2.4 million patients, only 17% had a recorded BMI in 2024. That’s not just a data completeness issue — it’s a bias issue. Our latest research at PHARMO (part of Lumanity), conducted with Amsterdam UMC, analyzed 12 years (2013–2024) of Dutch primary care data and found that missing BMI isn’t random.
What we found:
-
- BMI recording was more common among older patients, women, and those with chronic conditions.
-
- Unrecorded BMI was linked to younger age, an equal sex distribution, and lower recorded prevalence of cardiovascular, metabolic, and respiratory disease.
-
- When benchmarked against national statistics, in 2024, normal BMI was underrepresented (27.5% vs. 53.0%), overweight (40.7% vs. 31.0%) and obesity (30.8% vs. 13.8%) overrepresented compared to national figures.
Why it matters:
BMI is commonly used in real-world data to identify obesity, define treatment eligibility, and adjust for confounding in outcome studies. However, BMI is not routinely recorded and missingness is often systematic, so restricting analyses to patients with measured BMI can introduce selection bias, limit generalizability, and affect the validity of evidence informing your obesity program’s real-world story.
For sponsors developing anti-obesity medications, this pattern has direct implications for how you size eligible populations, design comparative effectiveness studies, and interpret comorbidity burden in RWE-based submissions.
The fix isn’t complicated, but it does require rigor.
Researchers studying obesity should consider inverse probability weighting and sensitivity analyses, and should transparently report recording rates and subgroup differences when interpreting their findings.
At PHARMO, we don’t just flag this bias: we’ve built the methodology to correct for it, using longitudinal GP data linked and benchmarked against national population statistics.
If you’re designing an obesity RWE study and want to make sure your evidence holds up to scrutiny, let’s talk.
Complete the form to access a copy of the poster presented during the ISPE Congress in Milan, August 29 – September 2, 2026.
Download the poster
"*" indicates required fields
