harwin de vries msc., prof. albert wagelmans, prof. joris ......optimizing population screening for...
TRANSCRIPT
Optimizing Population Screening for Infectious Diseases
Harwin de Vries MSc., Prof. Albert Wagelmans, Prof. Joris van deKlundert
June 7th, 2017
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 1 / 18
Introduction: 2 challenges
1. Linking decisions to outcomes:
Decisions Outcomes
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 2 / 18
Introduction: 2 challenges
1. Linking decisions to outcomes:
Decisions KPIs
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 3 / 18
Introduction: 2 challenges
1. Linking decisions to outcomes:
Decisions OutcomesKPIs
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 4 / 18
Introduction: 2 challenges
2. Transforming insight into better decisions:
Decisions Outcomes
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 5 / 18
Introduction: 2 challenges
2. Transforming insight into better decisions:
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 6 / 18
Problem Description
Human African Trypanosomiasis (HAT; sleeping sickness):
Slowly progressing parasitic diseaseTransmitted by Tsetse flyWHO: 30000 cases
Two disease stages (both appr. 1.5 years):1 Symptoms absent or non-specific2 Neurological problems
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 7 / 18
Problem Description
Human African Trypanosomiasis (HAT; sleeping sickness):
Slowly progressing parasitic diseaseTransmitted by Tsetse flyWHO: 30000 cases
Two disease stages (both appr. 1.5 years):1 Symptoms absent or non-specific2 Neurological problems
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 7 / 18
Problem Description
Population screening crucial:
Infected people do not observe disease/ do not obtain care
Source for infection of Tsetse
Result: HAT spreads silently
Current strategy (DRC)
30 mobile screening teams
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 8 / 18
Problem Description
Population screening crucial:
Infected people do not observe disease/ do not obtain care
Source for infection of Tsetse
Result: HAT spreads silently
Current strategy (DRC)
30 mobile screening teams
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 8 / 18
Problem Description
Main research question:
Given current screening capacity...
which villages to visit ...
at what time interval ...
to minimize the total averageexpected prevalence level?
?
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 9 / 18
Problem Description
Main research question:
Given current screening capacity...
which villages to visit ...
at what time interval ...
to minimize the total averageexpected prevalence level?
?
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 9 / 18
Problem Description
Main research question:
Given current screening capacity...
which villages to visit ...
at what time interval ...
to minimize the total averageexpected prevalence level?
?
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 9 / 18
Problem Description
Main research question:
Given current screening capacity...
which villages to visit ...
at what time interval ...
to minimize the total averageexpected prevalence level?
?
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 9 / 18
Problem Description
Main research question:
Given current screening capacity...
which villages to visit ...
at what time interval ...
to minimize the total averageexpected prevalence level?
?
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 9 / 18
Linking decisions to outcomes
Decisions Outcomes
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 10 / 18
Link Between Decisions and Prevalence
2004 2006 2008 2010 2012 2014 20160
0.001
0.002
0.003
0.004
0.005
0.006
0.007
0.008
0.009
0.01
Time
Expec
ted p
reval
ence
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 11 / 18
Transforming insight into better decisions
Decisions Outcomes
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 12 / 18
Frequency required for expected eradication
0
←4.9
100
1
2
10090
3
Required screening interval (years)
Sensitivity (%)
4
90
Participation (%)
5
8080
70 70
←0.84
1
1.5
2
2.5
3
3.5
4
4.5
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 13 / 18
Solution method for the general planning problem
Sophisticated methods:
Column generation
MIP
Local optimization
Simple policies:
Equalizing
Filtering
Differentiation
Burden level
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 14 / 18
Solution method for the general planning problem
Sophisticated methods:
Column generation
MIP
Local optimization
Simple policies:
Equalizing
Filtering
Differentiation
Burden level
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 14 / 18
Solution method for the general planning problem
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 15 / 18
Solution method for the general planning problem
0
20
40
60
80
100
120
140
160
180
Optimized Equalization Filtering Differentiation Burden-level
Average exp. # infected Final exp. # infected
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 16 / 18
Conclusions
Linking decisions to outcomes...
can be complex
becomes easier (more data)
yields important insights
Translating the link into decisions...
can be essential
can benefit both policy makers and people on the ground
does not always require advanced tools
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 17 / 18
Conclusions
Linking decisions to outcomes...
can be complex
becomes easier (more data)
yields important insights
Translating the link into decisions...
can be essential
can benefit both policy makers and people on the ground
does not always require advanced tools
De Vries, Wagelmans, Van de Klundert Optimizing Population Screening 17 / 18