Last update
Date: 28 September, 2024
Author: Emmanuel Kirui Barkacha
Introduction
Problem Statement and Objectives
Study Area
Required Data and Analysis
Malaria remains a significant global health problem, particularly in sub-Saharan Africa.
In Kenya, an estimated 27 million malaria cases and 12,600 deaths were reported in 2020.
Malaria risk in Kenya is heterogeneous, and its epidemiology is influenced mainly by
Altitude
Rainfall patterns and temperature
Previous interventions
Kenya has five malaria zones:Coastal endemic, Lake endemic, Seasonal transmission, Highlands epidemic and Low epidemic
Lake and coastal zones have year round transmission.
Highland and arid zones experience seasonal and epidemic malaria.
Kakamega county is classified under the lake endemic zone.
It experience intense malaria transmission year-round due to its favorable climatic conditions.
Despite substantial progress in reducing malaria morbidity and mortality in Kenya, the disease remains a significant public health problem, particularly in western regions such as Kakamega County, where malaria transmission is intense and highly heterogeneity. Traditional malaria control strategies face challenges due to varying local epidemiological and environmental conditions. While mathematical models like OpenMalaria which is an individual based model offer valuable insights for predicting the impact of interventions, they require accurate calibration using local data to be effective. Currently, there is a lack of enough detailed, locally calibrated models for malaria transmission in high-prevalence and high heterogeneity areas like Kakamega County, which limits the effectiveness of targeted interventions and policy decisions.
To calibrate the OpenMalaria model at admin 1 (county) level in a high spatial aggregation setting in Kakamega County.
To assess the additional value of calibrating the OpenMalaria model at admin 2 (sub-county) vs admin 1 (county).
To assess the spatial variability of malaria transmission within Kakamega County, identifying high-risk areas and factors contributing to this variability.
To evaluate the impact of the difference between the two calibrated models on the simulation of future interventions.
Kakamega County is located in the western region of Kenya. It is characterized by a diverse range of geographical and ecological features that make it a unique area for studying malaria transmission and control. The county lies between latitudes 0°07’30.0”N and 0°16’30.0”N and longitudes 34°32’30.0”E and 34°57’30.0”E. Kakamega is bordered by Vihiga County to the south, Siaya County to the west, Bungoma County to the north, and Nandi County to the east. The county covers an area of approximately 3,051.3 square kilometers.
The county is divided into 12 sub-counties: Lugari, Likuyani, Malava, Lurambi, Navakholo, Mumias West, Mumias East, Matungu, Butere, Khwisero, Shinyalu, and Ikolomani. Each sub-county exhibits unique demographic, socioeconomic, and environmental characteristics that influence malaria transmission dynamics.
Number of sub counties: 12
The table below shows the size of each sub-county from the smallest to the largest. Shinyalu is the largest Sub-county of the twelve with an approximate area of 445.5 Km2 whereas Ikolomani is the smallest with an approximate area of 143.6 Km2.
| SubCounty | Area_Km2 |
|---|---|
| Ikolomani | 143.6 |
| Khwisero | 145.6 |
| Mumias East | 149.2 |
| Lurambi | 161.7 |
| Mumias West | 165.3 |
| Butere | 210.4 |
| Navakholo | 258.0 |
| Matungu | 275.8 |
| Likuyani | 302.0 |
| Lugari | 367.0 |
| Malava | 427.2 |
| Shinyalu | 445.5 |
| County | 3,051.3 |
Kakamega County experiences a tropical rainforest climate, classified as Af under the Koppen climate classification. The county receives high annual rainfall, ranging from 1,280.1 mm to 2,214.1 mm, with two peak rainy seasons: the long rains from March to July and the short rains from December and February. The high rainfall and moderate temperatures, which average between 18°C and 29°C, create ideal conditions for the proliferation of Anopheles mosquitoes, the primary vectors of malaria. The county has also an average humidity of 67 per cent.
For the precipitation data we are using the chirps library to load it for Kakamega county for the period \(1^{st} \text{ January } 2010\) to \(31^{st} \text{ December } 2023\) coming from Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS) CHIRPS. This data from precipitation will be used as seasonality data as malaria transmission and precipitation are correlated. Since the chirps data is per day we will aggregate it to per month for each sub-county.
According to the 2019 Kenya Population and Housing Census Report by Kenya National Bureau of Statistics KNBS, the County’s population was 1,867,579 persons comprising of 897,133 males and 970,406 females, representing 48 percent male and 52 percent female. Malava Sub-County had the highest population of 238,330 persons while Shinyalu Sub County had the lowest population at 111,743 persons. This makes the County the fourth most populous county after Nairobi, Kiambu and Nakuru counties.
| SubCounty | Male | Female | Total |
|---|---|---|---|
| BUTERE | 73,634 | 80,463 | 154,097 |
| IKOLOMANI | 80,853 | 86,784 | 167,637 |
| KHWISERO | 53,670 | 59,803 | 113,473 |
| LIKUYANI | 73,710 | 78,341 | 152,051 |
| LUGARI | 90,884 | 98,016 | 188,900 |
| LURAMBI | 92,774 | 95,432 | 188,206 |
| MALAVA | 115,511 | 122,814 | 238,325 |
| MATUNGU | 78,793 | 88,143 | 166,936 |
| MUMIAS EAST | 55,895 | 60,953 | 116,848 |
| MUMIAS WEST | 54,915 | 60,438 | 115,353 |
| NAVAKHOLO | 73,275 | 80,695 | 153,970 |
| SHINYALU | 53,219 | 58,524 | 111,743 |
| KAKAMEGA | 897,133 | 970,406 | 1,867,539 |
For population projection we use the data from 2019 general knbs census and data from Kakaamega County Annual Development Plan (CADP) 2024/25. From the projection we can use the concept of exponential growth with the following formula
The annual growth rate \(r\) between two years can be calculated using the formula:
\[ \text{r} = \left( \frac{\text{P}_{\text{end}}}{\text{P}_{\text{start}}} \right)^{\frac{1}{n}} - 1 \]
where:
\(\text{P}_{\text{end}}\) = The population at the end of the year
\(\text{P}_{\text{start}}\) = The population at the start of the year
\(n\) = The number of years between the start and end year
Once we know the growth rate from above , we can project the population for any future year using the formula below.
\[ P_{\text{future}} = P_{\text{present}} \times (1 + r)^{t}\]
where:
\(P_{\text{future}}\) is the projected population.
\(P_{\text{present}}\) is the population at the base year (2019 in this case).
\(r\) is the annual growth rate.
\(t\) is the number of years from the present year to the future year.
By using these formulas, we can make informed projections about how Kakamega County’s population will change for our case from 2024 to 2030. The annual growth rate gives us a sense of the rate of growth, and the projection formula allows us to estimate future populations based on this growth rate.
| Region | Base_2019 | 2024 | 2025 | 2026 | 2027 | 2028 | 2029 | 2030 | AGR |
|---|---|---|---|---|---|---|---|---|---|
| Kakamega | 1,867,579 | 2,036,901 | 2,072,565 | 2,108,853 | 2,145,777 | 2,183,347 | 2,221,575 | 2,260,472 | 0.0175089 |
| Butere | 154,100 | 168,071 | 171,014 | 174,008 | 177,055 | 180,155 | 183,309 | 186,519 | 0.0175088 |
| Lurambi | 188,212 | 205,276 | 208,870 | 212,527 | 216,248 | 220,034 | 223,887 | 227,807 | 0.0175087 |
| Ikolomani | 167,641 | 182,840 | 186,041 | 189,298 | 192,613 | 195,985 | 199,416 | 202,908 | 0.0175086 |
| Malava | 238,330 | 259,938 | 264,489 | 269,120 | 273,832 | 278,626 | 283,505 | 288,468 | 0.0175087 |
| Shinyalu | 111,743 | 121,874 | 124,008 | 126,179 | 128,389 | 130,636 | 132,924 | 135,251 | 0.0175089 |
| Khwisero | 113,476 | 123,764 | 125,931 | 128,136 | 130,379 | 132,662 | 134,985 | 137,348 | 0.0175086 |
| Likuyani | 152,055 | 165,841 | 168,745 | 171,700 | 174,706 | 177,765 | 180,877 | 184,044 | 0.0175093 |
| Lugari | 188,900 | 206,041 | 209,651 | 213,325 | 217,063 | 220,866 | 224,736 | 228,675 | 0.0175228 |
| Matungu | 166,940 | 182,075 | 185,263 | 188,507 | 191,807 | 195,165 | 198,582 | 202,059 | 0.0175085 |
| Mumias East | 116,851 | 127,446 | 129,677 | 131,948 | 134,258 | 136,609 | 139,001 | 141,434 | 0.0175094 |
| Mumias West | 115,354 | 125,812 | 128,015 | 130,256 | 132,537 | 134,857 | 137,219 | 139,621 | 0.0175085 |
| Navakholo | 153,977 | 167,938 | 170,878 | 173,870 | 176,914 | 180,012 | 183,164 | 186,371 | 0.0175093 |
I received the Plasmodium falciparum Parasite Rate (PfPR) data from
Malaria Atlas Project (MAP) but it was provided by Swiss TPH team. The
Plasmodium falciparum parasite rate (PfPR) is a commonly reported index
of malaria transmission intensity. In our case we will be using PfPR as
our malaria prevalence. The data received is of admin 2 from 1980 to
2024 which is labled
00_PfPR_table_Africa_CHAI_admin2_1980-2024.csv.The PfPR
data contains PfPR_rmean(mean), PfPR_LCI (Lower Confidence Interval),
PfPR_median (Median) and PfPR_UCI (Confidence Interval) which is for the
age group 2-10 years.
For my analysis I subset the data to Kakamega subcounties for the year 2022.
| Name | Year | Age | Pop | PfPR_rmean | PfPR_LCI | PfPR_median | PfPR_UCI |
|---|---|---|---|---|---|---|---|
| Likuyani | 2022 | 2-10_years | 264921.4 | 0.0229792 | 0.0089153 | 0.0208298 | 0.0518890 |
| Butere | 2022 | 2-10_years | 248916.6 | 0.1320386 | 0.0695961 | 0.1316295 | 0.2166028 |
| Shinyalu | 2022 | 2-10_years | 190349.9 | 0.0789242 | 0.0405881 | 0.0775125 | 0.1348691 |
| Khwisero | 2022 | 2-10_years | 121061.9 | 0.1260298 | 0.0621587 | 0.1271901 | 0.2102835 |
| Lugari | 2022 | 2-10_years | 203342.0 | 0.0292948 | 0.0119859 | 0.0270091 | 0.0638100 |
| Lurambi | 2022 | 2-10_years | 288374.5 | 0.0644496 | 0.0327343 | 0.0630284 | 0.1080119 |
| Malava | 2022 | 2-10_years | 395393.0 | 0.0477680 | 0.0220321 | 0.0441859 | 0.0955322 |
| Matungu | 2022 | 2-10_years | 211416.4 | 0.1334700 | 0.0606450 | 0.1296337 | 0.2404565 |
| Mumias East | 2022 | 2-10_years | 142932.6 | 0.1036357 | 0.0526933 | 0.0978843 | 0.1763287 |
| Mumias West | 2022 | 2-10_years | 208798.7 | 0.1320427 | 0.0643413 | 0.1274528 | 0.2279786 |
| Navakholo | 2022 | 2-10_years | 252341.1 | 0.0755592 | 0.0364110 | 0.0715479 | 0.1448511 |
| Ikolomani | 2022 | 2-10_years | 282693.4 | 0.0501423 | 0.0244225 | 0.0470628 | 0.0908116 |
We can see from the plot bellow that though the all 12 sub county are in Kakamega but we can see each one of them is a having a different PfPR.
The Plasmodium falciparum Parasite Rate (PfPR) map represent an annual average of PfPR in 2022.
From the data I received we can observe that we have four interventions:
Insecticide-Treated Net (ITN)
Indoor Residual Spraying (IRS)
Seasonal Malaria Chemoprevention (SMC)
Artemisinin-based Monotherapy (AM)
I also had Data on malaria vaccine for some sub-counties in Kakamega. The data was from a website 25 More Sub-Counties in Western Kenya Receive Malaria Vaccine and a published paperMalaria vaccine coverage estimation using age-eligible populations and service user denominators in Kenya
The ITN data is labeled
00_ITN_table_Africa_CHAI_admin2_1980-2024.csv. The data
showed the mean ITN coverage rate for Admin2 which ranges from 1980 to
2024 for the all age groups.
The IRS data is labeled
00_IRS_table_Africa_CHAI_admin2_1980-2024.csv. The data
showed the mean IRS coverage rate for Admin2 which ranges from 1980 to
2024 for the all age groups. We can see from the data that 2014 was the
last time IRS was used ever since the the IRS coverage is zero.
The SMC data is labeled
00_SMC_table_Africa_CHAI_admin2_1980-2024.csv. The data
showed the mean SMC coverage rate for Admin2 which ranges from 1980 to
2024 for the all age groups. We can see from the data that there have
never been SMC in Kakamega County.
The Artemisinin-based Monotherapy data is labeled
00_AM_table_Africa_CHAI_admin2_1980-2024.csv. The data
showed the mean AM coverage rate for Admin2 which ranges from 1980 to
2024 for the all age groups. In our case we will use AM to show the
access to care rates Which we can see its 0.603466 in each of the
subcounty. Since Artemisinin-based combination therapies are recommended
as first line treatments for Plasmodium falciparum malaria we are going
to use the AM as the effective treatment coverage.
The pilot RTS,S/AS01 malaria vaccine implementation programme in Kenya was undertaken in eight counties in the Western region: Bungoma, Busia, Homa Bay, Kakamega, Kisumu, Migori, Siaya, and Vihiga. In Kakamega county the subcounties which was targeted are Mumias West,Butere,Khwisero,Navakholo and Malava. Since the vaccine has four doses we are going to us the coverage of the first dose only which was estimated to be 80% in the western region.
Malaria vector distribution is the geographic and environmental spread of different Anopheles species. The vectors thrive in diverse environment influenced by different environmental factors such as amount temperature, humidity, rainfall, and the availability of breeding sites. The distribution of these vectors is closely linked to the transmission intensity of malaria.
The three common species which we will use in our research are Anopheles arabiensis, Anopheles_funestus and Anopheles gambiae ss which are distributed differently in the twelve sub-couties in Kakamega county. The spacial data we are using was from MAP but shared by Swiss TPH.