Enterprise Data Warehouse Research Articles

Using real-world electronic health record data to predict the development of 12 cancer-related symptoms in the context of multimorbidity.

This study uses electronic health record (EHR) data to predict 12 common cancer symptoms, assessing the efficacy of machine learning (ML) models in identifying symptom influencers. We analyzed EHR data of 8156 adults diagnosed with cancer who underwent cancer treatment from 2017 to 2020. Structured and unstructured EHR data were sourced from the Enterprise Data Warehouse for Research at the University of Iowa Hospital and Clinics. Several predictive models, including logistic regression, random forest (RF), and XGBoost, were employed to forecast symptom development. The performances of the models were evaluated by F1-score and area under the curve (AUC) on the testing set. The SHapley Additive exPlanations framework was used to interpret these models and identify the predictive risk factors associated with fatigue as an exemplar. The RF model exhibited superior performance with a macro average AUC of 0.755 and an F1-score of 0.729 in predicting a range of cancer-related symptoms. For instance, the RF model achieved an AUC of 0.954 and an F1-score of 0.914 for pain prediction. Key predictive factors identified included clinical history, cancer characteristics, treatment modalities, and patient demographics depending on the symptom. For example, the odds ratio (OR) for fatigue was significantly influenced by allergy (OR = 2.3, 95% CI: 1.8-2.9) and colitis (OR = 1.9, 95% CI: 1.5-2.4). Our research emphasizes the critical integration of multimorbidity and patient characteristics in modeling cancer symptoms, revealing the considerable influence of chronic conditions beyond cancer itself. We highlight the potential of ML for predicting cancer symptoms, suggesting a pathway for integrating such models into clinical systems to enhance personalized care and symptom management.

Evaluation of a Multi-Component Phenotype Algorithm for Systemic Lupus Erythematosus across the PCORnet, OMOP, and i2b2 Common Data Models

The aim of this study was to evaluate the performance of a computable phenotype for systemic lupus erythematosus (SLE) patients when it is ported from a local data warehouse to the i2b2, OMOP, and PCORnet CDMs. We adapted the SLE phenotype to the Northwestern Medicine (NM) Enterprise Data Warehouse (EDW) and NU i2b2, OMOP, and PCORnet instances. Each of the phenotype’s 17 components were determined by a rules-based algorithm built on diagnosis, medication, lab, and procedure codes. We assessed the phenotype over 168 clinician-confirmed SLE patients and 100 healthy controls and calculated agreement between the EDW and CDMs for overall SLE classification, individual SLICC criteria membership, and individual code occurrence level aggregates - count, first and last date, lab values using Cohen’s kappa (Cκ) and intraclass correlation coefficients (ICC). For overall SLE classification, agreement of the OMOP and PCORnet datamarts with the EDW was high (Cκ 0.928, 0.802, respectively) while that of i2b2 was low (Cκ 0.328). For the panel of SLICC criteria, OMOP had high agreement for 14/17 SLICC criteria (Cκ 0.792-1.000). PCORnet had high agreement for 13/17 criteria (Cκ 0.708-1.000). i2b2 had high agreement for 10/17 criteria (Cκ 0.813-1.000) with the exception of lab-based criteria. While agreement between the EDW and CDMs at the level of individual codes was relatively poor, agreement at the levels of overall SLE classification and individual SLICC criteria were reasonably high for OMOP and PCORnet, suggesting that inconsistencies at a micro, code level become less apparent at the macro, phenotype level.

Incidence and risk factors of SARS-CoV-2 breakthrough infection in the early Omicron variant era among vaccinated and boosted individuals in Chicago.

SARS-CoV-2 vaccines are safe and effective against infection and severe COVID-19 disease worldwide. Certain co-morbid conditions cause immune dysfunction and may reduce immune response to vaccination. In contrast, those with co-morbidities may practice infection prevention strategies. Thus, the real-world clinical impact of co-morbidities on SARS-CoV-2 infection in the recent post-vaccination period is not well established. This study was performed to understand the epidemiology of Omicron breakthrough infection and evaluate associations with number of comorbidities in a vaccinated and boosted population. A retrospective clinical cohort study was performed utilizing the Northwestern Medicine Enterprise Data Warehouse. Our study population was identified as fully vaccinated adults with at least one booster. The primary risk factor of interest was the number of co-morbidities. The primary outcome was the incidence and time to the first positive SARS-CoV-2 molecular test in the Omicron predominant era. Multivariable Cox modeling analyses to determine the hazard of SARS-CoV-2 infection were stratified by calendar time (Period 1: January 1 -June 30, 2022; Period 2: July 1 -December 31, 2022) due to violations in the proportional hazards assumption. In total, 133,191 patients were analyzed. During Period 1, having 3+ comorbidities was associated with increased hazard for breakthrough (HR = 1.16 CI 1.08-1.26). During Period 2 of the study, having 2 comorbidities (HR = 1.45 95% CI 1.26-1.67) and having 3+ comorbidities (HR 1.73, 95% CI 1.51-1.97) were associated with increased hazard for Omicron breakthrough. Older age was associated with decreased hazard in Period 1 of follow-up. Interaction terms for calendar time indicated significant changes in hazard for many factors between the first and second halves of the follow-up period. Omicron breakthrough is common with significantly higher risk for our most vulnerable patients with multiple co-morbidities. Age plays an important role in breakthrough infection with the highest incidence among young adults, which may be due to age-related behavioral factors. These findings reflect real-world differences in immunity and exposure risk behaviors for populations vulnerable to COVID-19.

Identifying Barriers to Timely Follow-up After Elevated PSA Screening: A Retrospective Analysis of a Large Healthcare System

ObjectivesTo examine elevated PSA follow-up within our system and identify areas for improvement in the timely diagnosis of prostate cancer. MethodsWe queried the Mass General Brigham’s Enterprise Data Warehouse from 2018-2021, identifying patients with elevated PSA and documented time to follow-up. Timely follow-up was defined as having a urologist appointment, prostate biopsy, or prostate magnetic resonance imaging within 6 months from diagnosis. We stratified the location of elevated PSA diagnosis to academic medical centers vs. community sites. Univariable and multivariable analyses were performed to identify factors impacting follow-up. ResultsWe included 28,346 patients, with 50.30%, 15.02%, and 34.69% receiving timely, untimely, and no follow-up during the study period, respectively. In multivariable analysis, patients seen at academic medical centers were more likely to receive follow-up care (OR=1.39, 95%CI 1.30-1.48). In a sensitivity analysis including two of our largest community hospitals as part of academic medical facilities, those following up at our main sites showed even higher odds of timely follow-up (OR=1.61, 95%CI 1.51-1.73). ConclusionOur study observed variations in follow-up rate between our academic medical centers and community sites. This finding highlights the need for efforts to improve consistency and timeliness of prostate cancer follow-up care across all facilities. By addressing interfacility disparities, we can facilitate the delivery of timely care to all patients.

Reducing Intravenous Contrast Utilization for CT: A Health System–Wide Intervention With Sustained Impact

ObjectiveTo determine the volume of intravenous iodinated contrast media used for CT before, during, and after the global iohexol shortage over a total of 17 months at a multisite health system. MethodsThis retrospective study included all patients who underwent CT at a large health system with 12 sites. Standardized contrast doses for 13 CT examinations were implemented May 23, 2022. Mean contrast utilization per CT encounter was compared between three periods (preintervention: January 1, 2022, to May 22, 2022; intervention: May 23, 2022, to September 11, 2022; postintervention: September 12, 2022, to June 30, 2023). Contrast doses and CT encounter data were extracted from the enterprise data warehouse. Categorical variables were compared with a χ2 test, and continuous variables were compared with a two-tailed t test. Multivariable linear regression assessed significance, with coefficients noted to determine magnitude and direction of effect. ResultsPreintervention, there were 152,009 examinations (87,722 with contrast [57.7%]); during the intervention, there were 120,031 examinations (63,217 with contrast [52.7%]); and during the postintervention, there were 341,862 examinations (194,231 with contrast [56.8%]). Preintervention, mean contrast dose was 89.3 mL per examination, which decreased to 78.0 mL after standardization (Δ of −12.7%) (P < .001). This decrease continued throughout the intervention and persisted in the postintervention period (80.4 mL; Δ −10.0%, P < .001). On multivariable analysis, patient weight, sex, and performing site were all associated with variations in contrast dose. Most but not all sites (9 of 12) sustained the decreased contrast media dose in the postintervention period. DiscussionImplementing standardized contrast media dosing for commonly performed CT examinations led to a rapid decrease in contrast media utilization, which persisted over 1 year.

A Calculated Risk: Evaluation of QTc Drug-Drug Interaction (DDI) Clinical Decision Support (CDS) Alerts and Performance of the Tisdale Risk Score Calculator.

A risk factor for a potentially fatal ventricular arrhythmia Torsade de Pointes is a prolongation in the heart rate-corrected QT interval (QTc) ≥ 500milliseconds (ms) or an increase of ≥60ms from a patient's baseline value, which can cause sudden cardiac death. The Tisdale risk score calculator uses clinical variables to predict which hospitalized patients are at the highest risk for QTc prolongation. To determine the rate of overridden QTc drug-drug interaction (DDI)-related clinical decision support (CDS) alerts per patient admission and the prevalence by Tisdale risk score category of these overridden alerts. Secondary outcome was to determine the rate of drug-induced QTc prolongation (diQTP) associated with overrides. Our organization's enterprise data warehouse was used to retrospectively access QTc DDI alerts presented for patients aged ≥18years who were admitted to Brigham and Women's Hospital during 2022. The QTc DDI CDS alerts were included if shown to a physician, fellow, resident, physician assistant, or nurse practitioner when entering the order in inpatient areas for patients with a length of stay of at least 2days. Variables collected for the Tisdale calculator included age, sex, whether patient was on a loop diuretic, potassium level, admission QTc value, admitting diagnosis of acute myocardial infarction, sepsis, or heart failure, and number of QTc-prolonging drugs given to the patient. A total of 2649 patients with 3033 patient admissions had 18,432 QTc DDI alerts presented that were overridden. An average of 3 unique QTc DDI alerts were presented per patient admission and the alerts were overridden an average of 6 times per patient admission. Overall, 6% of patient admissions were low risk (score ≤ 6), 64% moderate risk (score 7-10), and 30% high risk (score ≥ 11) of QTc prolongation. The most common QTc DDI alerts overridden resulting in an diQTP were quetiapine and propofol (11%) and amiodarone and haloperidol (7%). The diQTP occurred in 883 of patient admissions (29%) and was more frequent in those with higher risk score, with 46% of patient admissions with diQTP in high risk, 23% in moderate risk, and 8% in low risk. Use of the Tisdale calculator to assess patient-specific risk of QT prolongation combined with CDS may improve overall alert quality and acceptance rate, which may decrease the diQTP rate.

Implementation of Outpatient Automated Stewardship Information System (OASIS©) Audit and Feedback in Two Healthcare Systems

Background: Combating antibiotic resistance, exacerbated by widespread unnecessary outpatient antibiotic prescriptions, necessitates innovative stewardship solutions. Audit and feedback reports are effective but often resource heavy. We introduced a free, open-source system, Outpatient Automated Stewardship Information System (OASIS©), for automating the creation and distribution of recurring audit and feedback reports to clinicians to improve antibiotic prescribing. Methods: We used mixed methods to evaluate implementation of OASIS© across 11 clinics at Denver Health and Hospital Authority (DHHA) and Children’s Hospital Colorado (CHCO) from July 2022 to August 2023. Both sites host their own Epic® electronic healthcare and enterprise data warehouse systems. R statistical software was utilized to retrieve and process the data needed to create individual[HCM1] clinician audit and feedback reports with peer comparison. Reports were provided for 1) antibiotics prescribed for respiratory diagnoses, 2) antibiotics prescribed for respiratory diagnoses where antibiotics are never indicated, 3) first-line antibiotic prescribing for acute otitis media (AOM), and 4) five-day duration of antibiotics for children two years and older with AOM. Feedback reports for each metric were emailed to clinicians for three consecutive months. The primary outcome was adaptations needed to implement OASIS©. Secondary outcomes included fidelity (measured by email readership), time to set up and maintain the program, and barriers and facilitators to implementation (assessed by four qualitative interviews with OASIS© stakeholders). Results: The most significant adaptations made pertained to the automation of OASIS© reports for organizations not using R for data retrieval and reporting, setting up OASIS© specific email addresses, and validating clinician fidelity via read receipts. Fidelity was higher at DHHA (91-100%) compared to CHCO (10-30%). When interviewed, data analysts expressed that time for initial setup ranged from 1-6 hours. After reporting was automated, the estimated monthly time to send reports was 10 minutes. Views on setup complexity were split, but all recognized the readability of the reports and OASIS©’s value for improving prescribing behaviors. The greatest barriers to implementation included obtaining analytic resources for initial setup and the need to download additional R packages. No interviewee had prior experience creating audit and feedback reports. Conclusions: Implementing OASIS© requires addressing system diversity and knowledge gaps in outpatient informatics and antibiotic stewardship. Despite these challenges, the tool proved efficient and beneficial for monitoring and reporting antimicrobial prescribing. This free tool could likely be effectively disseminated to other health systems given the limited time and resources required for adaptations, setup, and monitoring. [HCM1]I

Understanding enterprise data warehouses to support clinical and translational research: impact, sustainability, demand management, and accessibility.

Healthcare organizations, including Clinical and Translational Science Awards (CTSA) hubs funded by the National Institutes of Health, seek to enable secondary use of electronic health record (EHR) data through an enterprise data warehouse for research (EDW4R), but optimal approaches are unknown. In this qualitative study, our goal was to understand EDW4R impact, sustainability, demand management, and accessibility. We engaged a convenience sample of informatics leaders from CTSA hubs (n = 21) for semi-structured interviews and completed a directed content analysis of interview transcripts. EDW4R have created institutional capacity for single- and multi-center studies, democratized access to EHR data for investigators from multiple disciplines, and enabled the learning health system. Bibliometrics have been challenging due to investigator non-compliance, but one hub's requirement to link all study protocols with funding records enabled quantifying an EDW4R's multi-million dollar impact. Sustainability of EDW4R has relied on multiple funding sources with a general shift away from the CTSA grant toward institutional and industry support. To address EDW4R demand, institutions have expanded staff, used different governance approaches, and provided investigator self-service tools. EDW4R accessibility can benefit from improved tools incorporating user-centered design, increased data literacy among scientists, expansion of informaticians in the workforce, and growth of team science. As investigator demand for EDW4R has increased, approaches to tracking impact, ensuring sustainability, and improving accessibility of EDW4R resources have varied. This study adds to understanding of how informatics leaders seek to support investigators using EDW4R across the CTSA consortium and potentially elsewhere.

Use of Electronic Patient Messaging by Pregnant Patients Receiving Prenatal Care at an Academic Health System: Retrospective Cohort Study.

The COVID-19 pandemic accelerated telemedicine and mobile app use, potentially changing our historic model of maternity care. MyChart is a widely adopted mobile app used in health care settings specifically for its role in facilitating communication between health care providers and patients with its messaging function in a secure patient portal. However, previous studies analyzing portal use in obstetric populations have demonstrated significant sociodemographic disparities in portal enrollment and messaging, specifically showing that patients who have a low income and are non-Hispanic Black, Hispanic, and uninsured are less likely to use patient portals. The study aimed to estimate changes in patient portal use and intensity in prenatal care before and during the pandemic period and to identify sociodemographic and clinical disparities that continued during the pandemic. This retrospective cohort study used electronic medical record (EMR) and administrative data from our health system's Enterprise Data Warehouse. Records were obtained for the first pregnancy episode of all patients who received antenatal care at 8 academically affiliated practices and delivered at a large urban academic medical center from January 1, 2018, to July 22, 2021, in Chicago, Illinois. All patients were aged 18 years or older and attended ≥3 clinical encounters during pregnancy at the practices that used the EMR portal. Patients were categorized by the number of secure messages sent during pregnancy as nonusers or as infrequent (≤5 messages), moderate (6-14 messages), or frequent (≥15 messages) users. Monthly portal use and intensity rates were computed over 43 months from 2018 to 2021 before, during, and after the COVID-19 pandemic shutdown. A logistic regression model was estimated to identify patient sociodemographic and clinical subgroups with the highest portal nonuse. Among 12,380 patients, 2681 (21.7%) never used the portal, and 2680 (21.6%), 3754 (30.3%), and 3265 (26.4%) were infrequent, moderate, and frequent users, respectively. Portal use and intensity increased significantly over the study period, particularly after the pandemic. The number of nonusing patients decreased between 2018 and 2021, from 996 of 3522 (28.3%) in 2018 to only 227 of 1743 (13%) in the first 7 months of 2021. Conversely, the number of patients with 15 or more messages doubled, from 642 of 3522 (18.2%) in 2018 to 654 of 1743 (37.5%) in 2021. The youngest patients, non-Hispanic Black and Hispanic patients, and, particularly, non-English-speaking patients had significantly higher odds of continued nonuse. Patients with preexisting comorbidities, hypertensive disorders of pregnancy, diabetes, and a history of mental health conditions were all significantly associated with higher portal use and intensity. Reducing disparities in messaging use will require outreach and assistance to low-use patient groups, including education addressing health literacy and encouraging appropriate and effective use of messaging.

Automating External Flat File Integration in SAP BW/4HANA: A Cost-Effective ABAP Approach

With SAP BW/4HANA as their enterprise data warehouse, organizations face challenges loading external data like CSV files into the system efficiently. Manual methods using GUIs are slow, error-prone, and lack scalability. This paper presents an automation solution using custom ABAP programming and standard BW/4HANA capabilities like process chains and advanced datastore objects (ADSOs) to orchestrate and load files. An ABAP program handles file discovery, logging, triggering process chains, monitoring status, and error handling. Benefits seen in production deployments include reduced load times and manual efforts, and improved data quality. While requiring upfront development effort, this approach provides a flexible, cost-effective data integration solution leveraging SAP's native tools. Key technical aspects include modular program design, use of SAP standard APIs, logging/status tracking, and BW process chain integration. Limitations are scalability for large volumes and complex transformations. The solution offers organizations an automation template for their BW/4HANA loading needs using in-house resources.

Natural language processing to identify lupus nephritis phenotype in electronic health records

BackgroundSystemic lupus erythematosus (SLE) is a rare autoimmune disorder characterized by an unpredictable course of flares and remission with diverse manifestations. Lupus nephritis, one of the major disease manifestations of SLE for organ damage and mortality, is a key component of lupus classification criteria. Accurately identifying lupus nephritis in electronic health records (EHRs) would therefore benefit large cohort observational studies and clinical trials where characterization of the patient population is critical for recruitment, study design, and analysis. Lupus nephritis can be recognized through procedure codes and structured data, such as laboratory tests. However, other critical information documenting lupus nephritis, such as histologic reports from kidney biopsies and prior medical history narratives, require sophisticated text processing to mine information from pathology reports and clinical notes. In this study, we developed algorithms to identify lupus nephritis with and without natural language processing (NLP) using EHR data from the Northwestern Medicine Enterprise Data Warehouse (NMEDW).MethodsWe developed five algorithms: a rule-based algorithm using only structured data (baseline algorithm) and four algorithms using different NLP models. The first NLP model applied simple regular expression for keywords search combined with structured data. The other three NLP models were based on regularized logistic regression and used different sets of features including positive mention of concept unique identifiers (CUIs), number of appearances of CUIs, and a mixture of three components (i.e. a curated list of CUIs, regular expression concepts, structured data) respectively. The baseline algorithm and the best performing NLP algorithm were externally validated on a dataset from Vanderbilt University Medical Center (VUMC).ResultsOur best performing NLP model incorporated features from both structured data, regular expression concepts, and mapped concept unique identifiers (CUIs) and showed improved F measure in both the NMEDW (0.41 vs 0.79) and VUMC (0.52 vs 0.93) datasets compared to the baseline lupus nephritis algorithm.ConclusionOur NLP MetaMap mixed model improved the F-measure greatly compared to the structured data only algorithm in both internal and external validation datasets. The NLP algorithms can serve as powerful tools to accurately identify lupus nephritis phenotype in EHR for clinical research and better targeted therapies.

Transforming Primary Healthcare Services with Centralized Health Intelligence: A Case Study from Qatar

Centralized Health Intelligence (CHI) represents an ecosystem where data, technology, and expertise converge to elevate the standards of healthcare services. In this case study, we explore the pivotal role of CHI in reshaping the healthcare paradigm at the Primary Health Care Corporation (PHCC) in Qatar. Adopting clinical information systems led to a wealth of data, necessitating the transformation of the conventional health information management team into the Business and Health Intelligence (BHI) Department. The establishment of a centralized enterprise data warehouse (EDW) and the use of business intelligence tools by the BHI Department further contributed to improved data governance. A meticulous Centralized Health Intelligence Framework was devised to ensure the effective use of PHCC's data assets. This framework encompasses policies and procedures related to data stewardship, information needs assessment, data classification, data privacy, and data literacy requirements. Through this study, we demonstrate how PHCC has harnessed CHI to redefine healthcare delivery by providing relevant stakeholders with easy access to advanced data analytics and data-driven decision-making tools. Implementing CHI infrastructure presents many challenges, including financial, technical, and organizational hurdles. Despite these challenges, CHI significantly benefits relevant stakeholders by providing enhanced access to health data, fostering active engagement. However, addressing data comprehension, security, and privacy concerns remains critical.

Models for building an enterprise data Warehouse

We live in an era of big data where data is generated continuously and at an exponential rate, so we need a place to collect, store and process it. Data can be generated by devices and sensors either remotely or in real time. The place where data can be collected, stored and processed is called a data warehouse. The term «data warehouse» was coined in the 1980s to describe a technology that allows companies to integrate data from many different sources into a single model. Data integration was seen as the most promising means of supporting deep insights into operational processes and identifying the latest opportunities to improve efficiency and profitability at the organizational level. An equally important role of data warehouses was seen as structural optimization by eliminating redundant decision support systems, most of which relied in one way or another on organizational performance data from the same centralized source. In contrast, the concept of a single data warehouse promised to eliminate redundancy and ensure data consistency, making it more suitable for optimizing management decisions at the enterprise level. This article discusses the concept of data warehouse, identifies its main advantages, reviews and compares enterprise data warehouse models and data platform on the example of Sberbank.

Linking Patient Encounters across Primary and Ancillary Electronic Health Record Systems: A Comparison of Two Approaches.

To achieve scientific goals, researchers often require integration of data from a primary electronic health record (EHR) system and one or more ancillary EHR systems used during the same patient care encounter. Although studies have demonstrated approaches for linking patient identity records across different EHR systems, little is known about linking patient encounter records across primary and ancillary EHR systems. We compared a patients-first approach versus an encounters-first approach for linking patient encounter records across multiple EHR systems. We conducted a retrospective observational study of 348,904 patients with 533,283 encounters from 2010 to 2020 across our institution's primary EHR system and an ancillary EHR system used in perioperative settings. For the patients-first approach and the encounters-first approach, we measured the number of patient and encounter links created as well as runtime. While the patients-first approach linked 43% of patients and 49% of encounters, the encounters-first approach linked 98% of patients and 100% of encounters. The encounters-first approach was 20 times faster than the patients-first approach for linking patients and 33% slower for linking encounters. Findings suggest that common patient and encounter identifiers shared among EHR systems via automated interfaces may be clinically useful but not "research-ready" and thus require an encounters-first linkage approach to enable secondary use for scientific purposes. Based on our search, this study is among the first to demonstrate approaches for linking patient encounters across multiple EHR systems. Enterprise data warehouse for research efforts elsewhere may benefit from an encounters-first approach.

Short stay unit led by pediatric hospital medicine advanced practice providers.

In response to a critical pediatric bed shortage in 2022, an urgent process change was required to provide safe and timely medical care. We proposed a pilot for an advanced practice provider (APP)-run short stay unit (SSU) for select pediatric hospital medicine (PHM) patients. To compare length of stay (LOS) and cost before and after implementation of a PHM APP-led SSU pilot at a tertiary pediatric hospital. Single-center prospective pilot observational cohort study with historical control. Observation encounters for patients with asthma, bronchiolitis, croup, and dehydration were included. An independent-practice model SSU staffed by APPs on the PHM service. SSU encounters (September 1, 2022-December 1, 2022) were compared against pre-SSU encounters (September 1, 2021-August 31, 2022). Cohorts were described with summary statistics. SSU encounters were matched against pre-SSU encounters based on demographics and diagnosis, and the average effect of treatment was calculated. LOS was abstracted from the enterprise data warehouse and standardized unit cost from the Pediatric Health Information Systems database. There were 1110 encounters included, 155 in the SSU cohort and 955 in the pre-SSU cohort:24.2% asthma, 30.8% bronchiolitis, 8.3% croup, and 36.7% dehydration. Median (interquartile range) unit LOS decreased from 21 (16-26) to 18 (10-22) h, p < .001. Cost decreased from $3593 ($3031-$4560) to $2958 ($2278-$3856), p < .001. After matching, the average treatment effect was reduction of 3.88 h (95% confidence interval [CI] 1.91-5.85) and $593 (95% CI $348-$839). There were no significant differences in 7-day ED revisit rates.

Length of stay is related to inpatient antipsychotic use in Alzheimer type dementia at a large US based tertiary care hospital.

AbstractBackgroundAntipsychotic(AP) medications carry a black box warning for increased risk of death in the elderly and should be used judiciously in a dementia population. Studies have shown that encounters with healthcare facilities increase the risk of AP exposure1. We wanted to know how admissions at our institution impacted antipsychotic use in patients with Alzheimer’s disease. Here we describe the results of a retrospective analysis of inpatient antipsychotic use in patients with Alzheimer’s disease at Northwestern Memorial Hospital(NMH), Chicago IL.MethodOur cohort was generated using NMH’s Enterprise Data Warehouse, including patients over 65 with a chart diagnosis of AD admitted to the hospital for over 48hrs between May 2005 and March 2018. We excluded patients with schizophrenia, schizotypal disorder, bipolar disorder, or psychotic illness due to primary psychiatric disease, enrollment in hospice, presence of tracheostomy or PEG‐tube and pervasive developmental disorder.Result1712 patients were admitted for a total of 4603 admissions amounting to 25,721 hospital days during the period of analysis. Demographic information is provided in Table 1. A higher proportion of admissions were four days or longer in encounters with an antipsychotic(Figure 1). Though longer LOS becomes progressively less common, this trend continued through a LOS of 20 days. When comparing LOS for encounters with an AP in the home medication list to encounters without, there was no impact on LOS. However, when comparing encounters with an AP prescribed while inpatient to encounters without, inpatient AP administration was associated with a longer length of stay (Table 2).ConclusionThese findings indicate that a longer LOS is associated with antipsychotic use. Because a longer length of stay was associated with new antipsychotic use but not the continuation of a home antipsychotic, our findings suggest that length of stay may contribute to behavioral and psychological symptoms of dementia requiring pharmacological treatment.1. Koponen M, Lavikainen P, Taipale H, et al. Accumulation of Hospital Days Among Antipsychotic Initiators With Alzheimer’s Disease. J Am Med Dir Assoc. 2019;20(12):1488‐1494.e3. https://doi.org/10.1016/j.jamda.2019.07.009

Low Inter-Rater Reliability of Calculating the 4T's Score for Heparin Induced Thrombocytopenia

Introduction Heparin induced thrombocytopenia (HIT) is an immune-mediated drug reaction that can cause thromboembolism in the setting of thrombocytopenia following heparin exposure. The “4T's score” has been validated to determine the pre-test probability of HIT and to assist in decision making around ordering testing for HIT. The 4T's scoring system requires an individual clinician to determine each component of the score. The objective of our study is to investigate the inter-rater reliability of calculating the 4T's score among clinicians. Methods Through retrospective query of Northwestern Enterprise Data Warehouse, we identified patients who had a HIT antibody (Ab) test ordered between 10/2019 and 10/2022 after implementation of a clinical decision support (CDS) tool that asked clinicians to calculate a 4T's score as part of HIT PF-4 Ab orders. From this cohort, an independent clinician randomly selected 15 patients. Four raters, including an attending hematologist, a hospital medicine attending, a hematology/oncology fellow, and an internal medicine resident, performed manual chart review of the randomly selected subjects to calculate a 4T's score. Data collected for each case included individual components of the 4T's score and overall 4T's score from each rater. In addition, we compared raters' scores to the 4T's score entered by the ordering clinician for the patient. We then categorized the numerical scores into validated pre-test probability categories: 0-3 as low, 4-5 as intermediate, and ≥6 as high. Inter-rater reliability for the categories was calculated using the Fleiss kappa statistic. Our study was approved by the Northwestern University Institutional Review Board. Results Of 15 cases selected, 5 each were scored as low, intermediate, and high probability 4T's scores by the ordering clinician (associated with the HIT Ab order). The overall agreement between the score categories for the 5 clinicians (4 raters and the ordering clinician) was 50.7%, with a Fleiss Kappa statistic of 0.26 (95% CI [0.07-0.45]), indicating poor inter-rater reliability (Figure 1). Excluding the ordering clinician, the overall agreement in score category between the 4 raters was 58.9%, with a Fleiss Kappa statistic of 0.38 (95% CI [0.14-0.62]). The same 4T's score pre-test probability category was calculated by the 4 raters in 5 cases, with only one case in which all four raters calculated the same numerical 4T's score. There was one case in which all 5 clinicians calculated the same 4T's score probability category, but none with the same numerical 4T's score. The hematology/oncology fellow had highest inter-rater agreement with the original clinical (Kappa 0.30 [-0.09-0.69]), whereas the internal medicine attending and resident had the lowest (both with Kappa 0.00 [-0.37-0.37]). Of the 15 4T's scores, 12/15 ordered by the hematology attending, 5/15 by the internal medicine attending, 12/15 by the hematology/oncology fellow, and 11/15 by the internal medicine resident were lower than those calculated by the original ordering clinicians. Conclusion Our study demonstrates poor inter-rater reliability of HIT 4T's score calculation, across levels of training and specialty. Importantly, poor inter-rater reliability was seen across 4T's categories, which has implications for clinical management of patients undergoing evaluation for HIT. This suggests that different strategies are necessary to help clinicians better use the 4Ts score.

2706. Effectiveness and Safety of Shingrix Vaccination in Solid Organ Transplant Recipients at a Large Urban Transplant Center

Abstract Background The varicella zoster virus (VZV) recombinant subunit vaccine, Shingrix, was found to be safe in solid organ transplant recipients (SOTR) and is now approved and recommended for all immunocompromised patients. Given limited efficacy data, we preferred a single center retrospective study to demonstrate safety and effectiveness of the Shingrix vaccine in solid organ transplant recipients. Methods After IRB approval, clinical data from all SOTR at Northwestern Memorial Hospital (NMH) who received the Shingrix vaccine after their transplant, between January 1, 2017 and March 1, 2022 was collected from our Enterprise Data Warehouse (EDW). Extracted data includes age, race, ethnicity, date of transplant, organ transplanted, and date(s) of vaccine administration(s). Clinical outcomes were assessed by diagnosis of acute allograft rejection and/or HZ made post-vaccine administration. Descriptive statistics were utilized. This project is supported by the IDSA G.E.R.M. program. Results 309 of 338 eligible SOTR were included in the analysis of this study (characteristics and reasons for exclusion in Table 1). Only 1 (0.32%) patient experienced a new episode of acute allograft rejection within 3 months of Shingrix administration. 18/309 (5.8%) were diagnosed with HZ after vaccination with Shingrix. This represents a reduction in VZV reactivation events compared to the 9.1% incidence found in the 2021 meta-analysis of HZ in the general solid organ transplant population [Kwon, Transpl Infect Dis., 2021].Table 1.Characteristics of 338 solid organ transplant patients who received Shingrix vaccine. Conclusion The Shingrix vaccine, when administered to SOTR post-transplant, is safe and likely effective in reducing the rate of herpes zoster. This data supports the current recommendation to vaccinate all SOTR with Shingrix. Larger, multi-center data are needed to inform the true efficacy of the vaccine in SOTRs. Disclosures All Authors: No reported disclosures

A Year of Inpatient Telehealth: Lessons From the COVID-19 Pandemic.

Relaxation of telehealth regulation enforcement during the coronavirus disease 2019 pandemic opened the door to massive expansion. Here we describe inpatient telehealth usage across a pediatric academic hospital during the first year of the pandemic. We created hospital bed-specific inpatient telehealth accounts and monitored their use over a 1 year period using data from our video conferencing vendor. We matched data with our enterprise data warehouse based on session date and time to identify patients who participated in telehealth. We performed secondary analysis of all video conferences to identify additional multidisciplinary team and family meetings that did not leverage the bed-specific telehealth accounts. We hosted 6931 inpatient telehealth sessions associated with 1648 unique patients. Hospitalized patients participating in telehealth sessions were older and had markedly longer length of stay compared with those who did not use telehealth (median age 12 vs 8 years, P < .001; median length of stay 9.03 vs 2.03 days, P < .001). There were 2006 charges for telehealth sessions, half of which were from psychiatry providers. Secondary analysis revealed an additional 1132 sessions used for interdisciplinary team or family meetings. Clinicians used inpatient telehealth to support care of hospitalized pediatric patients during the coronavirus disease pandemic, particularly for mental health care and family meetings. These findings suggest ongoing opportunities for inpatient telehealth systems beyond the pandemic.

Real-world genomic testing and treatment patterns of newly diagnosed adult acute myeloid leukemia patients within a comprehensive health system.

We evaluated the frequency of genomic testing and treatment patterns by age category in patients with newly diagnosed (ND) acute myeloid leukemia (AML) treated in both academic- and community-based health systems within a single Midwestern State. Retrospective analysis of data from the Indiana University Health System Enterprise Data Warehouse and two local cancer registries, of 629 patients aged ≥18 years with ND AML during 2011-2018. Primary outcome variables were, proportion of patients with genomic analysis and frequency of mutations. Chemotherapy was categorized as "standard induction" or "other chemotherapy"/targeted therapy, and hypomethylating agents. Overall, 13% of ND AML patients between 2011 and 2018 had evidence of a genomic sequencing report with a demonstrated increase to 37% since 2016. Genomic testing was more likely performed in patients: aged ≤60 years than >60 years (45% vs. 30%; p = 0.03), treated in academic versus community hospitals (44% vs. 26%; p = 0.01), and in chemotherapy recipients than non-therapy recipients (46% vs. 19%; p < 0.001). Most common mutations were ASXL1, NPM1, and FLT3. Patients ≥75 years had highest proportion (46%) of multiple (≥3) mutations. Overall, 31.2% of patients with AML did not receive any therapy for their disease. This subgroup was older than chemotherapy recipients (mean age: 71.4 vs. 55.7 years, p < 0.001), and was highest (66.2%) in patients ≥75 years. Our results highlight the unmet medical need to increase access to genomic testing to afford treatment options, particularly to older AML patients in the real-world setting, in this new era of targeted therapies.