RESUMO
BACKGROUND & AIMS: Guidelines recommend use of risk stratification scores for patients presenting with gastrointestinal bleeding (GIB) to identify very-low-risk patients eligible for discharge from emergency departments. Machine learning models may outperform existing scores and can be integrated within the electronic health record (EHR) to provide real-time risk assessment without manual data entry. We present the first EHR-based machine learning model for GIB. METHODS: The training cohort comprised 2,546 patients and internal validation of 850 patients presenting with overt GIB (hematemesis, melena, hematochezia) to emergency departments of 2 hospitals from 2014-2019. External validation was performed on 926 patients presenting to a different hospital with the same EHR from 2014-2019. The primary outcome was a composite of red-blood-cell transfusion, hemostatic intervention (endoscopic, interventional radiologic, or surgical), and 30-day all-cause mortality. We used structured data fields in the EHR available within 4 hours of presentation and compared performance of machine learning models to current guideline-recommended risk scores, Glasgow-Blatchford Score (GBS) and Oakland Score. Primary analysis was area under the receiver-operating-characteristic curve (AUC). Secondary analysis was specificity at 99% sensitivity to assess proportion of patients correctly identified as very-low-risk. RESULTS: The machine learning model outperformed the GBS (AUC=0.92 vs. 0.89;p<0.001) and Oakland score (AUC=0.92 vs. 0.89;p<0.001). At the very-low-risk threshold of 99% sensitivity, the machine learning model identified more very-low-risk patients: 37.9% vs. 18.5% for GBS and 11.7% for Oakland score (p<0.001 for both comparisons). CONCLUSIONS: An EHR-based machine learning model performs better than currently recommended clinical risk scores and identifies more very-low-risk patients eligible for discharge from the emergency department.
RESUMO
BACKGROUND AND OBJECTIVES: Reducing the maximum red blood cell (RBC) shelf-life is under consideration due to potential negative effects of older blood. An assessment of the impacts of this change on blood supply chain management is evaluated. MATERIALS AND METHODS: We performed a simulation study using data from 2017 to 2018 to estimate the outdate rate (ODR), STAT order and non-group-specific RBC transfusion at two Canadian health authorities (HAs). RESULTS: Shortening shelf-life from 42 to 35 and 28 days led to the following: ODRs (in percentage) in both HAs increased from 0.52% (95% confidence interval [CI] 0.50-0.54) to 1.32% (95% CI 1.26-1.38) and 5.47% (95% CI 5.34-5.60), respectively (p < 0.05). The estimated yearly median of outdated RBCs increased from 220 (interquartile range [IQR] 199-242) to 549 (IQR 530-576) and 2422 (IQR 2308-2470), respectively (p < 0.05). The median number of outdated redistributed units increased from 152 (IQR 136-168) to 356 (IQR 331-369) and 1644 (IQR 1591-1741), respectively (p < 0.05). The majority of outdated RBC units were from redistributed units rather than units ordered from the blood supplier. The estimated weekly mean STAT orders increased from 11.4 (95% CI 11.2-11.5) to 14.1 (95% CI 13.1-14.3) and 20.9 (95% CI 20.6-21.1), respectively (p < 0.001). The non-group-specific RBC transfusion rate increased from 4.7% (95% CI 4.6-4.8) to 8.1% (95% CI 7.9-8.3) and 15.6% (95% CI 15.3-16.4), respectively (p < 0.001). Changes in ordering schedules, decreased inventory levels and fresher blood received simulated minimally mitigated these impacts. CONCLUSION: Decreasing RBC shelf-life negatively impacted RBC inventory management, including increasing RBC outdating and STAT orders, which supply modifications minimally mitigate.
