LOW-FAT DIET PRACTICES OF OLDER WOMEN - PREVALENCE AND IMPLICATIONS FOR DIETARY ASSESSMENT

Objective To evaluate the importance of information on low-fat diet pr actices and consumption of reduced-fat foods for accurate assessment o f energy and fat intakes using a semiquantitative food frequency quest ionnaire (FFQ). Subjects Subjects were 7,419 women, aged 50 to 79 year s, who filled out an FFQ as part of eligibility screening for a diet m odification component and/or a hormone replacement trial in a multicen ter study of chronic disease prevention in postmenopausal women (Women 's Health Initiative). Statistical analysis For 26 FFQ questions, we r ecoded the low-fat diet choices of participants to a high-fat counterp art and recalculated energy and fat intakes. We than determined the de crease in energy and nutrient estimates attributable to adding low-fat options to the FFQ. Results Low-fat diet practices were widespread in this population. For example, 69% of respondents rarely or never ate skin on chicken, 76% rarely or never ate fat on meat, 36% usually dran k nonfat milk, 52% usually ate low-fat or fat-free mayonnaise, 59% ate low-fat chips/snacks, and 42% ate nonfat cheese. These low-fat choice s had substatial effects on energy and nutrient estimates. Absolute de creases (and mean percentage decreases) for energy and nutrient measur es attributable to adding low-fat diet options to the FFQ were 196 kca l (11.4%) energy, 9 percentage points in percentage energy from fat (2 2.3%). 23.2 g fat (29.0%), and 9.6 g saturated fat (32.5%). Black and Hispanic women and women of lower socioeconomic status reported signif icantly fewer low-fat diet practices than white women and women of hig her socioeconomic status. Conclusion Failure to collect information on low-fat diet practices with an FFQ will result in an upward bias in e stimates of energy and fat intake, and the amount of error will vary b y the personal characteristics of respondents.