The applications of capture-recapture models to epidemiological data

Area of Science:

• Epidemiology
• Biostatistics
• Population Dynamics

Background:

• Capture-recapture methodology, initially for animal populations, is adapted for human studies.
• Ascertainment data from multiple incomplete lists often leads to underestimation of target population size.
• Traditional merging of lists can introduce downward bias by missing individuals not captured in any list.

Purpose of the Study:

• To review closed capture-recapture models for estimating human population size from incomplete lists.
• To introduce methods that correct for under-ascertainment bias in epidemiological data.
• To demonstrate the application of capture-recapture techniques using real-world datasets.

Main Methods:

• Review of closed capture-recapture models including ecological, log-linear, and sample coverage approaches.
• Discussion of methods for handling list dependence and heterogeneity.
• Demonstration of the CARE (Capture-Recapture) software with four diverse datasets.

Main Results:

• Capture-recapture methods can effectively estimate the number of missed individuals under specific assumptions.
• The CARE program facilitates the application of these methods to real data.
• Examples illustrate the utility of capture-recapture for correcting under-ascertainment in hepatitis A, diabetes, spina bifida, and congenital anomaly data.

Conclusions:

• Capture-recapture methodology offers a robust approach to overcome under-ascertainment bias in population size estimation.
• The presented models and software provide valuable tools for epidemiological research.
• Awareness of methodology limitations and cautionary remarks is essential for accurate application.