(Re-)match: adjusting for matching factors in case–control studies may be unnecessary or insufficient