Document 15O6R3xm2LGO4bxdxMQnGQO8a

With the available sample of 264 cases and 264 controls, we would have a 62% chance of detecting a 2-fold relative risk. If there is a 50% increase in the attack rate among the exposed, we would have a 27% chance of detecting this difference. These power figures are not particularly encouranging. In addition, it may well be that these relative risks apply only to women with a specific level or type .of exposure. Since the available measure of ex posure is so coarse, the relative risks of "exposed" women based on aerial spraying data might be much lower. ' In that case, the statistical power would be even lower. It is important that the limits of such an epidemiologic study are identified in the beginning. We feel that if a study is undertaken, the casecontrol approach is the best approach. If you decide to go ahead with such a study, the Birth Defects Branch would be willing to assist the State Health Department and the PARC Board in planning and implementing the study. 25195 10 07.