Measurement error correction in exploiting gene-environment independence in family-based case-control studies