Influence of sampling on estimates of clustering and recent transmission of Mycobacterium tuberculosis derived from DNA fingerprinting techniques

Article Abstract:

In interpreting results of studies on clustering of Mycobacterium tuberculosis isolates, especially if sampling fractions are small, much caution is in order. It has been demonstrated, using stochastic simulation models based on hypothetical and real populations, that an influence from incomplete sampling on estimates of clustering obtained in tuberculosis studies exists. As a sampling fraction becomes greater, the proportion of isolates identified as clustered goes up and variance of the estimated proportion clustered goes down. Cluster size is a factor. When underestimation of clustering for a sampling fraction is larger, variability in results is greater for populations with small clusters than for populations with the same number of subjects in larger clusters. Availability of DNA fingerprinting techniques for M. tuberculosis has resulted in efforts to estimate the extent of recent transmission in populations. The underlying assumption has been that groups of tuberculosis patients with identical isolates, groups called clusters, are likely to reflect recently acquired infections.

South Africa, Tuberculosis, Statistical sampling, Sampling (Statistics), Epidemiology

---

Comparison of active and cancer registry-based follow-up for breast cancer in a prospective cohort study

Article Abstract:

Relying only on active follow-up or only on passive follow-up for breast cancer cases may introduce some kinds of biases in evaluating risk factors. Combing the two approaches seems a better strategy in prospective cohort studies. The relative effectiveness of two follow-up designs in prospective cohort studies has been compared. The approaches are the active and the passive, the former based on direct contact with study subjects, the latter on record linkages with population-based cancer registries. The comparison was done using available information from the New York State Cancer Registry (NYSCR) and the New York Univ. Women's Health Study (WHS). Cases occurred 1985-92 and follow-up was considered fairly complete by the NYSCR and the WHS. Of the 12,947 cohort members who gave a NY address, 303 cases pathologically confirmed were identified using active follow-up and 284 using record links.

Usage, Breast cancer, Science, Scientific method, Tumors, Cohort analysis, Statistical services

---

Mismeasurement and the resonance of strong confounders: correlated errors

Article Abstract:

Correlation of measurement errors in epidemiology can increase difficulty of evaluating structures in which measurement error and confounding exist. Observed correlations among risk factors can be greater than, less than or opposite to true correlations. The situation in which a strong risk factor, or confounder, and an inconsequential but suspected risk factor (confounded) are measured with errors that are correlated has been studied. The situation seems especially likely to be a found in the nutritional epidemiology area. Error correlation seems not to add much measurement error as a source of bias in estimating significance of a strong risk factor.

Measurement, Risk factors (Health), Food habits, Questionnaires

Similar abstracts:

• Abstracts: The impact on families of pediatric resident departure from a continuity clinic practice. The Pediatric Continuity Experience: Do Academic Sites Still Have a Role?
• Abstracts: Immunization status and birth order. Immunization status as determined by patients' hand-held cards vs medical records
• Abstracts: A balancing act. Cost-Effectiveness of Antiseptic-Impregnated Central Venous Catheters for the Prevention of Catheter-Related Bloodstream Infection
• Abstracts: Troubleshooting: delts and traps. Troubleshooting: triceps
• Abstracts: The association between midlife blood pressure levels and late-life cognitive function: the Honolulu-Asia Aging Study

This website is not affiliated with document authors or copyright owners. This page is provided for informational purposes only. Unintentional errors are possible.