Transmission dynamics
A risk score says where malaria is likely to be more common. The transmission data says when in the year cases are expected to be highest, and how measured positivity has moved over time. It is available from the transmission endpoints.
The transmission season
Section titled “The transmission season”For every LGA and state, AI4M gives the months when reported cases are expected to be highest.
How it is worked out. Each area’s usual rainfall by calendar month is moved one month later and scaled to its highest month. The high season is the run of months at half the peak or more.
Why one month later. In the three states with public monthly data, reported cases are tied most closely to rainfall one to two months earlier. Rain creates breeding sites, mosquitoes mature, and infection takes time to develop.
How it was checked. The expected season was compared with seven independent monthly records of malaria cases in six states: facility reports from Kwara, Nasarawa and Zamfara, and published studies from Borno, Enugu, Kwara and the FCT.
| Measure | Result |
|---|---|
| Agreement of the shape of the year (1 is perfect) | 0.88, median |
| How far the centre of the season was off | 0.6 months, on average |
| How far the single peak month was off | 1.1 months, on average |
The season as a whole is more reliable than the single peak month.
Why not read the season from the monthly scores. The risk model is trained on surveys carried out mostly between August and December, so it learns differences between places far better than movement within a year. Against the same records, the monthly scores agreed at 0.55 and ran about a month and a half early. Use the score to compare places, and the season for timing.
What the season does not say
Section titled “What the season does not say”- It does not say how high cases will be, only when they are expected to be highest.
- Malaria is transmitted all year. In the three states with monthly data, cases in the driest months are still more than half their peak.
- It is a typical year. A late or failed rainy season will move it.
Survey trend
Section titled “Survey trend”For the country and each state, AI4M gives the share of children aged 6 to 59 months who tested positive at each national survey: 2010, 2015, 2018 and 2021.
| Survey | All locations | Rural | Urban |
|---|---|---|---|
| 2010 | 0.47 | 0.53 | 0.32 |
| 2015 | 0.43 | 0.53 | 0.24 |
| 2018 | 0.38 | 0.47 | 0.25 |
| 2021 | 0.38 | 0.44 | 0.26 |
These are rates across the survey locations AI4M uses, weighted by children tested. They are not the surveys’ official national estimates, which apply sampling weights. State figures for 2010 and 2015 rest on few children and are returned with a range.
The three-state transmission model
Section titled “The three-state transmission model”For Kwara, Nasarawa and Zamfara, AI4M fits a compartmental model: people move between susceptible, infectious and immune, and transmission follows standing water from rainfall. It is fitted to monthly reported cases for 2015 to 2023.
| Kwara | Nasarawa | Zamfara | |
|---|---|---|---|
| Agreement with reported cases | 0.87 | 0.87 | 0.81 |
| Months in which infection is growing | April to September | May to October | May to September |
| Share of transmission that is year-round | 78% | 67% | 89% |
Projecting a year ahead, it was level with a three-year seasonal average and better than last year’s figure, in a final year (2024) scored once.
It is research, not an operational forecast:
- It covers three states at state level, and says nothing about other states or LGAs.
- Reported cases rise with testing and reporting as well as with infection.
- It assumes how long an infection and immunity last, and uses no mosquito data.
Other diseases
Section titled “Other diseases”AI4M covers malaria only.