On the Use of Surrogate Respondents for Controls in a Case‐Control Study of Alzheimer's Disease