Validity of calculating fatty acid intake from mixed diets.

One hundred and seventy six weighed duplicate diets were collected over 16 consecutive days from 11 subjects. Analyses of their fatty acid composition were used to asses the validity of food composition tables. Four different calculating techniques were employed. Using the published data produced correlation coefficients of 0.29 between analysed and calculated polyunsaturated/saturated fatty acid ratios, whilst the addition of new analytical data and recoding fried foods produced a correlation coefficient of 0.56. The latter method also decreased the mean difference between analysed and calculated polyunsaturated/saturated consumption, when compared with the standard procedure.