Malaria transmission in Africa: Its relationship with yellow fever and measles.

BACKGROUND: Malaria has been strongly linked to the transmission and pathophysiology of some viral diseases. Malaria and vaccine-preventable diseases often co-exist in endemic countries but the implication of their co-existence on their transmission dynamics and control is poorly understood. The study aims to evaluate the relationships between the incidence of malaria and cases of measles and yellow fever in Africa.
METHODS: The malaria incidence, death due to malaria, measles and yellow fever data were sourced from the WHO database. Poisson and zero-inflated time-trend regression were used to model the relationships between malaria and the two vaccine-preventable diseases. P-values <0.05 were considered statistically significant.
RESULTS: A significant negative relationship existed between malaria incidence and measles cases (P<0.05), however, malaria showed a positive relationship with yellow fever (P<0.05). The relationships between death due to malaria and measles/yellow fever cases followed similar trends but with a higher level of statistical significance (P<0.001).
CONCLUSIONS: Malaria varied negatively with measles cases but positively with yellow fever. The relationships observed in this study could be important for the management of malaria and the studied vaccine-preventable diseases. Increase vaccination coverage and/or malaria treatment could modulate the direction of these relationships.

Introduction

Malaria is the most important parasitic disease in the tropical and sub-tropical regions of the world. In 2020, the World Health Organization (WHO) reported 241 million cases and 627 thousand deaths from malaria [1]; this is an increase from an estimated 227 million cases and 558 thousand deaths recorded in 2019 [2]. Africa is the most affected region of the world; accounting for more than 95% malaria cases and associated burdens, and deaths [2]. Currently, the public health intervention measures recommended by the WHO to control and prevent the transmission of malaria include; prompt detection of infection, treatment of clinical and severe malaria, vector control, and chemoprevention in high-risk groups [3].

Yellow fever like malaria, is a vector-borne disease caused by the yellow fever virus of the genus Flavivirus [3]. With estimated 51 000 to 380 000 severe cases of yellow fever, and 19 000 to 180 000 deaths per year in Africa [4], the continent is the epicenter of the disease recording 90% of the total global cases of the disease [5]. Approximately half of the infected populations are asymptomatic and a further third suffer from a mild illness. Only about 12% of yellow fever infections present a severe disease in the form of jaundice and/or haemorrhage with 30–60% death rate [6]. Measles, like malaria and yellow fever, is another disease with serious public health implications for children. Although the global transmission of measles has been greatly reduced by vaccination about 60 years ago, it continues to be one of the major causes of mortality among children in sub-Saharan Africa [7].

Previous studies have related malaria with some viral infections, especially HIV and hepatitis [8-10], and these studies suggested the possible impact of malaria on the dynamics of transmission and morbidities associated with viral infections. Despite the dreadfulness of malaria and viral agents in the current study, studies on concomitant presentation of these diseases are very rare. Although few epidemiological studies are available for malaria and yellow fever [11,12], none is available for malaria and measles. Other studies involving the triads included the effect of chloroquine prophylaxis on yellow fever virus replication and vaccine antibody response [13,14], and impact of placental malaria on infant immunity to measles [15]. An experimental data indicated that chloroquine may inhibit yellow fever virus in vitro [13], but the only available clinical evidence showed that chloroquine preventive treatment against malaria had no effect on antibody response to yellow fever vaccination [14]. In a similar vein, malaria during pregnancy was reported to impair the transplacental transfer of measles antibody, thus, suggesting that malaria control during pregnancy could have substantial benefits in the mother and the baby including reduction in the risk of measles in early infancy [15].

Considering the similar geographical distributions between malaria and vaccine preventable diseases (yellow fever virus and measles) and the possible associations that could exist between them, analysis of relationships between them using robust secondary data could provide more information on the trend of these relationships. The relationships could inform policy on what approach could be best adopted for the various preventive or interventional programmes related to these diseases.

The aim of the current study is therefore to determine the magnitude and direction of the trend of malaria incidence, death due to malaria, measles cases and yellow fever cases in selected African countries. Also to investigate the effect of measles and yellow fever cases on malaria incidence and death due to malaria, respectively.

Material and methods

Data and scope

Available data on malaria incidence, death due to malaria, measles and yellow fever was downloaded from public domain (https://www.who.int/data/gho) for all African countries. Countries with a well filled data were arranged in a separate Microsoft excel sheet for each variable which produced 47 countries for malaria incidence, 46 countries for death due to malaria, 54 countries for measles cases and 22 countries for yellow fever cases. Each of the Microsoft excel sheets was imported to STATA software package for trend analysis. Prior to estimation of regression model, the data was merged into two STATA data file. The first file contains malaria incidence, measles and yellow fever while the second contain death due to malaria, measles cases and yellow fever cases. We ensured that only information with uniform country and time were retained in the merged files. The malaria incidence and death due to malaria files used regression had 49 countries (2000–2018) and 46 countries (2000–2017), respectively.

Regression analyses

Count data are known to be discrete variables and not continuous, thus, the traditional regression model that assumes that outcome or response variable comes from a normal distribution is not tenable bearing in mind that discrete variable is highly skewed. This study used a discrete probability distribution that treats the response variable of count observation as a Poisson process with probability mass function given as

This bound indicates that the distribution takes on nonnegative random numbers. The quantity λ is the mean of the Poisson distribution. This distribution has a unique property of equi-dispersion that the mean, λ = E(y) is equal to variance, V(y). However, the assumption of equi-dispersion is highly restrictive and hardly satisfied in real-life application as shown here that the mean each variable considered is not equal to the variance. Two major problems that are associated with Poisson regression are over-dispersion and under-dispersion which connote situations where V(y) > E(y) and V(y) < E(y), respectively. Sellers and Shmueli [16] provided a more general count distribution that captured a wide range of dispersion which they titled Conway–Maxwell-Poisson (COM-Poisson) and Huang [17] provided a COM-Poisson Double Generalized Linear Model that addressed the problem of both under and overdispersion simultaneous.

Poisson regression is achieved by expressing λ in Eq (1) as a function of some explanatory variables through a log link function. This can be mathematically expressed as:

Given the density function of Eq (1) type and set of independent variable X, the maximum likelihood estimates of regression parameter β is defined as;

Count data most often have excess zero cases and if not adjusted for can lead to inaccurate regression estimate and misleading conclusion. Lambert [18] developed a zero-inflated Poisson (ZIP) regression that corrects for the occurrence of excess zero cases in count data. The ZIP has become a popular choice in empirical problems relating to count data. The ZIP model is often appealing, first, since it divides the dependent variable into two subpopulations where the first one takes on the value of zero with probability π while the second one is Poisson distributed with probability 1 − π [19]. Due to this setup, the ZIP model may be used when the data contain an excess amount of zeros. Second, this model is popular because it accounts for overdispersion (meaning that the variance of the dependent variable exceeds the mean value), which leads to an underestimation of the variance of the estimated coefficients when applying the standard Poisson regression model [19]. The ZIP divides the count data into two non-overlapping subpopulations using the equation:

In this expression q is the ith row of the Q, which is the data matrix for the logit model and γ corresponds to a (p + 1) × 1 vector of coefficients. Also, , where x is the ith row of X, which is the data matrix for the ZIP model while β corresponds to a (p + 1) × 1 vector of coefficients. The most common method of estimating β is to apply the ML method. By defining 1(yi = 0) as an indicator function that takes on the value of 1 if y = 0, the following joint likelihood should then be maximized:

This complex likelihood function is estimated through the modification of the simplex method of Nelder and Mead [20] proposed by Kibria et al. [19] in which the start-up values come from the iterative weighted least-squares algorithm of the individual Poisson and logit estimation. The modification to this method applied the estimator specified in (6).

is the estimate obtained using the simplex method. Furthermore, the shrinkage parameter k may take on values between zero and infinity, and when k equals zero, we have . When k is greater than zero, we have . Since is, on average, too long in the presence of multicollinearity, is expected to perform better than .

Statistical analysis

The statistical analysis was carried out using five statistical software packages. SPSS version 23 was used for data management, cleaning and descriptive analysis. STATA version 13 and R version 4.0.3 were used to estimate the Poisson and zero-inflated Poisson (ZIP) regression model. Because all the studied variables are cases and rates which are regarded as count data, the Panel Poisson regression model was adopted to understudy these relationships. Where there are observed excess zero in the cases of event considered in the study, the zero-inflated Poisson (ZIP) regression that is robust to the occurrence of zero was adopted to achieve our specified objectives. GeoDA 1.9 was used to prepare the shapefile and merging of the estimated annual average change (EAAC) while QGIS version 3.14 was used for geospatial mapping of the EAAC of each variable. P-values <0.05 were considered statistically significant.

Results

Time trend Poisson and zero-inflated Poisson model was fitted for malaria incidences (47 countries), death due to malaria (46 countries), measles cases (54 countries) and yellow fever cases (22 countries). Poisson regression was used for 46(97.87%), 46(100.0%) and 47(87.04%) for malaria incidence, death due to malaria and measles cases while zero-inflated Poisson was used in 1(2.13%), 7(12.06%) and 22(100.0%) countries where there were clear occurrences of excess zero in the malaria incidence, measles cases and yellow fever cases respectively.

Malaria incidence reduced significantly in 38(80.9%) countries, increased significantly in 5(10.6%) and increased insignificantly in 4(8.51%) of out the 47 countries observed in the study over time. Death due to malaria decreased significantly in 20(43.5%) countries but increased significantly in 24(52.2%) of the 46 countries included for the study over time. Also, of the total 54 countries studied for measles cases, 49(90.7%) showed significant decrease trend, 4(7.4%) significant increase trend and 1(1.85%) insignificant time trend coefficient. Yellow fever cases decreased significantly in 13(59.1%), increased significantly in 3(13.6%) countries while the remaining countries 6(27.27%) showed an insignificant trend (Table 1).

Table 1

Poisson and zero inflated time trend regression.

Region	Country	Malaria Incidence			Death due to Malaria			Measles Cases			Yellow fever Cases
		Method	Estimate	Significance	Method	Estimate	Significance	Method	Estimate	Significance	Method	Estimate	YFTD	Significance
North	Algeria	P	-0.088	0.901	P	0.257	0.023	P	-0.082	< 0.001
North	Egypt				P	-0.514	0.025	P	-0.045	< 0.001
North	Libya							P	-0.041	< 0.001
North	Morocco	ZIP	-0.405	0.976	P	0.13	0.383	P	-0.17	< 0.001
North	Sudan	P	-0.046	< 0.001	P	-0.144	< 0.001	P	-0.111	< 0.001
North	Tunisia							P	-0.064	< 0.001
East	Burundi	P	-0.05	< 0.001	P	-0.171	0.021	P	-0.135	< 0.001
East	Comoros	P	-0.084	< 0.001				ZIP	-0.087	< 0.001
East	Djibouti	P	0.144	< 0.001	P	-0.055	0.087	P	-0.117	< 0.001
East	Eritrea	P	0.046	< 0.001	P	-0.018	0.003	P	-0.061	< 0.001
East	Ethiopia	P	-0.074	< 0.001	P	0.093	< 0.001	P	-0.039	< 0.001
East	Kenya	P	-0.08	< 0.001	P	0.165	< 0.001	P	-0.175	< 0.001	ZIP	-0.119	1	0.007
East	Madagascar	P	0.024	< 0.001	P	0.019	< 0.001	P	-0.042	< 0.001
East	Malawi	P	-0.036	< 0.001				P	-0.126	< 0.001
East	Mauritius			< 0.001				ZIP	0.042	< 0.001
East	Mozambique	P	-0.025	< 0.001	P	0.125	< 0.001	P	-0.066	< 0.001
East	Rwanda	P	0.078	< 0.001	P	0.168	< 0.001	P	-0.168	< 0.001
East	Seychelles							ZIP	-0.006	0.084
East	Somalia	P	-0.104	< 0.001	P	0.078	< 0.001	P	0.063	< 0.001
East	South Sudan	P	-0.029	< 0.001	P	-0.255	< 0.001	P	0.058	< 0.001
East	Uganda	P	-0.038	< 0.001	P	-0.011	< 0.001	P	-0.04	< 0.001	ZIP	-0.071	1	< 0.001
East	United Republic of Tanzania	P	-0.07	< 0.001	P	0.013	< 0.001	P	-0.139	< 0.001
East	Zambia	P	-0.049	< 0.001	P	0.101	< 0.001	P	-0.128	< 0.001
East	Zimbabwe	P	-0.033	< 0.001	P	0.145	< 0.001	ZIP	-0.077	< 0.001
Central	Angola	P	-0.035	< 0.001	P	0.035	< 0.001	P	-0.084	< 0.001	ZIP	0.11	2	< 0.001
Central	Cameroon	P	-0.034	< 0.001	P	-0.026	< 0.001	P	-0.107	< 0.001	ZIP	-0.032	1	< 0.001
Central	Central African Republic	P	-0.016	< 0.001	P	-0.165	< 0.001	P	-0.065	< 0.001	ZIP	-0.017	1	0.605
Central	Congo	P	-0.033	< 0.001	P	-0.078	< 0.001	P	-0.116	< 0.001	ZIP	-0.236	1	< 0.001
Central	Democratic Republic of the Congo	P	-0.028	< 0.001	P	-0.075	< 0.001	P	0.062	< 0.001	ZIP	-0.136	1	< 0.001
Central	Equatorial Guinea	P	-0.02	< 0.001	P	-0.161	< 0.001	P	-0.101	< 0.001
Central	Gabon	P	-0.007	0.017	P	0.192	< 0.001	P	-0.085	< 0.001	ZIP	-0.042	1	< 0.001
Central	Sao Tome and Principe	P	-0.213	< 0.001	P	0.314	< 0.001	ZIP	-0.088	< 0.001
West	Benin							P	-0.099	< 0.001	ZIP	-0.061	1	0.002
West	Burkina Faso	P	-0.02	< 0.001	P	-0.006	< 0.001	P	-0.056	< 0.001	ZIP	0.003	2	0.266
West	Cape Verde	P	-0.016	0.785	P	0.088	< 0.001	P	-0.077	< 0.001
West	Chad	P	-0.018	< 0.001	P	-0.109	< 0.001	P	-0.033	< 0.001	ZIP	0.268	2	< 0.001
West	Côte d’Ivoire	P	-0.03	< 0.001	P	0.045	< 0.001	P	-0.075	< 0.001	ZIP	0.004	2	0.408
West	Gambia	P	-0.068	< 0.001	P	0.036	< 0.001	ZIP	-0.085	< 0.001
West	Ghana	P	-0.027	< 0.001	P	0.034	< 0.001	P	-0.097	< 0.001	ZIP	-0.033	1	< 0.001
West	Guinea	P	-0.013	< 0.001	P	-0.042	< 0.001	P	-0.057	< 0.001	ZIP	-0.069	1	< 0.001
West	Guinea-Bissau	P	-0.096	< 0.001	P	0.066	< 0.001	P	-0.075	< 0.001	ZIP	0.03	2	0.301
West	Liberia	P	-0.018	< 0.001	P	-0.079	< 0.001	P	-0.065	< 0.001	ZIP	-0.22	1	< 0.001
West	Mali	P	-0.002	0.312	P	-0.036	< 0.001	P	-0.089	< 0.001	ZIP	-0.072	1	< 0.001
West	Mauritania	P	-0.075	< 0.001	P	-0.035	< 0.001	P	-0.106	< 0.001	ZIP	-0.047	1	0.056
West	Niger	P	0.008	< 0.001	P	-0.023	< 0.001	P	-0.048	< 0.001
West	Nigeria	P	-0.025	< 0.001	P	-0.02	< 0.001	P	-0.05	< 0.001	ZIP	-0.052	1	< 0.001
West	Senegal	P	-0.106	< 0.001	P	0.086	< 0.001	P	-0.099	< 0.001	ZIP	-0.026	1	< 0.001
West	Sierra Leone	P	-0.01	< 0.001	P	-0.108	< 0.001	P	-0.056	< 0.001	ZIP	0.151	2	< 0.001
West	Togo	P	-0.026	< 0.001	P	0.013	< 0.001	P	-0.149	< 0.001	ZIP	0.039	2	0.080
South	Botswana	P	-0.188	< 0.001	P	0.077	< 0.001	P	-0.099	< 0.001
South	Eswatini	P	-0.037	0.265	P	0.165	< 0.001	ZIP	-0.122	< 0.001
South	Lesotho							P	-0.135	< 0.001
South	Namibia	P	-0.095	< 0.001	P	0.29	< 0.001	P	-0.136	< 0.001
South	South Africa	P	-0.063	0.030	P	0.019	< 0.001	P	-0.096	< 0.001

The spatial distribution of malaria incidence in African countries is presented in Fig 1. Malaria incidence average annual trend decreased significantly in several African countries except in Algeria, Cape Verde, Mali and Eswatini with time trend regressions -0.088, -0.016, -0.002 and -0.037, respectively, which showed a non-significant trend (P>0.05). A significant increase in malaria incidence was recorded in Djibouti (0.144), Eritrea (0.046), Madagascar (0.024), Rwanda (0.078), and Niger (0.008) (P<0.05) (Table 1).

Fig 1

Spatial Distribution of Malaria (A) and Death due to Malaria (B) in Africa.

The spatial distribution of malaria-associated death in African countries is presented in Fig 1. Algeria with trend regression (0.257) recorded the highest increase in death due to malaria while Egypt (-0.514), on the other hand, recorded the highest decrease in death due to malaria (P<0.05). Several countries in East Africa recorded a significant increase in malaria-associated death while many countries in Central Africa however, showed a significant decrease in malaria-associated death (P<0.05). All the countries in the South African region recorded a significant increase in death due to malaria (P<0.05) (Table 1).

The spatial distribution of measles in African countries is presented in Fig 2. Majority of the African countries recorded a significant decrease in measles cases except in Mauritius (0.042), Somalia (0.063), South Sudan (0.058), and Democratic Republic of Congo (0.062) (P<0.05). Seychelles (-0.006) showed a non-significant decrease in measles cases annual trend (P>0.05) (Table 1). The spatial distribution of yellow fever cases is presented in Fig 2. Chad with trend regression 0.268 recorded the highest significant increase in yellow fever cases while Congo (-0.236) showed the highest significant decrease in yellow fever cases in Africa (P<0.05). The Central Africa Republic (-0.017), Burkina Faso (0.003), Côte d’Ivoire (0.004), Guinea-Bissau (0.03), Mauritania (-0.047), and Togo (0.039) recorded non-significant yellow fever cases trends (Table 1).

Fig 2

Spatial Distribution of Measles (A) and Yellow Fever Cases (B) in Africa.

The Panel Poisson Regression presented in Table 2 showed the effect of Measles and Yellow Fever in each of African Region and Africa as a whole. Malaria incidence increased significantly with measles cases in East Africa while significant negative relationship occurred between the two in West and South African countries (P<0.001). Overall, a significant negative relationship existed between malaria incidence and measles cases (P<0.05). While a positive relationship existed between malaria incidence and yellow fever cases in West Africa (P<0.01), a negative relationship occurred between the two in South Africa (P<0.001) (Table 2). Overall, malaria incidence was significantly positively related to yellow fever in Africa (P<0.05).

Table 2

Malaria incidence and death due to random effect panel model.

Variables		North	East	Central	West	South	Overall
Malaria incidence	Measles	0.0000074	0.0000053***	0.00000015	-0.00000084***	-0.00057***	-0.00000042**
		(0.52)	(5.73)	(0.65)	(-4.24)	(-6.21)	(-2.86)
	Yellow Fever	0	0.000346	0.0000428	0.000114**	-0.548***	0.0000786**
		(.)	(1.27)	(1.22)	(3.27)	(-4.42)	(3.20)
	trend	-0.0356***	-0.0224***	-0.0308***	-0.0237***	-0.0762***	-0.0262***
		(-6.44)	(-16.11)	(-33.47)	(-34.84)	(-8.11)	(-52.68)
	Constant	4.691***	5.472***	5.960***	5.956***	3.107***	5.653***
		(4.15)	(22.07)	(45.80)	(25.24)	(4.74)	(30.10)
	Observations	27	123	152	264	43	609
Death due to malaria	Measles	-0.000016	-0.000014***	-0.0000031***	-0.0000014***	-0.00038***	-0.0000043***
		(-0.92)	(-59.09)	(-77.37)	(-16.74)	(-20.32)	(-125.04)
	Yellow fever	0	0.00137***	0.0000890***	-0.0000127	-0.356***	0.000156***
		(.)	(15.31)	(9.79)	(-1.10)	(-12.81)	(20.57)
	Trend	-0.0790***	-0.129***	0.00648***	0.00381***	-0.172***	-0.0472***
		(-3.86)	(-377.34)	(20.67)	(11.20)	(-59.53)	(-270.38)
	Constant	5.748***	9.715***	8.456***	7.560***	6.553***	8.567***
		(4.48)	(22.44)	(16.29)	(24.47)	(10.05)	(33.46)
	Observations	15	88	118	224	39	484

t statistics in parentheses.

* p < 0.05,

** p < 0.01,

*** p < 0.001.

Death due to malaria was significantly negatively related to measles cases in Africa (P<0.001). A negative relationship existed between the two West and South African countries (P<0.001). Overall, a significant positive relationship existed between death due to malaria and yellow fever in Africa (P<0.001) (Table 2).

Discussion

Malaria is still being actively transmitted across African countries, however, observations from our study showed that the various interventional programmes could be yielding expected results as there was a significant decrease in malaria incidence in about 81% of the observed African countries. Increase insecticide-treated nets (ITNs) coverage across African countries over the years may be responsible for this malaria decline [21,22]. The use of ITNs may result in a decline in the population of indoor Anopheles mosquitoes by triggering behavioural changes in biting and resting of mosquitoes including a shift from an indoor to an outdoor environment [23]. Besides the increase in ITNs coverage, changes in socio-ecological conditions in many malaria-endemic areas of Africa including changes in temperature, deforestation [24], increase use of agricultural pesticides with mosquitocidal properties [25], improved house constructions [26] could result in a decreased density of mosquitoes.

It is however important to stress that the recent WHO published report showed increase in malaria cases and associated deaths after 2019 [27]. The reasons for this have been accrued to plateaued malaria interventional funding, increase in drug and insecticide resistance, and climate change which threatens to push malaria transmission to new regions [28]. COVID-19 has further disrupted progress and implementation of various malaria intervention programs [28].

The significant decrease in malaria transmission could have culminated in the 43.5% significant reduction in death associated with malaria in Africa. However, the 52.2% significant increase in malaria-associated deaths in Africa could have resulted from poor access to and low-quality health services, and increased resistance of Plasmodium falciparum to first-line drugs [29,30]. While the incidence of malaria may correlate with death associated with the disease, the malaria-associated deaths could be difficult to assess in the actual sense [31,32].

The higher significant decrease in the incidence of measles in many African countries compared with that of yellow fever could portray a probable relaxed public health alert system towards prevention of yellow fever. Importantly, the immunization coverage of yellow fever appeared to be lower than that of measles in many African countries. For example in Nigeria, the mean official yellow fever vaccine (YFV) coverage was 56.3% while that of the first dose of measles-containing vaccine (MCV1) was 65.0%, for the Democratic Republic of Congo, YFV and MCV1 were 80.67 and 87.67% respectively [33,34]. In Togo however, the mean immunization coverage for YFV (82.83%) and MCV1 (82.75%) was similar [35].

While our models showed a generalized negative relationship between the incidence of malaria and measles cases, the factors responsible for this relationship cannot be readily explained at the moment. The usual administration of preventive chemotherapy such as chloroquine in malaria highly endemic areas of Africa may inhibit the replication of measles virus in concomitant infection, thus, the observed negative relationship. One study in Ghanaian pregnant women population has also associated maternal malaria preventive treatment with increasing infant immunity against measles [14]. A closer look at the differences in the direction of the relationship at the regional levels however suggested that in Eastern Africa, malaria transmission could impact measles transmission. While the reason for these conflicting observations in Eastern Africa cannot be readily explained at the moment, well designed clinical studies in all the regions could resolve these conflicting observations. Other unknown factors other than preventive chemotherapy could play some roles.

The overall positive relationship between malaria and yellow fever could be as a result of co-existence of mosquito vectors of the two diseases in similar ecological zones. In addition to Aedes mosquitoes, Anopheles mosquitoes have been identified in similar habitats in Africa [36,37]. Although our study showed decreased in the incidence of malaria over the years in many countries in West Africa, the African region is one of the most affected regions of the world contributing more than 30% of the world’s malaria-associated deaths [38]. This could be difficult to explain, but complex immune modulation which could influence the susceptibility and immunopathology of yellow fever in malaria-endemic areas may be responsible for the positive relationship in the incidence of the two diseases in the region. The reverse relationship trend between the two diseases in countries in the Southern Africa region may be due to lower malaria transmission in many countries in the region. Overall similar relationship trends observed between malaria incidence and yellow fever/measles were also observed between death due to malaria and the two vaccine-preventable diseases in Africa. These can be adduced to the aforementioned reasons highlighted for malaria incidence and yellow fever/measles.

Conclusions

Our study showed that significant relationships exist between malaria and yellow fever/measles. The direction of the relationships differed, with yellow fever showing positive and measles negative relationships with malaria incidence and malaria-related deaths, respectively. The relationships observed in this study could be important for the management of malaria and the studied vaccine-preventable diseases. Increase vaccination coverage or malaria treatment could shape the direction of these relationships. Case-control studies are recommended to critically understudy the relationships between malaria, measles and yellow fever.

(XLSX)

Click here for additional data file.

5 Jan 2022

6 Jan 2022

Response to comments

We thank you for your efforts both in giving our manuscript quality review and expediting the process. As suggested and requested by the Editor and the Reviewer, we have provided responses to the concerns raised. Our responses are appended below:

Editor’s comments

1. We have now adapted the manuscript as suggested to the journal’s style

2. Online profile is now update to include the ORCID number of the corresponding

3. We have now removed the phrase “data not shown” and associated sentence from the manuscript since the data are not a core part of the research being presented in our study.

4. Figures 1 and 2 are original and we generated by the software QGIS (version 3.14). This was stated in the statistical analyses. So, they have no copyright issues.

5. Data availability statement has been revised and the website included. The file containing the minimal data is also included as supplementary file.

Reviewer #1: General comments

This is an important subject in Africa, the continent with the highest burden of malaria, but of recently experiencing outbreaks due to viral diseases.

Line 69: Authors should change the word “Precious” to “Previous”

Response – Thank you for the observation. It has now been corrected

Line 207. The statement is not correct. Rephrase to read “Fever in each of African Region and Africa as a whole”

Response – Thank you. This is now corrected as requested

Line 236: The word ITN has been introduced without being defined anywhere in the manuscript

Response – Thank you for the observation, it is now defined at first mention (please see line 224)

Lines 236-237: The statement must be supported by appropriate References

Response – Thank you, we have now included references

Line 276: The mention of Culex mosquito is irrelevant and confusing. The mosquito species neither transmit malaria no Yellow fever.

Response – Thank you, we have deleted Culex mosquitoes as advised

Submitted filename: Response to Reviewers Comments.docx

Click here for additional data file.

17 Mar 2022

Submitted filename: PONE-D-21-27941_R1 (1).pdf

Click here for additional data file.

25 Mar 2022

Response to reviewer’s comments

Abstract

Comment - Do you think that having only few studies indicate the relevance of your study? Better to show existing gaps that assured the valuable of your study in this section?

Response – Thank you, we have modified the statement (see lines 32-33).

Comment - please Indicate data management and clearance methods if the author implemented prior to data analysis

Response – Thank you for the comment. While we did not include this because of words limit in abstract section, we instead given the detail in the materials and methods section

Comment - Please indicate the significant limit of your analysis.

Response – Thank you. This is now included (Line 39)

Comment - The author should conclude the main findings of this work before indicating the relevance of this work.

Response – Thank you, this is now revised accordingly

Introduction

The background section of your text is somewhat shallow. So, I suggest the author to re-write the background section of the text in detail based on the following structure if the author agreed with this structure.

1. malaria is xxx and Malaria causes X and Y cases and problems globally and /or in africa..... Cuurently, WHO and/or other stakeholders recommend XXX public health intervention measures to control and prevent or eliminate or eradicate....malaria.

2. However, the co-existence of XXX viral infection in similar geographical location with malaria

3. Then if their any available challenges due to having similar geographical distribution..may in relation to current prevention and control mechanism, may be in relation to vaccine effectiveness....may be in related to treatment failure or anyother.

Then gabs and the relevance your finding to fill such available gabs

Response – Thank you for these important observations. We have included more information as suggested to include the gaps and the justifications for the study (see the red highlights).

Comments - I am feeling there is something cracked between the previous paragraph and this one. So, I recommend the author to incorporate few transition sentences prior to define yellow fever (i;e. a sentences which act as bridge between malaria you stated in the first paragraph and yellow fever in this paragraph to keep idea flow better.

Response – Thank you, we have now rephrased for flow of idea (see Line 61)

Comment - Please add a reference for the impact of placental malaria on the infant immunity to measle.

Response – Thank you, reference now included (see ref. 15 in line 78).

Comment – ‘’Because all the studied variables are cases and rates which are regarded as count data, a Panel Poisson regression model was adopted to understudy these relationships. Where there are observed excess zero in the cases of event considered in the study, a zero-inflated Poisson (ZIP) regression that is robust to the occurrence of zero is adopted to achieve our specified objectives’’. Such statements should be incorporated in the statistical analysis sub-heading of the tesx

Response – We have now included the statement in the statistical analysis sub-section.

Materials and methods (line 167-171).

Comments - How could manage and associate the findings of malaria with yellow fever and measles in this paper since the author collected the recorded data in different time period. So, did the author consider this time period variation while the author conclude the association of malaria incident and death with yellow fever and measles?

Response: Thank you for your comments

There are two main analyses in the study. The first analysis was the determination of annual trend change for each Malaria incidence, Death due to malaria, Measles and Yellow fever. Here, changes of these four variables were evaluated over time for each country considered in the study and each dataset were treated separately for this analysis. The time horizon for malaria incidence and death due to the malaria trend equation was 2000-2018 whereas the time horizon for Measles and Yellow fever trend equation was 1980 and 2019. This decision was based on the availability of the data obtained from the secondary source mentioned in the manuscript.

The second analysis was a regression equation where the effect of both Measles and Yellow fever on each of the Malaria incidence and Death due to malaria was investigated. Two data files were used for this analysis bearing in mind that in any regression problem there must be a pair of x and y variables. The first data file had Malaria as the dependent variable while both Measles and Yellow fever were the independent variables. The time horizon for this regression was 2000-2018 since information for malaria incidence was not available for 1980-1999 and 2019. Similarly, the second data file contained Death due to malaria as the dependent variable with Measles and Yellow fever as independent variables. The time horizon for this analysis was 2000-2017

Comment - Despite you are indicating the significance and insignificance of your finding in your result and discussion sub-heading, the author failed to indicate the significance limit of your analysis. So please add here the significance limit for your statistical analysis output.

Response – Thank you for the observation. We have now included the significance limit (see line 173).

Results

Comment – Figures quality should be improved

Response - Thank you. We have now used the PACE website provided by Plos to help ensure that figures meet PLOS requirements and improve the quality.

Discussion

Comment - According to the recent World Health Organization (WHO) published report showed, malaria and related deaths have been increased after 2019. If so, How could you harmonize your conclusion based on your finding with the WHO report?

Response – Thank you for this important observation. We have tried to provide reasons for your concern in the MS as follow “It is however important to stress that the recent WHO published report showed increase in malaria cases and associated deaths after 2019 [27]. The reasons for this have been accrued to plateaued malaria interventional funding, increase in drug and insecticide resistance, and climate change which threatens to push malaria transmission to new regions [28]. COVID-19 has further disrupted progress and implementation of various malaria intervention programs [28]”. Please see lines 256-260.

Comment - Are sure does modification of mosquito biting behavior is a factor for decline of malaria? I need more justification for this statement if your answer is yes.

Response – Thank you for your comment. This is an observation from another author which we referenced. Yes, we agree with the statement as the inability of the mosquitoes to access humans for blood meal indoor due to hindrance caused by ITN could trigger a behavioural change in mosquitoes to source for blood meal outside where there is no ITN.

Comment - would you mention few socio-ecological conditions which favour for the reduction of density of mosquito? Please elaborate it by giving example?

Response – Thank you for the observation. The examples are already given in the text. Please see the sentence below.

“changes in socio-ecological conditions in many malaria-endemic areas of Africa including changes in temperature, deforestation [22], increase use of agricultural pesticides with mosquitocidal properties [23], improved house constructions [24] could result in a decreased density of mosquitoes”.

Comment - Your justification is insuffeicent? please discuss in detail the possible reasons why malaria associated deaths is high in Africa unlike other past of the coninent.

Response – Thank you for the comment. However, the factors already highlighted are the major reasons we believe to be associated with high malaria deaths in Africa.

Comment - Are you sure do the incidence of malaria and measles case showed a positive relationship based on your finding? But you indicated in line 197-198, o]a significant negative relationship existed between malaria and measles cases (P<0.05). If so, why the author failed to keep the consistence of the findings? The author should give a clear information for this contradictory conclussion? Moreover, the author expected to discuss possible reasons why this findings happen?

Response – Thank you for this very important observation. This is a mistake from our end. We have now corrected it. We have also changed the direction of the discussion based on this change (see lines 276-287).

Comment - Why the author need to discuss the incidence of malaria in association with HIV and hebatitis? Does this is the authors objective? How could you relate it?

Response – Thank you for the observation. Since there were no enough reports to discuss the association between malaria and the studied vaccine related viruses we thought discussing the relationships between malaria and other viral agents could shed some light on the relationships between malaria and measles/YFV. Since the reviewer is not satisfied with this, we have now expunged the sentences and the references.

Comment - Are you sure does the region shared 30% of malaria associated death?

Response – Thank you for the comment. Based on the reference WHO report cited, the total malaria-related deaths in West Africa Region is more than 30%. Nigeria alone accounts for 25%.

Submitted filename: Plos One Revision2.docx

Click here for additional data file.

31 Mar 2022

1 Apr 2022

Why did the authors failed to specify the exact statistical output of P-value rather than indicated as < 0.05 in the table 1? Unless and until the author provides a satisfactory explanation for why the P-value was not indicated, I will be skeptical of the data analysis's reliability.

Response – Thank you for your comment. We merely supplied the level of significance (P<0.05 or P>0.05) in the previous version. However, we have now provided the exact P value as requested by the reviewer.

Submitted filename: Response to Reviewers Comments R2.docx

Click here for additional data file.

18 Apr 2022

18 Apr 2022

Response to reviewer

Reviewer #2: First and foremost, I'd like to thank the authors for their success in producing a scientifically soundable article while taking into account my previous comments and suggestions. Having said that, I'd like to make a suggestion for the authors in table one.

In table one, when the author made a change after considering my previous comment, the p-value was written as o.ooo, which is not a scientifically sound way of writing. To the best of my knowledge, P-value cannot be written as 0.000 as you did in table 1. Even if the statistical software produces a p value of 0.000, the exact value cannot be equals to zero. You can verify it by double-clicking on it, which will display the actual value of P. In such cases, P value should be expressed as < o.oo1. The logic behind for such expression is the observed value can never be zero due to the presence of random error, which is unavoidable and uncontrollable. Thus, I recommend the authors to express the P-value as < 0.001 in table 1 rather than simply put 0.000. However, it should only be done if and only if the author agrees with my comment.

Response – Thank you for this important comment. We agree with the reviewer on this point. So, we have now replaced P value 0.000 with <0.001 as suggested.

Submitted filename: Response to Reviewers Comments R3.docx

Click here for additional data file.

22 Apr 2022

Malaria transmission in Africa: its relationship with yellow fever and measles

PONE-D-21-27941R4

Dear Dr. Oyeyemi ,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Luzia Helena Carvalho, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments: