Number of households and assessments required for accurate measurement of selected nutrients in foods in the home

Measurement of the home food environment is of interest to researchers because it likely impacts food intake and is a feasible target for obesity prevention and treatment interventions. We measured all foods in the homes of 80 African American first‐time mothers and determined 6 nutrient‐related attributes. Sixty‐three households were measured 3 times, 11 were measured twice and 6 were measured once, producing 217 inventories collected at approximately 2 month intervals. Following log transformations, number of foods, total energy, dietary fiber, and fat required only 1 measurement in a household for the observed value to lie within 10% of the true value and also achieved a correlation of 0.8 between the observed and true value. For percent energy from fat and energy density 10 and 5 measurements were required, respectively, for accuracy within 10%. Three and 2 repeated measurements, respectively, were needed to achieve a correlation of 0.8. We calculated the number of households required to detect a given difference (50% of a standard deviation) in the means of two groups with customary levels of power and significance. A sample size of 126 was needed with 1 measurement, and 3 measurements per household reduced this number by 14 to 48. Nutrient‐related characteristics of homes appeared relatively stable over a 6 month period, and repeated measures may not be needed. NCI R21CA125735