Figures and data in Relationship between changing malaria burden and low birth weight in sub-Saharan Africa: A difference-in-differences study via a pair-of-pairs approach

Figures
Tables
Additional files

3 figures, 11 tables and 2 additional files

Figures

Figure 1

Download asset Open asset

Figure 2

Download asset Open asset

Formed quadruples (pairs of pairs) of matched high-low and high-high pairs of clusters.

In Step 1, pairs of clusters from the early and late time periods are matched on geographic proximity and categorized as ‘high-high’ (comparison, or control) or ‘high-low’ (treated). In Step 2, pairs of high-high clusters are matched with pairs of high-low clusters based on cluster-level sociodemographic characteristics. The difference-in-differences estimate of the coefficient of changing malaria burden on the low birth weight rate is based on comparing (**D–C**) to (**B–A**).

Figure 3

Download asset Open asset

The estimated low birth weight rate of each cluster within the 219 high-high pairs and the 219 high-low pairs.

The estimated low birth weight rate for each cluster are obtained from averaging over all the 500 imputed data sets of the 18,112 individual records. We draw a line to connect two paired clusters (one early year cluster and one late year cluster). Box plots for the low birth weight rates are also shown. Two of the four outliers of the late year clusters among the high-low pairs (i.e. the top four late year clusters in terms of low birth weight rate among the high-low pairs) may result from their extremely small within-cluster sample sizes (no more than three individual records for both two clusters).

Tables

Table 1

The 19 selected sub-Saharan African countries along with their chosen early/late years of malaria prevalence (i.e. estimated parasite rate ${𝑃𝑓PR}_{2 - 10}$ ) and IPUMS-DHS early/late years.

Note that some DHS span over two successive years.

Country	Malaria prevalence		IPUMS-DHS
Country	Early year	Late year	Early year	Late year
Benin	2001	2012	2001	2011–12
Burkina Faso	2003	2010	2003	2010
Cameron	2004	2011	2004	2011
Congo Democratic Republic	2007	2013	2007	2013–14
Cote d’Ivoire	2000	2012	1998–99	2011–12
Ethiopia	2000	2010	2000	2010–11
Ghana	2003	2014	2003	2014
Guinea	2005	2012	2005	2012
Kenya	2003	2014	2003	2014
Malawi	2000	2010	2000	2010
Mali	2001	2012	2001	2012–13
Namibia	2000	2013	2000	2013
Nigeria	2003	2013	2003	2013
Rwanda	2005	2014	2005	2014–15
Senegal	2005	2010	2005	2010–11
Tanzania	2000	2015	1999	2015–16
Uganda	2000	2011	2000–01	2011
Zambia	2007	2013	2007	2013–14
Zimbabwe	2005	2015	2005–06	2015

Table 2

Summary of the Bayesian logistic regression model fitted over records with observed birth weight which is used to predict missing low birth weight indicators.

Predictor	Posterior mean	Posterior std	z-score	p-value
(Intercept)	1.916	0.628	3.051	0.002**
Mother’s age (linear term)	−0.207	0.045	−4.562	<0.001***
Mother’s age (quadratic term)	0.003	0.001	3.987	<0.001***
Wealth index	0.060	0.037	1.591	0.112
Child’s birth order (linear term)	−0.989	0.338	−2.925	0.003**
Child’s birth order (quadratic term)	0.211	0.086	2.447	0.014*
0 - rural; 1 - urban	0.126	0.103	1.214	0.225
Mother’s education level	−0.226	0.062	−3.633	<0.001***
Child is boy	−0.068	0.083	−0.815	0.415
Mother is married or living together	−0.173	0.117	−1.482	0.138
Indicator of antenatal care	−0.046	0.093	−0.493	0.622
Indicator of low birth size	2.410	0.090	26.776	<0.001***
Indicator of large birth size	−1.387	0.129	−10.786	<0.001***

Table 3

An interpretation of the coefficients of the intercept term and the three indicators defined in model (1) (i.e. the $k_{0}, k_{1}, k_{2}, k_{3}$ ) within each matched quadruple.

The coefficient of the low malaria prevalence indicator (i.e. the k₁) incorporates the information of the magnitude of the effect of changing malaria burden (from high to low) on the low birth weight rate.

Cluster	Prevalence	Time	Pair	Coefficients	Within-pair	Between-pair
Cluster	Prevalence	Time	Pair	Coefficients	Contrast	Contrast
1	High	Early	High-low	$k_{0} + k_{3}$	$k_{1} + k_{2}$	$k_{1}$
2	Low	Late	High-low	$k_{0} + k_{1} + k_{2} + k_{3}$	$k_{1} + k_{2}$
3	High	Early	High-high	$k_{0}$	$k_{2}$
4	High	Late	High-high	$k_{0} + k_{2}$	$k_{2}$

Table 4

The mean Haversine distance of the early year clusters and late year clusters is 24.1 km among the 219 high-low pairs of clusters, and 28.7 km among the 219 high-high pairs of clusters.

The within-pair longitudes’ and latitudes’ correlations between the paired early year and late year clusters among the high-low and high-high pairs all nearly equal one. The mean values of the longitudes, the latitudes, the annual malaria prevalence (i.e. ${𝑃𝑓PR}_{2 - 10}$ ) measured at the early year, denoted as ${𝑃𝑓PR}_{2 - 10}$ (early), and at the late year, denoted as ${𝑃𝑓PR}_{2 - 10}$ (late), of the paired early year clusters (clusters sampled at the early year) and late year clusters (clusters sampled at the late year) among the 219 high-low and 219 high-high pairs of clusters used for the statistical inference respectively. Note that an early year cluster has a late year ${𝑃𝑓PR}_{2 - 10}$ and a late year cluster has an early year ${𝑃𝑓PR}_{2 - 10}$ since the MAP data contain ${𝑃𝑓PR}_{2 - 10}$ for each location and for each year between 2000 and 2015.

	High-low pairs		High-high pairs
Mean within-pair haversine distance	24.1 km		28.7 km
Within-pair correlation of longitude	0.9999		0.9996
Within-pair correlation of latitude	0.9998		0.9997
	Longitude	Latitude	${𝑃𝑓PR}_{2 - 10}$ (early)	${𝑃𝑓PR}_{2 - 10}$ (late)
Early clusters among high-low pairs	16.92	−1.15	0.52	0.17
Late clusters among high-low pairs	16.88	−1.15	0.48	0.12
Early clusters among high-high pairs	19.15	0.43	0.51	0.47
Late clusters among high-high pairs	19.13	0.46	0.53	0.49

Table 5

Balance of each covariate before matching (BM) and after matching (AM).

We report the mean of each covariate (including early and late years) for high-low and high-high pairs of clusters, before and after matching. We also report each absolute standardized difference (Std.dif) before and after matching.

	Before matching		After matching		Std.dif
	High-low	High-high	High-low	High-high	BM	AM
	(410 pairs)	(540 pairs)	(219 pairs)	(219 pairs)
Urban/rural (early)	0.44	0.20	0.26	0.26	0.53	0.00
Urban/rural (late)	0.60	0.21	0.37	0.32	0.85	0.09
Toilet facility (early)	0.88	0.60	0.82	0.79	0.86	0.10
Toilet facility (late)	0.94	0.69	0.90	0.88	0.90	0.10
Floor material (early)	1.90	1.68	1.60	1.67	0.31	0.10
Floor material (late)	2.22	1.79	1.92	1.87	0.59	0.07
Electricity (early)	0.36	0.12	0.17	0.16	0.70	0.02
Electricity (late)	0.54	0.18	0.33	0.30	0.99	0.10
Mother’s education (early)	1.00	0.36	0.69	0.64	1.36	0.10
Mother’s education (late)	1.23	0.42	0.87	0.83	1.78	0.10
Contraception indicator (early)	0.16	0.12	0.15	0.17	0.27	0.10
Contraception indicator (late)	0.22	0.18	0.24	0.26	0.23	0.10

Table 6

Inference with multiple imputation and mixed-effects linear probability model (1).

The unit of estimates and CIs is a percentage point.

Regressor	Estimate	95% CI	p-value
0 - high prevalence; 1 - low prevalence	−1.48	[−3.70, 0.74]	0.191
0 - early year; 1 - late year	−0.06	[−1.82, 1.69]	0.943
0 - high-high pairs; 1 - high-low pairs	0.21	[−1.40, 1.82]	0.797
Mother’s age (linear term)	−1.86	[−2.48, −1.23]	<0.001***
Mother’s age (quadratic term)	0.03	[0.02, 0.04]	<0.001***
Child’s birth order (linear term)	−13.91	[−18.49, −9.32]	<0.001***
Child’s birth order (quadratic term)	2.91	[1.82, 4.00]	<0.001***
Wealth index	0.09	[−0.38, 0.56]	0.709
0 - rural; 1 - urban	0.82	[−0.63, 2.27]	0.269
Mother’s education level	−2.02	[−2.82, −1.22]	<0.001***
Child is boy	−1.75	[−2.75, −0.74]	<0.001***
Mother is married or living together	−1.43	[−3.04, 0.19]	0.083
Antenatal care indicator	−0.96	[−2.06, 0.13]	0.085

Appendix 1—table 1

The early and late years coded in the IPUMS-DHS and GPS data sets.

	GPS data		Malaria prevalence		IPUMS-DHS
	Early	Late	Early	Late	Early	Late
Benin	2001	2012	2001	2012	2001	2011
Burkina Faso (BF)	2003	2010	2003	2010	2003	2010
Cameron (CM)	2004	2011	2004	2011	2004	2011
Congo Democratic Republic (CD)	2007	2013	2007	2013	2007	2013
Cote d’Ivoire (CI)	1998	2012	2000	2012	1998	2011
Ethiopia (ET)	2000	2010	2000	2010	2000	2011
Ghana (GH)	2003	2014	2003	2014	2003	2014
Guinea (GN)	2005	2012	2005	2012	2005	2012
Kenya (KE)	2003	2014	2003	2014	2003	2014
Malawi (MW)	2000	2010	2000	2010	2000	2010
Mali (ML)	2001	2012	2001	2012	2001	2012
Namibia (NM)	2000	2013	2000	2013	2000	2013
Nigeria (NG)	2003	2013	2003	2013	2003	2013
Rwanda (RW)	2005	2014	2005	2014	2005	2014
Senegal (SN)	2005	2010	2005	2010	2005	2010
Tanzania (TZ)	1999	2015	2000	2015	1999	2015
Uganda (UG)	2000	2011	2000	2011	2001	2011
Zambia (ZM)	2007	2013	2007	2013	2007	2013
Zimbabwe (ZW)	2005	2015	2005	2015	2005	2015

Appendix 1—table 2

The numbers of the high-high pairs of clusters and high-low pairs of clusters contributed by each of the 19 selected sub-Saharan African countries after the matching in Step 1 and Step 2.

We also summarize the total number of pairs of clusters after Step 1 matching in the first column.

Country	Step 1 matching			Step 2 matching
Country	Total pairs	High-high	High-low	High-high	High-low
Benin	247	29	6	4	6
Burkina Faso	400	150	0	19	0
Cameron	466	17	163	16	51
Congo Democratic Republic	300	11	55	11	24
Cote d’Ivoire	140	19	2	7	2
Ethiopia	539	0	0	0	0
Ghana	412	24	18	18	8
Guinea	295	47	12	10	12
Kenya	400	2	10	2	8
Malawi	560	96	15	81	15
Mali	402	101	21	17	19
Namibia	260	0	0	0	0
Nigeria	362	24	11	16	1
Rwanda	462	0	0	0	0
Senegal	376	0	0	0	0
Tanzania	176	0	68	0	57
Uganda	298	19	29	17	16
Zambia	319	1	0	1	0
Zimbabwe	398	0	0	0	0
Total	6812	540	410	219	219

Appendix 1—table 3

Summary of the low malaria prevalence indicators, the time indicators, the group indicators, the covariates, and the birth weight records among the 18,112 study individual records.

Variables	Percentages of some categories
Low malaria prevalence indicator	High prevalence (70.6%)
	Low prevalence (29.4%)
Time indicator	Early year (50.3%)
	Late year (49.7%)
Group indicator	High-high pairs (40.9%)
	High-low pairs (59.1%)
Mother’s age in years	≤19 (7.1%)
	20–29 (52.5%)
	30–39 (31.4%)
	≥40 (8.9%)
Wealth index	Poorest (20.2%)
	Poorer (23.3%)
	Middle (22.8%)
	Richer (20.4%)
	Richest (13.3%)
Child’s birth order	1 (21.5%)
	2–4 (46.0%)
	4+ (32.6%)
Urban or rural	Rural (77.1%)
	Urban (22.9%)
Mother’s education level	No education (36.6%)
	Primary (47.2%)
	Secondary or higher (16.2%)
Child’s sex	Female (49.3%)
	Male (50.7%)
Mother’s marital status	Never married or formerly in union (11.6%)
	Married or living together (88.4%)
Indicator of antenatal care	Yes (61.9%)
	No or missing (38.1%)
Self-reported birth size	Very small or smaller than average (13.0%)
	Average (45.5%)
	Larger than average or very large (41.5%)
Low birth weight indicator	Yes (4.6%)
	No (48.5%) or Missing (47.0%)

Appendix 2—table 1

Diagnostics for multiple imputation with the mixed-effects linear probability model.

We report the between-imputation variance (`Between var’), the within-imputation variance (‘Within var’), and the variance ratio: (between-imputation variance)/(within-imputation variance), denoted as `Var ratio’.

Regressor	Between var	Within var	Var ratio
0 - high prevalence; 1 - low prevalence	3.21 × 10⁻⁵	9.62 × 10⁻⁵	0.334
0 - early year; 1 - late year	2.20 × 10⁻⁵	5.81 × 10⁻⁵	0.379
0 - high-high pairs; 1 - high-low pairs	1.92 × 10⁻⁵	4.83 × 10⁻⁵	0.398
Mother’s age (linear term)	3.32 × 10⁻⁶	6.85 × 10⁻⁶	0.486
Mother’s age (quadratic term)	8.28 × 10⁻¹⁰	1.68 × 10⁻⁹	0.493
Child’s birth order (linear term)	1.60 × 10⁻⁴	3.87 × 10⁻⁴	0.413
Child’s birth order (quadratic term)	8.55 × 10⁻⁶	2.24 × 10⁻⁵	0.382
Wealth index	1.74 × 10⁻⁶	4.05 × 10⁻⁶	0.430
0 -rural; 1 - urban	1.27 × 10⁻⁵	4.21 × 10⁻⁵	0.303
Mother’s education level	4.56 × 10⁻⁶	1.20 × 10⁻⁵	0.380
Child is boy	7.12 × 10⁻⁶	1.91 × 10⁻⁵	0.373
Mother is married or living together	1.83 × 10⁻⁵	4.96 × 10⁻⁵	0.370
Antenatal care indicator	9.63 × 10⁻⁶	2.16 × 10⁻⁵	0.447

Appendix 3—table 1

The results of the sensitivity analyses for the coefficient of the low malaria prevalence indicator under various sensitivity parameters $(p_{1}, p_{2})$ divided into the four cases: Case 1: $p_{1} > 0, p_{2} > 0$ ; Case 2: $p_{1} > 0, p_{2} < 0$ ; Case 3: $p_{1} < 0, p_{2} > 0$ ; Case 4: $p_{1} < 0, p_{2} < 0$ .

The unit of estimates and CIs is a percentage point.

Case 1	$p_{2} = 5.0$			$p_{2} = 10.0$
Case 1	Estimate	95% CI	p-value	Estimate	95% CI	p-value
$p_{1} = 2.5$	$- 1.52$	$[- 3.74, 0.70]$	0.179	$- 1.56$	$[- 3.77, 0.66]$	0.168
$p_{1} = 5.0$	$- 1.57$	$[- 3.79, 0.66]$	0.167	$- 1.65$	$[- 3.86, 0.57]$	0.145
$p_{1} = 7.5$	$- 1.61$	$[- 3.83, 0.61]$	0.156	$- 1.73$	$[- 3.95, 0.48]$	0.125
$p_{1} = 10.0$	$- 1.65$	$[- 3.88, 0.57]$	0.145	$- 1.82$	$[- 4.04, 0.40]$	0.107
Case 2	$p_{2} = - 5.0$			$p_{2} = - 10.0$
Case 2	Estimate	95% CI	p-value	Estimate	95% CI	p-value
$p_{1} = 2.5$	$- 1.44$	$[- 3.66, 0.78]$	0.204	$- 1.39$	$[- 3.62, 0.83]$	0.219
$p_{1} = 5.0$	$- 1.40$	$[- 3.62, 0.83]$	0.218	$- 1.31$	$[- 3.53, 0.92]$	0.249
$p_{1} = 7.5$	$- 1.35$	$[- 3.58, 0.87]$	0.234	$- 1.22$	$[- 3.44, 1.00]$	0.282
$p_{1} = 10.0$	$- 1.31$	$[- 3.53, 0.92]$	0.250	$- 1.13$	$[- 3.36, 1.09]$	0.318
Case 3	$p_{2} = 5.0$			$p_{2} = 10.0$
Case 3	Estimate	95% CI	p-value	Estimate	95% CI	p-value
$p_{1} = - 2.5$	$- 1.44$	$[- 3.66, 0.78]$	0.204	$- 1.39$	$[- 3.61, 0.83]$	0.219
$p_{1} = - 5.0$	$- 1.39$	$[- 3.61, 0.83]$	0.219	$- 1.30$	$[- 3.52, 0.91]$	0.249
$p_{1} = - 7.5$	$- 1.35$	$[- 3.57, 0.87]$	0.234	$- 1.22$	$[- 3.43, 1.00]$	0.282
$p_{1} = - 10.0$	$- 1.31$	$[- 3.53, 0.92]$	0.250	$- 1.13$	$[- 3.35, 1.09]$	0.319
Case 4	$p_{2} = - 5.0$			$p_{2} = - 10.0$
Case 4	Estimate	95% CI	p-value	Estimate	95% CI	p-value
$p_{1} = - 2.5$	$- 1.52$	$[- 3.75, 0.70]$	0.179	$- 1.56$	$[- 3.79, 0.66]$	0.168
$p_{1} = - 5.0$	$- 1.57$	$[- 3.79, 0.66]$	0.167	$- 1.65$	$[- 3.87, 0.57]$	0.146
$p_{1} = - 7.5$	$- 1.61$	$[- 3.84, 0.61]$	0.156	$- 1.74$	$[- 3.96, 0.49]$	0.126
$p_{1} = - 10.0$	$- 1.66$	$[- 3.88, 0.57]$	0.145	$- 1.83$	$[- 4.05, 0.40]$	0.108

Additional files

Source code 1 The source code for producing the results in Figure 1, the results in Figure 3, the results in Tables 2, 3, 5 and 6, the results in Table 4, the results in Appendix 1—table 2, the results in Appendix 1 Table 3, and the results in Appendix 3—table 1 can be found respectively in 'Code for Figure 1.R', 'Code for Figure 3.R', 'Code for primary analysis.R', 'Code for Table 4.R', 'Code for Appendix 1 Table 2.R', 'Code for Appendix 1 Table 3.R', and 'Code for Sensitivity Analyses.R' in the source code files. The source code are also posted on GitHub (https://github.com/siyuheng/Malaria-and-Low-Birth-Weight; Heng, 2021a).: https://cdn.elifesciences.org/articles/65133/elife-65133-code1-v1.zip
Download elife-65133-code1-v1.zip
Transparent reporting form: https://cdn.elifesciences.org/articles/65133/elife-65133-transrepform-v1.docx
Download elife-65133-transrepform-v1.docx

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Article PDF

Open citations (links to open the citations from this article in various online reference manager services)

Mendeley

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

Siyu Heng
Wendy P O'Meara
Ryan A Simmons
Dylan S Small

(2021)

Relationship between changing malaria burden and low birth weight in sub-Saharan Africa: A difference-in-differences study via a pair-of-pairs approach

eLife 10:e65133.

https://doi.org/10.7554/eLife.65133

Figures

Work flow diagram of the study.

Formed quadruples (pairs of pairs) of matched high-low and high-high pairs of clusters.

The estimated low birth weight rate of each cluster within the 219 high-high pairs and the 219 high-low pairs.

Tables

The 19 selected sub-Saharan African countries along with their chosen early/late years of malaria prevalence (i.e. estimated parasite rate ${𝑃𝑓PR}_{2 - 10}$ ) and IPUMS-DHS early/late years.

Summary of the Bayesian logistic regression model fitted over records with observed birth weight which is used to predict missing low birth weight indicators.

An interpretation of the coefficients of the intercept term and the three indicators defined in model (1) (i.e. the $k_{0}, k_{1}, k_{2}, k_{3}$ ) within each matched quadruple.

The mean Haversine distance of the early year clusters and late year clusters is 24.1 km among the 219 high-low pairs of clusters, and 28.7 km among the 219 high-high pairs of clusters.

Balance of each covariate before matching (BM) and after matching (AM).

Inference with multiple imputation and mixed-effects linear probability model (1).

The early and late years coded in the IPUMS-DHS and GPS data sets.

The numbers of the high-high pairs of clusters and high-low pairs of clusters contributed by each of the 19 selected sub-Saharan African countries after the matching in Step 1 and Step 2.

Summary of the low malaria prevalence indicators, the time indicators, the group indicators, the covariates, and the birth weight records among the 18,112 study individual records.

Diagnostics for multiple imputation with the mixed-effects linear probability model.

Additional files

Source code 1

Transparent reporting form

Download links

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

Be the first to read new articles from eLife

Share this article

Cite this article

Work flow diagram of the study.

Formed quadruples (pairs of pairs) of matched high-low and high-high pairs of clusters.

The estimated low birth weight rate of each cluster within the 219 high-high pairs and the 219 high-low pairs.

The 19 selected sub-Saharan African countries along with their chosen early/late years of malaria prevalence (i.e. estimated parasite rate 𝑃𝑓PR2-10) and IPUMS-DHS early/late years.

Summary of the Bayesian logistic regression model fitted over records with observed birth weight which is used to predict missing low birth weight indicators.

An interpretation of the coefficients of the intercept term and the three indicators defined in model (1) (i.e. the k0,k1,k2,k3) within each matched quadruple.

The mean Haversine distance of the early year clusters and late year clusters is 24.1 km among the 219 high-low pairs of clusters, and 28.7 km among the 219 high-high pairs of clusters.

Balance of each covariate before matching (BM) and after matching (AM).

Inference with multiple imputation and mixed-effects linear probability model (1).

The early and late years coded in the IPUMS-DHS and GPS data sets.

The numbers of the high-high pairs of clusters and high-low pairs of clusters contributed by each of the 19 selected sub-Saharan African countries after the matching in Step 1 and Step 2.

Summary of the low malaria prevalence indicators, the time indicators, the group indicators, the covariates, and the birth weight records among the 18,112 study individual records.

Diagnostics for multiple imputation with the mixed-effects linear probability model.

The results of the sensitivity analyses for the coefficient of the low malaria prevalence indicator under various sensitivity parameters (p1,p2) divided into the four cases: Case 1: p1>0,p2>0; Case 2: p1>0,p2<0; Case 3: p1<0,p2>0; Case 4: p1<0,p2<0.

Source code 1

Transparent reporting form

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

The 19 selected sub-Saharan African countries along with their chosen early/late years of malaria prevalence (i.e. estimated parasite rate ${𝑃𝑓PR}_{2 - 10}$ ) and IPUMS-DHS early/late years.

An interpretation of the coefficients of the intercept term and the three indicators defined in model (1) (i.e. the $k_{0}, k_{1}, k_{2}, k_{3}$ ) within each matched quadruple.