Comparing methods for handling missing values in food-frequency questionnaires and proposing k nearest neighbours imputation: effects on dietary intake in the Norwegian Women and Cancer study (NOWAC)

Public Health Nutrition

Ida Scheel¹  Anette Hjartåker¹  Petter Laake¹  Eiliv Lund¹  Christine L Parr¹  Marit B Veierød¹ 

关键词: Food-frequency questionnaire;    Missing values;    Non-response;    Imputation;    Data quality;    Bias;   

DOI  :  10.1017/S1368980007000365

学科分类：卫生学

来源: Cambridge University Press

【 摘 要 】

ObjectiveTo investigate item non-response in a postal food-frequency questionnaire (FFQ), and to assess the effect of substituting/imputing missing values on dietary intake levels in the Norwegian Women and Cancer study (NOWAC). We have adapted and probably for the first time applied k nearest neighbours (KNN) imputation to FFQ data.DesignData from a recent reproducibility study were used. The FFQ was mailed twice (test–retest) about 3 months apart to the same subjects. Missing responses in the test FFQ were imputed using the null value (frequencies = null, amount = smallest), the sample mode, the sample median, KNN, and retest values.SettingA methodological substudy of NOWAC, a national population-based cohort.SubjectsA random sample of 2000 women aged 46–75 years was drawn from the cohort in 2002 (response 75%). The imputation methods were compared for 1430 women who completed at least 50% of the test FFQ.ResultsWe imputed 16% missing values in the overall test data matrix. Compared to null value imputation, the largest differences in estimated dietary intake were seen for KNN, and for food items with a high proportion of missing. Imputation with retest values increased total energy intake, indicating that not all missing values are caused by respondents failing to specify no consumption, and that null value imputation may lead to underestimation and misclassification.ConclusionMissing values in FFQs present a methodological challenge. We encourage the application and evaluation of newer imputation methods, including KNN, which may reduce imputation errors and give more accurate intake estimates.

【 授权许可】

Unknown