External Validation of a Clinical and Claims-Based Approach for Predicting 90-Day Post-Pulmonary Embolism Outcomes Among US Veterans
Bibliographic record
Abstract
Abstract Background: Both the simplified Pulmonary Embolism Severity Index (sPESI) and the multivariable In-hospital Mortality for Pulmonary embolism using Claims daTa (IMPACT) rule classify patients' risk of early post-pulmonary embolism (PE) complications. Objective: To externally validate sPESI and IMPACT for predicting 90-day all-cause mortality and readmission rates among PE patients treated within the Veterans Health Administration (VHA). Methods: We used VHA data from 10/1/2010-9/30/2015 to identify adult patients with: (1) ≥1 inpatient diagnosis for acute PE (International Classification of Diseases-9th Revision-Clinical Modification codes=415.1x), (2) continuous medical and pharmacy enrollment for ≥12-months prior to the index PE (baseline period), (3) a minimum of 90-days of post-event follow-up or until death (whichever came first), and (4) ≥1 claim for an anticoagulant during the index PE stay. Patients were excluded if they had a claim for PE or an anticoagulant during the baseline period. We classified patients as low-risk for early post-PE complications if their sPESI score=0 or their absolute in-hospital mortality risk estimated by IMPACT was <1.5% (the latter calculated using the formula: 1/(1 + exp(-x); where x = −5.833 + [0.026*age] + [0.402*myocardial infarction] + [0.368*chronic lung disease] + [0.464*stroke] + [0.638*prior major bleeding] + [0.298*atrial fibrillation] + [1.06 1*cognitive impairment] + [0.554*heart failure] + [0.364*renal failure] + [0.484*liver disease] + [0.523*coagulopathy] + [1.068*cancer]). Sensitivity, specificity, negative and positive predictive value (NPV and PPV) for all-cause mortality, all-cause readmission, and readmission for recurrent venous thromboembolism (VTE) or major bleeding at 90-days were reported with 95% confidence intervals (CIs) for sPESI and IMPACT tools. Results: Of6,746 eligible PE patients, 851 (12.6%) died and 1,359 (20.1%) were readmitted for any reason within 90-days. Hospitalization for recurrent VTE and major bleeding occurred in 375 (5.6%) and 116 (1.7%), respectively.sPESI classified 1,918 (28.4%) as low-risk, while 1,024 (15.2%) were low-risk per IMPACT. Both tools displayed sensitivity >90% and NPVs >96% for all-cause 90-day mortality, but low specificity and PPVs (Table). IMPACT's sensitivity for all-cause readmission was numerically higher than sPESI, but both had comparable NPVs. Similar trends were observed for accuracy in predicting readmissions due to recurrent VTE or major bleeding. Conclusion: In this external validation study utilizing VHA data, IMPACT classified patients for 90-day post-PE outcomes with similar accuracy as sPESI. While not recommended for prospective clinical decision-making, IMPACT appears useful for identification of PE patients at low-risk for early mortality or readmission in retrospective claims-based studies. Table. Test characteristics for sPESI and IMPACT for 90-day post-pulmonary embolism outcomes CI= confidence interval; IMPACT=In-hospital Mortality for Pulmonary embolism using Claims data; NPV=negative predictive value; PPV=positive predictive value; sPESI=simplified Pulmonary Embolism Severity Index; VTE=venous thromboembolism Table. Test characteristics for sPESI and IMPACT for 90-day post-pulmonary embolism outcomes CI= confidence interval; IMPACT=In-hospital Mortality for Pulmonary embolism using Claims data; NPV=negative predictive value; PPV=positive predictive value; sPESI=simplified Pulmonary Embolism Severity Index; VTE=venous thromboembolism Disclosures Kumar: Johnson & Johnson: Employment. Wells:Itreas: Other: Served on a Writing Committee; Janssen Pharmaceuticals: Consultancy; Bayer Healthcare: Other: Speaker Fees and Advisory Board; BMS/Pfizer: Research Funding. Peacock:Comprehensive Research Associates LLC: Equity Ownership; Cardiorentis: Consultancy, Research Funding; The Medicine's Company: Consultancy, Research Funding; Banyan: Research Funding; Emergencies in Medicine LLC: Equity Ownership; Abbott: Research Funding; Alere: Consultancy, Research Funding; Prevencio: Consultancy; Janssen: Consultancy, Research Funding; Portola: Consultancy, Research Funding; Pfizer: Research Funding; Roche: Research Funding; ZS Pharma: Consultancy, Research Funding; Ischemia Care: Consultancy; Phillips: Consultancy. Fermann:Janssen Pharmaceuticals: Other: Advisory Board, Speakers Bureau; Pfizer: Research Funding. Wang:Janssen Pharmaceuticals: Research Funding. Baser:Janssen Pharmaceuticals: Research Funding. Schein:Johnson & Johnson: Employment, Equity Ownership, Other: Own in excess of $10,000 of J&J stock. Crivera:Johnson & Johnson: Employment, Equity Ownership, Other: Owns excess of $10,000 in stock. Coleman:Boehringer-Ingelheim Pharmaceuticals, inc.: Consultancy, Research Funding; Bayer Pharmaceuticals AG: Consultancy, Research Funding; Janssen Pharmaceuticals: Consultancy, Research Funding.
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 | 0.041 | 0.084 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.002 | 0.002 |
| 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".