Menu
Menu
Menu
Menu
Background: One out of ten deaths of children under five are attributable to indoor air pollution. And Acute Respiratory Illness (ARI) is among the direct causes. Objective: This study showed the possibilities of characterizing indoor air pollution in West African Economic and Monetary Union (WAEMU) area and it also made it possible to estimate its impact on the occurrence of ARI in children under five. Methods: It has been a secondary analysis based on Demographic and Health Surveys (DHSs) from WAEMU countries’ data. “Household level of air pollution” is the created composite variable, from questions on the degradation factors of indoor air quality (domestic combustion processes) which served to characterize indoor air pollution and to measure its impact by a logistic regression. Results: Burkina Faso stands out with a greater number of households with a high level of pollution (63.7%) followed by Benin (43.7%) then Togo (43.0%). The main exposure factor “Household level of air pollution” was associated with ARI symptoms (Togo: prevalence = 51.3%; chi-squared test’s p-value < 0.001). Exposure to high level of pollution constitutes a risk (AOR [95 CI]), even though it is not significant (Ivory Coast: 1.29 [0.72–2.30], Senegal: 1.39 [0.94–2.05] and Togo: 1.15 [0.67–1.95]) and this could be explained by the high infectious etiology of the ARI.
This is a retrospective cross-sectional study, in which we carried out a secondary analysis on data from DHS conducted in WAEMU member states. WAEMU is composed of eight Sahelian countries, linked by a common currency and cultural traditions: these are Benin, Burkina Faso, Mali, Niger, Ivory Coast, Guinea-Bissau, Senegal and Togo [16, 17]. It covers an area of 3.5 million km2 and has more than 120 million inhabitants, 34.8% of whom live in urban areas with disparities between countries [18, 19]. Indeed, urban population is larger in Ivory Coast (53.8%), Senegal (46.5%) and Benin (44.6%) and lower in Niger (14.9%). In addition, Ivory Coast represents 20.6% of the total population of the area, followed by Niger with 17.3% [19]. WAEMU area faces challenges related to poverty, access to basic social services, high fertility and is characterized by a high infant mortality [19]. In sum, the study included 59,765 children (Benin: 12,432; Burkina Faso: 13,583; Ivory Coast: 6941; Mali: 9222; Senegal: 11,182; and Togo: 6405), and 65,705 households (Benin: 14,156; Burkina Faso: 14,424; Côte d'Ivoire: 9686; Mali: 9510; Senegal: 8380; and Togo: 9549). DHS are designed to be nationally representative and aimed to provide information on the characteristics of the population (family planning, maternal and child health, child survival status, HIV/AIDS, Sexually Transmitted Infections (STIs), reproductive health, nutritional status, etc.). Data were collected according to a complex multi-stage stratified cluster sampling design. At first, Enumeration Areas (EAs) were identified and then drawn from a list established during the last General Population and Housing Census (RGPH), then in each selected EA, a sample of households was drawn from an updated list. Survey participants included women aged 15 to 49, men aged 15 to 59, and children under five. As regards to the latter, their mothers were invited to provide information on their demographic characteristics as well as their health status. Four questionnaires were used for data collection: household questionnaire, female questionnaire, male questionnaire and biomarker questionnaire. Household questionnaire served as a tool for collecting information on household characteristics (main source of drinking water, type of toilet, hand washing equipment, source of lighting, fuels and cooking place, passive smoking, etc.). It also allowed to identify household members eligible for individual interviews and/or biological tests and measurements. A Biomarker questionnaire allowed informing the anthropometric measurements as well as results of tests carried out on blood samples [20–26]. Results presented in this paper are based on characteristics of households and children under five included in the sixth DHS (Burkina Faso, Ivory Coast) and seventh DHS (Benin, Senegal, Togo, Mali). Databases were obtained following a request and a justification of study from managers of the DHS program. Guinea Bissau is not concerned by the DHS program and is therefore excluded, as is Niger due to the unavailability of some variables of interest in the used database. “Household level of air pollution” is the created composite variable, from questions on the degradation factors of indoor air quality (domestic combustion processes) which served to characterize indoor air pollution and to measure its impact by a logistic regression. These questions were: “Does your household have electricity?”; “What type of fuel does your household mainly use for cooking?”; “Is the cooking usually done in the house in a separate building or outdoors?”; “How often does anyone smoke inside your house, would you say daily, weekly, monthly, less often than once a month, or never?”; “Do you currently smoke cigarettes every day, some days, or not at all?”. The possible answers to some of these questions were first grouped before being assigned a score. As regards to the type of cooking fuel, grouping is based on the work of Mishra et al. [27]. Three categories corresponding to high pollution fuels (wood, straw/shrubs/grass, agricultural crop or animal dung), medium pollution fuels (Kerosene, coal/lignite or charcoal), and low pollution fuels (electricity, Liquefied Petroleum Gas, natural gas, biogas) are indeed defined on the basis of the answers to this question. The scores assigned to these categories were 3, 2 and 1 respectively. The question on the smoking status of household members was also categorized into three modalities (never, sometimes and daily) with scores of 0, 1 and 2 respectively. “sometimes” was introduced as a new modality and includes the following responses: weekly, monthly, and less often than once a month. Concerning the mothers’ smoking status, the variable was binarized (yes/no) by regrouping under the “yes” modality, the following answers: every day or some days. Thus, the score assigned to this variable was 1 for “yes” and 0 for “no”. The same is applied to the availability of electricity, which was collected in a binary form. Also, the place of cooking was not recoded, and the answers were outdoors, in a separate building or in the house, corresponding respectively to the following scores: 1, 2 and 3. The maximum summation of the scores is 10. Subsequently, three levels of scores were defined for “Household level of air pollution”: low level corresponding to households with a score less than 4; medium level for those with a score between 4 and 6; and high level when the score is greater than 6. The second variable of interest is defined by symptoms of ARI and is used to characterize respiratory health of children. The definition proposed for this indicator has evolves over time and this work retained the DHS Statistics Guide’s latest definition. Symptoms of ARI in the child is defined as “short, rapid breathing which was chest-related and/or difficult breathing which was chest-related” during two weeks preceding the survey [28]. These symptoms were self-reported by children’s mothers. Moreover, DHS Statistics Guide also classified types of drinking water and sanitation facilities into one of the following: improved and unimproved [28]. This classification is based on guidelines of WHO/UNICEF’s Joint Monitoring Program for water supply and sanitation [29]. We did a a frequency measurement to describe households and children included in the study. The two main variables used for this purpose are: “Household level of air pollution” and symptoms of ARI. In addition to variables used to construct these indicators, other variables were included in this phase of the analysis. For households, the latter variables are relating to the access of water, hygiene and sanitation, as for children, they are: age, sex, birth weight and the mothers’ age and level of education. The second phase of analysis was carried out by measures of association using the chi-squared test and logistic regression. At this stage, symptoms of ARI are defined as dependent variable and “Household level of air pollution” as the main exposure factors. Other variables used in households and children’s description steps were also taken into account. A multivariate logistic regression model was fitted by including all variables significantly associated with the occurrence of ARI in any of the WAEMU member countries according to chi-squared test’s p-value. Adjusted Odds Ratios (AOR) were estimated from regression models as well as 95% confidence intervals (95% CI). All statistical analyses were carried out using the R software. The “survey” package is used to weigh all the observations in order to compensate for the oversampling of certain categories of respondents and to take into account the complexity of the sampling plan.
N/A
