Mixed
Electronic health record (EHR) data from the one-year period prior to initiating anti-obesity medication (AOM) contains sufficient multidimensional clinical signals to identify distinct obesity subtypes (clusters) with unique physiological profiles, enabling precision medicine approaches that outperform traditional BMI-based classifications.
If you are considering obesity medication, ask your doctor about 'deep phenotyping.' This means using your full medical history (labs, vitals, diagnoses) from the past year to group you into a specific 'obesity subtype.' This helps predict which medication will work best for you, rather than just using your BMI. It reduces the guesswork in choosing between different drugs like GLP-1 agonists or others.
Our analysis revealed the presence of clusters with distinct clinical significance, which could have implications in AOM treatment options.
Why this rating
Large cohort (32,969 patients), rigorous ML methodology (GRU-D-AE, GMM), reproducible clustering, but observational and proof-of-concept.
Source
Deep Phenotyping of Obesity: Electronic Health Record–Based Temporal Modeling Study
Xiaoyang Ruan et al. · Journal of Medical Internet Research · 2025
DOI 10.2196/70140
Related findings · Mixed
- Bariatric surgery (VSG and RYGB) is the most effective and sustainable treatment for obesity, producing marked and sustained weight loss that non-surgical interventions cannot match.Strong
- Severely protein-deficient diets (2–3% energy) induce lean body mass loss and metabolic imbalance, whereas adequate protein intake (0.66 g/kg/d minimum) is a prerequisite for maintaining muscle, bone, and physiological function.Strong
- For patients with Class III obesity (BMI >39.9 kg/m2), metabolic bariatric surgery (MBS) is significantly more effective than any currently approved medication, with BilioPancreatic Diversion (BPD) showing the highest estimated weight loss.Strong
This is one finding among thousands. Every one is graded and traced to its source, so you can see what the evidence actually supports. Browse the research →