Outbreak maths and study design, taught without the panic
Epidemiology is counterfactual thinking with a calculator: who got sick, who did not, and what the comparison actually shows. The exam is study design and bias; the rest is vocabulary.
Why the mainstream APIs fail here
Anything phrased as an outbreak scenario trips hosted models into pandemic-risk hedging, which is useless against the actual exam question of which study design fits and which bias explains the result. The coursework is not a biosecurity risk; it is a table of numbers from a made-up town.
How MagmaRouter is used for this
Give the scenario and ask which design, which bias, which measure of association, with the reasoning. Practise the outbreak-investigation steps in order, since the exam wants the order. Ask for made-up datasets to compute attack rates and risk ratios on until the arithmetic is boring.
A working loop
- Design questions first; they are the marks and the transferable skill.
- Compute on made-up tables until attack rate and risk ratio are automatic.
- For every result, name the bias that could have made it; the examiner will.
Prompts are not stored. A signed receipt of the hash comes back on every call, so you can prove later what MagmaRouter logged. Details are on the privacy page. Per-model prices are on /models and the pricing model is on /pricing.