight/obesity, wealth, residence

Predictive margins of overweight/obesity and alcohol consumption

In Fig. 2, the predictive margins indicated that longer durations of breastfeeding were protective factors against self-reported hypertension, regardless of obesity status or alcohol consumption. Women who breastfed for 12–24 months or over 24 months consistently had lower predicted probabilities of self-reported hypertension, compared to those who breastfed for less than six months (see Fig. 2). Despite the increased risk of self-reported hypertension among overweight or obese women and those who consumed alcohol, the predictive margins did not show a significant moderating effect of these factors on the association between breastfeeding duration and self-reported hypertension (see Fig. 3). This suggests that the protective effect of longer breastfeeding durations remained relatively stable across different obesity and alcohol consumption statuses.

Fig. 2.

Predictive margins of overweight/obesity in the association between breastfeeding duration and self-reported hypertension

Fig. 3.

Predictive margins of alcohol consumption in the association between breastfeeding duration and self-reported hypertension

Discussion

This study sought to investigate the association between breastfeeding duration and self-reported hypertension risk while considering the moderating effects of overweight/obesity and alcohol consumption. We observed from the analysis that prolonged duration of breastfeeding significantly reduced the odds of self-reported hypertensive by 40%. This protective effect of prolonged duration of breastfeeding on self-reported hypertension persisted even after controlling for key confounders including overweight/obesity status, alcohol consumption, age and educational level. The established association is consistent with Kirkegaard et al. [16] whose study revealed that prolonged breastfeeding reduces hypertension risk by 20–30%. It also aligns with the findings of Park and Choi [13] that found a significantly inverse association between prolonged breastfeeding duration and hypertension risk in Korea. From a biological standpoint, there is evidence supporting that lactation significantly increases the release of oxytocin, a hormone that has vasodilatory properties [26]. This implies that prolonged breastfeeding would equate to more release of oxytocin which has been documented to facilitate the relaxation of blood vessels, thereby reducing blood pressure [27, 28].

Despite the strong theoretical rationale for expecting an interaction between obesity and breastfeeding duration in relation to self-reported hypertension risk, our results did not support this hypothesis. Women who were overweight/obese had 78% increased odds of being hypertensive. Similar findings have been reported in Kenya [5] and Ghana [29]. However, the interaction terms between breastfeeding duration and obesity were not statistically significant. This suggests that the protective effect of prolonged breastfeeding on self-reported hypertension was not moderated by obesity in this population. This unexpected non-significant moderating role of overweight/obesity on the association between prolonged breastfeeding duration and hypertension corroborates Park and Choi’s study [12] that found similar associations. Our findings imply that breastfeeding confer protection against hypertension through mechanisms that are not significantly influenced by obesity-related metabolic changes. Future studies can consider adopting a randomized control trial or longitudinal design to further interrogate why overweight/obesity has no significant moderation effect on breastfeeding duration and hypertension.

Our findings also indicate that women who consumed alcohol were at an 85% higher risk of self-reported hypertension. Evidence from the predictive margins also show that consistently, the risk of self-reported hypertension was higher among those who consumed alcohol compared to those who had never consumed alcohol. This result is consistent with a previous study conducted in the Kiambu County, Kenya which found that individuals who consumed alcohol were thrice more likely to develop hypertension than those who did not consume it [6]. Anto et al. [30] also found similar positive association between alcohol consumption and hypertension. Interestingly, we found no significant moderation effect of alcohol consumption on the association between prolonged breastfeeding duration and hypertension. Perhaps, the interaction between breastfeeding and alcohol might be influenced by other unmeasured factors such as diet, physical activity, or genetic predispositions, which could confound the relationship and obscure any potential interaction. This can also be explained by the low proportion of women who consumed alcohol – the statistical power may have been reduced.

Implications for policy and practice

Based on the findings, policy makers in Kenya must consider the integration breastfeeding support into national health strategies aimed at hypertension prevention. There is a need to improve access to lactation support services, providing educational resources on the benefits of breastfeeding, and creating supportive environments that encourage breastfeeding in both healthcare settings and the community. In clinical practice, healthcare providers must emphasize the protect effects of prolonged breastfeeding in hypertension risk reduction as part of their antenatal and postnatal care health messaging. While the present findings provide population-level evidence of an inverse association between prolonged breastfeeding and hypertension among parous women, further research is required to confirm these associations using longitudinal designs. Prospective cohort studies that follow women from the postpartum period into later adulthood would be particularly valuable for establishing temporal ordering, reducing recall bias, and strengthening causal inference regarding the long-term cardiovascular effects of breastfeeding.

Strengths and limitations

The multi-stage sampling technique employed in the KDHS, coupled with the application of sample weights guarantees the generalizability of the findings to the larger population of parous women in Kenya. Also, this study may be the first of its kind to consider the moderating effects of overweight/obesity and alcohol consumption. Thus, contributing substantially to what is already known about breastfeeding, and hypertension risk factors. Yet, there exists some limitations. This includes that causal inferences cannot be established since the KDHS is based on cross-sectional study designs. Residual confounding may also be present due to unmeasured lifestyle and biological factors such as dietary intake, physical activity, lipid profiles, and metabolic markers, which were not available in the dataset but may influence both breastfeeding practices and hypertension risk. Another key limitation of this study is the reliance on self-reported hypertension, which captures diagnosed rather than true disease prevalence. Given that a substantial proportion of hypertensive individuals in Kenya remain unaware of their condition, the prevalence reported in this study is likely underestimated. This underdiagnosis may have introduced misclassification bias, potentially attenuating the observed associations. Additionally, the low prevalence of alcohol consumption in the study population limited the ability to robustly assess interaction effects between breastfeeding duration and alcohol use. As such, the absence of statistically significant moderation should not be interpreted as definitive evidence of no effect.

Conclusion

Our study suggests that prolonged breastfeeding may have a protective effect against hypertension among parous women. This protective effect of longer breastfeeding durations remains relatively stable across different obesity and alcohol consumption statuses. It is imperative for the Kenyan Ministry of Health to intensify its breastfeeding awareness. Health providers at antenatal and postnatal care centers must emphasize the protective effects of prolonged breastfeeding on physical health outcomes.

Abbreviations

AOR

Adjusted Odds Ratio

AIC

Akaike Information Criterion

COR

Crude Odds Ratio

CI

Confidence Interval

KNBS

Kenya National Bureau of Statistics

K-HMSF

Kenya Household Master Sample Frame

LMICs

Low-and-middle-income Countries

MoH

Ministry of Health

Ref

Reference Category

Author contributions

JO conceived and designed the study. JO and CA contributed to the design of the analysis. JO performed the formal analysis and provided methodological insights. JO, CA, KSD, SO and PAD drafted the initial manuscript. KSD and PAD supervised the research. All authors read, revised and approved the final manuscript for submission. JO had the responsibility of submitting the manuscript.

Funding

We had no funding.

Data availability

The datasets generated and/or analysed during the current study are available in the Measure DHS repository: [http://dhsprogram.com/data/available-datasets.cfm](http:/dhsprogram.com/data/available-datasets.cfm).

Declarations

Ethics approval and consent to participate

We did not need to seek ethical clearance because the KDHS dataset we used is publicly available. We obtained the datasets from the KDHS Program after completing the necessary registration and getting approval for their use. We followed all the ethical guidelines that pertain to using secondary datasets in research publications. Details of KDHS data and ethical standards followed can be found here: http://goo.gl/ny8T6X.

Consent for publication

None declared.

Competing interests

The authors declare no competing interests.