Mothers’ Socioeconomic Differentials and Management of Malaria in Nigeria

About 150 million Nigerians live in areas of intense malaria transmission. Malaria has the greatest prevalence, close to 50% in children aged 6 to 59 months. A review of literatures revealed that more than 80% of malaria episodes received treatment outside of the existing government health care system. This means that treatments are rarely sought at health care facilities and are most often inappropriate or delayed. Reasons underlying these practices range from mothers’ socioeconomic status to difficulty in accessing health care facilities. Therefore, this study re-examined whether mothers’ socioeconomic characteristics and barriers to access health care facilities are major factors that influence mothers’ choice of treatment and delays in seeking treatment of malaria among under-five children in Nigeria. The study used Nigeria Demographic and Health Survey kids recode dataset. The data were analyzed using STATA 12 software. The result showed significant relationship between religion (χ2 = 216.24, p < .05), education (χ2 = 257.55, p < .05), occupation status (χ2 = 21.88, p < .05), wealth index (χ2 = 207.08, p < .05), type of residence (χ2 = 18.56, p < .05), region of residence (χ2 = 350.82, p < .05), and type of treatment sought and delay in seeking treatment for malaria. Also, the likelihood of seeking medical and prompt treatment among mothers with four different barriers is significantly less (odds ratio = 0.53; p < .05; 95% confidence interval = [0.38- 0.75]), when compared with their counterparts without any barrier. The study concluded that mothers` socioeconomic status and access to health care facilities must be improved to ensure appropriate and prompt use of health care facilities for treatment of malaria among under-five children in Nigeria.

intensified effort and initiated the Integrated Management of Childhood Illnesses (IMCI) to assist developing countries reduce childhood mortality caused by most childhood killer diseases: diarrhea, acute respiratory infections, malaria, measles, and malnutrition. (WHO/UNICEF, 1997). Malaria control activities in Nigeria are planned and implemented through the primary health care (PHC) system (WHO, 2006). The strategy used for the implementation of the national antimalarial treatment policy is that of Roll Back Malaria (RBM). The RBM strategy for the control of malaria has as one of its key elements that patients with malaria should have access to appropriate and adequate treatment within 24 hr of the onset of symptoms. However, the use of health care facilities as the first resort for malaria and treatment/management has been shown to be low in many African countries, including Nigeria. Treatments are rarely sought at health care facilities and are most often inappropriate or delayed (Guyatt & Snow, 2004;Kofoed, Rodrigues, Hedegaard, Rombo, & Aaby, 2004;Müller, Traore, Becher, & Kouyate, 2003). Less than 15% of the malaria episodes treated at home is treated correctly. Most fevers in children are treated with simple fever drugs, such as paracetamol and aspirin, but not with antimalarial drugs. Even when anti-malarial drugs are purchased, they are commonly administered in inappropriate doses (WHO, 2004).
Meanwhile, studies have shown that existing interventions could prevent many deaths among children if they are presented for appropriate and timely care (Jones et al., 2003). The WHO estimates that seeking prompt and appropriate care could reduce child deaths due to malaria, diarrhea, and acute respiratory infections by 20% (WHO, 1999). However, there are controversial findings on whether the use of modern health services is often influenced by individual perceptions of the efficacy of modern health services, religious beliefs, residence, occupation, place of delivery, wealth status, level of education, and so on (Adetunji, 1996;Adeyemi, 2000;Ene-Obong, Uwaegbute, Iroegbu, & Amazigo, 1998;Fapohunda & Beth, 2004;Hill, Kendall, Arthur, Kirkwood, & Adjei, 2003;Mekonnen & Mekonnen, 2002;Pate, 2001;Sabitu, 2004;United Nations Population Fund [UNFPA], 2002, 2004. Therefore, this article examined the influence of mothers' socioeconomic characteristics and barriers to access health care facilities on mothers' choice of treatment and delays in seeking treatment of malaria among under-five children in Nigeria.

Research Questions
Against the above background, the fundamental questions that readily come to mind are as follows: Research Question 1: Do mothers' socioeconomic characteristics (age, level of education, occupation, religion, wealth index, etc.) have significant influence on type of treatment sought and prompt treatment of febrile illness? Research Question 2: Does barrier in accessing health care facilities have any influence on type of treatment and prompt treatment of malaria?

Data Collection Method
The Nigeria Demographic and Health Survey (NDHS) kids recode data set was used for this study. The survey was crosssectional. It was designed to provide specific information on population and health indicators at the national, zonal, and state levels. Information collected includes birth histories, in-depth demographic and socioeconomic information on illnesses, medical care, immunizations, and anthropometric details of children (National Population Commission [NPC] [Nigeria] & ICF Macro, 2009). Therefore, from the sampling frame of 33,385 of all married women interviewed in Nigeria, 18,028 women with at least a child aged 12 to 59 years were extracted, out of which 3,068 women whose child has had at least an history of malaria episode in the last 2 weeks before the survey were extracted for this study having applied the weighting factor.

Data Analysis
Having obtained the data set and extracted the eligible respondents, the data were analyzed using STATA 12 software. The analysis involved three stages. The first stage is univariate analysis, at this stage, mothers' background characteristics such as age, marital status, religion, education, place of residence, and so on, were examined. The bivariate analysis involved comparison and testing for significance between two variables such as mothers place of residence and type of treatment sought, mothers' education and type of treatment sought, and so on. The third stage involved the multivariate analysis that further analyzed the relationships between independent and dependent variables. At this stage, logistic regression models were used based on the selected variables that were significant at the bivariate level.
Therefore, the following regression models were developed to predict the likelihood of medical and prompt treatment of malaria.

Model 1. (Model built with socioeconomic factors and type of treatment sought)
where P is probability of accessing medical treatment for malaria, a is the intercept, b i s are the slopes, X 1 is respondents' religion, X 2 is respondents' level of education, X 3 is respondents' occupational status, X 4 is respondents' place of residence, X 5 is respondents' wealth quintile, and X 6 is respondents' barrier to access health care facilities (n = 6).
where P is the probability of accessing prompt treatment, a is the intercept, b 1 s are the slopes, X 1 is respondents' wealth index, X 2 is respondents' place of residence (n = 2). All the analysis was done at 95.0% significance level (p -value, < 0.05).

Measurement of Some Variables
• • Malaria: Fever was used as a proxy for malaria.
According to the NDHS, fever is the main symptom of malaria, the proportion of febrile children in the population is a proxy for assessing malaria prevalence, and any reduction in the malaria disease burden should lead to a reduction in the overall prevalence of fever (

Result of the Study
The Table 1 below presents the sociodemographic characteristics of the respondents. Majority of the mothers are within the age 25 and 39 (64.1%) while 95.1% of them are married with shared religious affiliation as Christian (43.1%) and Muslim (54.9%). Almost half (46.5%) of the women had no education, while 31.0% and 22.5% had primary and post secondary education, respectively. By occupation, more than half (67.3%) of the mothers were working at the time of the survey. About half (49.2%) of the respondents are within the lower wealth index. Also, three quarter (75.1%) of the respondents lives in the rural area as one quarter of their counterparts lives within the urban centers (24.9%). By region of residence, 22.0%, 30.0%, 14.0%, and 17.2% of the women lives in North-East, North-West, South-East, and South-South, respectively, while others lives in North-Central (8.3%) and South-West (8.5%), respectively. More than half (54.1%) of the children received medical treatment; about one third (27.6%) received traditional or home-based treatment whereas about one fifth received no treatment during the recent episode of malaria. In addition, more than half (52.7%) of the mothers delayed in seeking medical treatment for their children while less than half (47.3%) of their counterparts sought prompt medical treatment for their children. In assessing barrier to access health care, findings showed that about one third (31.9%) of mothers had no barrier, one fifth (21.0%) had permission as a constraint, and more than one tenth (14.5%) had permission and money as a barrier, while 23.3% and 9.3% of the rest had permission/money/distance/transportation as a barrier in accessing health care facilities. The Table 2 below presents the bivariate analysis of the relationship between mothers' socioeconomic characteristics, type of treatment sought, and delay in seeking treatment. From the table, findings revealed no significant relationship between marital status (χ 2 = 11.32, p > .05) and type of treatment sought. Across all ages (χ 2 = 16.40, p < .05), tendency to seek medical treatments increases consistently with age of mothers. The proportion of never married (65.1%), widowed, or divorced women (66.3%) who sought medical treatment for their children is 10% point greater than the proportion of their married women (53.6%) counterparts.
Findings on mothers' age (χ 2 = 1.96, p > .05) and marital status (χ 2 = 3.70, p > .05) show no significant relationship with delay in seeking treatment. Overall, 47.3% of mothers received prompt treatment for their children against their counterparts (52.7%) who delayed in seeking treatment for their children. By marital status, less than half of never married mothers (41.4%) and married mothers (47.3%) treat their children promptly within 24 hr of onset of symptoms as against more than half (57.2%) of widowed/divorced mothers who treat their children promptly. Also, there exists a significant relationship between religion (χ 2 = 216.24, p < .05), education (χ 2 = 257.55, p < .05), occupation status (χ 2 = 21.88, p < .05), wealth index (χ 2 = 207.08, p < .05), type of residence (χ 2 = 18.56, p < .05), region of residence (χ 2 = 350.82, p < .05), and type of treatment sought for malaria. The proportion of Christian mothers who sought medical treatment (66.2%) for their children is about 20% point greater than the proportion of their Muslim (44.6%) counterparts who sought medical treatment for their children. Meanwhile, of those who sought medical treatment, less than half (48.4%) of Christian mothers and 46.2% of Muslim mothers sought prompt medical treatment as against more than half (51.8% vs. 53.5%) of their counterparts who delayed in seeking treatment.
Type of treatment and delay in seeking treatment varies consistently with mothers' level of education: 69.0% for mothers with secondary education and above, 59.0% for those with primary education, and 49.0% for those with no education. The percentage of mothers who did not seek any treatment or sought treatment at home/traditional places is 10% point lower among mothers with primary (12.6% and 27.9%) education compared with their counterparts without any education (25.8% and 32.5%). Although not significant, but the proportion of mothers who gave prompt malaria treatment to their children increased consistently with level of education: 44.8% versus 55.2% of mothers with no education; 47.3% versus 52.7% of those with primary education; and 50.4% versus 49.6% of those with secondary or post secondary education sought prompt malaria treatment and delayed treatment, respectively, for their children.
By occupational status, 56.7% of working mothers against 48.8% of non-working mothers received medical treatment for their children while 17.3% and 26.0% of working mothers as against 20.3% and 31.0% of non-working mothers received no treatment and home/traditional treatment for their children. Meanwhile, the proportion of mothers with prompt malaria treatment increase slightly with occupational status: 47.8% for working mothers versus 46.0% for nonworking mothers.
Also, more than half, 59.5% and 68.2%, of mothers within the middle and higher wealth index against 43.0% of those within the lower wealth index sought medical treatment for their children. Of those who sought treatment, larger percentage of mothers (57.9%) in the lower wealth index delayed as against 45.0% and 50.1% of their counterparts within the middle and higher wealth index that delayed in seeking treatment for malaria. Meanwhile, 58.7% versus 52.6% of urban and rural women sought medical treatment, respectively, while the percentage of rural women who delayed in seeking treatment is about 10% point greater than the percentage of their urban counterparts that delayed in seeking medical treatment.
By region of residence, more than half of the mothers in the South-East (71.9%), South-South (66.3%), North-Central (60.2%), and South-West (60.1%) sought medical treatment for their children, while about one third of them in the North-East (24.2%), North-West (34.3%), North-Central (27.5%), and South-West (32.8%) received either home or traditional treatment for their children. Meanwhile, North-East (21.7%) and North-West (33.1), across all regions, signify a region with higher percentage of children without any treatment during the illness. As the numbers of difficulties faced in seeking treatment increase (57.7% for permission; 55.1% for permission and money; 52.1% for permission, money, and distance; and 35.3% for permission, money, distance, and transportation), the percentage of women who sought medical treatment for their children decreases, respectively. Also, the proportion of mothers whose children received prompt treatment varies consistently with numbers of difficulties encountered.
The Table 3 below presents the odds ratio (OR) of the general binary logistic regression for the likelihood of receiving medical and prompt treatment for children aged 12 to 59 months who had malaria within 2 weeks preceding the survey. The logistic regression was built with two models by selecting all mothers' socioeconomic factors that have a significant relationship with type of treatment and delay in treatment at the bivariate level of analysis. The first model therefore examined the likelihood of receiving medical treatment with the selected mothers' socioeconomic factors while the second model examined the odds of giving prompt treatment with the selected significant socioeconomic factors at the bivariate level of analysis having controlled for region of residence.
Findings from logistic regression at this multivariate level of analysis showed that the odds of seeking medical treat Meanwhile, the logistic regression at this level found no significant relationship between mothers' occupational status and likelihood of seeking medical and prompt treatment for childhood malaria. However, findings showed that the odds of seeking medical treatment among mothers who were not working (OR = 0.88; p > .05; CI = [0.57 -0.88]) is less when compared with that of their counterparts in the working class category. Furthermore, Islamic mothers (OR = 0.72; p < .05; CI = [0.55 -1.71]) and those in the other religion (OR = 0.98; p < .05; CI = [0.14 -2.27]) are significantly less likely to take their children for medical treatment compared with their Christian counterparts.
Also, the odds of giving medical treatment (OR = 0.34; p > .05; CI = [1.07 -1.69]) as well as prompt treatment (OR = 0.82; p > .05; CI = [0.63 -1.06]) among mothers in the rural area are significantly less when compared with the odds of giving medical treatment as well as prompt treatment among mothers in the urban centers.

Discussion of Findings
This study examined mothers' socioeconomic differentials and management of childhood malaria using the 2008 NDHS kids recode data set. Overall, 3,968 children had fever 2 weeks before the survey with majority of them within the age bracket 12 to 35 months. Averagely across all ages, exactly half of the children received treatment from medical centers, about one third received treatment at home or traditional health providers while about one fourth of the rest received no treatment at all. Meanwhile, against the fourth key element of RBM strategy of accessing appropriate and adequate treatment within 24 hr of onset of symptoms, a little more than half of the mothers delayed beyond 24 hr before seeking treatment for their children. In line with D' Souza's (2003) research findings, delay in seeking treatment could be influenced by past experience with similar illnesses which motivate mothers to play a "waiting game" to see if the illness subsides on its own, particularly in situations where the cost of care is an inhibitory factor (D'Souza, 2003).
Urban mothers were found to have better chance of giving their children medical and prompt treatment during malaria illness than their rural mother counterparts. Also, greater percentage of children within South-East, South-South, North-Central, and South-West regions received medical treatment when compared with other regions. Conversely, greater percentage of mothers in the North-East, North-West, and North-Central have the highest proportion of children with fever whose treatment were delayed or were not treated at al or treated at home with traditional herbal medicine. These findings are in line with Guyatt and Snow's (2004), Kofoed et al.'s (2004), and Müller et al.'s (2003) research findings, which concluded that malaria treatments are rarely sought at health care facilities and are most often inappropriate or delayed.
The significant relationship found between religion, medical choice of treating malaria, and prompt treatment of malaria corroborate earlier findings by Adamu (2001), Addai (2000), and Adetunji (1996), as this study revealed that Islamic mothers and those in the other religion are significantly less likely to give medical and prompt treatment to their children during malaria illness when compared with their Christian counterparts.
In line with earlier studies by Pérez-Cuevas et al. (1996) and WHO/UNICEF (2003), which associate maternal literacy and mothers' education with utilization of health care facilities, this study also revealed that type of treatment and prompt treatment of malaria varies monotonically with mothers' level of education. Mothers with no education or with primary education are less likely to seek medical and prompt treatment for their children compared with their counterparts with secondary education or more. Efforts therefore need to be intensified on health education with emphasis on the importance of identifying vital malaria signs and giving prompt and appropriate care to febrile children within 24 hr of onset of the symptoms as advocated by WHO (2004).
Also, this study showed that the proportion of mothers with prompt malaria treatment increased slightly with occupational status. That is, mothers who are currently working are more likely to seek medical and prompt treatment of malaria for their children. In addition, mothers' wealth index showed a significant relationship with type of treatment sought and delay in seeking treatment. Mothers within the middle and lower wealth index are significantly less likely to seek medical treatment; as they are also less likely to seek prompt treatment for their children when compared with their counterparts within the higher wealth index.
Furthermore, findings revealed that larger percentages of urban mothers sought medical treatment for their children when compared with their counterparts in the rural areas. The odds of giving medical treatment and prompt treatment among mothers in the rural area was found to be significantly less when compared with the odds of giving medical treatment and prompt treatment among mothers in the urban centers. In addition, our findings showed that rural mothers are more likely to delay in seeking treatment compared with their urban counterparts. These findings are in line with the studies conducted by Adeyemi (2000), Mekonnen and Mekonnen (2002), Sabitu (2004), and UNFPA (2002,2004), which concluded that there are differences in utilization of health care facilities and maternal health care services between urban and rural areas. Findings on barrier or difficulties in accessing health care facilities and prompt treatment revealed a significant relationship in line with research findings by Hill et al. (2003), Mermert and Nsabagasanyi (2002), and Namusobya (1998), where they concluded that lack of finance for medical consultation and treatment, distance from health care facilities, and so on, are important barriers to accessing health care facilities. Therefore, these research findings also concluded that difficulties/barriers to access health care facilities significantly decrease medical and prompt treatment of malaria in children.

Conclusion and Policy Recommendation
Having used both bivariate and multivariate level of analysis, findings from this study concluded that mothers' socioeconomic characteristics and barriers to access health care facilities are significant factors militating against medical and prompt treatment of malaria in Nigeria. As revealed by this study, about 50% of febrile children are never seen within health care facilities. They are either treated at home or not treated at all. Therefore, appropriate and prompt use of health care facilities for treatment of malaria within 24 hr of onset of symptoms are critical to the successful achievement of the United Nation's Millennium Development Goal 4, which calls for a two thirds reduction in child mortality by 2015.
Although, National Policy on Malaria Diagnosis and Treatment (Federal Ministry of Health, 2005 promotes the treatment of uncomplicated malaria at home with artemisinin-based combination therapies (ACTs). However, it is disappointing that many people did not know about ACTs, and many of those who know cannot afford it. Also, misdiagnosis and management of malaria for other febrile illnesses at home have been established by studies to be contributing to over-diagnosis of malaria and irrational use of anti-malarial drugs which have the potential for facilitating resistance to these drugs as emphasized by WHO (2010). Therefore, to reduce childhood mortality due to malaria, home management of malaria must be discouraged, "both by the ministry of health and the religious leaders," and mothers must be sensitized to respond promptly to treatment of malaria in their children within health care facilities. Also, modern, equipped, and affordable health care facilities must be increased in the rural areas. Most of the rural areas do not have access to good health care systems. Usually, there are no accessible roads to the health centers, which in turn are poorly equipped and have inadequate drugs for malaria treatment. This poses a serious threat to clinical management and treatment of malaria and also influence people who cannot afford anti-malarial drugs to tend toward self-medication with local herbs. Finally, as mothers' education has been associated with medical and prompt treatment of malaria, investing in women's education through free education as well as intensive campaign to motivate them to go for higher level of education will be a worthwhile effort. Higher education will enhance their wealth index and occupational status which will in turn have a positive influence on their ability to care for children.

Limitation of the Study
The study make use of a secondary data obtained from the 2008 Nigeria Demographic and Health Survey, which is a nationally representative data, and the author was not opportune to get the data in its natural settings. Thus, there are some limitations as predetermined by the data in terms of the questionnaire design and variable measured. For instance, mothers' knowledge of severity of malaria sickness as well as perceived benefit of prompt initiation of treatment was not measure and therefore cannot be measured by this study. Also, cases of missing values and high non-response rate are common in the data set, and this was responsible for some variation in the row totals in the "Analysis" section.

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) received no financial support for the research and/or authorship of this article.