Top Banner
Data Assimilation Methods in Parameter Estimation: An Application to Tuberculosis Transmission Model IIT Mandi, Himachal Pradesh Pankaj Narula and Arjun Bhardwaj Supervisors: Dr. Sarita Azad Dr. Ankit Bansal International Conference on Mathematical Techniques in Engineering Applications (ICMTEA 2013)
21

Bayesian Estimation of Reproductive Number for Tuberculosis in India

Jul 10, 2015

Download

Engineering

arjun_bhardwaj
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Bayesian Estimation of Reproductive Number for Tuberculosis in India

Data Assimilation Methods in Parameter Estimation: An Application to Tuberculosis Transmission Model IIT Mandi, Himachal Pradesh Pankaj Narula and Arjun Bhardwaj Supervisors: Dr. Sarita Azad Dr. Ankit Bansal

International Conference on Mathematical Techniques in Engineering

Applications (ICMTEA 2013)

Page 2: Bayesian Estimation of Reproductive Number for Tuberculosis in India

Outline

Epidemiology of Tuberculosis (TB)

Model Formulation and Parameters

Research Interests

Previous Work on TB

Estimation Methods

Results

Page 3: Bayesian Estimation of Reproductive Number for Tuberculosis in India

Epidemiology of Tuberculosis

TB is one of the most widespread infectious diseases, and a leading cause of global mortality.

Particularly, TB in India accounts for 25% of the world’s incident cases.

RNTCP is being implemented by Government of India in the country with DOTS strategy.

Page 4: Bayesian Estimation of Reproductive Number for Tuberculosis in India

The SIR Epidemic Model

S

Susceptible: can catch the disease

I

Infectious: have caught the disease and can spread it to susceptible

R

Recovered: have recovered from the disease and are immune.

dS dt

= – b S I

dI dt

= b S I – γI

dR dt

= γ I

S + I + R = 1

Page 5: Bayesian Estimation of Reproductive Number for Tuberculosis in India

Parameters of the Model

= The infection rate

= The Removal rate

= Fraction of infectious persons.

Basic reproduction number obtained as:

Average secondary number of infections caused by an infective in total susceptible population. An epidemic occur if .

Fraction of population needs to be vaccinated 1 − 1/R0.

0

pR

b

bp

0 1R

Page 6: Bayesian Estimation of Reproductive Number for Tuberculosis in India

Model Formulation

SILS Model of TB

Recovery rate is assumed to be 0.85.

Aim is to estimate β and p.

( )

(1 )

dS ISI L

dt N

dI pISI tL

dt N

dL p ISI tL

dt N

b

b

b

Page 7: Bayesian Estimation of Reproductive Number for Tuberculosis in India

Research Interests

Mathematical models, deterministic or statistical, are important tools to understand TB dynamics and analyse voluminous data collected by various agencies like WHO, RNTCP.

Challenge is to accurately estimate model parameters.

Parameters like infection rate measure the disease burden and evaluate the measures for control.

Page 8: Bayesian Estimation of Reproductive Number for Tuberculosis in India

Previous work On TB

Parameter Estimation of Tuberculosis

Transmission Model using Ensemble

Kalman Filter; Vihari et al. (2013)

Bayesian Melding Estimation of a

Stochastic SEIR Model, Hotta et al. (2010)

Tuberculosis in intra-urban settings: a

Bayesian approach; Souza et al. (2007)

Page 9: Bayesian Estimation of Reproductive Number for Tuberculosis in India

Methods of Parameter Estimation

Least Square

Maximum Likelihood Method

Ensemble Kalman Filter

Bayesian Melding

Page 10: Bayesian Estimation of Reproductive Number for Tuberculosis in India

Maximum Likelihood Method

To estimate a density function

whose parameters are

as an ML estimate of

( )p x

1

( ) ( / )n

i

i

L P x

1 2( , ,....., )t

m

arg max[ ( )]L

Page 11: Bayesian Estimation of Reproductive Number for Tuberculosis in India

Ensemble Kalman Filter (EnKf)

The EnKf is a MC approximation of the

Kalman filter.

It avoids evolving the covariance matrix of

the pdf of the state vector.

The basic idea is to predict the values first

and then to adjust it by actual value.

Page 12: Bayesian Estimation of Reproductive Number for Tuberculosis in India

Ensemble Kalman Filter

Forecast Step

Ensemble Mean

Error Matrices

Analysis Step

1j j jp p

t t tk k 1,2,3,....,j m

1 1

1

1j

mpp

t t

j

k km

1

1

1 1 1 1

1 1 1 1

[ ....... ]

[ ....... ]

q

t

q

t

ppp p p

k t t t t

ppp p p

y t t t t

E k k k k

E y y y y

( [ ] )k p j fj j jt t tt t tk k K y v y

Page 13: Bayesian Estimation of Reproductive Number for Tuberculosis in India

Bayesian Melding Method

Bayesian melding which observes the

existence of two priors, explicit and

implicit, on every input and output.

The technique works good with

stochastic and deterministic models with

in high dimensional parameter estimation.

Page 14: Bayesian Estimation of Reproductive Number for Tuberculosis in India

Bayesian Melding Method

These priors are coupled via logarithmic

pooling.

It calibrates the knowledge and

uncertainty of inputs and outputs of the

model.

The technique ignores the Borel paradox.

Page 15: Bayesian Estimation of Reproductive Number for Tuberculosis in India

Results

We have used BIP. Bayes.Melding package

to estimate trend of various parameters.

We have used Fitmodel for Monte Carlo

simulations, 2000 samples were discarded.

Prior distributions for parameters are

taken to be normal.

Page 16: Bayesian Estimation of Reproductive Number for Tuberculosis in India

Results

The parameter estimation framework

presented here captures seasonality well

in the data which could not be expected

from standard-likelihood methods.

The estimates presented here are verified

from three different approaches.

Page 17: Bayesian Estimation of Reproductive Number for Tuberculosis in India

Results

Comparison of parameters values

A- our results (EnKf)(2011) B- Our results (Bayesian Melding) C-Christopher Dye(2012) * 8 secondary infections per year.

Parameters A(2011) B C

β 1.72 1.84* 3.5

p 0.6 0.30 0.45

R0 1.29 0.69 0.78

Page 18: Bayesian Estimation of Reproductive Number for Tuberculosis in India

Results Comparison of estimated values of β for

highest infected state Manipur from three

different approaches Bayesian Melding

Mean value = 1.90

EnKf

Mean value = 2.31

Least square

Mean value = 2.32

Page 19: Bayesian Estimation of Reproductive Number for Tuberculosis in India

Results Estimated values of parameters for India

Ro< 1 which shows the disease is endemic in

the country

Seasonal trend

in the plot of

β and Ro

00.35 0.94R

1.3 2.54b

Page 20: Bayesian Estimation of Reproductive Number for Tuberculosis in India

Results

Ranges of R0

TB transmission across various Indian states

Page 21: Bayesian Estimation of Reproductive Number for Tuberculosis in India

THANKYOU