An NLP tool for data extraction from electronic health records: COVID-19 mortalities and comorbidities
Bibliographic record
Abstract
Background: The high infection rate, severe symptoms, and evolving aspects of the COVID-19 pandemic provide challenges for a variety of medical systems around the world. Automatic information retrieval from unstructured text is greatly aided by Natural Language Processing (NLP), the primary approach taken in this field. This study addresses COVID-19 mortality data from the intensive care unit (ICU) in Kuwait during the first 18 months of the pandemic. A key goal is to extract and classify the primary and intermediate causes of death from electronic health records (EHRs) in a timely way. In addition, comorbid conditions or concurrent diseases were retrieved and analyzed in relation to a variety of causes of mortality. Method: An NLP system using the Python programming language is constructed to automate the process of extracting primary and secondary causes of death, as well as comorbidities. The system is capable of handling inaccurate and messy data, this includes inadequate formats, spelling mistakes and mispositioned information. A machine learning decision trees method is used to classify the causes of death. Results: For 54.8% of the 1691 ICU patients we studied, septic shock or sepsis-related multiorgan failure was the leading cause of mortality. About three-quarters of patients die from acute respiratory distress syndrome (ARDS), a common intermediate cause of death. An arrhythmia (AF) disorder was determined to be the strongest predictor of intermediate cause of death, whether caused by ARDS or other causes. Conclusion: We created an NLP system to automate the extraction of causes of death and comorbidities from EHRs. Our method processes messy and erroneous data and classifies the primary and intermediate causes of death of COVID-19 patients. We advocate arranging the EHR with well-defined sections and menu-driven options to reduce incorrect forms.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.007 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.002 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 teacher head, 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".