Estimating Genotype Relative Risks in Case-Parental Control Studies: An Optimal Weighting Approach
碩士 === 國立臺灣大學 === 流行病學研究所 === 87 === Epidemiologists often examine the relation between susceptible gene and disease with case-control study design. Due to different allele frequencies among the races, drawing unsuitable controls may result in selection bias. Calculating the relative risk...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
1999
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/91903160656104188092 |
| Summary: | 碩士 === 國立臺灣大學 === 流行病學研究所 === 87 === Epidemiologists often examine the relation between susceptible gene and disease with case-control study design. Due to different allele frequencies among the races, drawing unsuitable controls may result in selection bias. Calculating the relative risks under the 'case-parental control study' design will circumvent this problem of selection bias. The 'genotype relative risks' (GRRs) can be estimated by iterative or non-iterative methods. There are three non-iterative methods proposed in the recent literatures, which have similar formula apart from the weighting constants. Here we propose a new non-iterative method, which has the 'optimal weighting constants'. Monte-Carlo simulation shows that the variances of the new estimators are smallest among the non-iterative methods.
|
|---|