Statistical analysis of passive surveillance disease registry data

Area of Science:

• Epidemiology
• Biostatistics

Background:

• Passive surveillance data, while useful, suffers from incomplete case ascertainment and lack of vital status follow-up.
• Estimating disease risk from passive surveillance data presents significant limitations.

Purpose of the Study:

• To assess the feasibility of estimating disease risk using passive surveillance data.
• To determine additional information required for accurate risk estimation.
• To develop methods for estimating absolute disease incidence rates by integrating passive surveillance with cohort data.

Main Methods:

• Developed analytical approaches for estimating absolute disease incidence rates by combining passive surveillance data with cohort studies.
• Investigated scenarios with known and unknown death rates from other causes.
• Applied methods to a registry of patients with artificial heart valves.

Main Results:

• Relative risks derived from passive surveillance data are generally biased, though bias is negligible for rare diseases.
• Combining passive surveillance data with cohort studies significantly enhances the efficiency of estimating absolute disease incidence rates.
• Efficiency gains are particularly notable when the exposure is rare and the cohort study is small relative to the registry size.

Conclusions:

• Passive surveillance data alone is insufficient for unbiased disease risk estimation.
• Integrating passive surveillance data with cohort studies offers substantial improvements in estimating absolute disease incidence rates.
• This combined approach provides a more robust method for epidemiological research, especially in registries with limited follow-up.