Research Article

Validation of a computer based system for assessing dietary intake.

Br Med J (Clin Res Ed) 1987; 295 doi: http://dx.doi.org/10.1136/bmj.295.6594.369 (Published 08 August 1987) Cite this as: Br Med J (Clin Res Ed) 1987;295:369
  1. J A Levine,
  2. A M Madden,
  3. M Y Morgan
  1. Academic Department of Medicine, Royal Free Hospital, London.

    Abstract

    Dietary intake was assessed in 50 patients in hospital by using a dietary history method and computer based system for data collection and standard food tables to calculate the composition of nutrients. The results were compared with those from a weighed assessment that was calculated by using both food tables and manufacturers' food analyses. The use of the food tables overestimated mean (SEM) individual nutrient intakes by between 2.5% (1.5%) and 15.5% (3.0%). The mean errors associated with the dietary history assessment varied from -23% (7.8%) for fat intake to +21.4% (8.5%) for carbohydrate intake. Overall, 30% of the assessments of total nutrient intakes that were calculated using this method were within -20% to +20% of actual values; 18% were within -10% to +10%. The mean errors associated with the computer based assessment varied from -1.0% (4.3%) for carbohydrate intake to +8.5% (3.4%) for protein intake. Overall, 56% of the assessments of total nutrient intakes were within -20% to +20% of actual intakes; 31% were within -10% to +10%. The computer based system provides an accurate, reproducible, convenient, and inexpensive method for assessing dietary intake.