Skip to main content
NCBI home page
PLOS Medicine logoLink to PLOS Medicine
. 2025 Jan 31;22(1):e1004516. doi: 10.1371/journal.pmed.1004516

Long-term drought and risk of infant mortality in Africa: A cross-sectional study

Pin Wang 1,2,3,*, Tormod Rogne 4,5, Joshua L Warren 5,6,7, Ernest O Asare 7,8, Robert A Akum 9, N’datchoh E Toure 10, Joseph S Ross 11,12,13, Kai Chen 1,2
Editor: Yuming Guo14
PMCID: PMC11785314  PMID: 39888958

Abstract

Background

As extreme events such as drought and flood are projected to increase in frequency and intensity under climate change, there is still large missing evidence on how drought exposure potentially impacts mortality among young children. This study aimed to investigate the association between drought and risk of infant mortality in Africa, a region highly vulnerable to climate change that bears the heaviest share of the global burden.

Methods and findings

In this cross-sectional study, we obtained data on infant mortality in 34 African countries during 1992–2019 from the Demographic and Health Surveys program. We measured drought by the standardized precipitation evapotranspiration index at a timescale of 24 months and a spatial resolution of 10 × 10 km, which was further dichotomized into mild and severe drought. The association between drought exposure and infant mortality risk was estimated using Cox regression models allowing time-dependent covariates. We further examined whether the association varied for neonatal and post-neonatal mortality and whether there was a delayed association with drought exposure during pregnancy or infancy.

The mean (standard deviation) number of months in which children experienced any drought during pregnancy and survival period (from birth through death before 1 year of age) was 4.6 (5.2) and 7.3 (7.4) among cases and non-cases, respectively. Compared to children who did not experience drought, we did not find evidence that any drought exposure was associated with an increased risk of infant mortality (hazard ratio [HR]: 1.02, 95% confidence interval [CI] [1.00, 1.04], p = 0.072). When stratified by drought severity, we found a statistically significant association with severe drought (HR: 1.04; 95% CI [1.01, 1.07], p = 0.015), but no significant association with mild drought (HR: 1.01; 95% CI [0.99, 1.03], p = 0.353), compared to non-exposure to any drought. However, when excluding drought exposure during pregnancy, the association with severe drought was found to be non-significant. In addition, an increased risk of neonatal mortality was associated with severe drought (HR: 1.05; 95% CI [1.01, 1.10], p = 0.019), but not with mild drought (HR: 0.99; 95% CI [0.96, 1.02], p = 0.657).

Conclusions

Exposure to long-term severe drought was associated with increased infant mortality risk in Africa. Our findings urge more effective adaptation measures and alleviation strategies against the adverse impact of drought on child health.

Author summary

Why was this done?

  • Climate change is projected to result in a higher frequency and intensity of drought events.

  • Previous studies have found rainfall anomaly to be associated with increased infant mortality.

  • The risk of drought exposure for infant health outcomes is not well understood.

What did the researchers do and find?

  • Employing an advanced measure of drought incorporating multiple climatic variables, we found no evidence that any drought exposure was associated with an increased risk of infant mortality in Africa.

  • When stratifying the analysis by drought severity, exposure to long-term severe drought was observed to be associated with higher infant mortality risk.

  • There is a delayed association between exposure to drought during pregnancy and risk of infant mortality.

What do these findings mean?

  • The protection of pregnant women and young children from drought could potentially help reduce the heavy toll of infant mortality in Africa.

  • Our study urges the prioritization of neonatal and pediatric care as well as the development of water, sanitation, and hygiene infrastructure and adaptation strategies in rural Africa.

  • This study may underestimate the potential impact of drought on mortality in early life given it could have already contributed to miscarriage or stillbirth.

Introduction

Infant mortality rate (IMR), defined as the number of deaths among children younger than 1 year per 1,000 live births, is an important measure of overall population health, with global rates decreasing from 64.6 deaths per 1,000 live birth in 1990 to 28.4 deaths per 1,000 in 2021 [1]. Despite this trend towards improvement, the world is still struggling to achieve the Sustainable Development Goal 3.2 on reducing under-five child mortality [2], of which infant mortality is the largest contributor [3].

Environmental exposures such as polluted air [4], unsafe water [5], and unsatisfactory sanitation [6] have been associated with an increased risk of infant mortality. The risks associated with drought are less well understood. Drought is a complex environmental phenomenon influenced by multiple climatological parameters, with low precipitation as the main driver [7]. Worldwide, climate change is projected to increase the occurrences of climate extremes, particularly in Africa, identified as one of the regions highly vulnerable to these changes due to high exposure and low adaptive capacity [8], including projections of increased frequency, heightened intensity, and prolonged duration of drought hazards in many parts of the world [9]. Drought affects human health through many interacting pathways [10,11]. In particular, studies have observed drought to be associated with water-borne [12] and vector-borne [13] diseases, cardiovascular [14] and respiratory [15] diseases, malnutrition [16], mental disorders [17], and mortality [18–23]. To our knowledge, only one previous study in Ethiopia looked into the quantitative impact of drought on under-five child mortality, with no significant result reported [24]. Furthermore, there is a paucity of evidence on the association between drought and infant mortality, although two studies from China and Brazil suggested such a link, in which drought exposure was characterized by rainfall fluctuations during the 12 months before birth [25,26]. Despite the lack of consensus on drought definition [27], there has been a growing recognition of the significance of incorporating temperature and other meteorological information, simultaneously taking water supply and demand into account when computing drought indices [7]. However, previous studies only included precipitation in drought characterization [25,26] and evidence on how such a more rigorously defined drought measure is associated with risk of infant deaths is missing.

In addition, over half of infant mortality is attributed to deaths during the first 27 days after birth [3]. However, it remains unclear whether exposure to drought impacts risk of neonatal and post-neonatal mortality differently. Furthermore, the indirect affecting pathway of drought on the infant through the mother during pregnancy may delay its association with subsequent deaths during infancy. However, the evidence on the potentially lagged association between drought and infant mortality is lacking.

To date, there is a scarcity of research investigating the relationship between drought and infant mortality in Africa, a region that bears the greatest burden of infant deaths worldwide [1] and where some profound drought events have occurred in the last century [28]. In Africa, infant mortality is associated not only with various demographic and socioeconomic characteristics, but also with local cultural settings and societal structural frameworks, such as differences in healthcare facilities provision between urban and rural areas, domestic healthcare expenditure, and foreign resources flowing into health [29,30], which remain challenges the continent still faces. In addition, this region of the world deserves careful scrutiny, not only because the human health effects of climate change are infrequently studied among African populations [31], but also because it is one of the fastest growing populations, expected to be home to a quarter of the world’s population by 2050 [32].

To fill these research gaps, by using data from 34 African countries and developing a complex drought indicator, we aimed to (1) examine the association between drought and risk of infant mortality; (2) investigate whether the association varies between neonatal and post-neonatal deaths; and (3) explore the potential delayed effect of drought exposure on infant mortality.

Methods

Study population and infant mortality

The Demographic and Health Surveys (DHS) program routinely conducts nationally representative household surveys in over 90 low- and middle-income countries (LMICs) [33]. We obtained individual-level mortality data of children born during 1992–2019 within the 5 years prior to survey administration in 34 African LMICs. Using a two-stage sampling process, DHS selects survey clusters stratified by geographic region and urban/rural residence, and interviews 20–30 randomly selected households per cluster [33]. Mothers of reproductive age (15–49 years) in these sampled households were queried about each birth during the past 5 years. The geocoded coordinate location of each survey cluster was recorded as the centroid of the interviewed households, which was randomly displaced by 2–10 km to protect anonymity by DHS design [34]. To compute a wealth index that is comparable across surveys, we also extracted data on household characteristics, including source of drinking water, type of toilet facilities, electricity, type of flooring, and possession of radio, television, phone (landline or cellphone), refrigerator, motorcycle, and car. Following the method described by Bendavid and colleagues [35], we then performed a principal component analysis, generated a wealth index, and employed the quintiles of this index in the subsequent analysis.

Drought and other climate data

We quantified long-term drought conditions by the standardized precipitation evapotranspiration index (SPEI), which offers a robust and reliable measure to account for both water supply and demand [7], at a timescale of 24 months. We have previously described in detail how to calculate drought indicators [12]. Briefly, we first downloaded gridded monthly meteorological data at a resolution of 0.1° (~10 × 10 km) during the study period from the fifth-generation European Centre for Medium-Range Weather Forecasts atmospheric reanalysis of the global climate (ERA5-Land). We then calculated the climatic water balance as the difference between the available water content of soil and vegetation, measured by precipitation, and potential evapotranspiration, calculated using maximum and minimum temperatures, dew point temperature, wind speed, solar radiation, air pressure, latitude, and elevation. The SPEI is a monthly measure accounting for the overall cumulative effects of climatic conditions during a preceding period (S1 Fig). In order to compute the SPEI at a timescale of 24 months (referred to as SPEI-24 hereinafter) for a given month during the study period, we first generated a time series by summing the water balance over the preceding 23 months up to the current month; then we transformed this series to a normal distribution with a mean of zero and standard deviation of one, according to a log-logistic distribution, to eliminate the effect of both the local season and climate patterns [36].

Next, we linked each child with the gridded monthly SPEI, mean temperature, and total rainfall based on the geographic location of each cluster and the month and year of the child’s birth. Therefore, the exposure period was defined as from the start of pregnancy to death or the 12th month after birth, depending on whether a child died before 1 year of age (every pregnancy was assumed to be full-term [9 months] due to largely missing information on gestational age). Finally, adapting from the classification from the Federal Office of Meteorology and Climatology MeteoSwiss, we defined three types of drought events based on their severity (any drought: SPEI ≤ −0.5; mild drought: −1.3 < SPEI ≤ −0.5; severe drought: SPEI ≤ −1.3) [12].

Statistical analysis

We used an extended Cox regression model with time-dependent covariates to examine the association between drought at a timescale of 24 months and risk of infant mortality. Previously applied in other infant death studies [37], this extended model allows for time-varying covariance caused by the change of a given covariate over time during the follow-up period [38]. It is necessary to reiterate that our explanatory variable, the monthly drought indicator, changed over time (indicating the presence or absence of drought conditions for each month from conception to infant death). Our time-dependent Cox regression analysis was an appropriate approach for handling the time-varying drought indicator because, during each month in which an infant death occurred, the model compared the current drought condition experienced by the deceased infant with the current drought conditions experienced by all other infants at risk at that time [39]. With its ability to incorporate time-varying covariates, this method has been widely adopted in both environmental and non-environmental epidemiological research [37,40].

Specifically, we included in the model a time-dependent indicator for drought exposure and a series of time-invariant covariates, including child’s sex, area of residence, mother’s education, and wealth quintile. Categorical birth month, a natural cubic spline of birth year with three degrees of freedom, and a random intercept for a composite indicator for country and survey cluster were also included to adjust for seasonality, long-term trend, and cross-location differences, respectively (S1 Method). To explore whether risk of neonatal and post-neonatal mortality was differently associated with drought exposure, we applied the same model for infant deaths during the first 27 days and from the 28th day through the 12th month after birth.

We then stratified the main analysis by several baseline characteristics by incorporating an interaction term between the binary indicator for any drought and each of the following factors: child’s sex (male and female), area of residence (urban and rural), mother’s education (no education and any education), household’s wealth status (lower as the 1st–2nd quintiles and higher as the 3rd–5th quintiles), year period of birth (1992–2005 and 2006–2019), and climate zone (tropical, temperate, and dry). The statistical significance of the difference between subgroups was indicated by the p-value of the interaction term.

We then performed several secondary analyses. We first evaluated the association of drought exposure with infant mortality by month of death (27 days after birth for neonatal mortality), for which two types of exposure were assessed, namely a monthly binary indicator and the number of drought months experienced during pregnancy or infancy. Second, we combined all post-neonatal deaths (between 28 and 364 days of age) and examined whether neonatal and post-neonatal mortality were differently associated with monthly exposure during pregnancy. With the same exposure assessment period for all infant deaths within a specific month, we employed mixed-effects logistic regression models to quantify these associations (S1 Method). We categorized children who died during a specific month/period as cases and children who survived beyond that month/period as non-cases. For the death-month-specific analysis, we excluded children who were alive but younger than the age of that month at the time of the interview. Additionally, children who had died before that month were accounted for in the models that examined the association in previous months, and thus were also excluded from the analysis for the relevant month.

Several sensitivity analyses were conducted to test the model robustness. First, we separately included a natural cubic spline of monthly total precipitation with three degrees of freedom as a time-varying covariate in the main model. Second, we removed the month of birth from the main model to evaluate the impact of seasonality. Third, we adopted a discrete-time complementary log–log binomial regression model given the similar relative hazard interpretation of its coefficients to the Cox regression model in time-to-event analysis [41]. Fourth, we performed an additional analysis to compare the association with exposure from birth through infant deaths to the association with exposure from the start of pregnancy through infant deaths (main analysis). In addition, to test the importance of including gestational exposure in our main model, we incorporated drought during pregnancy in the fourth sensitivity analysis starting from birth. Since there were nine drought indicators during pregnancy for each child, we computed two alternative variables to assess gestational drought in the models: (1) an indicator of the total number of drought months a child experienced during pregnancy as a continuous variable, and (2) an indicator of any drought month a child ever experienced during pregnancy as a binary variable.

Only any drought was evaluated in the subgroup analysis and secondary analysis. We applied DHS sampling weight in all models to take into account the influence of survey representations [42]. The significance level for all statistical tests was set at two-sided p < 0.05. While this study did not have a prospectively registered protocol and statistical analysis plan, all variable definitions and analyses were pre-specified over the course of communications between the study team members and during laboratory meetings. All data analyses were completed using R statistical software 4.3.1, with the SPEI package for drought exposure assessment, and the coxme and glmmTMB packages for regression analysis.

This study is reported as per the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline (S1 STROBE Checklist). The Yale Institutional Review Board determined this study as not-human-subject research (IRB ID: 2000036064); thus, ethics approval for this study was not required.

Results

Our study included a total of 850,924 children from 103 surveys conducted in 34 countries during 1992–2019. The overall IMR in surveyed participants in Africa was 59.2 per 1,000 live births (50,377 deaths) (Table 1), with the highest observed in Sierra Leone (82.3), followed by Mali (82.0) and Eswatini (80.8), and the lowest in Gabon (27.5) (Fig 1A). The average (standard deviation) count of months in which children encountered all, mild, or severe drought during both pregnancy and the survival period was as follows among cases: 4.6 (5.2), 3.3 (4.1), or 1.3 (3.1), respectively. In comparison, among non-cases (i.e., children survived 12 months after birth and alive infants at the time of interview), the respective number of months was 7.3 (7.4), 5.2 (5.6), or 2.0 (4.3). The spatial distribution of drought events exhibits significant variation across the continent (Fig 1B). Eswatini experienced the highest average number of drought months (135.5 months) (area-weighted mean across all 10-km grids), whereas these events were least frequent in Guinea (39.9 months) (Fig 1C). In addition, fewer mothers and children experienced severe droughts than mild droughts, and more drought events were observed for children born during 2006–2019 than for children born during 1992–2005 and for those living in the tropical zone than for those living in the temperate or dry zone (S1 Table). The number of infant deaths during the first 27 days was significantly higher than the following 11 months (S2 Table).

Table 1. Study population characteristics in 34 low- and middle-income African countries during 1992–2019 by baseline characteristics (total N = 850,924).

Missing (%)

Infant mortality

(N [%])

Infant mortality rate

(per 1,000 live births)
Case Non-case Unweighted Weighted
Total 50,377 800,547 59.2 59.2
Child’s sex 0
 Male 27,802 (55.2) 403,017 (50.3) 64.5 64.7
 Female 22,575 (44.8) 397,530 (49.7) 53.7 53.7
Mother’s education 0
 No education 25,973 (51.6) 370,310 (46.3) 65.5 65.2
 Primary 16,422 (32.6) 266,580 (33.3) 58.0 58.3
 Secondary 7,310 (14.5) 145,469 (18.2) 47.8 47.7
 Higher 671 (1.3) 18,133 (2.3) 35.7 41.0
Residence 0
 Urban 11,563 (23.0) 216,519 (27.0) 50.7 48.4
 Rural 38,814 (77.0) 584,028 (73.0) 62.3 63.2
Wealth quintile 11.1
 Lowest 11,093 (25.3) 153,525 (21.5) 67.4 67.3
 Second 9,529 (21.8) 135,105 (19.0) 65.9 66.7
 Middle 8,534 (19.5) 140,712 (19.7) 57.2 57.2
 Fourth 8,287 (18.9) 143,021 (20.1) 54.8 52.9
 Highest 6,368 (14.5) 140,505 (19.7) 43.4 44.5
Year of birth 0
 1992–2005 20,615 (40.9) 253,104 (31.6) 75.3 76.6
 2006–2019 29,762 (59.1) 547,443 (68.4) 51.6 51.0
Climate zone 0
 Tropical 30,515 (60.6) 472,541 (59.0) 60.7 58.7
 Temperate 7,633 (15.2) 130,152 (16.3) 55.4 65.2
 Dry 12,229 (24.3) 197,854 (24.7) 58.2 56.5
Open in a new tab

Fig 1. Infant mortality rate (deaths per 1,000 live births) (A), number of drought months (B), and country-specific distributions of the number of drought months across all 10-km grids, sorted by area-weighted mean value (diamonds) and labeled with number of survey clusters (C) in Africa during 1992–2019.

Fig 1

Open in a new tab

The base layer of the map was obtained from ArcGIS Hub (https://hub.arcgis.com/datasets/07610d73964e4d39ab62c4245d548625/).

Compared to children who did not experience drought, we did not find evidence that any drought exposure was associated with an increased risk of infant mortality (hazard ratio [HR]: 1.02; 95% confidence interval [CI] [1.00, 1.04], p = 0.072) (Fig 2). When the analysis was stratified by drought severity, a significant positive association was observed for severe drought (HR: 1.04; 95% CI [1.01, 1.07], p = 0.015), whereas a null association was observed for mild drought (HR: 1.01; 95% CI [0.99, 1.03], p = 0.353). This null association with mild drought was estimated for both post-neonatal mortality (HR: 1.02; 95% CI [0.99, 1.06], p = 0.203) and neonatal mortality (HR: 0.99; 95% CI [0.96, 1.02], p = 0.657). However, severe drought was found to be significantly associated with neonatal mortality (HR: 1.05; 95% CI [1.01, 1.10], p = 0.019), but not with post-neonatal mortality (HR: 1.04; 95% CI [0.99, 1.09], p = 0.120), compared to non-exposure to any drought.

Fig 2. Error bar charts for the associations between risk of infant mortality and exposure to long-term drought represented by 24-month standardized precipitation and evapotranspiration index by drought severity (any, mild, or severe) and type of infant mortality (neonatal or post-neonatal mortality).

Fig 2

Open in a new tab

Reference group is the children who did not experience droughts. Covariates in the models included child’s sex, area of residence, mother’s education, and wealth quintile, categorical birth month, a natural cubic spline of birth year with three degrees of freedom, and a random intercept for a composite indicator for country and survey cluster.

We found that the association was significantly stronger among children living in rural areas than those living in urban areas and among children living in tropical climate zones than those living in dry zones (Fig 3). We did not observe significant effect modification by other factors.

Fig 3. Associations between risk of infant mortality and exposure to long-term drought stratified by baseline characteristics, year period of birth, and climate zone.

Fig 3

Open in a new tab

Statistically significant pairwise differences (p < 0.05) are marked with an asterisk.

We found the association of monthly exposure to drought during pregnancy with neonatal mortality to be non-significant except for that during the 1st month of pregnancy (odds ratio: 1.04; 95% CI [1.01, 1.07], p = 0.012) (Figs 4A and S1, 9th month before birth). Furthermore, monthly drought exposure was associated with infant mortality risk during the 7th–8th month after birth in a more consistent and significant fashion (Fig 4A), and the significant increase in mortality risk was observed in the 3rd and 7th–8th month for the number of drought months during pregnancy and only in the 7th month for the number of drought months during infancy (Fig 4B). We also found that the association with all post-neonatal mortality was observed to be significant during the 2nd–4th and 8th–9th month of pregnancy (S2 Fig, 6th–8th and 1st–2nd month before birth).

Fig 4. Associations between infant deaths during each month within 1 year of age and long-term drought.

Fig 4

Open in a new tab

The exposure was measured by a binary drought indicator during each month before and after birth (the ninth and first month of drought exposure before birth represents the first and last month of pregnancy, respectively) (A) and by the number of drought months experienced before and after birth (from the first through the last month of pregnancy for prenatal exposure and from the month of birth through the relevant month examined for postnatal exposure) (B). Statistically significant results (p < 0.05) are marked with an asterisk.

The results from the sensitivity analysis suggested that our estimation was largely robust, although we observed slight attenuation of the association with all types of drought when applying complimentary log–log binary regression models (S3 Table). In addition, when we started the analysis from birth, the association with severe drought was found to be non-significant (S3 Table). However, when we included drought exposure during pregnancy as a covariate in this postnatal-period-only analysis, it showed a statistically significant association with our outcome (S4 Table), further demonstrating the importance of including this prenatal exposure to drought in our primary analysis.

Discussion

In this cross-sectional, multi-country study, we used an advanced drought measure and found no evidence that any drought exposure was associated with an increased risk of infant mortality in African LMICs. However, long-term severe drought exposure was observed to be significantly associated with an increased risk of infant mortality. A significantly stronger association was observed among children living in rural areas than among those in urban areas, and among children living in tropical zones compared to those in dry zones.

Researchers have investigated the relationship between drought and under-five child mortality, with no plausible association observed [24]. Compared to child mortality during 1–4 years after birth, infant mortality, the largest contributor of under-five mortality [3], is more likely to be biologically linked to drought exposure during pregnancy. However, there is still limited evidence on the negative impact of drought on infant deaths. Two studies in China and Brazil defined drought as rainfall shocks or fluctuations, which were calculated as the log deviation of total rainfall during the 12 months before birth relative to the historical average yearly total rainfall, and both studies reported increased risk of infant mortality associated with rainfall deficit [25,26]. However, no other climatic conditions were involved in the drought indicator calculation. Similarly, a global study employed positive or negative rainfall shocks, defined as one standard deviation above or below the historical average rainfall, and found a negative or positive association with IMR, respectively [43]. Our study, for the first time, assessed the relationship between infant mortality and drought considering both water supply and demand, and also found a heightened risk associated with long-term drought conditions. Moreover, by using a monthly drought indicator, instead of annual rainfall deficit employed by previous studies, our study was able to examine the critical window of the association between drought exposure and infant mortality.

The exact mechanism behind this association is still unknown. However, the interaction among drought, water, sanitation, and hygiene practices, and diarrhea may play a notable role in the effect pathway. Drought has been found to be associated with an increased risk of diarrhea [12], one of the leading causes of death in children under five. In addition, drought not only directly declines surface water and groundwater quantity through reduced precipitation and enhanced evapotranspiration, but also deteriorates water quality of freshwater systems via increased salinity, turbidity, algal level, and metal and pathogen concentration [44]. Inadequate access to water and consumption of unsafe drinking water could either directly link to infant deaths [45,46], or increase the likelihood of diarrhea, subsequently contributing to higher mortality. Furthermore, drought has a detrimental effect on crop yield and livestock, and plays a substantial role in driving land degradation and potentially triggering forced migration [47], posing a threat of food insecurity, which in turn leads to undernutrition for pregnant women and infant—an additional contributor to infant mortality. Specifically, suboptimal nutrition during pregnancy increases the risk of low birth weight, a well-documented factor in infant deaths [48]. In addition, poor nutritional status among children, characterized by underweight, stunting, and wasting, can also contribute to the burden of infant mortality [49].

There could be delays from drought exposure to diarrhea occurrences or poor water, sanitation, and hygiene practices, both of which could also directly or indirectly result in infant deaths. Indeed, we found the association of in-utero or postnatal drought exposure to be lagged for several months in our month-specific analysis. However, the significant association we observed between severe drought and neonatal mortality implied the stronger and more acute impact of long-term drought condition with greater intensity. To our knowledge, this is the first study looking into the delayed association between rigorously defined drought measure and infant health. Apart from the direct impact due to insufficient water supply, all indirect pathways linking drought to infant mortality contribute to the time lag in the association. Additionally, younger infants are likely to be less vulnerable due to breastfeeding, which may explain the null associations we observed for the first 2 months after birth. However, the reason behind the exhibited lag pattern (elevation in the association during the 3rd, 7th–8th, and 11th–12th month) is unknown and future studies are needed to further investigate the delayed effect of drought.

We observed a significantly stronger association among children residing in tropical zones than those living in dry zones. Interestingly, the association between drought exposure and diarrhea among children in tropical zones was found to be significantly weaker or null in our previous work [12]. The reason behind this discrepancy remains unclear. However, this contrast could stem from the more indirect and lagged influence of drought on infant mortality. With a delayed association, despite greater resources, tropical zones may still lack the capacity to adapt to more persistent drought conditions, compared to temperate and dry zones that have already learned to adapt to intermittent drought conditions. Rural areas were found to bear a higher burden of the impact from drought events than urban areas. Maternal and child healthcare facilities are likely to be absent or significantly fewer in rural settings, and poor road networks in these areas would further hinder accessibility to these facilities. Furthermore, there may be insufficient water, sanitation, and hygiene practices in the traditional childbirth environment in rural African regions [50,51], and drought exposure could further diminish the already inadequate washing practices, posing notably increased risk of life-threatening infections among newborns.

We acknowledge several limitations. First, our survival analysis started at the beginning of pregnancy, whereas our outcome of interest was infant mortality, which could not occur until birth. This misalignment may introduce immortal time bias, potentially over- or underestimating our investigated association. Second, despite a monthly measure, the calculated 24-month SPEI in our study incorporates multiple weather parameters to illustrate the overall dry condition for the preceding 24 months. However, drought and other unfavorable climatic conditions during this period could have already resulted in deaths in an earlier stage (i.e., miscarriage or stillbirth), introducing live-birth bias [52]. This study focusing on infant mortality might potentially underestimate the detrimental impact of drought on deaths in early life. Third, DHS randomly displaced all geocoded survey cluster locations by 2–10 km to protect privacy [34], and thus our drought exposure may be misclassified, particularly for children living in a cluster close to a border of a 10-km grid. Fourth, the definition of our drought indicator and its severity classification was derived from a single source, which may vary across locations and, as a result, may potentially limit the generalizability of our findings. Fifth, due to largely missing information on gestational age, we assumed the same pregnancy duration (i.e., 9 months) for all live births, which may also lead to drought exposure misclassification. Sixth, in our secondary analysis we examined the associations between infant deaths during each of the 12 months after birth and drought exposure during each of the 21 months before and after birth, which might increase the possibility of chance associations due to multiple comparisons. Last, nutritional status and diarrhea occurrences both significantly impact infant mortality risk. However, we were not able to unravel the interrelationship between these factors and drought since DHS only queries these variables for living children.

In summary, this study found that exposure to long-term severe drought was associated with increased risk of infant mortality. Comprehensive engagement, planning, and preparedness should be advanced among pregnant women and children that are disproportionately affected by drought, particularly in rural and tropical settings. Understanding the association between drought and infant mortality can guide a multi-faceted approach in Africa involving early warning systems, neonatal and pediatric care infrastructure planning, community education, food security program development, and water and sanitation interventions. With more frequent and intense drought events projected due to climate change, coupled with inadequate drought adaptation capacity in LMICs, the protection of pregnant women and young children from drought could potentially play a vital role in alleviating the substantial burden of infant mortality in Africa.

Supporting information

S1 STROBE Checklist. STROBE Statement—Checklist of items that should be included in reports of cross-sectional studies.

(DOCX)

pmed.1004516.s001.docx (32.7KB, docx)
S1 Method. Model specification of main and secondary analysis.

(DOCX)

pmed.1004516.s002.docx (24.3KB, docx)
S1 Table. Descriptive statistics for the number of drought months experienced by all included children by exposure window, year of birth, climate zone, and drought severity.

(DOCX)

pmed.1004516.s003.docx (28.9KB, docx)
S2 Table. Number of infant deaths (among 850,924 children) during each month within 1 year of age and number of months for SPEI-24 any drought experienced before and after birth.

SPEI: standardized precipitation evapotranspiration index.

(DOCX)

pmed.1004516.s004.docx (28.1KB, docx)
S3 Table. Association between long-term drought and risk of infant mortality in various sensitivity analyses.

(DOCX)

pmed.1004516.s005.docx (26.1KB, docx)
S4 Table. Estimates and uncertainties of drought exposure before and after birth when including gestational drought exposure in the sensitivity analysis of postnatal period only.

(DOCX)

pmed.1004516.s006.docx (26.2KB, docx)
S1 Fig. Illustration of calculation of SPEI at a timescale of 24 months for a given month from the start of pregnancy through death or the 12th month after birth.

Climate variables include maximum and minimum temperatures, dew point temperature, wind speed, solar radiation, and air pressure. SPEI: standardized precipitation evapotranspiration index.

(DOCX)

pmed.1004516.s007.docx (74.3KB, docx)
S2 Fig. Associations between neonatal and post-neonatal mortality and drought exposure by month of pregnancy.

(DOCX)

pmed.1004516.s008.docx (66KB, docx)

Acknowledgments

The authors would like to thank the DHS Program, ICF International for providing the data used in the analysis. The views expressed in this paper are those of the authors and not necessarily those of RGHI. The contents are solely the responsibility of the authors and do not necessarily represent the official views of NIH. JSR was an expert witness at the request of Relator’s attorneys, the Greene Law Firm, in a qui tam suit alleging violations of the False Claims Act and Anti-Kickback Statute against Biogen that was settled September 2022.

Abbreviations

CI

confidence interval

DHS

Demographic and Health Surveys

HR

hazard ratio

IMR

infant mortality rate

LMICs

low- and middle-income countries

SPEI

standardized precipitation evapotranspiration index

STROBE

Strengthening the Reporting of Observational Studies in Epidemiology

Data Availability

Data on infant mortality and socioeconomic status in this study were collected from the Demographic and Health Surveys program (https://dhsprogram.com/), which are freely available upon request. Publicly available meteorological records were obtained from the fifth generation ECMWF atmospheric reanalysis of the global climate (ERA5-Land, https://cds.climate.copernicus.eu/).

Funding Statement

PW, EOA, and KC received support from the Reckitt Global Hygiene Institute (https://rghi.org/; RGHI RIN: 2021-001). TR was funded by CTSA (Grant Number UL1 TR001863) from the National Center for Advancing Translational Science (NCATS; https://ncats.nih.gov/), a component of the National Institutes of Health (NIH). JSR currently receives research support through Yale University from Johnson and Johnson to develop methods of clinical trial data sharing, from the Food and Drug Administration for the Yale-Mayo Clinic Center for Excellence in Regulatory Science and Innovation (CERSI; https://www.fda.gov/science-research/advancing-regulatory-science/centers-excellence-regulatory-science-and-innovation-cersis) program (U01FD005938), from the Agency for Healthcare Research and Quality (R01HS022882) (https://www.ahrq.gov/), and from Arnold Ventures; formerly received research support from the Medical Device Innovation Consortium as part of the National Evaluation System for Health Technology (NEST; https://www.fda.gov/about-fda/cdrh-reports/national-evaluation-system-health-technology-nest) and from the National Heart, Lung and Blood Institute (https://www.nhlbi.nih.gov/) of the National Institutes of Health (NIH) (R01HS025164, R01HL144644). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

  • 1.World Health Organization. Indicators [cited 2023 Aug 15]. Available from: https://www.who.int/data/maternal-newborn-child-adolescent-ageing/indicator-explorer-new/mca/infant-mortality-rate-(per-1000-live-births)
  • 2.Sharrow D, Hug L, You D, Alkema L, Black R, Cousens S, et al. Global, regional, and national trends in under-5 mortality between 1990 and 2019 with scenario-based projections until 2030: a systematic analysis by the UN Inter-agency Group for Child Mortality Estimation. Lancet Glob Health. 2022;10(2):e195–e206. doi: 10.1016/S2214-109X(21)00515-5 ; PMCID: PMC8789561 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.United Nations Inter-agency Group for Child Mortality Estimation (UN IGME). Levels & trends in child mortality: report 2022, estimates developed by the United Nations Inter-agency Group for Child Mortality Estimation. New York: 2023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Pullabhotla HK, Zahid M, Heft-Neal S, Rathi V, Burke M. Global biomass fires and infant mortality. Proc Natl Acad Sci U S A. 2023;120(23):e2218210120. Epub 2023/05/30. doi: 10.1073/pnas.2218210120 ; PMCID: PMC10266003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.He G, Perloff JM. Surface water quality and infant mortality in China. Econ Dev Cul Change. 2016;65(1):119–39. doi: 10.1086/687603 [DOI] [Google Scholar]
  • 6.Ezeh OK, Agho KE, Dibley MJ, Hall J, Page AN. The impact of water and sanitation on childhood mortality in Nigeria: evidence from demographic and health surveys, 2003-2013. Int J Environ Res Public Health. 2014;11(9):9256–72. Epub 2014/09/05. doi: 10.3390/ijerph110909256 ; PMCID: PMC4199018 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Vicente-Serrano SM, Beguería S, López-Moreno JI. A multiscalar drought index sensitive to global warming: the standardized precipitation evapotranspiration index. J Climate. 2010;23(7):1696–718. doi: 10.1175/2009jcli2909.1 [DOI] [Google Scholar]
  • 8.Trisos CH, Adelekan IO, Totin E, Ayanlade A, Efitre J, Gemeda A, et al. Africa. Climate change 2022: impacts, adaptation and vulnerability. Contribution of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, UK and New York, NY, USA: Cambridge University Press; 2022. p. 1285–455. [Google Scholar]
  • 9.Naumann G, Alfieri L, Wyser K, Mentaschi L, Betts RA, Carrao H, et al. Global changes in drought conditions under different levels of warming. Geophys Res Lett. 2018;45(7):3285–96. doi: 10.1002/2017gl076521 [DOI] [Google Scholar]
  • 10.Salvador C, Nieto R, Vicente-Serrano SM, Garcia-Herrera R, Gimeno L, Vicedo-Cabrera AM. Public health implications of drought in a climate change context: a critical review. Annu Rev Public Health. 2023;44:213–32. Epub 2023/01/09. doi: 10.1146/annurev-publhealth-071421-051636 [DOI] [PubMed] [Google Scholar]
  • 11.Vicente-Serrano SM, Quiring SM, Peña-Gallardo M, Yuan S, Domínguez-Castro F. A review of environmental droughts: increased risk under global warming? Earth-Sci Rev. 2020;201:102953. doi: 10.1016/j.earscirev.2019.102953 [DOI] [Google Scholar]
  • 12.Wang P, Asare E, Pitzer VE, Dubrow R, Chen K. Associations between long-term drought and diarrhea among children under five in low- and middle-income countries. Nat Commun. 2022;13(1):3661. Epub 2022/06/30. doi: 10.1038/s41467-022-31291-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Lowe R, Lee SA, O’Reilly KM, Brady OJ, Bastos L, Carrasco-Escobar G, et al. Combined effects of hydrometeorological hazards and urbanisation on dengue risk in Brazil: a spatiotemporal modelling study. Lancet Planet Health. 2021;5(4):e209–e19. doi: 10.1016/S2542-5196(20)30292-8 [DOI] [PubMed] [Google Scholar]
  • 14.Phung D, Nguyen-Huy T, Tran NN, Tran DN, Doan VQ, Nghiem S, et al. Hydropower dams, river drought and health effects: a detection and attribution study in the lower Mekong Delta Region. Clim Risk Manag. 2021;32:100280. doi: 10.1016/j.crm.2021.100280 [DOI] [Google Scholar]
  • 15.Machado-Silva F, Libonati R, Melo de Lima TF, Bittencourt Peixoto R, de Almeida França JR, de Avelar Figueiredo Mafra Magalhães M, et al. Drought and fires influence the respiratory diseases hospitalizations in the Amazon. Ecol Indic. 2020;109:105817. doi: 10.1016/j.ecolind.2019.105817 [DOI] [Google Scholar]
  • 16.Lieber M, Chin-Hong P, Kelly K, Dandu M, Weiser SD. A systematic review and meta-analysis assessing the impact of droughts, flooding, and climate variability on malnutrition. Glob Public Health. 2022;17(1):68–82. Epub 2020/12/17. doi: 10.1080/17441692.2020.1860247 ; PMCID: PMC8209118 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Vins H, Bell J, Saha S, Hess JJ. The mental health outcomes of drought: a systematic review and causal process diagram. Int J Environ Res Public Health. 2015;12(10):13251–75. Epub 2015/10/22. doi: 10.3390/ijerph121013251 ; PMCID: PMC4627029 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Berman JD, Ebisu K, Peng RD, Dominici F, Bell ML. Drought and the risk of hospital admissions and mortality in older adults in western USA from 2000 to 2013: a retrospective study. Lancet Planet Health. 2017;1(1):e17–e25. Epub 2017/04/12. doi: 10.1016/S2542-5196(17)30002-5 ; PMCID: PMC5646697 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Salvador C, Nieto R, Linares C, Diaz J, Gimeno L. Quantification of the effects of droughts on daily mortality in Spain at different timescales at regional and national levels: a meta-analysis. Int J Environ Res Public Health. 2020;17(17):6114. Epub 2020/08/22. doi: 10.3390/ijerph17176114 ; PMCID: PMC7504151 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Abadi AM, Gwon Y, Gribble MO, Berman JD, Bilotta R, Hobbins M, Bell JE. Drought and all-cause mortality in Nebraska from 1980 to 2014: time-series analyses by age, sex, race, urbanicity and drought severity. Sci Total Environ. 2022;840:156660. Epub 2022/06/13. doi: 10.1016/j.scitotenv.2022.156660 [DOI] [PubMed] [Google Scholar]
  • 21.Lynch KM, Lyles RH, Waller LA, Abadi AM, Bell JE, Gribble MO. Drought severity and all-cause mortality rates among adults in the United States: 1968-2014. Environ Health. 2020;19(1):52. Epub 2020/05/18. doi: 10.1186/s12940-020-00597-8 ; PMCID: PMC7236144 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Salvador C, Vicedo-Cabrera AM, Libonati R, Russo A, Garcia BN, Belem LBC, et al. Effects of drought on mortality in macro urban areas of Brazil between 2000 and 2019. Geohealth. 2022;6(3):e2021GH000534. Epub 2022/03/01. doi: 10.1029/2021GH000534 ; PMCID: PMC8902811 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Gwon Y, Ji Y, Bell JE, Abadi AM, Berman JD, Rau A, et al. The association between drought exposure and respiratory-related mortality in the United States from 2000 to 2018. Int J Environ Res Public Health. 2023;20(12). Epub 2023/06/07. doi: 10.3390/ijerph20126076 ; PMCID: PMC10298035 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Delbiso TD, Altare C, Rodriguez-Llanes JM, Doocy S, Guha-Sapir D. Drought and child mortality: a meta-analysis of small-scale surveys from Ethiopia. Sci Rep. 2017;7(1):2212. Epub 2017/05/19. doi: 10.1038/s41598-017-02271-5 ; PMCID: PMC5438405 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Lin Y, Liu F, Xu P. Effects of drought on infant mortality in China. Health Econ. 2021;30(2):248–69. Epub 2020/11/09. doi: 10.1002/hec.4191 [DOI] [PubMed] [Google Scholar]
  • 26.Rocha R, Soares RR. Water scarcity and birth outcomes in the Brazilian semiarid. J Dev Econ. 2015;112:72–91. doi: 10.1016/j.jdeveco.2014.10.003 [DOI] [Google Scholar]
  • 27.Slette IJ, Post AK, Awad M, Even T, Punzalan A, Williams S, et al. How ecologists define drought, and why we should do better. Glob Chang Biol. 2019;25(10):3193–200. Epub 2019/07/06. doi: 10.1111/gcb.14747 [DOI] [PubMed] [Google Scholar]
  • 28.Held IM, Delworth TL, Lu J, Findell KL, Knutson TR. Simulation of Sahel drought in the 20th and 21st centuries. Proc Natl Acad Sci U S A. 2005;102(50):17891–6. Epub 2005/12/01. doi: 10.1073/pnas.0509057102 ; PMCID: PMC1312412 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Ekholuenetale M, Wegbom AI, Tudeme G, Onikan A. Household factors associated with infant and under-five mortality in sub-Saharan Africa countries. Int J Child Care Educ Policy. 2020;14(1):10. doi: 10.1186/s40723-020-00075-1 [DOI] [Google Scholar]
  • 30.Akinlo AE, Sulola AO. Health care expenditure and infant mortality in sub-Saharan Africa. J Policy Model. 2019;41(1):168–78. doi: 10.1016/j.jpolmod.2018.09.001 [DOI] [Google Scholar]
  • 31.Victoria LB, Ravi G, Joshua W, Kate N, Kai C, Joseph SR. Published research on the human health implications of climate change from 2012-2019: a cross-sectional study. medRxiv. 2022:2022.10.01.22280596. doi: 10.1101/2022.10.01.22280596 [DOI] [Google Scholar]
  • 32.United Nations Department of Economic and Social Affairs Population Division. World population prospects 2022: summary of results. 2022. [Google Scholar]
  • 33.Corsi DJ, Neuman M, Finlay JE, Subramanian SV. Demographic and health surveys: a profile. Int J Epidemiol. 2012;41(6):1602–13. Epub 2012/11/12. doi: 10.1093/ije/dys184 [DOI] [PubMed] [Google Scholar]
  • 34.Burgert CR, Colston J, Roy T, Zachary B. Geographic displacement procedure and georeferenced data release policy for the Demographic and Health Surveys. Calverton, MD, USA: ICF International; 2013. [Google Scholar]
  • 35.Bendavid E. Is health aid reaching the poor? Analysis of household data from aid recipient countries. PLoS One. 2014;9(1):e84025. doi: 10.1371/journal.pone.0084025 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Vicente-Serrano SM, National Center for Atmospheric Research Staff, editors). The climate data guide: standardized precipitation evapotranspiration index (SPEI) [updated 18 Jul 2015]. Available from: https://climatedataguide.ucar.edu/climate-data/standardized-precipitation-evapotranspiration-index-spei [Google Scholar]
  • 37.Platt RW, Joseph KS, Ananth CV, Grondines J, Abrahamowicz M, Kramer MS. A proportional hazards model with time-dependent covariates and time-varying effects for analysis of fetal and infant death. Am J Epidemiol. 2004;160(3):199–206. doi: 10.1093/aje/kwh201 [DOI] [PubMed] [Google Scholar]
  • 38.Zhang Z, Reinikainen J, Adeleke KA, Pieterse ME, Groothuis-Oudshoorn CGM. Time-varying covariates and coefficients in Cox regression models. Ann Transl Med. 2018;6(7):121. doi: 10.21037/atm.2018.02.12 ; PMCID: PMC6015946 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Therneau T, Crowson C, Atkinson E. Using time dependent covariates and time dependent coefficients in the Cox modelOct 26, 2024. Available from: https://cran.r-project.org/web/packages/survival/vignettes/timedep.pdf [Google Scholar]
  • 40.Lepeule J, Rondeau V, Filleul L, Dartigues JF. Survival analysis to estimate association between short-term mortality and air pollution. Environ Health Perspect. 2006;114(2):242–7. doi: 10.1289/ehp.8311 ; PMCID: PMC1367838 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Suresh K, Severn C, Ghosh D. Survival prediction models: an introduction to discrete-time modeling. BMC Med Res Methodol. 2022;22(1):207. Epub 2022/07/26. doi: 10.1186/s12874-022-01679-6 ; PMCID: PMC9316420 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Rutstein SO, Rojas G. Guide to DHS statistics. Calverton, MD: Demographic and Health Surveys; 2006. [Google Scholar]
  • 43.Ponnusamy S. Rainfall shocks, child mortality, and water infrastructure. Health Econ. 2022;31(7):1317–38. Epub 2022/04/06. doi: 10.1002/hec.4498 [DOI] [PubMed] [Google Scholar]
  • 44.Mosley LM. Drought impacts on the water quality of freshwater systems; review and integration. Earth Sci Rev. 2015;140:203–14. doi: 10.1016/j.earscirev.2014.11.010 [DOI] [Google Scholar]
  • 45.Cheng JJ, Schuster-Wallace CJ, Watt S, Newbold BK, Mente A. An ecological quantification of the relationships between water, sanitation and infant, child, and maternal mortality. Environ Health. 2012;11:4. Epub 2012/01/27. doi: 10.1186/1476-069X-11-4 ; PMCID: PMC3293047 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Hafeman D, Factor-Litvak P, Cheng Z, van Geen A, Ahsan H. Association between manganese exposure through drinking water and infant mortality in Bangladesh. Environ Health Perspect. 2007;115(7):1107–12. doi: 10.1289/ehp.10051 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Hermans K, McLeman R. Climate change, drought, land degradation and migration: exploring the linkages. Curr Opin Environ Sustain. 2021;50:236–44. doi: 10.1016/j.cosust.2021.04.013 [DOI] [Google Scholar]
  • 48.McCormick MC. The contribution of low birth weight to infant mortality and childhood morbidity. N Engl J Med. 1985;312(2):82–90. doi: 10.1056/NEJM198501103120204 [DOI] [PubMed] [Google Scholar]
  • 49.Black RE, Allen LH, Bhutta ZA, Caulfield LE, de Onis M, Ezzati M, et al. Maternal and child undernutrition: global and regional exposures and health consequences. Lancet. 2008;371(9608):243–60. doi: 10.1016/S0140-6736(07)61690-0 [DOI] [PubMed] [Google Scholar]
  • 50.Gon G, Restrepo-Mendez MC, Campbell OM, Barros AJ, Woodd S, Benova L, Graham WJ. Who delivers without water? A multi country analysis of water and sanitation in the childbirth environment. PLoS One. 2016;11(8):e0160572. Epub 2016/08/17. doi: 10.1371/journal.pone.0160572 ; PMCID: PMC4988668 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Arowosegbe AO, Ojo DA, Shittu OB, Iwaloye O, Ekpo UF. Water, sanitation, and hygiene (WASH) facilities and infection control/prevention practices in traditional birth homes in Southwest Nigeria. BMC Health Serv Res. 2021;21(1):912. Epub 2021/09/03. doi: 10.1186/s12913-021-06911-5 ; PMCID: PMC8417956 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Neophytou AM, Kioumourtzoglou MA, Goin DE, Darwin KC, Casey JA. Educational note: addressing special cases of bias that frequently occur in perinatal epidemiology. Int J Epidemiol. 2021;50(1):337–45. doi: 10.1093/ije/dyaa252 ; PMCID: PMC8453403 [DOI] [PMC free article] [PubMed] [Google Scholar]

Decision Letter 0

Katrien G Janin

14 Mar 2024

Dear Dr Wang,

Thank you for resubmitting your manuscript entitled "The drier, the deadlier: The association between long-term drought and risk of infant mortality in Africa" for consideration by PLOS Medicine.

Your manuscript and rebuttal letter has now been evaluated by the PLOS Medicine editorial staff and I am pleased to let you know that we would like to send your revised submission out for external peer review.

However, before we can send your manuscript to reviewers, we need you to complete your submission by providing the metadata that is required for full assessment. To this end, please login to Editorial Manager where you will find the paper in the 'Submissions Needing Revisions' folder on your homepage. Please click 'Revise Submission' from the Action Links and complete all additional questions in the submission questionnaire.

Please re-submit your manuscript within two working days, i.e. by Mar 18 2024 11:59PM.

Login to Editorial Manager here: https://www.editorialmanager.com/pmedicine

Once your full submission is complete, your paper will undergo a series of checks in preparation for peer review. Once your manuscript has passed all checks it will be sent out for review.

Feel free to email me at kjanin@plos.org if you have any queries relating to your submission.

Kind regards,

Katrien G. Janin, PhD

Senior Editor

PLOS Medicine

Decision Letter 1

Katrien G Janin

18 Apr 2024

Dear Dr. Wang,

Thank you very much for submitting your manuscript "The drier, the deadlier: the association between long-term drought and risk of infant mortality in Africa" (PMEDICINE-D-24-00825R1) for consideration at PLOS Medicine.

As you will see, the reviewers were very positive about the paper and the importance of these follow-up data, but they raised a number of questions about specific study details and presentation. After discussing the paper with the editorial team and an academic editor with relevant expertise, I’m pleased to invite you to revise the paper in response to the reviewers’ comments. We plan to send the revised paper to some of all of the original reviewers*, and of course we cannot provide any guarantees at this stage regarding publication.

When you upload your revision, please include a point-by-point response that addresses all of the reviewer and editorial points, indicating the changes made in the manuscript and either an excerpt of the revised text or the location (eg: page and line number) where each change can be found. Please submit a clean version of the paper as the main article file and a version with changes marked should as a marked-up manuscript. Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper.

We ask that you submit your revision by 9th of May. However, if this deadline is not feasible, please contact me by email, and we can discuss a suitable alternative.

Don’t hesitate to contact me directly with any questions (kjanin@plos.org). If you reply directly to this message, please be sure to ‘Reply All’ so your message comes directly to my inbox.

Kind regards,

Katrien

Katrien G. Janin, PhD

Associate Editor

PLOS Medicine

plosmedicine.org

-----------------------------------------------------------

Requests from the editors:

Data Availability: 

PLOS Medicine requires that the de-identified data underlying the specific results in a published article be made available, without restrictions on access, in a public repository or as Supporting Information at the time of article publication, provided it is legal and ethical to do so. Please see the policy at http://journals.plos.org/plosmedicine/s/data-availability and FAQs at http://journals.plos.org/plosmedicine/s/data-availability#loc-faqs-for-data-policy   

PLOS defines the “minimal data set” to consist of the data set used to reach the conclusions drawn in the manuscript with related metadata and methods, and any additional data required to replicate the reported study findings in their entirety. Authors do not need to submit their entire data set, or the raw data collected during an investigation. 

For each data source used in your study:  

a) If the data are freely or publicly available, note this and state the location of the data: within the paper, in Supporting Information files, or in a public repository (include the DOI or accession number). 

b) If the data are owned by a third party but freely available upon request, please note this and state the owner of the data set and contact information for data requests (web or email address). Note that a study author cannot be the contact person for the data. 

c) If the data are not freely available, please describe briefly the ethical, legal, or contractual restriction that prevents you from sharing it. Please also include an appropriate contact (web or email address) for inquiries (again, this cannot be a study author). 

Financial Disclosure

The funding statement should include: specific grant numbers, initials of authors who received each award, URLs to sponsors’ websites. Also, please state whether any sponsors or funders (other than the named authors) played any role in study design, data collection and analysis, the decision to publish, or preparation of the manuscript. If they had no role in the research, include this sentence: “The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Reporting guidance 

Please report your study according to the relevant guidance which can be found here https://www.equator-network.org/reporting-guidelines/ 

Given your study is a observational cross-sectional study, Please ensure that the study is reported according to the STROBE guideline, and include the completed STROBE checklist as Supporting Information. Please add the following statement, or similar, to the Methods: ""This study is reported as per the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline (S1 Checklist).The STROBE guideline can be found here: http://www.equator-network.org/reporting-guidelines/strobe/

When completing the checklist, please use section and paragraph numbers, rather than page numbers.

Did your study have a prospective protocol or analysis plan? Please state this (either way) early in the Methods section.

a) If a prospective analysis plan (from your funding proposal, IRB or other ethics committee submission, study protocol, or other planning document written before analyzing the data) was used in designing the study, please include the relevant prospectively written document with your revised manuscript as a Supporting Information file to be published alongside your study, and cite it in the Methods section. A legend for this file should be included at the end of your manuscript.

b) If no such document exists, please make sure that the Methods section transparently describes when analyses were planned, and when/why any data-driven changes to analyses took place.

c) In either case, changes in the analysis-- including those made in response to peer review comments-- should be identified as such in the Methods section of the paper, with rationale.

For all observational studies, in the manuscript text, please indicate: (1) the specific hypotheses you intended to test, (2) the analytical methods by which you planned to test them, (3) the analyses you actually performed, and (4) when reported analyses differ from those that were planned, transparent explanations for differences that affect the reliability of the study's results. If a reported analysis was performed based on an interesting but unanticipated pattern in the data, please be clear that the analysis was data-driven.

Statistical reporting 

As per standard PLOS Medical requirements, please provide 95% CIs and p values for all results where appropriate (including the abstract). We suggest reporting statistical information in the following format: ‘x’; (95% CI [‘y’,’ z’] p value) and use commas as opposed to hyphens (as these can be confused with negative values) to separate upper and lower bounds. For p values, please report as p<0.001 and where higher as 'p=0.002'. Please add the statistical method used to your method section. We also invite you to report p values to consistently to the third decimal digit - thousandths.

If applicable, please include any important dependent variables that are adjusted for in the analyses.

Please include the actual amounts and/or absolute risk(s) of relevant outcomes (including NNT or NNH where appropriate), not just relative risks or correlation coefficients. (example for absolute risks: PMID: 28399126). 

Title 

Please revise your title according to PLOS Medicine's style. Your title must be nondeclarative and not a question. It should begin with main concept if possible. "Effect of" should be used only if causality can be inferred, i.e., for an RCT. Please place the study design ("A randomized controlled trial," "A retrospective study," "A modelling study," etc.) in the subtitle (ie, after a colon). 

Abstract layout 

Please structure your abstract using the PLOS Medicine headings (Background, Methods and Findings, Conclusions). 

Author summary 

At this stage, we ask that you include a short, non-technical Author Summary of your research to make findings accessible to a wide audience that includes both scientists and non-scientists. The authors summary should consist of 2-3 succinct bullet points under each of the following headings: 

Why Was This Study Done? Authors should reflect on what was known about the topic before the research was published and why the research was needed. 

What Did the Researchers Do and Find? Authors should briefly describe the study design that was used and the study’s major findings. Do include the headline numbers from the study, such as the sample size and key findings.     

What Do These Findings Mean? Authors should reflect on the new knowledge generated by the research and the implications for practice, research, policy, or public health. Authors should also consider how the interpretation of the study’s findings may be affected by the study limitations. In the final bullet point of ‘What Do These Findings Mean?’, please describe the main limitations of the study in non-technical language. 

The Author Summary should immediately follow the Abstract in your revised manuscript. This text is subject to editorial change and should be distinct from the scientific abstract. Please see our author guidelines for more information: https://journals.plos.org/plosmedicine/s/revising-your-manuscript#loc-author-summary 

Introduction layout 

Please address past research and explain the need for and potential importance of your study. Indicate whether your study is novel and how you determined that. If there has been a systematic review of the evidence related to your study (or you have conducted one), please refer to and reference that review and indicate whether it supports the need for your study. 

Discussion layout 

Please present and organize the Discussion as follows: a short, clear summary of the article's findings; what the study adds to existing research and where and why the results may differ from previous research; strengths and limitations of the study; implications and next steps for research, clinical practice, and/or public policy; one-paragraph conclusion. 

Supplementary materials  

Please note that supplementary materials are not checked and will be posted as supplied by the authors. Therefore, please double check. Please cite your Supporting Information as outlined here: https://journals.plos.org/plosmedicine/s/supporting-information - Please note you may use almost any description as the item name of your supporting information as long as it contains an "S" and number. For example, “S1 Appendix” and “S2 Appendix,” “S1 Table” and “S2 Table. Please ensure each supplementary material has a call out (link) from your main manuscript.

-----------------------------------------------------------

Comments from the reviewers:

Reviewer #1: The authors have addressed my minor comments. I have nothing further to add.

Reviewer #2: The revisions and the rebuttal letter have answered all my queries in a good and well-organised way. The article was good also before this round of revisions, but is now even better. I am also glad to see that the team has discussed the study and the findings with colleagues from SSA and invited them to take part in the work.

Nothing more to add from my side. Important work.

Reviewer #3: Thanks for revising the paper based on the feedback from the reviewers. There have been significant improvements to the methods of the paper, but here are some remaining issues I found.

1. Line 155-160: "In order to compute the SPEI at a timescale of 24 months (referred to as SPEI-24 hereinafter) for a given month during the study period, we first generated a time series by summing the water balance over the preceding 23 months up to the current month; then we transformed this series to a normal distribution with a mean of zero and standard deviation of one to eliminate the effect of both the local season and climate patterns."

This is still very confusing. I spent time reading the SPEI documentation. The documentation suggests that an essential step of fitting a probability distribution to the 24-month historical water balance data is missing here. This is integral for deriving the CDF used in the SPEI formula: SPEI = Z(CDF(WB)/(1-CDF(WB))). It is crucial to specify which probability distribution was selected to model the observed 24-month SPEI time series. Commonly, distributions such as the log-logistic, or Gamma, are employed. After fitting the appropriate distribution, the values are then transformed into a Z-score, reflecting a normal distribution with a mean of 0 and a standard deviation of 1. Given the prominence of SPEI as the major exposure metric in this study, it is imperative that the method of calculating this index is delineated with clarity and precision in the paper.

2. "..There should be no such bias in the current analysis because our start of "follow-up" and start of exposure are both the start of pregnancy (time zero). Although each fetus had to survive through the entire pregnancy to be eligible to be included in the study sample,…"

What you're describing here aligns exactly with the concept of immortal time bias. In the context of your study, the fetus is 'immortal' from conception to birth, yet this period is still included as exposure time in the analysis. Actually, I believe the exposure period should start at birth, as fetal demise is not considered child mortality anyway. I would highly recommend you change the exposure start time for all the analyses in the paper.

Any attachments provided with reviews can be seen via the following link:

[LINK]

Decision Letter 2

Katrien G Janin

5 Jul 2024

Dear Dr Wang,

Many thanks for submitting your manuscript "The drier, the deadlier—the association between long-term drought and risk of infant mortality in Africa: A cross-sectional study" (PMEDICINE-D-24-00825R2) to PLOS Medicine. The paper has been reviewed by subject experts and a statistician; their comments are included below and can also be accessed here: [LINK]

Thank you for your detailed response to the editors' and reviewers' comments. I have discussed the paper with my colleagues, and it has also been seen again by the statistical original reviewer. The changes made to the paper were mostly satisfactory. However, the editorial team concurs with the statistical reviewer that issue around survival bias needs to be resolved and implemented as suggested by the reviewer. Therefore, we ask you to carefully address the comments in a further revision to preclude the need for further major revisions and satisfy the reviewers and editors. When submitting your revised paper, please again include a detailed point-by-point response to the comments.

When you upload your revision, please include a point-by-point response that addresses all of the reviewer and editorial points, indicating the changes made in the manuscript and either an excerpt of the revised text or the location (eg: page and line number) where each change can be found. Please also be sure to check the general editorial comments at the end of this letter and include these in your point-by-point response. When you resubmit your paper, please include a clean version of the paper as the main article file and a version with changes tracked as a marked-up manuscript. It may also be helpful to check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper.

We ask that you submit your revision by Jul 26 2024 11:59PM. However, if this deadline is not feasible, please contact me by email, and we can discuss a suitable alternative.

Don't hesitate to contact me directly with any questions (kjanin@plos.org).

Best regards,

Katrien

Katrien Janin, PhD

Associate Editor

PLOS Medicine

kjanin@plos.org

-----------------------------------------------------------

Comments from the reviewers:

Reviewer #3: I am very confused by the authors' response regarding the survival bias, particularly concerning the time-zero of the study. Previously, the authors stated, "...There should be no such bias in the current analysis because our start of 'follow-up' and start of exposure are both the start of pregnancy (time zero)...". However, in the current response, they said, "...However, in our study, our exposure starts from the start of pregnancy, and the time zero is birth (time zero is LATER than the start of exposure)," which contradicts their earlier statement.

If time zero is indeed at birth, then the follow-up time should start at birth. In that case, there would be no survival bias, and that's why I suggested changing the time zero to birth (instead of the start of pregnancy) for their survival analyses. However, if time zero is the start of pregnancy, the period between the start of pregnancy and birth is immortal time. It seems the authors may not fully understand immortal bias. Immortal bias occurs when follow-up time is counted during a period when the outcome of interest cannot happen. Misalignment between time zero and treatment assignment can cause immortal bias, but this is not the definition of survival bias. If the authors start counting time from the the start of pregnancy, then there is survival bias. The authors seem to not fully understand the concepts, and I am very confused by their contradictory statements...

For clarification, here is a definition of time zero and follow-up time (time-to-event): "Time zero, or the time origin, is the time at which participants are considered at-risk for the outcome of interest... Follow-up time is measured from time zero (the start of the study or from the point at which the participant is considered to be at risk) until the event occurs, the study ends, or the participant is lost, whichever comes first." (https://sphweb.bumc.bu.edu/otlt/mph-modules/bs/bs704_survival/BS704_Survival_print.html#:~:text=individual%20being%20followed.-,Follow%20Up%20Time,(i.e.%2C%20at%20enrollment )

Any attachments provided with reviews can be seen via the following link: [LINK]

--------------------------------------------------------- ---

Decision Letter 3

Katrien G Janin

10 Sep 2024

Dear Dr. Wang,

Many thanks for submitting your revised manuscript (PMEDICINE-D-24-00825R3) to PLOS Medicine. I’m contacting you on behalf of my colleague Katrien, who is no longer with the journal. The revised manuscript has been reviewed again by the statistical reviewer, and given the differing views of the original statistical reviewer and the authors, we also solicited the opinion of an independent statistician to contribute additional insight into the analysis and the question of immortal time bias. As you will see below (and attached), the original statistical reviewer and the arbitrating statistical reviewer (Reviewer 4) agree that the current statistical approach in the paper is problematic. Indeed, the arbitrating statistician does not feel that the analysis effectively accounts for the impact of the exposure (drought) during pregnancy (the reviewers’ comments can also be accessed here: [LINK]). More broadly, this reviewer does not feel that the conclusions as currently presented are sufficiently supported by the data.

In order to continue forward with the manuscript in view of the ongoing statistical questions, we must ask you to revised the paper again in response to the statisticians’ comments and the editorial points outlined below. Please note that the response and revisions must be satisfactory from a statistical standpoint for us to ultimately accept the paper for publication and that we will only consider one final revision.

As before, when you upload your revision, please include a point-by-point response that addresses all of the reviewer and editorial points, indicating the changes made in the manuscript and either an excerpt of the revised text or the location (eg: page and line number) where each change can be found. When you resubmit your paper, please include a clean version of the paper as the main article file and a version with changes tracked as a marked-up manuscript. It may also be helpful to check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper.

We ask that you submit your revision by October 1st. However, if this deadline is not feasible or you have any questions, please contact me directly by email (hvanepps@plos.org).

Kind regards,

Heather

Heather Van Epps, PhD

Executive Editor, PLOS Medicine

hvanepps@plos.org

[on behalf of]

Katrien Janin, PhD

Associate Editor

PLOS Medicine

kjanin@plos.org

-----------------------------------------------------------

Comments from Editors:

1. We appreciate the inclusion of the potential for immortal time bias as a limitation of the study. However, we feel that some additional consideration is needed in terms of the statistical approach given the stated importance of the pregnancy period as part of the exposure (noting that the editors agree with the notion that exposure to draught conditions during pregnancy could affect the developing fetus) and in view of the statistician’s comment that this exposure has not been accounted for in the model.

2. More detail is needed around the sensitivity analysis shown in figure S3, as noted by the arbitrating statistician.

3. Regarding the conclusions of the manuscript, I have discussed this with the editorial team, and we agree with the arbitrating statistical reviewer that the data presented in the paper do not support the strong conclusions. As noted by the statistician, in view of the results of the standard null hypothesis significance testing with a critical alpha of 0.05, the statement in the abstract should be amended to “Compared to children who did not experience drought, any drought exposure was *not* associated with an increased risk of infant mortality…” (or to the language suggested in the review). We also feel that the discordant result of the severe draught effect should be noted in the Abstract in the interest of full and transparent reporting.

4. The title of the manuscript should be changed to comply with PLOS Medicine style (it should not be declamatory) and to accurately reflect the data presented in the paper. We suggest “Long-term drought and risk of infant mortality in Africa: A cross-sectional study”.

5. We intend to send the revised paper to the statisticians for a final look. As noted above, it will be important for the paper to be acceptable from a statistical perspective for us to proceed to acceptance/publication.

-----------------------------------------------------------

Comments from the reviewers:

Reviewer #3:

I really appreciate the detailed explanations and the efforts the authors have made to revise the paper. However, I still do not agree with the decision to include the time between the start of pregnancy and the time of birth in the time-to-event outcome.

First, I fully agree with the authors' argument that "maternal exposure to drought during pregnancy is potentially associated with infant mortality through the mother." However, the time of exposure does not need to coincide with time-zero in a survival analysis. It can, but it does not have to. For example, you can examine the relationship between exposure to second-hand smoke before age 10 and "the risk of developing asthma after age 10." That's completely valid. However, the time-zero must start at age 10, not age 9, because the outcome is defined as "the risk of developing asthma <after> age 10." Any time before age 10 is "immortal time" because, by definition, it is impossible for the person to have the outcome.

For infant mortality, by definition, it is the death of a newborn <after birth> and before age 1. Therefore, "infant mortality" cannot occur before birth. Please note that the concept of time-zero in a survival analysis is related to the outcome (i.e., time to the event, which is infant mortality) rather than to the exposure. I believe the authors are still confused on this point.

It is encouraging to hear that using the time of birth as time-zero does not significantly change the results (in fact, by removing the immortal time period, I would expect the relationship to be stronger). Therefore, I recommend reporting those results using the time of birth as time-zero. Why do we have to introduce immortal time bias when it can be avoided?

Reviewer #4: See attachment

Michael Dewey

Any attachments provided with reviews can be seen via the following link: [LINK]

Attachment

Submitted filename: wang.pdf

pmed.1004516.s012.pdf (74.9KB, pdf)

Decision Letter 4

Heather Van Epps

6 Nov 2024

Hi Pin,

Thank you for getting back to me and apologies for the delay as I was out on vacation. Thank you also for updating the manuscript following our conversation. I have issued a major revision outcome so that you may upload your revised manuscript as well as any other supporting materials; the team will make a final decision based on this information and any additional consultation that we need.

Best wishes

Rebecca

Rebecca Kirk, MBA, PhD (she/her)

Associate Editorial Director

PLOS

Attachment

Submitted filename: wangrereview.pdf

pmed.1004516.s014.pdf (67.9KB, pdf)

Decision Letter 5

Heather Van Epps

12 Dec 2024

Dear Dr. Wang,

Thank you for re-submitting your manuscript "Long-term drought and risk of infant mortality in Africa: A cross-sectional study" (PMEDICINE-D-24-00825R5) and for your patience with what has been an unusually protracted process. We appreciate your continued willingness to engage with us and discuss the details of the statistical approach with the editors and statistical reviewers; this process has has resulted in increased clarity and improvements to the paper.

I have discussed the paper with my colleagues and one of the statistical reviewers, and I’m pleased to say that provided the remaining editorial and production issues are dealt with we are planning to accept the paper for publication in the journal. The remaining issues that need to be addressed are listed at the end of this email. Please take these into account before resubmitting your manuscript:

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

In revising the manuscript for further consideration here, please ensure you address the specific points made by the editors. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments and the changes you have made in the manuscript. Please submit a clean version of the paper as the main article file. A version with changes marked must also be uploaded as a marked up manuscript file.

Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. If you haven't already, we ask that you provide a short, non-technical Author Summary of your research to make findings accessible to a wide audience that includes both scientists and non-scientists. The Author Summary should immediately follow the Abstract in your revised manuscript. This text is subject to editorial change and should be distinct from the scientific abstract.

We expect to receive your revised manuscript within 1 week. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

Please note, when your manuscript is accepted, an uncorrected proof of your manuscript will be published online ahead of the final version, unless you've already opted out via the online submission form. If, for any reason, you do not want an earlier version of your manuscript published online or are unsure if you have already indicated as such, please let the journal staff know immediately at plosmedicine@plos.org.

In view of the upcoming holidays, we ask that you submit the final revision by Monday, January 6th. If you have any questions in the meantime, please contact me directly at hvanepps@plos.org. Otherwise, we look forward to receiving the revised manuscript..

Kind regards,

Heather

Heather Van Epps, PhD

Executive Editor

PLOS Medicine

hvanepps@plos.org

------------------------------------------------------------

Requests from Editors:

1. Abstract, line 67-68: please amend the wording from “…and a non-significant association with mild drought” to “but no significant association with mild drought”

2. Author summary: for full transparency, please include a bullet point to report the overarching finding that there was no evidence that any drought exposure was associated with an increased risk of infant mortality. It would make sense for this to precede the point about the association with severe drought.

3. Methods, line 163: please insert “et al” after the author name (Bendavid).

4. Results, line 276-278: please rephrase this sentence so that it’s consistent with the abstract. Ie, “Compared to children who did not experience drought, we did not find evidence that any drought exposure was associated with an increased risk of infant mortality (hazard ratio [HR]: 1.02; 95% confidence interval [CI] [1.00, 1.04], p=0.072).”

5. Results, line 281-283: please modify the following sentence to report lack of an association rather than a non-significant association. “A stronger but non-significant association with mild drought was estimated for postneonatal mortality (HR: 1.02; 95% CI [0.99, 1.06], p=0.203) than for neonatal mortality (HR: 0.99; 95 % CI [0.96, 1.02], p=0.657).”

6. Discussion: For completeness, we ask that you report the topline finding of no association between any drought exposure and infant mortality, as part of the initial paragraph of the Discussion.

7. References: please format the references according to PLOS style. In the text, references should be cited in square brackets (e.g., “We used the techniques developed by our colleagues [19] to analyze the data.” Full details on reference formatting can be found here: https://journals.plos.org/plosmedicine/s/submission-guidelines#loc-references.

8. Please remove the data availability statement from the main paper; this information will be included with the article meta-data (completed as part of the submission process).

9. Figure 1: Please confirm that the appropriate usage rights apply to the use of this map. Please see our guidelines for map images: https://journals.plos.org/plosmedicine/s/figures#loc-maps

10. Figure 2: please define all elements of box plots in the figure caption - center line, box limits and whiskers.

11. The supplemental file appears not to have been included in the most recent revision (these files are unfortunately not carried over from one revision to the next). Please ensure that a final version of the supplement is uploaded with this final revision.

Decision Letter 6

Heather Van Epps

19 Dec 2024

Dear Dr Wang, 

On behalf of my colleagues and the Academic Editor, Yuming Guo, I am pleased to inform you that we have agreed to publish your manuscript "Long-term drought and risk of infant mortality in Africa: A cross-sectional study" (PMEDICINE-D-24-00825R6) in PLOS Medicine.

Before your manuscript can be formally accepted you will need to complete some formatting changes, which you will receive in a follow up email. Please be aware that it may take several days for you to receive this email; during this time no action is required by you. Once you have received these formatting requests, please note that your manuscript will not be scheduled for publication until you have made the required changes.

In the meantime, please log into Editorial Manager at http://www.editorialmanager.com/pmedicine/, click the "Update My Information" link at the top of the page, and update your user information to ensure an efficient production process. 

PRESS

We frequently collaborate with press offices. If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximise its impact. If the press office is planning to promote your findings, we would be grateful if they could coordinate with medicinepress@plos.org. If you have not yet opted out of the early version process, we ask that you notify us immediately of any press plans so that we may do so on your behalf.

We also ask that you take this opportunity to read our Embargo Policy regarding the discussion, promotion and media coverage of work that is yet to be published by PLOS. As your manuscript is not yet published, it is bound by the conditions of our Embargo Policy. Please be aware that this policy is in place both to ensure that any press coverage of your article is fully substantiated and to provide a direct link between such coverage and the published work. For full details of our Embargo Policy, please visit http://www.plos.org/about/media-inquiries/embargo-policy/.

Thank you again for submitting to PLOS Medicine. We look forward to publishing your paper and hope that you will consider us for future publications.

Kind regards,

Heather

Heather Van Epps, PhD 

Senior Editor 

PLOS Medicine

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    S1 STROBE Checklist. STROBE Statement—Checklist of items that should be included in reports of cross-sectional studies.

    (DOCX)

    pmed.1004516.s001.docx (32.7KB, docx)
    S1 Method. Model specification of main and secondary analysis.

    (DOCX)

    pmed.1004516.s002.docx (24.3KB, docx)
    S1 Table. Descriptive statistics for the number of drought months experienced by all included children by exposure window, year of birth, climate zone, and drought severity.

    (DOCX)

    pmed.1004516.s003.docx (28.9KB, docx)
    S2 Table. Number of infant deaths (among 850,924 children) during each month within 1 year of age and number of months for SPEI-24 any drought experienced before and after birth.

    SPEI: standardized precipitation evapotranspiration index.

    (DOCX)

    pmed.1004516.s004.docx (28.1KB, docx)
    S3 Table. Association between long-term drought and risk of infant mortality in various sensitivity analyses.

    (DOCX)

    pmed.1004516.s005.docx (26.1KB, docx)
    S4 Table. Estimates and uncertainties of drought exposure before and after birth when including gestational drought exposure in the sensitivity analysis of postnatal period only.

    (DOCX)

    pmed.1004516.s006.docx (26.2KB, docx)
    S1 Fig. Illustration of calculation of SPEI at a timescale of 24 months for a given month from the start of pregnancy through death or the 12th month after birth.

    Climate variables include maximum and minimum temperatures, dew point temperature, wind speed, solar radiation, and air pressure. SPEI: standardized precipitation evapotranspiration index.

    (DOCX)

    pmed.1004516.s007.docx (74.3KB, docx)
    S2 Fig. Associations between neonatal and post-neonatal mortality and drought exposure by month of pregnancy.

    (DOCX)

    pmed.1004516.s008.docx (66KB, docx)
    Attachment

    Submitted filename: Response to reviewers_revised.docx

    pmed.1004516.s009.docx (38.7KB, docx)
    Attachment

    Submitted filename: Response_to_reviewers_revised_auresp_1.docx

    pmed.1004516.s010.docx (38.6KB, docx)
    Attachment

    Submitted filename: Response to reviewers_2nd.docx

    pmed.1004516.s011.docx (38.9KB, docx)
    Attachment

    Submitted filename: Response to reviewers_3rd.docx

    pmed.1004516.s013.docx (61.2KB, docx)
    Attachment

    Submitted filename: wang.pdf

    pmed.1004516.s012.pdf (74.9KB, pdf)
    Attachment

    Submitted filename: Response to reviewers_4th.docx

    pmed.1004516.s015.docx (34.6KB, docx)
    Attachment

    Submitted filename: wangrereview.pdf

    pmed.1004516.s014.pdf (67.9KB, pdf)
    Attachment

    Submitted filename: Response to reviewers_5th.docx

    pmed.1004516.s016.docx (132.2KB, docx)
    Attachment

    Submitted filename: Response to editors_6th.docx

    pmed.1004516.s017.docx (31.5KB, docx)

    Data Availability Statement

    Data on infant mortality and socioeconomic status in this study were collected from the Demographic and Health Surveys program (https://dhsprogram.com/), which are freely available upon request. Publicly available meteorological records were obtained from the fifth generation ECMWF atmospheric reanalysis of the global climate (ERA5-Land, https://cds.climate.copernicus.eu/).

    Articles from PLOS Medicine are provided here courtesy of PLOS

    RESOURCES