Identification of Gender Differences in Acute Myocardial Infarction Presentation and Management at Aga Khan University Hospital-Pakistan: Natural Language Processing Application in a Dataset of Patients With Cardiovascular Disease
Bibliographic record
Abstract
BACKGROUND: Ischemic heart disease is a leading cause of death globally with a disproportionate burden in low- and middle-income countries (LMICs). Natural language processing (NLP) allows for data enrichment in large datasets to facilitate key clinical research. We used NLP to assess gender differences in symptoms and management of patients hospitalized with acute myocardial infarction (AMI) at Aga Khan University Hospital-Pakistan. OBJECTIVE: The primary objective of this study was to use NLP to assess gender differences in the symptoms and management of patients hospitalized with AMI at a tertiary care hospital in Pakistan. METHODS: We developed an NLP-based methodology to extract AMI symptoms and medications from 5358 discharge summaries spanning the years 1988 to 2018. This dataset included patients admitted and discharged between January 1, 1988, and December 31, 2018, who were older than 18 years with a primary discharge diagnosis of AMI (using ICD-9 [International Classification of Diseases, Ninth Revision], diagnostic codes). The methodology used a fuzzy keyword-matching algorithm to extract AMI symptoms from the discharge summaries automatically. It first preprocesses the free text within the discharge summaries to extract passages indicating the presenting symptoms. Then, it applies fuzzy matching techniques to identify relevant keywords or phrases indicative of AMI symptoms, incorporating negation handling to minimize false positives. After manually reviewing the quality of extracted symptoms in a subset of discharge summaries through preliminary experiments, a similarity threshold of 80% was determined. RESULTS: Among 1769 women and 3589 men with AMI, women had higher odds of presenting with shortness of breath (odds ratio [OR] 1.46, 95% CI 1.26-1.70) and lower odds of presenting with chest pain (OR 0.65, 95% CI 0.55-0.75), even after adjustment for diabetes and age. Presentation with abdominal pain, nausea, or vomiting was much less frequent but consistently more common in women (P<.001). "Ghabrahat," a culturally distinct term for a feeling of impending doom was used by 5.09% of women and 3.69% of men as presenting symptom for AMI (P=.06). First-line medication prescription (statin and β-blockers) was lower in women: women had nearly 30% lower odds (OR 0.71, 95% CI 0.57-0.90) of being prescribed statins, and they had 40% lower odds (OR 0.67, 95% CI 0.57-0.78) of being prescribed β-blockers. CONCLUSIONS: Gender-based differences in clinical presentation and medication management were demonstrated in patients with AMI at a tertiary care hospital in Pakistan. The use of NLP for the identification of culturally nuanced clinical characteristics and management is feasible in LMICs and could be used as a tool to understand gender disparities and address key clinical priorities in LMICs.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.008 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".