Development and Validation of Case-Finding Algorithms to Identify Pancreatic Cancer in the Veterans Health Administration.

BACKGROUND: Survival in pancreatic ductal adenocarcinoma (PDAC) remains poor due to late diagnosis. Electronic Health Records (EHRs) can be used to study this rare disease, but validated algorithms to identify PDAC in the United States EHRs do not currently exist. AIMS: To develop and validate an algorithm using Veterans Health Administration (VHA) EHR data for the identification of patients with PDAC. METHODS: We developed two algorithms to identify patients with PDAC in the VHA from 2002 to 2023. The algorithms required diagnosis of exocrine pancreatic cancer in either ≥ 1 or ≥ 2 of the following domains: (i) the VA national cancer registry, (ii) an inpatient encounter, or (iii) an outpatient encounter in an oncology setting. Among individuals identified with ≥ 1 of the above criteria, a random sample of 100 were reviewed by three gastroenterologists to adjudicate PDAC status. We also adjudicated fifty patients not qualifying for either algorithm. These patients died as inpatients and had alkaline phosphatase values within the interquartile range of patients who met ≥ 2 of the above criteria for PDAC. These expert adjudications allowed us to calculate the positive and negative predictive value of the algorithms. RESULTS: Of 10.8 million individuals, 25,533 met ≥ 1 criteria (PPV 83.0%, kappa statistic 0.93) and 13,693 individuals met ≥ 2 criteria (PPV 95.2%, kappa statistic 1.00). The NPV for PDAC was 100%. CONCLUSIONS: An algorithm incorporating readily available EHR data elements to identify patients with PDAC achieved excellent PPV and NPV. This algorithm is likely to enable future epidemiologic studies of PDAC.

Decreasing alloimmunization-specific mortality in sickle cell disease in the United States: Cost-effectiveness of a shared transfusion resource.

Red blood cell alloimmunization and consequent delayed hemolytic transfusion reaction (DHTR) incidence and mortality in patients with sickle cell disease (SCD) are high. A shared transfusion resource has decreased both in other countries, while in the United States cost concerns persist. We conducted a Markov cohort simulation of a birth cohort of alloimmunized patients with SCD to estimate lifetime DHTR incidence, DHTR-specific mortality, quality-adjusted life expectancy (QALE), and costs with the implementation of a shared transfusion resource to identify antibody history versus without (i.e., status quo). We conducted our analysis using a lifetime analytic time horizon and from a United States health system perspective. Implementation of shared transfusion resource projects to decrease cumulative DHTR-specific mortality by 26% for alloimmunized patients with SCD in the United States, relative to the status quo. For an average patient population of 32 000, this intervention would generate a discounted increment of 4000 QALYs at an incremental discounted cost of $0.3 billion, resulting in an incremental cost-effectiveness ratio of $75 600/QALY [95% credible interval $70 200-81 400/QALY]. The results are most sensitive to the baseline lifetime medical expenditure of patients with SCD. Alloantibody data exchange is cost-effective in 100% of 10 000 Monte Carlo simulations. The resource would theoretically need a minimum patient population of 1819 patients or cost no more than $5.29 million annually to be cost-effective. By reducing DHTR-specific mortality, a shared transfusion resource in the United States projects to be a life-saving and cost-effective intervention for patients with SCD in the United States.

Crawling toward obsolescence: The extended lifespan of amylase for pancreatitis.

The correlation between hyperamylasemia and acute pancreatitis was discovered in 1929, yet another test, lipase, was shown to provide better diagnostic performance in the late 1980s and early 1990s. Subsequent studies demonstrated co-ordering amylase with lipase did not provide additional benefit, only added cost. We sought to investigate the impact of studies advocating for the obsolescence of amylase on its clinical demand. We reviewed 1.3 million reportable results for amylase over 14 years (2009-2022). The trend in utilization of amylase over this period declined by 66% along a linear trajectory (R2 = 0.97). Despite demand for amylase decreasing by an average of 17,003 tests per year, the last year of the study (2022) recorded over 100,000 results for amylase. By interpolating the decline of amylase until the utilization reached zero, we calculated amylase orders will continue for 6 more years until 2028. Tests for creatinine and lipase changed <3% over the same period. Despite a multitude of studies advocating for the obsolescence of amylase, robust demand continues. Many important clinical guidelines, a source many practicing physicians rely on, have yet to acknowledge the preference for lipase over amylase. They frequently treat the two tests as equivalent, neglecting their head-to-head comparison studies and subsequent studies advocating against co-ordering both tests simultaneously. To expedite the obsolescence of amylase, which we anticipate lasting 46 years in our case study from its initial call for obsolescence to the last orders placed, metrics created specifically to monitor the utilization of unnecessary tests are also needed.

Testing and Case Rates of Gonorrhea, Chlamydia, Syphilis, and HIV among People with Substance Use Disorders in the Veterans Health Administration.

BACKGROUND: Little is known about national patterns of sexually transmitted infection (STI) testing and infections among people with substance use disorders (SUDs). METHODS: This study used a national retrospective analysis of people with SUDs receiving healthcare in the Veterans Health Administration in 2019 (N = 485,869). We describe testing rates, test positivity, and case rates for gonorrhea, chlamydia, syphilis, and HIV among individuals with alcohol, opioid, cocaine, and noncocaine stimulant use disorders in a national cohort of Veterans Health Administration patients. RESULTS: Test and case rates for all STIs were highest among people with noncocaine stimulant use. People with alcohol use disorder had the lowest testing rates but intermediate incidence for all STIs. People with multiple SUDs had higher incidence of all STIs than those with single SUDs. Mental health diagnoses and houselessness were common. The HIV test positivity was 0.14% to 0.36% across SUD groups. CONCLUSIONS: Sexually transmitted infection testing rates between SUD groups were discordant with their respective case rates. High STI rates in people with SUDs suggest a need for more comprehensive testing, particularly for those with noncocaine stimulant use and those with comorbid houselessness or mental health diagnoses.

Machine learning to develop a predictive model of pressure injury in persons with spinal cord injury.

STUDY DESIGN: A 5-year longitudinal, retrospective, cohort study. OBJECTIVES: Develop a prediction model based on electronic health record (EHR) data to identify veterans with spinal cord injury/diseases (SCI/D) at highest risk for new pressure injuries (PIs). SETTING: Structured (coded) and text EHR data, for veterans with SCI/D treated in a VHA SCI/D Center between October 1, 2008, and September 30, 2013. METHODS: A total of 4709 veterans were available for analysis after randomly selecting 175 to act as a validation (gold standard) sample. Machine learning models were created using ten-fold cross validation and three techniques: (1) two-step logistic regression; (2) regression model employing adaptive LASSO; (3) and gradient boosting. Models based on each method were compared using area under the receiver-operating curve (AUC) analysis. RESULTS: The AUC value for the gradient boosting model was 0.62 (95% CI = 0.54-0.70), for the logistic regression model it was 0.67 (95% CI = 0.59-0.75), and for the adaptive LASSO model it was 0.72 (95% CI = 0.65-80). Based on these results, the adaptive LASSO model was chosen for interpretation. The strongest predictors of new PI cases were having fewer total days in the hospital in the year before the annual exam, higher vs. lower weight and most severe vs. less severe grade of injury based on the American Spinal Cord Injury Association (ASIA) Impairment Scale. CONCLUSIONS: While the analyses resulted in a potentially useful predictive model, clinical implications were limited because modifiable risk factors were absent in the models.

PROSER: A Web-Based Peripheral Blood Smear Interpretation Support Tool Utilizing Electronic Health Record Data.

OBJECTIVES: Peripheral blood smear (PBS) interpretation represents a cornerstone of pathology practice and resident training but has remained largely static for decades. Here, we describe a novel PBS interpretation support tool. METHODS: In a mixed-methods quality improvement study, a web-based clinical decision support (CDS) tool to assist pathologists in PBS interpretation, PROSER, was deployed in an academic hospital over a 2-month period in 2022. PROSER interfaced with the hospital system's electronic health record and data warehouse to obtain and display relevant demographic, laboratory, and medication information for patients with pending PBS consults. PROSER used these data along with morphologic findings entered by the pathologist to draft a PBS interpretation using rule-based logic. We evaluated users' perceptions of PROSER with a Likert-type survey. RESULTS: PROSER displayed 46 laboratory values with corresponding reference ranges and abnormal flags, allowed for entry of 14 microscopy findings, and computed 2 calculations based on laboratory values; it composed automated PBS reports using a library of 92 prewritten phrases. Overall, PROSER was well received by residents. CONCLUSIONS: In this quality improvement study, we successfully deployed a web-based CDS tool for PBS interpretation. Future work is needed to quantitatively evaluate this intervention's effects on clinical outcomes and resident training.

Associations between ABO non-identical platelet transfusions and patient outcomes-A multicenter retrospective analysis.

BACKGROUND: Due to platelet availability limitations, platelet units ABO mismatched to recipients are often transfused. However, since platelets express ABO antigens and are collected in plasma which may contain ABO isohemagglutinins, it remains controversial as to whether ABO non-identical platelet transfusions could potentially pose harm and/or have reduced efficacy. STUDY DESIGN AND METHODS: The large 4-year publicly available Recipient Epidemiology and Donor Evaluation Study-III (REDS-III) database was used to investigate patient outcomes associated with ABO non-identical platelet transfusions. Outcomes included mortality, sepsis, and subsequent platelet transfusion requirements. RESULTS: Following adjustment for possible confounding factors, no statistically significant association between ABO non-identical platelet transfusion and increased risk of mortality was observed in the overall cohort of 21,176 recipients. However, when analyzed by diagnostic category and recipient ABO group, associations with increased mortality for major mismatched transfusions were noted in two of eight subpopulations. Hematology/Oncology blood group A and B recipients (but not group O) showed a Hazard Ratio (HR) of 1.29 (95%CI: 1.03-1.62) and intracerebral hemorrhage group O recipients (but not groups A and B) showed a HR of 1.75 (95%CI: 1.10-2.80). Major mismatched transfusions were associated with increased odds of receiving additional platelet transfusion each post-transfusion day (through day 5) regardless of the recipient blood group. DISCUSSION: We suggest that prospective studies are needed to determine if specific patient populations would benefit from receiving ABO identical platelet units. Our findings indicate that ABO-identical platelet products minimize patient exposure to additional platelet doses.

Assessment of the Frequency, Phenotypes, and Outcomes of Acute Liver Injury Associated with Amoxicillin/Clavulanate in 1.4 Million Patients in the Veterans Health Administration.

INTRODUCTION: Drug-induced liver injury is a significant health issue, yet the exposure-based incidence remains to be characterized. OBJECTIVE: We aimed to assess the frequency, phenotypes, and outcomes of acute liver injury associated with amoxicillin/clavulanate using a large electronic health record system. METHODS: Using the Veterans Health Administration electronic health record system, we developed the framework to identify unexplained acute liver injury, defined by alanine aminotransferase and/or alkaline phosphatase elevation temporally linked to prescription records of amoxicillin/clavulanate, a major culprit of clinically significant drug-induced liver injury, excluding other competing causes. The population was subcategorized by pre-existing liver conditions and inpatient status at the time of exposure for the analysis. RESULTS: Among 1,445,171 amoxicillin/clavulanate first exposures in unique individuals [92% men; mean age (standard deviation): 59 (15) years], 6476 (incidence: 0.448%) acute liver injuries were identified. Of these, 4427 (65%) had alternative causes, yielding 2249 (incidence: 0.156%) with unexplained acute liver injuries. The incidence of unexplained acute liver injury was lowest in outpatients without underlying liver disease (0.067%) and highest in inpatients with pre-existing liver conditions (0.719%). Older age, male sex, and American Indian or Alaska Native (vs White) were associated with a higher incidence of unexplained acute liver injury. Cholestatic injury affected 74%, exhibiting a higher frequency with advanced age, inpatient exposure, and pre-existing liver conditions. Hepatocellular injury with bilirubin elevation affected 0.003%, with a higher risk at age >45 years. During a 12-month follow-up, patients with unexplained acute liver injury had a higher adjusted overall mortality risk than those without evident acute liver injury. CONCLUSIONS: This framework identifies unexplained acute liver injury following drug exposure in large electronic health record datasets. After validating in other systems, this framework can aid in deducing drug-induced liver injury in the general patient population and regulatory decision making to promote drug safety and public health.

Screening for Latent Infections Among Users of High-Risk Immunosuppressants: A Cross-Sectional Analysis From the Veterans Health Administration Healthcare System.

OBJECTIVES: Guidelines recommend screening for latent hepatitis B virus (HBV), hepatitis C virus (HCV), and tuberculosis (TB) before initiating biologics or targeted synthetic disease-modifying antirheumatic drugs (b/ts DMARDs) to avoid reactivation of life-threatening infections. The extent to which such screening occurs in the national Veterans Health Administration (VA) healthcare system is unknown. METHODS: Using data from the Veterans Affairs' (VA) Corporate Data Warehouse, we performed a cross-sectional analysis of veterans receiving b/ts DMARDs between October 1, 2017, and September 30, 2019. We calculated the proportion of patients with screening completed for latent HBV, HCV, and TB between October 1, 1999 and September 30, 2019. Patient characteristics associated with complete screening were evaluated using mixed-effects multivariate logistic regression models. We also examined facility-level factors associated with high versus lower performance. RESULTS: A total of 51,764 unique patients from 129 VA facilities received b/ts DMARDs from 2017 to 2019. Of these, 63% had complete screening. Among the 11,006 patients identified as new users, 64% had complete screening. Higher screening rates were observed among Hispanic/Latinx and Black/African American patients, users of B-cell therapies, and patients who had seen oncology subspecialists. Substantial variation was observed across facilities, with complete screening ranging from 13% to 98% of patients. Higher screening rates were associated with highly complex, urban, and higher-volume facilities. CONCLUSIONS: Approximately two-thirds of veterans taking b/ts DMARDs have received guideline-recommended screening for HBV, HCV, and TB, but substantial facility variation was observed. Performance measures, robust multidisciplinary workflows, and electronic health record-based tools to feed information back to providers may improve screening rates for low-performing facilities.

Combining Charlson comorbidity and VACS indices improves prognostic accuracy for all-cause mortality for patients with and without HIV in the Veterans Health Administration.

Introduction: As people age with HIV (PWH), many comorbid diseases are more common than among age matched comparators without HIV (PWoH). While the Veterans Aging Cohort (VACS) Index 2.0 accurately predicts mortality in PWH using age and clinical biomarkers, the only included comorbidity is hepatitis C. We asked whether adding comorbid disease groupings from the Charlson Comorbidity Index (CCI) improves the accuracy of VACS Index. Methods: To maximize our ability to model mortality among older age groups, we began with PWoH in Veterans Health Administration (VA) from 2007-2017, divided into development and validation samples. Baseline predictors included age, and components of CCI and VACS Index (excluding CD4 count and HIV RNA). Patients were followed until December 31, 2021. We used Cox models to develop the VACS-CCI score and estimated mortality using a parametric (gamma) survival model. We compared accuracy using C-statistics and calibration curves in validation overall and within subgroups (gender, age

Sexually Transmitted Infection Testing in the National Veterans Health Administration Patient Cohort During the Coronavirus Disease 2019 Pandemic.

Background: We performed a retrospective study of chlamydia, gonorrhea, syphilis, and human immunodeficiency virus (HIV) testing in the Veterans Health Administration (VHA) during 2019-2021. Methods: We determined the annual number of chlamydia, gonorrhea, syphilis, and HIV tests from 2019 through 2021 using electronic health record data. We calculated rates by age, birth sex, race, census region, rurality, HIV status, and use of preexposure prophylaxis. Results: The VHA system experienced a 24% drop in chlamydia/gonorrhea testing, a 25% drop in syphilis testing, and a 29% drop in HIV testing in 2020 versus 2019. By the conclusion of 2021, testing rates had recovered to 90% of baseline for chlamydia/gonorrhea, 91% for syphilis, and 88% for HIV. Declines and subsequent improvements in sexually transmitted infection (STI) testing occurred unequally across age, sex, race, and geographic groups. Testing for all 4 STIs in 2021 remained below baseline in rural Veterans. Excluding those aged <25 years, women experienced a steeper decline and slower recovery in chlamydia/gonorrhea testing relative to men, but quicker recovery in HIV testing. Asian Americans and Hawaiian/Pacific Islanders had a steeper decline and a slower recovery in testing for chlamydia/gonorrhea. Black and White Veterans had slower recovery in HIV testing compared with other race groups. People living with HIV experienced a smaller drop in testing for syphilis compared with people without HIV, followed by a near-total recovery of testing by 2021. Conclusions: After dramatic reductions from 2019 to 2020, STI testing rates returned to near-baseline in 2021. Testing recovery lagged in rural, female, Asian American, Hawaiian/Pacific Islander, and Black Veterans.

Gonorrhea and Chlamydia Testing and Case Rates Among Women Veterans in the Veterans Health Administration.

BACKGROUND: United States (US) rates of sexually transmitted infection (STI) in women, especially gonorrhea and chlamydia, have increased over the past decade. Women Veterans may be at increased risk for STIs due to high rates of sexual trauma. Despite the availability of effective diagnostic tests and evidence-based guidelines for annual screening among sexually active women under age 25, screening rates for gonorrhea and chlamydia remain low in the US and among Veterans. OBJECTIVE: To examine patient characteristics and health system factors associated with gonorrhea and chlamydia testing and case rates among women Veterans in the Veterans Health Administration (VHA) in 2019. DESIGN: We performed a retrospective cohort study of all women Veterans in VHA care between January 1, 2018, and December 31, 2019. PARTICIPANTS: Women Veteran patients were identified as receiving VHA care if they had at least one inpatient admission or outpatient visit in 2019 or the preceding calendar year. KEY RESULTS: Among women under age 25, 21.3% were tested for gonorrhea or chlamydia in 2019. After adjusting for demographic and other health factors, correlates of testing in women under age 25 included Black race (aOR: 2.11, CI: 1.89, 2.36), rural residence (aOR: 0.84, CI: 0.74, 0.95), and cervical cancer screening (aOR: 5.05, CI: 4.59, 5.56). Women under age 25 had the highest infection rates, with an incidence of chlamydia and gonorrhea of 1,950 and 267 cases/100,000, respectively. Incidence of gonorrhea and chlamydia was higher for women with a history of military sexual trauma (MST) (chlamydia case rate: 265, gonorrhea case rate: 97/100,000) and those with mental health diagnoses (chlamydia case rate: 263, gonorrhea case rate: 72/100,000.) CONCLUSIONS: Gonorrhea and chlamydia testing remains underutilized among women in VHA care, and infection rates are high among younger women. Patient-centered, system-level interventions are urgently needed to address low testing rates.

Graphical analysis of guideline adherence to detect systemwide anomalies in HIV diagnostic testing.

BACKGROUND: Analyses of electronic medical databases often compare clinical practice to guideline recommendations. These analyses have a limited ability to simultaneously evaluate many interconnected medical decisions. We aimed to overcome this limitation with an alternative method and apply it to the diagnostic workup of HIV, where misuse can contribute to HIV transmission, delay care, and incur unnecessary costs. METHODS: We used graph theory to assess patterns of HIV diagnostic testing in a national healthcare system. We modeled the HIV diagnostic testing guidelines as a directed graph. Each node in the graph represented a test, and the edges pointed from one test to the next in chronological order. We then graphed each patient's HIV testing. This set of patient-level graphs was aggregated into a single graph. Finally, we compared the two graphs, the first representing the recommended approach to HIV diagnostic testing and the second representing the observed patterns of HIV testing, to assess for clinical practice deviations. RESULTS: The HIV diagnostic testing of 1.643 million patients provided 8.790 million HIV diagnostic test results for analysis. Significant deviations from recommended practice were found including the use of HIV resistance tests (n = 3,007) and HIV nucleic acid tests (n = 16,567) instead of the recommended HIV screen. CONCLUSIONS: We developed a method that modeled a complex medical scenario as a directed graph. When applied to HIV diagnostic testing, we identified deviations in clinical practice from guideline recommendations. The model enabled the identification of intervention targets and prompted systemwide policy changes to enhance HIV detection.

Development and Implementation of a Standard Format for Clinical Laboratory Test Results.

OBJECTIVES: Surprisingly, laboratory results, the principal output of clinical laboratories, are not standardized. Thus, laboratories frequently report results with identical meaning in different formats. For example, laboratories report a positive pregnancy test as "+," "P," or "Positive." To assess the feasibility of a widespread implementation of a result standard, we (1) developed a standard result format for common laboratory tests and (2) implemented a feedback system for clinical laboratories to view their unstandardized results. METHODS: In the largest integrated health care system in America, 130 facilities had the opportunity to collaboratively develop the standard. For 15 weeks, clinical laboratories received a weekly report of their unstandardized results. At the study's conclusion, laboratories were compared with themselves and their peers by metrics that reflected their unstandardized results. RESULTS: We rereviewed 156 million test results and observed a 51% decline in the rate of unstandardized results. The number of facilities with fewer than 23 unstandardized results per 100,000 (Six Sigma σ > 5) increased by 58% (52 to 82 facilities; ß = 1.79; P < .001). CONCLUSIONS: This study demonstrated significant improvement in the standardization of clinical laboratory results in a relatively short time. The laboratory community should create and promulgate a standardized result format.

Human herpesvirus 8-negative effusion-based large B-cell lymphoma: a distinct entity with unique clinicopathologic characteristics.

Rare cases of human herpesvirus 8 (HHV8)-negative effusion-based large B-cell lymphoma (EB-LBCL) occur in body cavities without antecedent or concurrent solid mass formation. In contrast to HHV8 + primary effusion lymphoma (PEL), EB-LBCL has no known association with HIV or HHV8 infection. However, the small sample sizes of case reports and series worldwide, especially from non-Japanese regions, have precluded diagnostic uniformity. Therefore, we conducted a retrospective, multi-institutional study of 55 cases of EB-LBCL and performed a comprehensive review of an additional 147 cases from the literature to identify distinct clinicopathologic characteristics. In our study, EB-LBCL primarily affected elderly (median age 80 years), immunocompetent patients and manifested as lymphomatous effusion without a solid component. The lymphomatous effusions mostly occurred in the pleural cavity (40/55, 73%), followed by the pericardial cavity (17/55, 31%). EB-LBCL expressed CD20 (53/54, 98%) and PAX5 (23/23, 100%). Most cases (30/36, 83%) were of non-germinal center B-cell subtype per the Hans algorithm. HHV8 infection was absent (0/55, 0%), while Epstein-Barr virus was detected in 6% (3/47). Clinically, some patients were managed with drainage alone (15/34, 44%), while others received rituximab alone (4/34, 12%) or chemotherapy (15/34, 44%). Eventually, 56% (22/39) died with a median overall survival (OS) of 14.9 months. Our findings were similar to those from the literature; however, compared to the non-Japanese cases, the Japanese cases had a significantly higher incidence of pericardial involvement, a higher rate of chemotherapy administration, and longer median OS. Particularly, we have found that Japanese residence, presence of pericardial effusion, and absence of MYC rearrangement are all favorable prognostic factors. Our data suggest that EB-LBCL portends a worse prognosis than previously reported, although select patients may be managed conservatively. Overall, EB-LBCL has distinct clinicopathologic characteristics, necessitating the establishment of separate diagnostic criteria and consensus nomenclature.

Robust Hepatitis A Vaccination Response Within the United States Veterans Health Administration in the Wake of State Outbreaks.

We assessed hepatitis A (HepA) vaccine receipt among susceptible individuals in outbreak and matched nonoutbreak states. Difference-in-differences models and multivariable logistic regression were used to compare HepA vaccination rates in these states. In the postoutbreak year, there was a 112% increase in HepA vaccinations in outbreak states versus a 6% decrease in nonoutbreak states. Differences persisted in our multivariable model (adjusted odds ratio = 2.53; 95% confidence interval = 2.45, 2.61). HepA vaccination rates increased dramatically in outbreak states, but many individuals susceptible to hepatitis A virus remain unvaccinated. (Am J Public Health. 2022;112(7):990-994. https://doi.org/10.2105/AJPH.2022.306845).

Unexpectedly low hemoglobin A1c in a patient with chronic lymphocytic leukemia.

Glycated hemoglobin (HbA1c) assays are currently utilized to monitor patients with diabetes mellitus type 2 (T2DM). These assays employ various methods, some of which are more prone to interference than others. Commonly recognized causes of interference include hemoglobin variants and conditions that result in reduced red blood cell survival such as hemolytic anemia and certain medications. Enzymatic assays represent one of the predominant methods for HbA1c testing for practical reasons. Herein, we describe a potentially novel interference in an enzymatic HbA1c assay by neoplastic lymphocytes in a patient with chronic lymphocytic leukemia. We hypothesize that the marked number of neoplastic lymphocytes are interfering in the enzymatic steps inherent to this assay, resulting in a discordantly low HbA1c result. Awareness of this possibility and further investigation into the precise mechanism by which this interference is occurring are warranted.

Clinical knowledge extraction via sparse embedding regression (KESER) with multi-center large scale electronic health record data.

The increasing availability of electronic health record (EHR) systems has created enormous potential for translational research. However, it is difficult to know all the relevant codes related to a phenotype due to the large number of codes available. Traditional data mining approaches often require the use of patient-level data, which hinders the ability to share data across institutions. In this project, we demonstrate that multi-center large-scale code embeddings can be used to efficiently identify relevant features related to a disease of interest. We constructed large-scale code embeddings for a wide range of codified concepts from EHRs from two large medical centers. We developed knowledge extraction via sparse embedding regression (KESER) for feature selection and integrative network analysis. We evaluated the quality of the code embeddings and assessed the performance of KESER in feature selection for eight diseases. Besides, we developed an integrated clinical knowledge map combining embedding data from both institutions. The features selected by KESER were comprehensive compared to lists of codified data generated by domain experts. Features identified via KESER resulted in comparable performance to those built upon features selected manually or with patient-level data. The knowledge map created using an integrative analysis identified disease-disease and disease-drug pairs more accurately compared to those identified using single institution data. Analysis of code embeddings via KESER can effectively reveal clinical knowledge and infer relatedness among codified concepts. KESER bypasses the need for patient-level data in individual analyses providing a significant advance in enabling multi-center studies using EHR data.

Transfusion practices for pediatric oncology and hematopoietic stem cell transplantation patients: Data from the National Heart Lung and Blood Institute Recipient Epidemiology and Donor Evaluation Study-III (REDS-III).

BACKGROUND: To evaluate transfusion practices in pediatric oncology and hematopoietic stem cell transplant (HSCT) patients. STUDY DESIGN AND METHODS: This is a multicenter retrospective study of children with oncologic diagnoses treated from 2013 to 2016 at hospitals participating in the National Heart Lung and Blood Institute Recipient Epidemiology and Donor Evaluation Study-III. Transfusion practices were evaluated by diagnosis codes and pre-transfusion laboratory values. RESULTS: A total of 4766 inpatient encounters of oncology and HSCT patients were evaluated, with 39.3% (95% confidence interval [CI]: 37.9%-40.7%) involving a transfusion. Red blood cells (RBCs) were the most commonly transfused component (32.4%; 95% CI: 31.1%-33.8%), followed by platelets (22.7%; 95% CI: 21.5%-23.9%). Patients in the 1 to <6 years of range were most likely to be transfused and HSCT, acute myeloid leukemia, and aplastic anemia were the diagnoses most often associated with transfusion. The median hemoglobin (Hb) prior to RBC transfusion was 7.5 g/dl (10-90th percentile: 6.4-8.8 g/dl), with 45.7% of transfusions being given at 7 to <8 g/dl. The median platelet count prior to platelet transfusion was 20 × 109 /L (10-90th percentile: 8-51 × 109 /L), and 37.9% of transfusions were given at platelet count of >20-50 × 109 /L. The median international normalized ratio (INR) prior to plasma transfusion was 1.7 (10-90th percentile: 1.3-2.7), and 36.3% of plasma transfusions were given at an INR between 1.4 and 1.7. DISCUSSION: Transfusion of blood components is common in hospitalized pediatric oncology/HSCT patients. Relatively high pre-transfusion Hb and platelet values and relatively low INR values prior to transfusion across the studied diagnoses highlight the need for additional studies in this population.