Identifiability and Model Misspecification for Modelling Recurrent Infections Using Routine Health Care Data.

Area of Science:

• Epidemiology
• Mathematical Modeling
• Infectious Diseases

Background:

• Time to recurrence data from infectious diseases offers insights into transmission dynamics.
• Routinely collected health records are a potential data source but pose analytical challenges due to unobserved patient dropout.
• Standard epidemiological methods like parametric survival analysis are unsuitable for data with unobserved dropout.

Purpose of the Study:

• To explore the feasibility of analyzing routinely collected time to recurrence data.
• To develop and validate a method for interrogating such data by calibrating mechanistic transmission models with explicit dropout mechanisms.
• To accurately estimate key epidemiological quantities from complex, real-world datasets.

Main Methods:

• Calibrated mechanistic transmission models with integrated dropout mechanisms.
• Identified specific model structures and parameter ranges for precise and accurate estimation.
• Applied the developed method to real-world malaria infection data from Papua, Indonesia.

Main Results:

• The method precisely and accurately estimated epidemiological quantities.
• Successfully estimated forces of infection for various malaria species.
• Quantified dropout and recrudescence rates for P. falciparum and treatment success probabilities.

Conclusions:

• The developed method is feasible for analyzing routinely collected time to recurrence data, even with unobserved dropout.
• This approach enhances the utility of existing and new datasets for public health research.
• The findings have the potential to improve understanding and control of infectious diseases.