Desrosiers et al 1 state that they have observed an increased prevalence of neural tube defects (NTDs) in the offspring of women exposed to chlorinated solvents during the periconceptional period. This observation is based on an adjusted OR of 1.96 for the exposed cases compared with the controls. This estimate of effect is only as valid as the exposure estimate, but exposure was not measured. Instead, the exposure is expressed as a probability, on the basis of a job-exposure matrix. Cases and controls were classified under any one of five categories according to the …