Study population and sample attrition. Women without a survey weight were added to enrich the sample for specific hypotheses (for example, twin or sibling studies) and are not considered part of the core representative sample or women over-sampled by design