Related Experiment Videos
N-estimation from retrospectively ascertained events with applications to AIDS
1Department of Biostatistics, Johns Hopkins University, School of Hygiene and Public Health, Baltimore, Maryland 21205.
Biometrics
|March 1, 1992
Summary
This study introduces a semiparametric method for estimating event occurrences in retrospective studies, particularly for analyzing blood transfusion-related Acquired Immunodeficiency Syndrome (AIDS) cases. The method provides conservative estimates of event numbers in given intervals.
Area of Science:
- Biostatistics
- Epidemiology
- Retrospective Study Design
Background:
- Retrospective studies often sample individuals based on specific criteria.
- Understanding event occurrences requires robust estimation methods, especially when data is limited to specific cohorts.
Purpose of the Study:
- To develop a semiparametric method for estimating the expected number of specified events in retrospective studies.
- To apply and analyze this method using Acquired Immunodeficiency Syndrome (AIDS) blood transfusion cases.
Main Methods:
- Employed a semiparametric estimation approach for N-estimation.
- Developed statistical properties of the derived estimates.
- Applied the method to analyze blood transfusion-associated AIDS cases from the Centers for Disease Control.
Main Results:
- The semiparametric method provides conservative estimates, generally underestimating the actual number of events.
- Statistical properties of the estimates were theoretically developed.
- Detailed analysis of blood transfusion AIDS cases was performed.
Conclusions:
- The proposed semiparametric method offers a viable approach for estimating event counts in retrospective sampling scenarios.
- The conservative nature of the estimates is an inherent limitation of semiparametric and nonparametric models.
- The method is applicable to real-world epidemiological data, such as AIDS case analysis.