Environmental determinants of patterns and trends of the occurrence of unstable malaria in Kericho District, Kenya

Abstract

This study investigated environmental determinants of patterns and trends of occurrences of unstable malaria burdens in Kericho district of Kenya, a formerly malaria free highland district in early 1960s but classified by Kenya government in 2001 as malaria epidemic prone district where fatal malaria epidemics have reemerged seasonally each year since 1980s and represents a leading cause of morbidity and mortality in the district. To achieve this objective patterns and trends of malaria burdens; malaria hospital admissions, morbidity and mortality, climatic elements; and households' characteristics and district health delivery system were investigated. Secondary data on malaria burdens and climatic element were obtained from Kericho district health facilities and Kericho meteorological station respectively for the synoptic period 1988-2005. Primary households' characteristics were obtained using questionnaires from a randomly sampled households (N=301) apportioned to all four Agro-Ecological Zones (AEZs) study sites in seven administrative divisions based on 1999 national population census. Useful information on health efficiency delivery system in the district was obtained from key medical professionals respondents to questionnaires in Kericho district health facilities. The data were analysed displayed and interpreted using statistical methods; tables, graphs, means, percentages, ranges, correlation, regression analysis and tests of significance (t-test) and Analysis of Variance (ANOVA). Tables and time series graphs, were used to study emerging patterns and trends in malaria burdens and climatic elements over time. Correlation and regression methods were used to determine the relationship between malaria burden and climatic and socio-economic data. The West was used to establish the significance of computed correlation coefficients between malaria morbidity and each of the household's variables. Archview Geographic Information System (GIS) software was used to create a malaria zonation map of the distribution of incidences of malaria morbidity in the district and SWOT analysis was used to analyse the effectiveness of malaria control, prevention and management algorithms in the district. Findings emerging from this study showed that outpatient malaria cases grew from 19643 cases in 1989 to 124408 cases in 1994 or a growth rate of 106.7% per annum and growth rate of 101.9% per annum in 1995-2002. Malaria hospital admissions grew from 10.27% per annum and a growth rate of 42.8% per annum in 1988-1994 and 1995-2002 respectively. Malaria deaths in Kericho district main hospital grew from 23 cases in 1988 to 34 in 1994, a growth rate of 6.9% per annum and in 1995-2002 malaria deaths grew from 84 to 160 cases or an annual increase of 18.1 %. Overall total 70611 malaria hospital admissions were recorded in health facilities in the district and resulted in 1476 deaths or 21 malaria deaths in every 1000 malaria admissions. Malaria hospital admissions were seasonal in patterns of occurrences with peak occurrences in June- July each year and generally malaria hospital admissions increased gradually between April and July and gradually decreased in August-September to March each year. Case fatality rates (1988-2002) ranged between 0.02% to 4.2% or 10 cases to 24 cases per 1000 malaria admissions in 1988-1990 and 2 to 21 cases per 1000 malaria admissions in1990-2002. Trend in malaria hospital admissions was on upward trend in 1988-2002 with a growth rate of 111.13% while malaria death rate grew by 10.486% per annum and West revealed that trend in observed malaria hospital admissions was significant at 0.05 thus the null hypothesis stated that there was no significant difference in temporal trend discernible in occurrence of malaria morbidity in the district was rejected at 0.05 level. The mean reporting rate by health institutions changed from 27.2% in 1999 to 67.1% in 2003 and proportion (%) of population who needed treatment for malaria and actually received it increased from 10.8% to 22% in 1999-2004 in the district. In 1988-2005 the annual mean temperature, rainfall and relative humidity were 17.6°C,153mm and 64.6% respectively while monthly means were17.7°C, 162.3mm and 65% respectively. Results showed emerging positive trend in temperature with a change of 0.8°C but no trend in monthly values. Negative trend in mean annual rainfall was observed with change of 20mm and a negative growth of 0.78% during 1988-2005. The regression model regarding malaria admissions produced coefficient of determination R2=0.3052 i.e. 30.52% of the change in time accounted for 30.52% variation in malaria burdens (admissions) while time accounted for 34.79% of the variation in malaria deaths due to the disease and the West revealed that the difference in trends in malaria hospital admissions in 1988-1995 and 1995-2002 was significant at 0.05 level while difference in trends in malaria deaths during the same periods was significant at 0.05 levels. Thus the null hypothesis that there are no significant trends discernible in occurrence of malaria burdens and change in time in the district was rejected at 0.05 level. Regression results revealed that change in time (1988-2002) accounted for 95.35% variation in mean annual temperature, 7.9% and 0.83% variations in mean annual rainfall and relative humidity respectively. Results revealed that malaria morbidity correlated strongly (0.943) with frequency of visits to the nearest health facility and this value being significant at 0.05 level (2 tailed test). However malaria morbidity lowly (0.134) correlated with family size. Malaria morbidity in a household correlated positively (0.321) with distance to the nearest health facility and this value being significant at 0.01 level (2-tailed test), positive sign implied that distance from a household influenced malaria morbidity and might have led to delay in prompt treatment of malaria thus an increase of malaria transmission in a household. Family income correlated negatively (-0.114) with family size, a correct sign implied that a low family income was associated with a large family size. Family income correlated negatively with each of the variables maternal age and maternal education with values of 0.205 and 0.238 respectively and both values being significant at 0.01 level (2-failed test). However, family income did not have effect on malaria occurrences in a household. GIS results created 5 malaria zones in the district based on annual malaria morbidly cases as; very high malaria zone. 10855-15029 cases, high malaria zone 7917-10834 cases, medium zone 6239-7916 cases, low zone 5367-6238 cases and least malaria zone had below 5366 malaria cases annually. Overall 60% of respondents interviewed (N= 301) purchased anti-malaria drugs from local kiosks/shops without the prescription from qualified health personnel. Only 21.2% of all 250 medical staff in the district public health facilities received on job training on malaria control and information technology during 2000-2005. Less than 12% of the households' respondents (N= 301) interviewed used ITNS regularly. The uncollected used plastic/polythene materials were observed in 60% of 301 households' compounds visited in 2006 where 70% of all households' respondents knew and used local herbs/plants polyganum, AcaciaSiebariana, Cyathilaennacea, SP and Septenitrionalis to treat malaria and more than 61% knew Tegetes minuta repulsed away mosquitoes. The efficacy of anti-malaria drugs: Coartem, Helfan, Paluther and Cotexcin ranged between 50% and 80% in 1999-2005. SWOT analysis results have revealed future strategies for effective malaria control programmes in the district. Overall strategies opportunities and strengths in malaria control exist and should be stepped up in the district to meet the long term strategies and to be in line with NMCS, 2001-2010 framework. The study concludes that to prevent and control malaria, more malaria educational programmes, improvement of accessibility to health facilities provision of free ITNS and improvement of environmental sanitation should be stepped up as well as full implementation of malaria control programmes according to ministry of health, 2001-2010 framework.