PURPOSE Severe and febrile neutropenia present serious hazards to patients with cancer undergoing chemotherapy. We seek to develop a machine learning–based neutropenia prediction model that can be used to assess risk at the initiation of a chemotherapy cycle. MATERIALS AND METHODS We leverage rich electronic medical records (EMRs) data from a large health care system and apply machine learning methods to predict severe and febrile neutropenic events. We outline the data curation process and challenges posed by EMRs data. We explore a range of algorithms with an emphasis on model interpretability and ease of use in a clinical setting. RESULTS Our final proposed model demonstrates an out-of-sample area under the receiver operating characteristic curve of 0.865 (95% CI, 0.830 to 0.891) in the prediction of neutropenic events on the basis of only 20 clinical features. The model validates known risk factors and offers insight into potential novel clinical indicators and treatment characteristics that elevate risk. It relies on factors that are directly extractable from EMRs, provided a tool can be easily integrated into existing workflows. A cost-based analysis provides insight into optimal risk thresholds and offers a framework for tailoring algorithms to individual hospital needs. CONCLUSION A better understanding of neutropenic risk on an individual level enables a more informed approach to patient monitoring and treatment decisions.
1510 Background: Despite curative or disease-controlling roles in AML/MDS and MM, access to allogeneic (allo) and autologous (auto) hematopoietic stem cell transplantation (SCT) remains far from universal. Socioeconomic status (SES) and geographic distance from SCT centers have been shown to be barriers to SCT access. In 2016, Hartford HealthCare (HHC) and the Memorial Sloan Kettering Cancer Center (MSK) pioneered a Shared-Care Model (SCM) to streamline access to allo and auto SCT at MSK, featuring a dedicated nurse SCT coordinator, shared hematology tumor boards, MSK-led didactics for HHC providers, and an electronic health record sharing pipeline. We sought to determine if this has improved access to SCT for HHC patients. Methods: A retrospective chart review was conducted of HHC patients aged 18-70 with new diagnoses of AML, MDS, and MM between 2016 and 2020. Socioeconomic status (SES) was estimated by 9-digit zip-code using the Area Deprivation Index (ADI), shown to be a surrogate for healthcare access. Referral or not to a SCT center, referral to MSK through the SCM, and reasons for non-referral were abstracted from the medical record. For patients referred for SCT at MSK, we also captured the number of peri-SCT days in New York City (NYC) and number of subsequent MSK and HHC clinic visits/hospitalizations within 1-year post-SCT. Results: A total of 126 patients was included, with 81 (64%) treated for AML/MDS and 45 (36%) for MM. The median age was 60 years (interquartile range [IQR]: 53-66). The majority were white (n = 101, 80%) followed by 10% (n = 13) Black/African American; 10% (n = 12) were of Hispanic ethnicity. The median ADI percentile was 38 (IQR: 20-51; higher percentiles reflect decreased SES). The median ADI for MSK SCT referrals from New York, New Jersey, and Connecticut 2016-2020 for the same indications was 19 (IQR: 10-30, p < 0.001). A total of 90 patients (71%) were referred to SCT centers. Leading reasons for no referral were favorable-risk disease (n = 10), goals of care (n = 9), and death prior to referral (n = 5); 3 patients were not referred due to comorbidities/performance status. No differences were found between patients referred to MSK vs. other centers. Thirty-four HHC patients were referred to MSK (21 AML/MDS, 13 MM), vs. 3 between 2010 and 2015. Twelve patients underwent allo SCT, with median 97 days in NYC (range: 68-247); 8 underwent auto SCT, with median 21 days in NYC (range: 15-48). Conclusions: Our findings show the feasibility of a shared-care model between a non-SCT-providing large regional hospital system and a major academic transplantation center. Close collaboration between institutions may minimize time patients spend away from home. The SES of HHC referrals was lower than the general MSK population, suggesting that a shared-care model may facilitate access to SCT for patients with previous barriers for this potentially curative therapy.
31 Background: Neutropenic fever is a medical emergency. Delays in treatment can lead to increase in morbidity, mortality, and increase length of stay. The American Society of Clinical Oncology currently recommends that antibiotics be prescribed within 60 minutes of triage. Literature review shows through a multidisciplinary effort involving the ED, lab, oncology, and pharmacy significant improvement in time to antibiotics can be achieved. Since many patients with neutropenic fever present with sepsis, these guidelines also will need to be followed. Methods: Three PDSA cycles were conducted. The first involved education of the ED staff on the importance of treating neutropenic fever and using the correct antibiotic. The second PDSA cycle involved the laboratory and the calling of critical white counts and low neutrophil counts. The third PDSA involves patient education on the importance of temperature monitoring and reporting they are on chemotherapy to ED staff. Results: Baseline data show only 33% of patients receive the correct antibiotic and the average time to administration is 3 hours and 41 minutes. Results of the quality improvement project show a substantial improvement in time to antibiotic administration to 1 hour 58 minutes and an increase in the percentage of patients who receive the correct antibiotic. The time from the specimen received in the lab until critical called also improved from 1 hour 14 minutes to 18.5 minutes. Conclusions: This quality improvement led to a significant improvement in time to correct antibiotics, but several additional steps need to be taken to meet ASCO guidelines. [Table: see text]
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.