Capture-recapture, epidemiology, and list mismatches: two lists

Area of Science:

• Epidemiology
• Biostatistics
• Population Dynamics

Background:

• Capture-recapture methods are widely used in epidemiology for estimating population sizes from incomplete lists.
• A significant challenge is the presence of errors in data collection and matching, leading to inaccurate population estimates.
• These errors can cause individuals to be missed across multiple lists, underestimating the true population size.

Purpose of the Study:

• To develop and evaluate methods for estimating population size and list error probabilities using a tag loss concept.
• To investigate the impact of heterogeneity on error rates in capture-recapture models.
• To apply these novel methods to real-world data for diabetic population estimation.

Main Methods:

• Adapted tag loss concepts from animal population studies to epidemiological data.
• Developed statistical models to estimate probabilities of list errors and total population size for two independent lists.
• Examined the influence of individual heterogeneity on the accuracy of error estimation.

Main Results:

• Applied the developed methods to a dataset of diabetic individuals from surveys and doctor's records.
• Found high rates of errors in list matching and data entry.
• Demonstrated that ignoring these errors leads to a substantial overestimation of the total diabetic population.

Conclusions:

• Accounting for list errors is crucial for accurate population size estimation in epidemiological studies.
• The proposed methods provide a robust approach to simultaneously estimate population size and error rates.
• Failure to address data errors can lead to significantly inflated estimates, impacting public health resource allocation.