Related Experiment Videos
Agreement in self-reported personal risk factor information collected by different modes in Mongolia
B Tserenpuntsag1, K Nelson, O Lamjav
1Department of Epidemiology and Biostatistics, School of Public Health, University at Albany, State University of New York, School of Public Health, Albany, NY, USA. bxt07@health.state.ny.us
International Journal of STD & AIDS
|December 24, 2008
Summary
Comparing survey methods for HIV risk behaviors in Mongolia, this study found that while prevalence estimates varied, the association between behaviors and hepatitis B infection remained consistent. Data collection method impacts prevalence but not necessarily the observed risk associations.
Area of Science:
- Epidemiology
- Public Health
- Infectious Disease Transmission
Background:
- Accurate data collection on personal risk behaviors is crucial for understanding HIV and blood-borne pathogen transmission.
- Developing countries face unique challenges in epidemiological data gathering.
Purpose of the Study:
- To compare the effectiveness of self-administered paper surveys versus face-to-face interviews for collecting personal risk behavior data.
- To assess how data collection methods influence prevalence estimates and the association between risk factors and blood-borne infections in Mongolia.
Main Methods:
- An epidemiological study involving 2504 blood donors in Mongolia.
- Comparison of data from 2250 paper surveys and 923 interviews, with 669 participants completing both.
- Statistical analysis using McNemar's test and Kappa statistics to compare responses.
Main Results:
- Prevalence of alcohol use and smoking was higher in interviews than surveys, with lower Kappa coefficients (e.g., 0.61 for alcohol use).
- Kappa coefficients for health-care factors ranged from 0.57 to 0.81.
- Despite variations in prevalence based on data collection mode, the association between behaviors and hepatitis B surface antigen was not substantially different.
Conclusions:
- Misclassification of risk behaviors is probable regardless of the data collection method used.
- Biased prevalence estimates from different data collection methods did not significantly alter the observed associations between risk factors and blood-borne infections in this study.
