Background
One of the United Nation’s sustainable development goals for 2030 is to end the epidemics of AIDS, tuberculosis, malaria and other neglected tropical diseases [
1]. Malaria continues to be one of the major concerns for public health in Africa. Sub-Saharan Africa presents a disproportionately high portion of the global malaria burden. In 2015, the region was home to 88% of malaria cases and 90% of malaria deaths. The number of malaria deaths globally was 438,000 in 2015 (Range 236,000–635,000) [
2]. According to the Mozambican Ministry of Health, the country recorded over six million cases of malaria in 2015 [
3] and deaths due to malaria (per 100,000) was 42.75 in 2013 [
4]. Malaria killed 3245 people and is the second cause of death in the country, at 19.2% [
5].
Chimoio is a municipality in the central region of Mozambique where the incidence of malaria is 20.1%, and the attributable factor 16% [
6]. Malaria is the fourth leading cause of deaths in Chimoio at 9.4% [
7]. Climatic factors such as temperature, relative humidity, precipitation and evaporation influences the lifecycle and development of both the mosquito vector and the parasite [
8].
Understanding the trends and variation of deaths is of paramount importance for precision public health. Precision public health is a relatively new concept and its ultimate goal is to develop and implement health interventions that can benefit the right population at the right time [
9]. Civil registration constitutes the most timely and accurate source of information on mortality and causes of death. In Mozambique registration of deaths falls under the Ministry of Justice [
10].
Describing trends and characteristics of malaria mortality, and its relationship with climate factors, can assist in monitoring and planning resource needs of the health system and municipal management.
Geographic information systems can help to describe variations in malaria mortality and this is important to identify areas at high risk, to assist in designing appropriate interventions, or lead to further investigations to identify important risk factors [
11].
As stated elsewhere, the best ways to help the living is by counting the dead. Few data on malaria mortality, trends and characteristics of malaria death exist in Mozambique, particularly in Chimoio. The few existent data are from hospitals and do not represent the entire community.
The objective of this study is to determine malaria mortality trends, characterize malaria mortality, describe its spatial distribution and variation in Chimoio, and verify its relationship with climate parameters to help local authorities in programmatic malaria activities for the prevention and eradication of the disease.
Discussion
The civil registration covers all registered malaria mortality cases from hospitals and from private residences. In this study, 78% of malaria deaths occurred in hospitals and the 22% at private residences. A previous study in Chimoio reported that in all-cause deaths, 86.1% of the deaths took place in hospitals and 11.7% at private residences [
7]. Malaria deaths at private residences is two times greater than in all-cause of death in Chimoio. These disparities can probably be because malaria patients delay the treatment of the disease resulting in fatalities.
Trend analysis indicates that in Chimoio, cases of deaths, and malaria deaths are increasing over the years, contrary to reports in Kwazulu Natal [
18], Malawi [
11], and Tanzania [
19] that reported decreasing cases in malaria mortality. The malaria crude mortality rate per 100,000 was 51 per 100,000, higher than the national Mozambique figure of 42.75 for 2014 [
3]. In terms of malaria mortality by gender, there was no difference between malaria deaths in females and males. Similar results were reported previously by [
17]. The results disagree with the findings in Kwazulu Natal and Sudan that reported higher mortality from malaria in males than in females [
18,
20]. There is evidence that suggests that given equal exposure, adult men and women are equally vulnerable to malaria except for pregnant women [
20]. In this study, 25% of malaria deaths occur at the age of 2, and 75% of malaria deaths at the age 43. The results are in concordance with a report on all causes of death carried out in Chimoio [
7].
Age category 0 comprises 3% of the Chimoio population and recorded 9% of malaria deaths while, age category 1–4 comprises 13% of the Chimoio population, and recorded 25% of malaria deaths. This can be due to the lack of immunity in the first years of life. Similar results were reported in another seven African countries and Bangladesh [
11,
21‐
25]. From the age of 45 onwards the proportion of deaths by malaria and, all-cause mortality is almost the same.
Malaria mortally was significantly different between month and years. Similar results of seasonality were reported in Ethiopia and Burkina Faso [
24,
25] and were related to climatic conditions. January, February and March presented the highest percentage of mortality from malaria decreasing thereafter. This peak period occurs 2 months after the rainy season onset.
There was no difference in times of death from malaria in Chimoio, and this result clearly contradicts a previous report on all-cause mortality in Chimoio, that indicates that peak mortality occurs between 3:00 and 4:00 a.m. [
7]. This result suggests that malaria deaths can occur at any time contrary, to other deaths that were found to peak from 3:00 to 4:00 a.m. in Chimoio.
The centre of town (low density) presents a low malaria crude mortality rate, 0–27 per 100,000 and the rural “Bairros” a very high crude mortality rate, over 80 per 100,000. This can be due to the fact that the centre has better health facilities and infrastructures which means the residents are better-off than in rural areas. Some rural neighbourhoods present low malaria mortality rates. This can be attributed to the fact those areas have poor accessibility and the residents carry out their burials without Civil Registration.
The onset of rain occurs in mid-November. This indicates that malaria occurrence has a strong association with rainfall 6–8 weeks before, coinciding, with the malaria cycle’s three components: (i) the growth of the
Anopheles female mosquito from egg to adult to parasite transmission; (ii) the development of the
Plasmodium parasites (gametocyte to sporozoites) that are able to infect humans; and (iii) the incubation period in the human host from infection to malaria symptoms [
22]. Thus, malaria occurrence peaks can be expected 45–60 days after the onset of rain. Similar results were also found in Mozambique [
4] and South Africa [
18]. Increased precipitation can provide more breeding sites for mosquitoes, however excess rain can also destroy breeding sites [
26].
ARMA (2,0)(1,0)
12 fitted the data well although it was not able to capture the sudden change occurring during malaria outbreaks. Introducing the intervention effect allowed for a better fit of death peaks and the seasonal ARMA model with intervention reduced root mean square error by almost 25%. Other studies reported ARIMA (2,1,1) in Zambia, ARIMA (1,0,0) in Burundi [
27], and India [
14] with comparable results.
Besides the possibility that the malaria mortality was under-reported, especially in the rural areas, another limitation of this study is that it did not take into consideration malaria intervention factors such as bed net distribution and improvement of health coverage. Despite the limitations, one great strength of the study is that this is the first specific study in malaria mortality using civil registration data in Chimoio. More data from other additional data from other parts of the country are needed to generalize the results to the national level.
Authors’ contributions
JLF contributed with study design, data analysis and interpretation, manual writing and figures, JMM and MP contributed in the critical revision and, SZ contributed to data collection. All authors read and approved the final manuscript.