Matched controls in a case-control study. Does matching by doctor’s list mean matching by relative deprivation?