Theses / Dissertations >
College of Health Sciences >
Please use this identifier to cite or link to this item:
|Title: ||Spatial analysis of under-5 malaria in Ghana (2008 GDHS) for public health intervention evaluation|
|Authors: ||Atinuke, Olusola Adebanji|
|Issue Date: ||18-Jan-2017|
|Abstract: ||Malaria control programmes adopt national strategies without recourse to individual, social and location factors that affect the dynamics of the infection transmission. Spatial relationships between locations are often unaccounted for with the number of cases assumed to be randomly occurring. This renders inappropriate a one-size-fits all intervention approach for a parsimonious allocation of limited resources for the design, implementation and evaluation of malaria control programmes. To accomplish the desired outcomes, interventions must reach the subgroups of the populations where malaria prevalence is at its highest. This can only be achieved if malaria control interventions take into cognisance factors that ensure equity in program design, implementation, monitoring and evaluation. Impact assessment requires nationally representative data that are often prohibitively expensive. To circumvent this, nationally representative routine data can easily be adopted
Methods: This study utilizes the GDHS 2008 data for the identification of disease patterns for under-5 malaria using individual child, maternal and household socio-economic and geographic variables as predictor variables. These variables were used to model the likelihood of U-5 malaria and to develop the risk/prevalence maps to explain the risk and spatial dependencies of the incidence of malaria between households. The chi-square test of independence, binary logistic regression and spatial statistics analytical tools were utilized.
Results: Logistic regression model showed age group 5, anaemia, type of residence and epidemiological zones as significant (p <.05) with odds ratios less than 1 of the incidence of U-5 malaria. Model overall percentage performance was 74.7%. Moran’s index for non-randomness in occurrence was significant (p <.02) and LISA cluster maps showed significant clustering between 45 clusters with most of them in the southern part of Ghana.
The malaria risk maps showed 4 dense clustering of low prevalence centres clustered around hot spots in Northern, Ashanti, Western and Greater Accra regions. Identifying these high prevalence foci is important to the success of any public health intervention or vector control. These are important conduits for malaria transmission. The SEM was found to give a better fit than the SLM based on the AIC and Schwarz criteria.
Conclusion: The use of statistical models for predicting probability of under-5 malaria could be adopted as a household risk index for children’s exposure to malaria. The spatial clustering and risk maps have been utilized to depict distribution of the observed pattern of under-5 malaria prevalence and demonstrated as an efficient tool for quick identification hot and cold spots of disease prevalence. This can provide valuable insight into the underlying mechanisms driving incidence and transmission of malaria in children|
|Description: ||Thesis submitted to the Department of Population and Reproductive Health, College of Health Sciences, School of Public Health in partial fulfilment of the requirements for the Award of the degree of Master of Public Health in Population, Family and Reproductive Health, 2016|
|Appears in Collections:||College of Health Sciences|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.