Research proposal writer
CHAPTER FOUR
PRESENTATION OF RESULTS
4.0 Introduction
This chapter discusses the data analysed from secondary data source on analysis of malaria prevalence in children aged less than 5 years in Uganda from 1989 to 2018, The data where subject to transformation using SPSS, where the mean, standard deviation, correlations, Regression, Forecasting and ANOVA. Presentation and interpretation were based on specific objectives to address the problem.
Unit Root Tests
Unit root tests were carried out using the augmented Dickey-Fuller test statistic. This was carried out to check whether the series were stationary (integrated) or not.
Unit root tests of the series 1989-2018
| Variable in level | Variable in 1st difference | |||
| ADF | Critical value (5%) | ADF | Critical value (5%) | |
| RURAL | -2.010219 | -3.6852 | -6.527535 | -3.6959 |
| URBAN | -0.155668 | -3.029 | -4.478099 | -3.6959 |
In the table 4, the urban and rural are stationary in the levels because the ADF statistics are greater than critical values while after the 1st difference they are non stationary because the ADF statistic are less than the critical values.
Histogram normality test
Source: primary data
A regression was run and on clicking on the view-residual test-histogram-normality test, the histogram is bell-shaped, suggesting a normal curve shape, and the jarque-bera statistics has high p-value of 1.779 indicating that the errors in the regression are normal that is to say; the jarque-bera statistics probability of 0.410820 is greater than zero and it has a percentage of 41% greater than 10%(41%>10%) thus the errors in the regression are normal.
Comparison of Malaria Cases Residence
This analysis was done by the researcher to compare the malaria cases between rural areas and urban areas by determining the mean and standard deviation, this was to enable the researcher identify which region had the highest malaria prevalence against the other as shown in the table below;
This was done to answer the second hypothesis;
HO : Rural does not have the same malaria prevalence as urban
Ha : Rural has the same malaria prevalence as urban
The comparison is done by comparing the mean and standard deviation of rural and urban places of residence as shown in the table below.
Comparison of Malaria Cases Residence
| N | Minimum | Maximum | Sum | Mean | Std. Deviation | |
| Rural | 31 | 1698 | 3952 | 92635 | 2988.23 | 457.718 |
| Urban | 31 | 216 | 793 | 9744 | 314.32 | 142.437 |
Source: primary data
The study shows that rural residence had mean total malaria cases of 2988.23 with a standard deviation of 457.718 while urban residence had mean of 314.32 with a standard deviation of 142.437. This implies that there was much variation in the data in rural centers compared to urban areas for the last 31 years and the prevalence was high in rural areas compared to urban areas for last 31 years, These results in the study indicates majority of the malaria cases occurs in the rural areas than in urban areas. This also implies that in Urban areas there is better health facilities than in rural areas therefore the malaria cases were lower this is also in line with Warrell et al,( 2002) who states that Malaria is a major cause of morbidity and mortality worldwide, especially in young African children under five years especially in rural areas because of lack the necessary facilities and limited number of hospitals especially in rural African communities and the prevalence of rural areas is high because of high illiteracy rates and lack of knowledge in rural areas.
Comparison of Malaria Cases by Region
The table below gives a comparison of the malaria cases (prevalence) among different regions across the Country-Uganda that was collected from time series data for last 31 years.
This was carried out to answer the third objective.
Ho: There is no difference in malaria prevalence by region
Ha: There is difference in malaria prevalence by region
The comparison in this table was done using mean and standard deviation in which the area with the highest mean had the highest malaria cases while low mean value indicated low prevalence of malaria cases.
Comparison of Malaria Cases by Region
| N | Minimum | Maximum | Sum | Mean | Std. Deviation | |
| north | 31 | 541 | 1330 | 27977 | 902.48 | 158.922 |
| east | 31 | 390 | 1304 | 31683 | 1022.03 | 192.211 |
| west | 31 | 263 | 1132 | 17920 | 578.06 | 184.080 |
| central | 31 | 546 | 1236 | 24404 | 787.23 | 154.660 |
| Valid N (leastwise) | 31 |
Source: ministry of Health Uganda
From the above table the study revealed that Eastern region had the highest registered number of malaria cases with highest mean of 1022.03 and standard deviation of 192.211 followed by Northern with a mean of 902.48 and the standard deviation 158.922. Western and central regions had mean malaria cases of 578.06 and 787.23 with the standard deviation of 184.080 and 154.660 respectively. Despite eastern region registering the highest number of malaria there was little variation in the data and therefore there is difference in malaria cases by region.
From the above results majority of malaria cases were in eastern Uganda this is because eastern Uganda has not been able to engage into the preventive malaria programmes this explains the reason for the high malaria prevalence for children under the age of 5 years, this is also in line with (US PMI, 2009) which states that in an area where there is little malaria prevention like encouraging people to sleep under treated mosquito nets has escalated malaria cases in such regions.
| Model | Unstandardized Coefficients | Standardized Coefficients | t | Sig. | ||
| B | Std. Error | Beta | ||||
| 1 | (Constant) | 604.537 | 176.294 | 3.429 | .002 | |
| north | -.366 | .202 | -.376 | -1.806 | .082 | |
| east | -.062 | .112 | -.077 | -.552 | .585 | |
| west | .996 | .195 | 1.186 | 5.118 | .000 | |
According to the table above it indicates that when malaria cases in the north, east and west equals to zero then malaria cases is 604.537, while a unit increase in malaria cases in central will on average lead to 0.366 decrease in the north, this is statistically insignificant since the P-value, (0.082) >0.05 thus the null hypothesis is accepted, therefore there is no difference in malaria prevalence across the north and central regions.
A unit increase in central will on average lead to (0.062) decrease of malaria in the East, this is statistically insignificant since the P-value (0.585)>0.05, the null hypothesis is accepted, hence there is no difference in malaria cases across the east and central regions.
A unit increase in central will on average lead to (0. 996) increase of malaria cases in the west, this is statistically significant since the P-value (0.000)<0.05, the null hypothesis is rejected, hence there is a difference in malaria cases across the west and central regions.
This implied that the central has lower cases of malaria compared to north, east and west, then reject null hypothesis and conclude that there is a difference in malaria prevalence by region, this is also in line with (Okello et al., 2006) who states that the climate in Uganda allows stable, year round malaria transmission with relatively little seasonal variability in most areas. Malaria is highly endemic in Uganda affecting approximately 90% of the 34 million population , most of the regions of Uganda especially eastern and north are face too much malaria prevalence as compared to the other parts of the country though the whole country faces significvant amounts of malaria cases.
Correlations of Malaria Cases by Residence
The relationship of malaria prevalence by residence in Uganda for the study that was conducted, this was carried out in order to ensure that the researcher is able to determine the relationship between malaria prevalence in urban and rural areas.
Ho: There is no difference in malaria prevalence by residence.
Ha: There is difference in malaria prevalence by residence.
The correlation table below indicates the malaria prevalence on the places of residence either rural or urban however the null hypothesis was rejected incase the P-Value was greater than 0.05, therefore the null is rejected, otherwise it is accepted.
Correlations of Malaria Cases by Residence
| rural | urban | ||
| rural | Pearson Correlation | 1 | .553** |
| Sig. (2-tailed) | .001 | ||
| N | 31 | 31 | |
| urban | Pearson Correlation | .553** | 1 |
| Sig. (2-tailed) | .001 | ||
| N | 31 | 31 | |
| **. Correlation is significant at the 0.01 level (2-tailed). | |||
Source: Primary Data
There is a high significant and positive relationship of malaria cases between rural and urban parts of Uganda. The results are significant since sig 0.001<0.05. This results indicates that malaria cases in both regions are high and there is a high positive relationship between the regions in terms of malaria prevalence, these results are also in line with WHO (2013) and the Global Malaria Action Plan 3.4 billion people (half the world’s population) live in areas at risk of malaria transmission in 106 countries and territories .In 2012, malaria caused an estimated 207 million clinical episodes, and 627,000 deaths. An estimated 91% of deaths in 2010 were in the African Region and both people in urban and rural areas are at risk of getting malaria unless the government in their respective governments.
Regression of total malaria cases on rural and urban
The table below shows Regression of total malaria cases on rural and urban in Uganda for the study that was conducted, this was carried out in order to enable the researcher determine percentage contribution of the independent variables on the dependent variable and also determine variations in the total malaria cases in rural and urban.
This was in order to determine the hypothesis.
Ho : Rural does not have the same malaria cases like urban.
Ha : Rural has the same malaria cases like urban.
The regression analysis was carried out to determine the relationship between total malaria cases and malaria cases by region.
The higher the R-square value the more the relationship and the lower the R-square Value the lower the relationship.
Regression of total malaria cases on rural and urban
| Model Summaryb | ||||||||||
| Model | R | R Square | Adjusted R Square | Std. Error of the Estimate | Change Statistics | Durbin-Watson | ||||
| R Square Change | F Change | df1 | df2 | Sig. F Change | ||||||
| 1 | .996a | .992 | .992 | 49.580 | .992 | 1822.962 | 2 | 28 | .000 | 2.233 |
| a. Predictors: (Constant), urban, rural | ||||||||||
| b. Dependent Variable: ttmlc (Total malaria cases) | ||||||||||
Source: ministry of health Uganda
The R-value tells us about correlation coefficient (0. 992) means that there is a very high positive relationship between malaria cases by region and total malaria cases. R Square value explains the percentage contribution of the independent variables on the dependent variable.
Therefore approximately 99.2% of the variations in the total malaria cases are explained by the changes in the malaria cases of both rural and urban areas. This view is also in line with (Moh, 2014) which states that there is very high malaria prevalence in both rural; and urban areas and therefore the government ought to increase more investments in the fight against malaria and ensure the citizens are aware of the preventive malaria programmes to eliminate the increase in the malaria volumes in the country.
The regression analysis was carried out to calculate if places of residence have an effect on the total malaria cases, this was carried out to answer the two hypotheses.
Ho : Places of residence do not have an effect on total malaria cases.
Ha : Place of residence have an influence on total malaria cases.
The Anova indicates whether places of residence have an effect on the total malaria cases. This was determined by P-value, when the P-value is less than 0.05 the we reject the null hypothesis and otherwise we accepted.
The ANOVA table below explains the overall significant of the model.
| ANOVAa | ||||||
| Model | Sum of Squares | df | Mean Square | F | Sig. | |
| 1 | Regression | 8962409.258 | 2 | 4481204.629 | 1822.962 | .000b |
| Residual | 68829.580 | 28 | 2458.199 | |||
| Total | 9031238.839 | 30 | ||||
| a. Dependent Variable: ttmlc(total malaria cases) | ||||||
| b. Predictors: (Constant), urban, rural | ||||||
Source: Ministry of Health Uganda
Basing on the hypothesis that the independent variables have no effect on malaria cases, since the sig value 0.00 <0.05 we reject the null hypothesis and conclude that the independent variables (places of residence) have an effect on the total malaria cases. This is also in line with (Makundi et al, 2007) it was reported that the burden of malaria is greatest among poor people, imposing significant direct and indirect costs on individuals and households and pushing households into in a vicious circle of disease and poverty. Furthermore vulnerable households with little coping and adaptive capacities are particularly affected by malaria.
The table is used to explain the effect of places of residence on malaria cases, this was done by using coefficient tests, the model is estimated as Y=B+B1X1+B2X2. This was specifically to enable the researcher determine malaria prevalence in both rural and urban areas of the country.
Where B is constant, B1 is coefficient for rural , x1 is rural areas and B2 is the coefficient for urban ,x2 is urban areas and Y is the malaria prevalence at the time in a given region and residence.
| Model | Unstandardized Coefficients | Standardized Coefficients | t | Sig. | 95.0% Confidence Interval for B | |||
| B | Std. Error | Beta | Lower Bound | Upper Bound | ||||
| 1 | (Constant) | -1.849 | 2.668 | -.693 | .511 | -8.158 | 4.460 | |
| rural | 1.001 | .001 | .809 | 903.798 | .000 | .999 | 1.004 | |
| urban | .995 | .004 | .251 | 280.500 | .000 | .987 | 1.004 | |
| a. Dependent Variable: ttmlc | ||||||||
From the table, the coefficients are -1.849, 1.001, and .995 for B, B1, and B2 respectively. All the effects are significant since the sig-values for both rural and urban are less than 0.05. In summary residences from urban areas have higher chances of contracting malaria compared to their village counterparts this is also in line with (WMR, 2015), which indicates that, malaria transmission occurs in five of the six WHO regions, with Europe remaining free. Globally, an estimated 3.2 billion people continue to be at risk of being infected with malaria and developing disease, and 1.2 billion are at high risk .more than 1 in 1000 are at a high chance of getting malaria in a year. There were 214 million cases globally in 2015, of which 88% were from the African region, 10% from SE Asia region and 2% from Eastern Mediterranean region. There were an estimated 438000 deaths, 90% from Africa, 7% from SE Asia region and 2% from Eastern Mediterranean region.(WMR, 2015)in comparison, 198 million infections and 584 000 deaths were estimated in 2013.
4.1.6 Forecasting model for the year 2019
The researcher forecasted total malaria cases for 2019, this was done to determine the future malaria prevalence for the year 2019.
Forecasting model for the year 2019
The model from the above table was obtained by regressing total malaria cases against time to be able to forecast malaria cases for the 2019. The model is written as Y=B+B1X1 where B is constant B1 is the coefficient of years. Hence the model is fitted as Y=-40825.148+22.065X1
Autoregressive Integrated Moving Average forecasting
The table below shows the unit root tests for malaria prevalence by residence in Uganda for the study that was conducted, this was carried out in order to ensure that the researcher is able to determine if the time series was stationary or non stationary hence use the unit root test.
This was to answer the hypothesis;
Ho : There was no stationarity
Autoregressive Integrated Moving Average forecasting
The forecasting of the model required to first test the principle of stationarity in order to find out if the time series was stationary or non-stationary hence use the unit root test. From the above results the Dickey Fuller value (18)>stationary R-squared (0.749) therefore we accept the hypothesis and conclude that there is a unit root hence non stationary of the time series data.
Correlation between the regions
The table below shows correlation between the different regions in Uganda for the study that was conducted, this was carried out in order to enable the researcher determine relationship between the four regions across the country namely; north, east, west and central as shown in the modle summary table, Anova table and table for the Coefficients as shown below.
This was in order to determine the hypothesis.
Ho : There is no difference in malaria prevalence by region
Ha : There is difference in malaria prevalence by region.
The researcher used the correlation coefficient to determine malaria prevalence across different regions, this was specifically to determine malaria prevalence in central east and west using the northern region as a constant, the researcher used the P-value, when the P-Value is less than 0.05 we reject the null hypothesis and other wise accept the hypothesis.
Correlation between the regions
| Model Summary | |||||||||
| Model | R | R Square | Adjusted R Square | Std. Error of the Estimate | Change Statistics | ||||
| R Square Change | F Change | df1 | df2 | Sig. F Change | |||||
| 1 | .937a | .877 | .816 | 106.419 | .877 | 14.287 | 3 | 6 | .004 |
| a. Predictors: (Constant), central, east, west b. | |||||||||
From the figure above the results indicates that the R Square 0.877 implying that 87.7 of malaria prevalence is due the difference in the regions, this is also indicated by significance level of 0. .004 <0.05, therefore the results rejects the null hypothesis and the conclusion that there is a difference in malaria prevalence among the regions. This is also in line with (Collins et al, 1997& Yadav et al., 1999) who states that socio-economic conditions of the community have direct bearing on the problem of malaria in Uganda some very poor parts of the country experience more cases of malaria cases. Ignorance and impoverished conditions of people contribute in creating source and spread of malaria and hinder disease control strategy. This was also evidenced by Filmer (2002) that high costs of malaria treatment may lead to delays in treatment seeking behavior, whereby he found that the poorest groups in a society did not seek care as much as the non-poor, and did so at lower level public facilities.
