Calibrating self-reported BMI in national surveillance: impact on obesity misclassification and socioeconomic inequalities in Portugal
This study demonstrates that applying calibration equations to self-reported anthropometric data in Portugal corrects the systematic underestimation of obesity prevalence and refines the assessment of socioeconomic inequalities, particularly among women, thereby enhancing the validity of national obesity surveillance.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer
Imagine trying to map the health of a nation by asking people to guess their own height and weight. It is a common shortcut for governments and researchers because measuring millions of people is expensive and time-consuming. However, human memory is not a tape measure, and our self-perception is often colored by how we wish to be seen. People tend to say they are taller than they are and lighter than they are, a habit that skews the data used to track obesity. When these numbers are off, the picture of public health becomes distorted, making it difficult to see who is struggling with weight issues and how social factors like income or education influence those struggles. If the data is flawed, the policies built to fix the problem may miss their mark entirely.
In Portugal, researchers set out to fix this distortion using a unique opportunity provided by a national survey. The study team had access to data from over 3,400 adults who had both told the researchers their height and weight and then had those same measurements taken by trained staff. This allowed them to see exactly where the self-reported numbers went wrong. They found a clear pattern: people consistently overestimated their height and underestimated their weight. This error was not random; it grew larger as people got older and as their actual body weight increased. Women were particularly likely to underreport their weight, while men often overreported their height. The study also revealed that these mistakes were not spread evenly across society; they varied depending on where a person lived and their level of education.
To correct these errors, the researchers developed a set of adjustment rules, or calibration equations, based on simple information that is already collected in most surveys: a person's age, sex, where they live, and their education level. By feeding the self-reported numbers into these rules, they could generate a "calibrated" estimate that was much closer to the actual measured weight and height. The results were striking. Before the adjustment, the self-reported data suggested that fewer people were obese than was actually the case. After applying the calibration, the estimated obesity rates rose to match the real measurements almost perfectly. For women, the calibrated data showed an obesity rate of 26.0 percent, nearly identical to the 26.2 percent found in the physical measurements. For men, the calibrated figure was 21.7 percent, very close to the measured 22.5 percent.
The most significant impact of this correction was seen when the researchers looked at inequality. They examined how obesity rates differed between people with different levels of education, income, and employment. The study found that relying on uncorrected self-reports had been hiding the true extent of these gaps, but in different ways for men and women. For women, the adjustment actually made the gap between the most and least educated groups appear slightly wider, suggesting that previous estimates had underestimated the disparity. For men, the opposite happened; the uncorrected data had made the inequality look larger than it truly was, and the calibration brought the numbers down to a more accurate level. The patterns for income and employment were less affected by the correction, but the education-related gaps shifted noticeably.
This work demonstrates that while self-reported data is convenient, it carries a hidden cost when used to understand social inequalities. The researchers showed that by using simple mathematical adjustments based on common demographic details, national health surveys can recover a much truer picture of obesity without needing to physically measure every single participant. This approach does not just fix the total number of people with obesity; it clarifies who is most at risk and how social circumstances shape those risks. The study suggests that health officials should consider adding small groups of people with measured data to their surveys, allowing them to build these correction tools and ensure that the story told by the numbers reflects reality.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.