Search

PURPOSE Venous thromboembolism (VTE), especially pulmonary embolism (PE) and lower extremity deep vein thrombosis (LE-DVT), is a serious and potentially preventable complication for patients with cancer undergoing systemic therapy. METHODS Using retrospective data from patients diagnosed with incident cancer from 2011-2020, we derived a parsimonious risk assessment model (RAM) using least absolute shrinkage and selection operator regression from the Harris Health System (HHS, n = 9,769) and externally validated it using the Veterans Affairs (VA) health care system (n = 79,517). Bootstrapped c statistics and calibration curves were used to assess external model discrimination and fit. Dichotomized risk strata using integer scores were created and compared against the Khorana score (KS). RESULTS Incident VTE and PE/LE-DVT at 6 months occurred in 590 (6.2%) and 437 (4.6%) patients in HHS and 4,027 (5.1%) and 3,331 (4.2%) patients in the VA health care system. Assessed at the time of systemic therapy initiation, the new RAM included components of the KS with the modified cancer subtype, cancer staging, systemic therapy class, history of VTE, history of paralysis/immobility, recent hospitalization, and Asian/Pacific Islander race. The c statistic was 0.71 in HHS and 0.68 in the VA health care system (compared with 0.65 and 0.60, respectively, for KS). Furthermore, the new RAM appropriately reclassified 28% of patients and increased the proportion of VTEs in the high-risk group from 37% to 68% in the validation data set. CONCLUSION The novel RAM stratified patients with cancer into a high-risk group with 8%-10% cumulative incidence of VTE and 7% PE/LE-DVT at 6 months ( v 3% and 2%, respectively, in the low-risk group). The model had improved performance over the original KS and doubled the number of VTE events in the high-risk stratum. We encourage additional external validation from prospective studies. [Media: see text]

Pancreatic cancer is an aggressive disease that typically presents late with poor outcomes, indicating a pronounced need for early detection. In this study, we applied artificial intelligence methods to clinical data from 6 million patients (24,000 pancreatic cancer cases) in Denmark (Danish National Patient Registry (DNPR)) and from 3 million patients (3,900 cases) in the United States (US Veterans Affairs (US-VA)). We trained machine learning models on the sequence of disease codes in clinical histories and tested prediction of cancer occurrence within incremental time windows (CancerRiskNet). For cancer occurrence within 36 months, the performance of the best DNPR model has area under the receiver operating characteristic (AUROC) curve = 0.88 and decreases to AUROC (3m) = 0.83 when disease events within 3 months before cancer diagnosis are excluded from training, with an estimated relative risk of 59 for 1,000 highest-risk patients older than age 50 years. Cross-application of the Danish model to US-VA data had lower performance (AUROC = 0.71), and retraining was needed to improve performance (AUROC = 0.78, AUROC (3m) = 0.76). These results improve the ability to design realistic surveillance programs for patients at elevated risk, potentially benefiting lifespan and quality of life by early detection of this aggressive cancer.

We aimed to evaluate and compare the performance of multiple myeloma (MM) selection algorithms for use in Veterans Affairs (VA) research.Using the VA Corporate Data Warehouse (CDW), the VA Cancer Registry (VACR), and VA pharmacy data, we randomly selected 500 patients from 01/01/1999 to 06/01/2021 who had (1) either one MM diagnostic code OR were listed in the VACR as having MM AND (2) at least one MM treatment code. A team reviewed oncology notes for each veteran to annotate details regarding MM diagnosis and initial treatment within VA. We evaluated inter-annotator agreement and compared the performance of four published algorithms (two developed and validated external to VA data and two used in VA data).A total of 859 patients were reviewed to obtain 500 patients who were annotated as having MM and initiating MM treatment in VA. Agreement was high among annotators for all variables: MM diagnosis (98.3% agreement, Kappa = 0.93); initial treatment in VA (91.8% agreement; Kappa = 0.77); and initial treatment classification (87.6% agreement; Kappa = 0.86). VA Algorithms were more specific and had higher PPVs than non-VA algorithms for both MM diagnosis and initial treatment in VA. developed the "VA Recommended Algorithm," which had the highest PPV among all algorithms in identifying patients diagnosed with MM (PPV = 0.98, 95% CI = 0.95-0.99) and in identifying patients who initiated their MM treatment in VA (PPV = 0.93, 95% CI = 0.90-0.96).Our VA Recommended Algorithm optimizes sensitivity and PPV for cohort selection and treatment classification. This article is protected by copyright. All rights reserved.

The Surveillance, Epidemiology, and End Results (SEER) program reflects a third of the population of the United States. However, SEER may not be generalizable to the veteran population. Because veterans comprise a high-risk population, this discrepancy may limit our understanding of the epidemiology of melanoma in such high-risk populations.To assess differences in demographics, tumor characteristics, and melanoma-specific survival (MSS) in veterans compared to the general population.Data were collected from the Veterans Affairs Cancer Registry (VACR) and SEER (18 registries) from 2009 to 2017.We identified 15,334 veterans and 166,265 SEER patients with melanoma. Veterans were more likely to present with regional or distant disease (17.5% vs 13.0% in SEER). In VACR relative to SEER, the 5-year MSS was lower across all ages, except those diagnosed at ≥80 years. From 2009 to 2017, MSS by stage was lower across all stages in VACR. However, for stage IV melanomas diagnosed in 2015 to 2017 compared to 2011-2014, 2-year MSS increased from 37.8% to 51.5% in VACR versus 36.4% to 44.8% in SEER.Unique veteran demographics and missing data inherent to VACR.Compared to SEER, veterans with melanoma were diagnosed at later stages; however, both exhibited recent improvement in stage IV MSS.

Background We previously reported a transient increased risk of hospitalization after mRNA vaccination among patients with prior SARS-CoV-2, absolute risk ~ 1:1000. Here, we extend and expand this analysis to evaluate the impact of prior infection on hospitalization after a third (booster) dose. Methods Nationwide, retrospective cohort study of hospitalization among US Veterans who received a third dose of mRNA vaccine between 7/1/2021-2/28/2022. Daily rates of incident hospitalization were compared before and after booster doses, stratified by history of SARS-CoV-2. Results 1,632,806 patients received a third dose, including 90,174 with a history of SARS-CoV-2 infection. Hospitalization rates were unchanged before and after the booster dose among patients with (112.3/100,000 post-dose versus 100.2/100,000 pre-dose, p = 0.24) or without previous infection (32.1/100,000 post-dose versus 31.3/100,000 pre-dose, p = 0.71). Among 241 patients hospitalized after receipt of the initial vaccination, 90 received a booster, and none of these 90 patients were hospitalized. Conclusions There was not convincing evidence for increased hospitalizations shortly after booster vaccines, including in patients with a history of SARS-CoV-2 infection who required hospitalization after their initial vaccine. The size and design of the study prevent strong conclusions about absence of risk.

BackgroundCardiovascular disease (CVD) is the leading cause of death in men with prostate cancer (PC). Accumulated stress plays an important role in CVD development. The cumulative burden of chronic stress and life events can be measured using allostatic load (AL).MethodsThe initial cohort included males aged 18 years and older diagnosed with PC (2005-2019). AL was modeled as an ordinal variable (0-11). Fine-Gray competing risk regressions measured the impact of precancer diagnosis AL and postdiagnosis AL in 2-year major cardiac events (MACE). The effect of AL changes over time on MACE development was calculated via piecewise Cox regression (before, and 2 months, 6 months, and 1 year after PC diagnosis).ResultsWe included 5261 PC patients of which 6.6% had a 2-year MACE. For every 1-point increase in AL before and within 60 days after PC diagnosis, the risk of MACE increased 25% (adjusted hazard ratio [aHR] =1.25, 95% confidence interval [CI] = 1.18 to 1.33) and 27% (aHR = 1.27, 95% CI = 1.20 to 1.35), respectively. Using AL as a time-varying exposure, the risk of MACE increased 19% (aHR = 1.19, 95% CI = 1.11 to 1.27), 22% (aHR = 1.22, 95% CI = 1.14 to 1.33), 28% (aHR = 1.28, 95% CI = 1.23 to 1.33), and 31% (aHR = 1.31, 95% CI = 1.27 to 1.35) for every 1-point increase in AL before, 2 months after, 6 months after, and 1 year after PC diagnosis, respectively.ConclusionAL and its changes over time are associated with MACE in PC patients, suggesting a role of a biological measure of stress as a marker of CVD risk among men with PC.

Background. Death within 30 days of a positive test for SARS-CoV-2 infection is used by some agencies in definitions of death from COVID-19. With omicron variants predominating in 2022, widespread immunity, and routine screening of asymptomatic patients in some healthcare systems, there is cause to systematically evaluate the proportion of deaths within 30 days of a positive test are COVID-19-related. Methods. Nationwide records in the Veterans Health Administration were used to identify all patients who died within 30 days after positive tests for SARS-CoV-2. Randomly-selected records of previously-vaccinated patients with positive tests in January-February, 2022, or June-August, 2022, underwent review for classification of death as caused by or contributed to by COVID-19, focused on death within 30 days. Multivariable logistic regression was used to determine demographic and clinical variables associated with classification of death as COVID-related, which included cause or contribution. Results. Among vaccinated patients who died within 30 days of a positive test in January-February, 2022 (omicron BA.1 predominance), death was COVID-related in 103/150 cases (69%), as the cause in 55% and contributor in 14%. In June-August, 2022 (BA.5 predominance), death was COVID-related in 70/150 cases (47%), as the cause in 22% and contributor in 25%. Variables positively associated with COVID-related death included medications for COVID-19, hypoxemia, immune-suppressive drugs, and age, whereas advanced cancer, time since positive test, and June-August time period had negative associations. Conclusions. As of mid-2022, “death within 30 days” alone no longer provided an accurate estimate of deaths from COVID-19 in a US healthcare system with routine asymptomatic screening. Over time, the proportion of cases where SARS-CoV-2 was a contributor, rather than the primary cause of death, has increased.

With growing knowledge of multiple myeloma (MM) pathobiology and the introduction of novel therapies, risk stratification continues to evolve. Myeloid-derived suppressor cells and tumor-associated macrophages, derived from peripheral blood monocytes, support malignant plasma cell proliferation in the bone marrow. Since peripheral blood absolute monocyte count (AMC) is thought to could reflect the bone marrow microenvironment, we sought to evaluate the prognostic significance of AMC in MM. We retrospectively analyzed 10,822 patients with newly-diagnosed MM between 2000 and 2019 at Veteran's Administration hospitals. We obtained AMC closest to diagnosis and every 3 months thereafter up to 2.5 years. Patients were stratified into 4 groups: low, normal, elevated, and severely elevated AMC (AMC 0.2, 0.2 to0.8, 0.8-1.25 and1.25 x103/mm3, respectively). Abnormal AMC at diagnosis was observed in 25.3% of patients and was associated with inferior overall survival (OS). In patients with low, elevated, severely elevated, and normal AMC, At diagnorespectively, sis, median OS at diagnosis was 2.3, 2.7, 3.1, 3.6 years (p0.001), and at 2.5 years median OS at 2.5 years was 2.0, 2.6, 3.4, 3.9 years (p0.001) in patients with low, elevated, severely elevated, and normal AMC, respectively. Patients with normal AMC at diagnosis who developed an abnormal AMC1 year after diagnosis also had inferior OS relative to patients who maintained a normal AMC. Abnormal AMC was also associated with inferior OS independent of validated prognostic markers including the international staging system stage ISS Stage, and high lactate dehydrogenase LDH. Our findings provide novel clues for future prospective studies on the functional role of monocytes in multiple myeloma, which could be a readily available metric for risk stratification.

ImportanceWith a large proportion of the US adult population vaccinated against SARS-CoV-2, it is important to identify who remains at risk of severe infection despite vaccination.ObjectiveTo characterize risk factors for severe COVID-19 disease in a vaccinated population.Design, Setting, and ParticipantsThis nationwide, retrospective cohort study included US veterans who received a SARS-CoV-2 vaccination series and later developed laboratory-confirmed SARS-CoV-2 infection and were treated at US Department of Veterans Affairs (VA) hospitals. Data were collected from December 15, 2020, through February 28, 2022.ExposuresDemographic characteristics, comorbidities, immunocompromised status, and vaccination-related variables.Main Outcomes and MeasuresDevelopment of severe vs nonsevere SARS-CoV-2 infection. Severe disease was defined as hospitalization within 14 days of a positive SARS-CoV-2 diagnostic test and either blood oxygen level of less than 94%, receipt of supplemental oxygen or dexamethasone, mechanical ventilation, or death within 28 days. Association between severe disease and exposures was estimated using logistic regression models.ResultsAmong 110 760 patients with infections following vaccination (97 614 [88.1%] men, mean [SD] age at vaccination, 60.8 [15.3] years; 26 953 [24.3%] Black, 11 259 [10.2%] Hispanic, and 71 665 [64.7%] White), 10 612 (9.6%) had severe COVID-19. The strongest association with risk of severe disease after vaccination was age, which increased among patients aged 50 years or older with an adjusted odds ratio (aOR) of 1.42 (CI, 1.40-1.44) per 5-year increase in age, such that patients aged 80 years or older had an aOR of 16.58 (CI, 13.49-20.37) relative to patients aged 45 to 50 years. Immunocompromising conditions, including receipt of different classes of immunosuppressive medications (eg, leukocyte inhibitor: aOR, 2.80; 95% CI, 2.39-3.28) or cytotoxic chemotherapy (aOR, 2.71; CI, 2.27-3.24) prior to breakthrough infection, or leukemias or lymphomas (aOR, 1.87; CI, 1.61-2.17) and chronic conditions associated with end-organ disease, such as heart failure (aOR, 1.74; CI, 1.61-1.88), dementia (aOR, 2.01; CI, 1.83-2.20), and chronic kidney disease (aOR, 1.59; CI, 1.49-1.69), were also associated with increased risk. Receipt of an additional (ie, booster) dose of vaccine was associated with reduced odds of severe disease (aOR, 0.50; CI, 0.44-0.57).Conclusions and RelevanceIn this nationwide, retrospective cohort of predominantly male US Veterans, we identified risk factors associated with severe disease despite vaccination. Findings could be used to inform outreach efforts for booster vaccinations and to inform clinical decision-making about patients most likely to benefit from preexposure prophylaxis and antiviral therapy.

Natural language processing of medical records offers tremendous potential to improve the patient experience. Sentiment analysis of clinical notes has been performed with mixed results, often highlighting the issue that dictionary ratings are not domain specific. Here, for the first time, we re-calibrate the labMT sentiment dictionary on 3.5M clinical notes describing 10,000 patients diagnosed with lung cancer at the Department of Veterans Affairs. The sentiment score of notes was calculated for two years after date of diagnosis and evaluated against a lab test (platelet count) and a combination of data points (treatments). We found that the oncology specific labMT dictionary, after re-calibration for the clinical oncology domain, produces a promising signal in notes that can be detected based on a comparative analysis to the aforementioned parameters.