Introduction

The Health Survey for England (HSE) is the main data source for monitoring overweight and obesity in the general population in England. Height and weight have been directly measured during the interviewer visit in each year of the HSE series, and these values are used to calculate body mass index (BMI). In each survey between 2011 and 2016, the Health Survey also recorded self-reported height and weight, and these can be used to calculate BMI. Within this report, these are referred to as ‘measured BMI’ and ‘BMI from self-report’ respectively.

Epidemiological studies that have explored differences between self-reported and measured height and weight have shown that self-report measures introduce an element of bias (reporting bias). Adults on average overestimate height and underestimate weight compared with measured values.

The HSE 2011-2016 data provide an opportunity to compare self-reported and interviewer-measured height and weight and identify groups of individuals who may under- or over-estimate their height and weight when self-reporting. By modelling the associations between self-reported and interviewer-measured height and weight, HSE data can be used to develop correction factors (prediction equations) that adjust self-reported values of height and weight so that they more closely approximate measured values. Such equations can be transferred for use with surveys in England which collect only self-report data on height and weight, but caveats to their use should be considered.

Further information about the Health Survey for England series

Health Survey for England series

Partly due to the high cost of directly measuring height and weight, many surveys rely on self-reported data on height and weight to calculate BMI, which is then used to classify participants into BMI groups (underweight, normal (or healthy) weight, overweight, or obese). However, reviews by Connor Gorber et al. (2007) and Maukonen, Männistö and Tolonen (2018) showed that adults in a wide range of different populations tend to overestimate their height and underestimate their weight, compared with measured values. This results in a systematic underestimation of BMI and obesity.

The Health Survey for England (HSE) is one of a small number of nationally representative surveys that includes measurements of height and weight, taken by trained interviewers. Self-reported data on height and weight were collected in each HSE between 2011 and 2016. The questions were asked early in the interview, before measurements were taken towards the end of the visit.

In this report, these data have been used to compare self-reported and interviewer-measured values. This has enabled the identification of the characteristics of individuals who were more likely to under- or over-estimate their height and weight when self-reporting. Based on these, prediction equations have been developed for the purpose of adjusting self-reported data on height and weight to correct for such bias.

The work was carried out in three stages.

First, systematic literature reviews were undertaken on the socio-demographic and health-related factors associated with differences between self-reported and measured height and weight (and BMI derived from these), and the changes over time in misreporting.

Second, the HSE 2011-2016 data were analysed to identify which of the factors documented in the literature, or other variables we identified ourselves, were associated with the gap between self-reported and interviewer-measured height and weight.

The final stage involved modelling the associations between self-reported and interviewer-measured height and weight to develop prediction equations that could be used to adjust self-reported values of height and weight to more closely approximate measured values of height and weight. Using the interviewer-measured values as the ‘gold standard’, the effects of applying the models were evaluated according to how well they classified individuals into different BMI categories.

The report concludes by discussing a number of caveats for applying the prediction equations to other surveys.