Dataset Information

Evaluation of measurement error in 24-hour dietary recall for assessing sodium and potassium intake among US adults - National Health and Nutrition Examination Survey (NHANES), 2014.

ABSTRACT:

Background

Understanding measurement error in sodium and potassium intake is essential for assessing population intake and studying associations with health outcomes.

Objective

The aim of this study was to compare sodium and potassium intake derived from 24-h dietary recall (24HDR) with intake derived from 24-h urinary excretion (24HUE).

Design

Data were analyzed from 776 nonpregnant, noninstitutionalized US adults aged 20-69 y who completed 1-to-2 24HUE and 24HDR measures in the 2014 NHANES. A total of 1190 urine specimens and 1414 dietary recalls were analyzed. Mean bias was estimated as mean of the differences between individual mean 24HDR and 24HUE measurements. Correlations and attenuation factors were estimated using the Kipnis joint-mixed effects model accounting for within-person day-to-day variability in sodium excretion. The attenuation factor reflects the degree to which true associations between long-term intake (estimated using 24HUEs) and a hypothetical health outcome would be approximated using a single 24HDR: values near 1 indicate close approximation and near 0 indicate bias toward null. Estimates are reported for sodium, potassium, and the sodium: potassium (Na/K) ratio. Model parameters can be used to estimate correlations/attenuation factors when multiple 24HDRs are available.

Results

Overall, mean bias for sodium was -452 mg (95% CI: -646, -259), for potassium -315 mg (CI: -450, -179), and for the Na/K ratio -0.04 (CI: -0.15, 0.07, NS). Using 1 24HDR, the attenuation factor for sodium was 0.16 (CI: 0.09, 0.21), for potassium 0.25 (CI:0.16, 0.36), and for the Na/K ratio 0.20 (CI: 0.10, 0.25). The correlation for sodium was 0.27 (CI: 0.16, 0.37), for potassium 0.35 (CI: 0.26, 0.55), and for the Na/K ratio 0.27 (CI: 0.13, 0.32).

Conclusions

Compared with 24HUE, using 24HDR underestimates mean sodium and potassium intake but is unbiased for the Na/K ratio. Additionally, using 24HDR as a measure of exposure in observational studies attenuates the true associations of sodium and potassium intake with health outcomes.

SUBMITTER: Va P

PROVIDER: S-EPMC6537943 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundA dietary screener questionnaire (DSQ) was used to assess dietary outcomes among children in the Healthy Communities Study (HCS), a study of the relationships between programs and policies to prevent child obesity and child diet, physical activity and weight outcomes.MethodsTo compare dietary intake estimates derived from the DSQ against those from the Automated Self-Administered 24-Hour Recalls for Children (ASA24-Kids) among children, a measurement error model, using structural equation modelling, was utilized to estimate slopes, deattenuated correlation coefficients, and attenuation factors by age and sex, ethnicity, and BMI status.Participants/settingA randomly selected sub-sample of HCS participants aged 4-15 years in 130 communities throughout the U.S. who completed the DSQ and up to two ASA24-Kids recalls (n?=?656;13% of HCS participants).ResultsFor most nutrient/foods examined, the DSQ yielded larger mean intake estimates than the ASA24-Kids, and agreement between the two measures varied by food/nutrient, age and sex, ethnicity, and BMI category. Deattenuated correlation coefficients of 0.4 or greater were observed for added sugars from SSBs (0.54), fruits and vegetables (0.40), and dairy foods (0.50). Lower deattenuated correlation coefficients were seen for total added sugars (0.37), whole grains (0.34), and fiber (0.34). Attenuation factors were most severe for total added sugars intake among overweight children, and for several other dietary outcomes among children aged 9-11 years.ConclusionsThe DSQ was found to be a tool with acceptable agreement with the ASA24-Kids for measuring multiple dietary outcomes of interest in the HCS, although there may be potential due to measurement error to underestimate results (bias towards the null). In future studies, measurement error modelling and regression calibration may be possible solutions to correct for bias due to measurement error in most food/nutrient intake estimates from the DSQ when used among children.

Project description:BackgroundMonitoring salt consumption in children is essential for informing and implementing public health interventions to reduce children's salt intake. However, collection of 24-hour urines, considered as the most reliable approach, can be especially challenging to school children. This study aimed to assess the agreement between 24-hour urine (24hrU) and 24-hour food recall (24hrFR) in: (1) estimating salt intake in children; (2) classifying salt intakes above the recommended upper level set for children, and; (3) estimating change in mean salt intake over time.MethodsThis study utilised data from two cross-sectional surveys of school children aged 8 to 12 years living in the state of Victoria, Australia. A single 24hrU and 24hrFR were collected from each participant. Suspected inaccurate urine collections and implausible energy intakes were excluded based on pre-defined criteria. The agreement between the two methods was assessed using Bland-Altman methodology, the intraclass correlation coefficient (ICC), and the kappa statistic. The difference between the measured change in salt intake over time using 24hrU and 24hrFR was derived using mixed effects linear regression analysis.ResultsA total of 588 participants provided a 24hrU and 24hrFR. Overall, there was no meaningful difference in mean estimated salt intake between the two methods (- 0.2 g/day, 95% CI - 0.5 to 0.1). The Bland-Altman plot showed wide 95% limits of agreement (- 7.2 to 6.8). The ICC between the two methods was 0.13 (95% CI 0.05 to 0.21). There was poor interrater reliability in terms of classifying salt intake above the recommended upper level for children, with an observed agreement of 63% and kappa statistic of 0.11. The change in mean salt intake over time was 0.2 g/day (- 0.4 to 0.7) based on 24hrU, and 0.5 g/day (- 0.0 to 1.1) based on 24hrFR, with a difference-in-differences of 0.4 g/day (- 0.3 to 1.1).Conclusions24hrFR appears to provide a reasonable estimate of mean salt intake as measured by 24hrU in Australian school children. However, similar to previous observations in adults, and of studies exploring other alternative methods for estimating salt intake, 24hrFR is a poor predictor of individual-level salt intake in children.

Project description:BACKGROUND:Cigarette smoking is common in persons living with hepatitis C (hepatitis C+), but national statistics on this harmful practice are lacking. A better understanding of smoking behaviors in hepatitis C+ individuals may help in the development of targeted treatment strategies. METHODS:We extracted data from the National Health and Nutrition Examination Survey (NHANES) between 1999 and 2014. Hepatitis C+ were compared with hepatitis C- adults in the entire sample and in the subset of current smokers. Measures included demographics, current smoking, cigarettes/day, nicotine dependence, other tobacco use, substance use, and medical and psychiatric comorbidities. RESULTS:Complete smoking and hepatitis C virus (HCV) data were available for 39,472 (90.1%) of 43,793 adult participants in NHANES during the study years. Hepatitis C+ smoked at almost triple the rate of hepatitis C- adults (62.4% vs 22.9%), with no significant difference between hepatitis C+ men and women (64.5% vs 58.2%). Hepatitis C+ smokers were more likely to smoke daily than hepatitis C- smokers (87.5% vs 80.0%), but had similar levels of nicotine dependence. Hepatitis C+ smokers were more likely to be older (mean age: 47.1 vs 41.5 years), male (69.4% vs 54.4%), Black (21.2% vs 12.1%), less educated (any college: 31.8% vs 42.9%), poor (mean family monthly poverty index: 1.80 vs 2.47), uninsured (43.9% vs 30.4%), use drugs (cocaine: 11.1% vs 3.2%; heroin: 4.0% vs 0.6%), and be depressed (33.2% vs 13.5%). Multivariate analyses revealed significant associations of both hepatitis C infection and cigarette smoking with current depression and hypertension. CONCLUSIONS:There is a cigarette smoking epidemic embedded within the hepatitis C epidemic in the United States. The sociodemographic profile of hepatitis C+ smokers suggests that the implementation of effective tobacco treatment will be challenging. Thoughtful treatment strategies that are mindful of the unique characteristics of this group are needed.