Structure of dietary measurement error: Results of the OPEN Biomarker Study. American Journal of Epidemiology, 158, 14-21

Biometry Research Group, Division of Cancer Prevention, National Cancer Institute, Bethesda, MD 20892-7354, USA.
American Journal of Epidemiology (Impact Factor: 5.23). 08/2003; 158(1):14-21; discussion 22-6.
Source: PubMed


Multiple-day food records or 24-hour dietary recalls (24HRs) are commonly used as "reference" instruments to calibrate food frequency questionnaires (FFQs) and to adjust findings from nutritional epidemiologic studies for measurement error. Correct adjustment requires that the errors in the adopted reference instrument be independent of those in the FFQ and of true intake. The authors report data from the Observing Protein and Energy Nutrition (OPEN) Study, conducted from September 1999 to March 2000, in which valid reference biomarkers for energy (doubly labeled water) and protein (urinary nitrogen), together with a FFQ and 24HR, were observed in 484 healthy volunteers from Montgomery County, Maryland. Accounting for the reference biomarkers, the data suggest that the FFQ leads to severe attenuation in estimated disease relative risks for absolute protein or energy intake (a true relative risk of 2 would appear as 1.1 or smaller). For protein adjusted for energy intake by using either nutrient density or nutrient residuals, the attenuation is less severe (a relative risk of 2 would appear as approximately 1.3), lending weight to the use of energy adjustment. Using the 24HR as a reference instrument can seriously underestimate true attenuation (up to 60% for energy-adjusted protein). Results suggest that the interpretation of findings from FFQ-based epidemiologic studies of diet-disease associations needs to be reevaluated.

Download full-text


Available from: Raymond Carroll, Feb 19, 2015
  • Source
    • "Thus, the correlation between urinary and dietary cadmium may be lacking because these methods assessed cadmium exposure over different periods of time. Overall, measurement error in assessment of dietary cadmium would be nondifferential in this prospective cohort study and could have introduced substantial bias toward a finding of no association (Freedman et al. 2011; Kipnis et al. 2003). "
    [Show abstract] [Hide abstract]
    ABSTRACT: In vitro and animal data suggest cadmium, a heavy metal that contaminates some foods and tobacco plants, is an estrogenic endocrine disruptor. Elevated estrogen exposure is associated with breast, endometrial, and ovarian cancer risk. Examine the association between dietary cadmium intake and risk of these cancers in the large, well-characterized Women's Health Initiative (WHI). 155,069 postmenopausal women, age 50-79, enrolled in the WHI clinical trials or observational study participated in this study. We estimated dietary cadmium consumption by combining baseline food frequency questionnaire responses with US Food and Drug Administration data on food cadmium content. Participants reported incident invasive breast, endometrial, or ovarian cancer and WHI centrally adjudicated all cases through August 2009. We applied Cox regression to estimate adjusted hazard ratios (HRs) and 95% confidence intervals (95% CIs) for each cancer, comparing quintiles of energy-adjusted dietary cadmium intake. Over an average of 10.5 years, 6,658 invasive breast cancers, 1,198 endometrial cancers, and 735 ovarian cancers were reported. We observed no statistically significant associations between dietary cadmium and risk of any of these cancers after adjustment for potential confounders including total dietary energy intake. Results did not differ in any sub-group of women examined. We found little evidence that dietary cadmium is a risk factor for breast, endometrial, or ovarian cancers in postmenopausal women. Misclassification in dietary cadmium assessment may have attenuated observed associations.
    Environmental Health Perspectives 03/2014; 122(6). DOI:10.1289/ehp.1307054 · 7.98 Impact Factor
  • Source
    • "However, results have thus far been inconsistent [8], likely due to varied study methods. These studies rely on patient self-reporting and food diaries, which are both techniques fraught with error, bias, and inaccuracy [64] [65] [66] [67], with the strongest bias towards patient recall of high-fat foods [8]. Epidemiologic data have shown a potential link between dietary fat and breast cancer risk, which has led dietary fat to come under the most scrutiny by prospective studies. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Metabolic syndrome, which can include weight gain and central obesity, elevated serum insulin and glucose, and insulin resistance, has been strongly associated with breast cancer recurrence and worse outcomes after treatment. Epidemiologic and prospective data do not show conclusive evidence as to which dietary factors may be responsible for these results. Current strategies employ low-fat diets which emphasize supplementing calories with increased intake of fruit, grain, and vegetable carbohydrate sources. Although results thus far have been inconclusive, recent randomized trials employing markedly different dietary strategies in noncancer patients may hold the key to reducing multiple risk factors in metabolic syndrome simultaneously which may prove to increase the long-term outcome of breast cancer patients and decrease recurrences. Since weight gain after breast cancer treatment confers a poor prognosis and may increase recurrence rates, large-scale randomized trials are needed to evaluate appropriate dietary interventions for our breast cancer patients.
    09/2012; 2012(11):506868. DOI:10.1155/2012/506868
  • Source
    • "Measurement error that occurs when assessing person's usual diet with a Food Frequency Questionnaire generally involves bias related to true intake in addition to random variation [15]. Bias related to true intake often occurs as the “flattened-slope phenomenon” (consumers with high levels of intake tend to underreport and low-level consumers tend to overreport), which biases the beta-coefficient upward. "
    [Show abstract] [Hide abstract]
    ABSTRACT: We have previously shown that physical activity predicts intra-abdominal adipose tissue (IAT), but it is unknown whether energy intake predicts IAT independently of physical activity in a community-based, naturalistic environment. The association of energy intake with IAT was explored cross-sectionally in women, recruited between 2002 and 2005 for a study of fat patterning in midlife. IAT at L4-L5 vertebral interspace was assessed by computed tomography, energy intake by the Block Food Frequency Questionnaire, and physical activity by the Kaiser Physical Activity Survey. Linear regression models were used for the principal analyses. Among the 257 women, 48% were African American and 52% were Caucasian. Women were 52 ± 3 years old, and 49% were postmenopausal. Every 500 kcal increase in energy intake was associated with a 6% higher IAT (P = 0.02), independent of physical activity (P = 0.02), after adjustment for ethnicity, menopausal status, age, smoking, income, and DXA-assessed percent body fat. Energy intake had a significant interaction with ethnicity (P = 0.02), but not with physical activity. Models using the IAT to subcutaneous abdominal adipose tissue ratio as an outcome had similar associations. In conclusion, self-reported EI was associated with preferential IAT accumulation in midlife women, independent of physical activity. This association was significantly stronger in Caucasian than African American women. Future longitudinal studies are needed to explore lifestyle predictors of IAT accumulation during the menopausal transition.
    Journal of obesity 06/2012; 2012(2090-0708):567320. DOI:10.1155/2012/567320
Show more