Characterizing the clinical profile and prevalence of people with diabetes attended in the hospital setting by using unstructured healthcare data and natural language processing: the Diabetic@ study
Authors' conclusion
NLP and ML for profiling DM using EHRs unstructured data are helpful, but additional data and better EHR documentation are crucial.
Study context
- Population
- Individuals with diabetes mellitus (DM) identified from electronic health records (EHRs) of eight Spanish hospitals (2013–2018)of 2,582,778 individuals
- Participants
- 638,730 individuals
- Exposure
- sT1DM vs sT2DM
- Outcome
- Prevalence and clinical spectrum of DM in hospitals
Methodological detailsmulticentre, natural language processing and machine learning of unstructured EHR data
- Multicentre
- Natural language processing and machine learning of unstructured EHR data
Main finding
As reported in the abstract
Prevalence of DM type
Observational data: associations, not effects.
Further findings
DM type as registered in EHRs
- Unregistered type (UrDM)
- 75.4%
- T2DM
- 21.3%
- T1DM
- 3.3%
Machine learning (ML) reclassification of UrDM
- Reclassified as T2DM
- 93.5%
- Reclassified as T1DM
- 6.5%
Missing data
- Relevant variables (anthropometrics, laboratory values, treatments) missing
- Over 50%
Comorbidities and complications · 60 months
- Major comorbidities: hypertension, dyslipidemia, chronic kidney disease (CKD), ischemic heart disease, chronic HF
- CKD and chronic HF were the most frequent complications for sT1DM/sT2DM at 60 months
Original publication
Characterizing the clinical profile and prevalence of people with diabetes attended in the hospital setting by using unstructured healthcare data and natural language processing: the Diabetic@ study. Diabetes Res Clin Pract. Published 2 May 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →