Comparative Validity of a Food Frequency Questionnaire for Adults with IDDM

  1. Leigh Blizzard
  1. Menzies Centre for Population Health Research, University of Tasmania Hobart, Australia 7001
  1. Address correspondence and reprint requests to Malcolm Riley, BSC, Menzies Centre for Population Health Research, GPO Box 252C, Hobart, Tasmania 7001, Australia. E-mail: M.Riley{at}


OBJECTIVE To investigate the characteristics of a food frequency questionnaire (FFQ) in measuring dietary intake in an adult insulin-dependent diabetes mellitus (IDDM) population. FFQs have been widely used in developed countries to assess usual dietary intake; however, information regarding the application of the method to individuals advised to follow a specific dietary regimen (such as people with IDDM) is scarce.

RESEARCH DESIGN AND METHODS The measurement of energy and macronutrients by an interviewer-administered FFQ was assessed in 84 adults, who were representative of adults with IDDM in Tasmania, Australia, by comparing it with 2 days of weighed dietary records. Mean daily energy and macronutrient intakes by each method were compared, and Pearson correlation coefficients were calculated for both unadjusted and energy-adjusted macronutrient intakes.

RESULTS The ratio of within-person to between-person variance in dietary estimates from weighed records was 0.53 for energy, 0.70 for fat, 2.55 for protein, 0.50 for carbohydrate, 0.78 for saturated fat, and 3.56 for dietary cholesterol. The correlation coefficient between the FFQ and a 2-day weighed dietary record for the same nutrients ranged from 0.38 for protein intake to 0.60 for saturated fat intake. The correlation coefficient between the FFQ and true usual dietary intake was estimated to be 0.60 for energy-adjusted fat intake, 0.36 for energy-adjusted protein intake, 0.72 for energy-adjusted carbohydrate intake, 0.55 for energy-adjusted saturated fat intake, and 0.77 for energy-adjusted cholesterol intake.

CONCLUSIONS People with IDDM had a low ratio of within-person to between-person variance in daily energy, fat, carbohydrate, saturate fat, and cholesterol intake. The performance of the FFQ in the IDDM population is consistent with similar questionnaires in nondiabetic populations; however, there was a significant decrease in the dietary energy estimate upon readministration of the questionnaire. Validation studies such as this are important to provide information to guide the design and analysis of investigations that use dietary questionnaires.

| Table of Contents