Microepidemics in the Carriage of Streptococcus Pneumoniae

1
Micro-epidemics in the Carriage of Streptococcus
Pneumoniae

• Fabian Hoti, Kari Auranen

2
Outline

• Introduction
• Pneumococcus
• Vaccination
• The Pneumocarr project
• Modelling transition dynamics
• The PNCeuro data
• Micro-epidemics
• Characterizing quantities
• Preliminary Results

3
Introduction

• Streptococcus pneumoniae (Pnc) is a bacterial
  pathogen that accounts for over one million
  annual deaths in children
• Pneumococcal diseases include meningitis,
  pneumonia, and otitis media.
• Pneumococcus is part of the natural flora in
  human nasopharynx
• Pneumococcus has 90 different serotypes capsules
  that the human immune system may react to

4
Introduction vaccination

• A conjugate vaccine providing partial protection
  against 7 serotypes has been licensed
• Introduction of the conjugate vaccine in US
  revealed a strong herd immunity effect as well as
  evidence on serotype replacement
• Herd immunity reduction in exposure due to less
  potential spreading routes
• Serotype replacement removed serotypes are
  replaced by others

5
The Pneumocarr project

• A five year project founded by the Grand
  Challenges in Global Health Initiative
• The main goal of the project is to establish
  colonization as a measure of the effectiveness of
  vaccination.
• Reduction in carriage ? reduction in disease
• Serotype specific serological correlates of
  protection (antibodies)
• Carriage is very common ? less study individuals
  needed? reduced cost

6
Modeling goals

• To understand the transition dynamics of
  serotype carriage (acquisition/clearance)
• To understand the role of between serotype
  competition (replacement)
• To understand the micro-epidemic nature of the
  carriage episodes (this talk)

7
Cross-sectional swab data

• Problems
• Acquisition and clearance times unknown
• All short carriage episodes are not detected

8
Cross-sectional swab data

• Possible solution
• Characterize episodes by acquisition and
  clearance rates
• Model acquisition and clearance rates by
  parameterized models
• Continuous time approach (Auranen, 2000
  Cauchemez, 2006)
• Carriage episodes are handled by latent processes
  (Bayesian data-augmentation)
• A Markov chain Monte Carlo algorithm is used to
  sample over possible carriage episodes and
  produce estimates of the model parameters

9
Transition intensities

• Individual based transition rates (age, exposure)
• Carriage of one serotype reduces/increases the
  colonization rates of other serotypes

Serotype 1
Competition parameter
Non-carrier
Serotype 2
10
PNCeuro data

• a KTL data set
• 3 day care centres in Tampere, Finland
• Index children family members employees
• 10 monthly nasopharyngeal swabs

11
Composition of Cohort
12
Distribution of Family Compositions
13
Number of swabs in DCCs
14
Micro-epidemic behavior

• The dominant types 9V,19F, and18C were detected
  in one dcc only
• 18C and 9V are rare types
• 23F was not detected at all (usually one of the
  most prevalent)

15
Micro-epidemics in DCCs

• Serotype specific quantities describing the
  micro-epidemics
• The stationary prevalence of individual carriers
  (ps)
• The mean waiting time until next micro-epidemic
  (Ts-)
• The stationary prevalence of DCCs carrying (Ps)
• The mean duration of a micro-epidemic (Ts)
• The mean size of a micro-epidemic (ms)
• The rate a typical carrier of serotypes s makes
  infectious contacts to the community (?s)
• The average number of infectious contacts
  emanated from a single outbreak (As)

16
Micro-epidemics

• The mean waiting time until next micro-epidemic
  (Ts-)

Proportion of carriers of other serotypes
Proportion of non carriers
17
Micro-epidemics in DCCs

• The mean duration of a micro-epidemic (Ts)

18
Micro-epidemics in DCCs

• The rate a typical carrier of serotype s makes
  infectious contacts to the community (?s)
• We assume the population consists of n DCCs sized
  N only
• Thus the total force of infection generated by
  the DCCs must equal the total acquisition
  pressure on the DCCs

19
Micro-epidemics in DCCs

• The average number of infectious contacts
  emanated from a single outbreak (As)

Average duration
Average number of carriers in a micro-epidemic
20
Micro-epidemics in DCCs
21

• Thanks !!
• www.pneumocarr.org

22
Structured Population (3)

• The acquisition rate of serotype s for individual
  i at time t can be modeled as

Community
Within day care centre
Within family
23
Structured Population (4)

• To consider age-specific susceptibility and
  infectiousness, age-specific parameters can be
  introduced
• For an individual in age group b the
  acquisition rate of serotype s from the mixing
  group family is

Number of family members in age group a at time t
Number of age groups