Comparison of an Innovative Static Model With Dynamic Adjustment With a Dynamic Model To Assess the Public Health Impact of TAK-003 Vaccination: Case Studies in Thailand and Brazil
Author(s)
Elizaveta Kharitonova, MSc1, Jing Shen, PhD2, Anna Tytula, MSc3, Justyna Zawieja, MSc3, Yaneth Gil Rojas, MSc4, Samuel Aballea, PhD1, Riona Hanley, MSc2;
1Putnam (external consultant), Paris, France, 2Takeda Pharmaceuticals International AG, Zürich, Switzerland, 3Putnam, Krakow, Poland, 4Putnam, London, United Kingdom
1Putnam (external consultant), Paris, France, 2Takeda Pharmaceuticals International AG, Zürich, Switzerland, 3Putnam, Krakow, Poland, 4Putnam, London, United Kingdom
Presentation Documents
OBJECTIVES: Dynamic transmission models are considered the standard approach to evaluate vaccination, though complex and computationally intensive; static models do not capture the indirect effect of vaccination but are easier to use. An innovative static model with a dynamic component (dyna-static model) was developed for the TAK-003 dengue vaccine; this study aims to validate the dyna-static model by comparing it with the dynamic model.
METHODS: The dynamic transmission model was developed with all four dengue serotypes and explicitly modelled host and vector populations. The dyna-static model used a Markov structure with 16 health states capturing up to four consecutive infections. Both models differentiated infection severity and incorporated transient cross-protection and higher probability of severe disease with secondary infections. In the dyna-static model the probability of infection at each cycle was determined based on the number of infectious units (to approximate changing transmission intensity and capture indirect effect) predicted in the previous cycle. Different magnitude of indirect effects was approximated by varying vaccination strategy, relative transmissibility of symptomatic dengue, and efficacy against asymptomatic dengue. Thailand and Brazil were used as case studies.
RESULTS: The results of the dyna-static and dynamic models were generally close for timeframes above 10 years. The average difference between the two models across multiple outcomes (symptomatic cases of different severity and disability-adjusted life years) and each timeframe between 11-30 years, did not exceed 3% for Thailand and 5% for Brazil. The maximum absolute deviation of 16% for Thailand and 18% for Brazil was observed for the scenario with 10 additional cohorts vaccinated (i.e. large degree of indirect effects).
CONCLUSIONS: In both case studies, the dyna-static model yielded results that were comparable to those from the dynamic model, affirming its applicability as a substitute for the dynamic model to inform the impact of TAK-003 vaccination and program design.
METHODS: The dynamic transmission model was developed with all four dengue serotypes and explicitly modelled host and vector populations. The dyna-static model used a Markov structure with 16 health states capturing up to four consecutive infections. Both models differentiated infection severity and incorporated transient cross-protection and higher probability of severe disease with secondary infections. In the dyna-static model the probability of infection at each cycle was determined based on the number of infectious units (to approximate changing transmission intensity and capture indirect effect) predicted in the previous cycle. Different magnitude of indirect effects was approximated by varying vaccination strategy, relative transmissibility of symptomatic dengue, and efficacy against asymptomatic dengue. Thailand and Brazil were used as case studies.
RESULTS: The results of the dyna-static and dynamic models were generally close for timeframes above 10 years. The average difference between the two models across multiple outcomes (symptomatic cases of different severity and disability-adjusted life years) and each timeframe between 11-30 years, did not exceed 3% for Thailand and 5% for Brazil. The maximum absolute deviation of 16% for Thailand and 18% for Brazil was observed for the scenario with 10 additional cohorts vaccinated (i.e. large degree of indirect effects).
CONCLUSIONS: In both case studies, the dyna-static model yielded results that were comparable to those from the dynamic model, affirming its applicability as a substitute for the dynamic model to inform the impact of TAK-003 vaccination and program design.
Conference/Value in Health Info
2025-05, ISPOR 2025, Montréal, Quebec, CA
Value in Health, Volume 28, Issue S1
Code
EPH55
Topic
Epidemiology & Public Health
Topic Subcategory
Public Health
Disease
STA: Vaccines