Using random-forest multiple imputation to address bias of self-reported anthropometric measures, hypertension and hypercholesterolemia in the Belgian health interview survey

Abstract

Background In many countries, the prevalence of non-communicable diseases risk factors is commonly assessed through self-reported information from health interview surveys. It has been shown, however, that self-reported instead of objective data lead to an underestimation of the prevalence of obesity, hypertension and hypercholesterolemia. This study aimed to assess the agreement between self-reported and measured height, weight, hypertension and hypercholesterolemia and to identify an adequate approach for valid measurement error correction. Methods Nine thousand four hundred thirty-nine participants of the 2018 Belgian health interview survey (BHIS) older than 18 years, of which 1184 participated in the 2018 Belgian health examination survey (BELHES), were included in the analysis. Regression calibration was compared with multiple imputation by chained equations based on parametric and non-parametric techniques. Results This study confirmed the underestimation of risk factor prevalence based on self-reported data. With both regression calibration and multiple imputation, adjusted estimation of these variables in the BHIS allowed to generate national prevalence estimates that were closer to their BELHES clinical counterparts. For overweight, obesity and hypertension, all methods provided smaller standard errors than those obtained with clinical data. However, for hypercholesterolemia, for which the regression model’s accuracy was poor, multiple imputation was the only approach which provided smaller standard errors than those based on clinical data. Conclusions The random-forest multiple imputation proves to be the method of choice to correct the bias related to self-reported data in the BHIS. This method is particularly useful to enable improved secondary analysis of self-reported data by using information included in the BELHES. Whenever feasible, combined information from HIS and objective measurements should be used in risk factor monitoring.

The potential impact fraction of population weight reduction scenarios on non-communicable diseases in Belgium: application of the g-computation approach

“…All missing values of the covariates included in the imputation models were imputed in the same process. Details on the application of this correction method in the BHIS is found in a previous publication [ 37 ]. The number of iterations of the random-forest multiple was set to 500 and the defined number of trees was set to 100.…”

“…This study aims to use a g-computation approach to quantify the effects of different population-based weight reduction interventions on important NCDs in Belgium in a multi-exposure context (taking into account lifestyle, metabolic, and environmental exposures). The research relies on cross-sectional data from the Belgian Health Interview Survey and Health Examination Surveys, addressing measurement bias due to self-reported health and anthropometric data through a random-forest multiple imputation method [ 37 ]. Additionally, this paper aims to provide a didactic application of the g-computation approach to assess PIF from cross-sectional data.…”

The potential impact fraction of population weight reduction scenarios on non-communicable diseases in Belgium: Application of the g-computation approach

“…Four NCDs were considered: diabetes, hypertension, cardiovascular disease (CVD) and musculoskeletal (MSK) disease. The biases related to self-reported height, weight, waist circumference, diabetes and hypertension were addressed using information from the BELHES and a random-forest multiple imputation (43). The variables used to construct these indicators are displayed in Table 1.…”

“…All missing values of the covariates included in the imputation models were imputed in the same process. Details on the application of this correction method in the BHIS is found in a previous publication (43). The number of iterations of the random-forest multiple was set to 500 and the de ned number of trees was set to 100.…”

“…This study also represents the rst application of the random-forest multiple imputation method to address the bias related to self-reported health and anthropometric data in the BHIS. This method has been recently identi ed as a more adequate approach for valid measurement error correction in comparison to regression calibration (43). Whenever feasible, self-reported information from health interview surveys should be combined with objective information from health examination surveys to address the bias related to self-reported anthropometric data and therefore provide more accurate PIF.…”