In annecori/mRIIDSprocessData: Epidemiological forcasting using real-time data

options(htmltools.dir.version = FALSE)

Our Brave New World

.center[]

---

Big Brother Is Watching

.center[]

Crisis Response

.center[]

mRIIDS - Mapping the Risk of International Infectious Disease Spread

.center[]

---

Project partners - ProMed, HealthMap and HealthSites

.center[]

Project Schematic

.center[.largeimg[]]

---

Model

Excpected number of cases at a location is influenced by

• local past incidence;

• pathogen characteristics - reproduction number, serial interval;

• local health access quality;

• movement between neighboring locations

$$I_{t, j} \sim Pois\left( \sum_{i = 1}^{n} {\left( p_{i \rightarrow j} R_{t, i} \sum_{s = 1}^{t}{I_{t - s, i} w_{s}}\right)} \right),$$ where $$R_{t, i} := f(haq_i, R_0, t).$$

$haq_i$ is a measure of health access quality index at a given location.

---

Movement between spatial units

Approximated by a phenomenological model (e.g. gravity¹ or radiation model²), or informed by other sources such as air or road traffic data.

$$p_{i \rightarrow j} = (1 - p_{stay}^i)\frac{\phi_{i \rightarrow j}}{\sum_{x}{\phi_{i \rightarrow j}}}$$

.footnote[[1] Grosche, T., Rothlauf, F., & Heinzl, A. (2007). Gravity models for airline passenger volume estimation. Journal of Air Transport Management, 13(4), 175-183.

[2] Simini, F., González, M. C., Maritan, A., & Barabási, A. L. (2012). A universal model for mobility and migration patterns. Nature, 484(7392), 96.]

Models of movement

Gravity model

$$\phi_{ij} = k\frac{N_i^{\alpha} N_j^{\beta}}{d_{ij}^{\gamma}}$$ Radiation model $$\phi_{ij} = {\phi}i \frac{N_i N_j}{(N_i + s{ij})(N_i + N_j + s_{ij})}$$

---

HealthMap and ProMed data pre-processing

.center[]

Incidence from WHO, HealthMap and ProMed

.center[]

---

Reproduction Number from WHO, HealthMap and ProMed

.center[]

---

Parameter estimation - prior on Reproduction Number

.center[]

---

MCMC time!

.center[]

---

Posterior Distribution of Reproduction Number

.center[]

---

Prediction using data from HealthMap

.center[]

---

Prediction using data from HealthMap

.center[]

Prediction - at district level (WHO data)

.center[]

---

Next steps

• Include health care capacity in the model.

• Informative priors for model parameters.

• Model averaging to account for model uncertainty.

• How best to choose the window for estimating the reproduction number?

annecori/mRIIDSprocessData documentation built on May 29, 2019, 1:16 p.m.