Natural language inference for curation of structured clinical registries from unstructured text.

OBJECTIVE: Clinical registries-structured databases of demographic, diagnosis, and treatment information-play vital roles in retrospective studies, operational planning, and assessment of patient eligibility for research, including clinical trials. Registry curation, a manual and time-intensive process, is always costly and often impossible for rare or underfunded diseases. Our goal was to evaluate the feasibility of natural language inference (NLI) as a scalable solution for registry curation. MATERIALS AND METHODS: We applied five state-of-the-art, pretrained, deep learning-based NLI models to clinical, laboratory, and pathology notes to infer information about 43 different breast oncology registry fields. Model inferences were evaluated against a manually curated, 7439 patient breast oncology research database. RESULTS: NLI models showed considerable variation in performance, both within and across fields. One model, ALBERT, outperformed the others (BART, RoBERTa, XLNet, and ELECTRA) on 22 out of 43 fields. A detailed error analysis revealed that incorrect inferences primarily arose through models' tendency to misinterpret historical findings, as well as confusion based on abbreviations and subtle term variants common in clinical text. DISCUSSION AND CONCLUSION: Traditional natural language processing methods require specially annotated training sets or the construction of a separate model for each registry field. In contrast, a single pretrained NLI model can curate dozens of different fields simultaneously. Surprisingly, NLI methods remain unexplored in the clinical domain outside the realm of shared tasks and benchmarks. Modern NLI models could increase the efficiency of registry curation, even when applied "out of the box" with no additional training.

Cardiovascular Disease and Severe Hypoxemia Are Associated With Higher Rates of Noninvasive Respiratory Support Failure in Coronavirus Disease 2019 Pneumonia.

Acute hypoxemic respiratory failure is the major complication of coronavirus disease 2019, yet optimal respiratory support strategies are uncertain. We aimed to describe outcomes with high-flow oxygen delivered through nasal cannula and noninvasive positive pressure ventilation in coronavirus disease 2019 acute hypoxemic respiratory failure and identify individual factors associated with noninvasive respiratory support failure. DESIGN: Retrospective cohort study to describe rates of high-flow oxygen delivered through nasal cannula and/or noninvasive positive pressure ventilation success (live discharge without endotracheal intubation). Fine-Gray subdistribution hazard models were used to identify patient characteristics associated with high-flow oxygen delivered through nasal cannula and/or noninvasive positive pressure ventilation failure (endotracheal intubation and/or in-hospital mortality). SETTING: One large academic health system, including five hospitals (one quaternary referral center, a tertiary hospital, and three community hospitals), in New York City. PATIENTS: All hospitalized adults 18-100 years old with coronavirus disease 2019 admitted between March 1, 2020, and April 28, 2020. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: A total of 331 and 747 patients received high-flow oxygen delivered through nasal cannula and noninvasive positive pressure ventilation as the highest level of noninvasive respiratory support, respectively; 154 (46.5%) in the high-flow oxygen delivered through nasal cannula cohort and 167 (22.4%) in the noninvasive positive pressure ventilation cohort were successfully discharged without requiring endotracheal intubation. In adjusted models, significantly increased risk of high-flow oxygen delivered through nasal cannula and noninvasive positive pressure ventilation failure was seen among patients with cardiovascular disease (subdistribution hazard ratio, 1.82; 95% CI, 1.17-2.83 and subdistribution hazard ratio, 1.40; 95% CI, 1.06-1.84, respectively). Conversely, a higher peripheral blood oxygen saturation to Fio2 ratio at high-flow oxygen delivered through nasal cannula and noninvasive positive pressure ventilation initiation was associated with reduced risk of failure (subdistribution hazard ratio, 0.32; 95% CI, 0.19-0.54, and subdistribution hazard ratio 0.34; 95% CI, 0.21-0.55, respectively). CONCLUSIONS: A significant proportion of patients receiving noninvasive respiratory modalities for coronavirus disease 2019 acute hypoxemic respiratory failure achieved successful hospital discharge without requiring endotracheal intubation, with lower success rates among those with comorbid cardiovascular disease or more severe hypoxemia. The role of high-flow oxygen delivered through nasal cannula and noninvasive positive pressure ventilation in coronavirus disease 2019-related acute hypoxemic respiratory failure warrants further consideration.

Individuals with Down syndrome hospitalized with COVID-19 have more severe disease.

PURPOSE: Rare genetic conditions like Down syndrome (DS) are historically understudied. Infection is a leading cause of mortality in DS, along with cardiac anomalies. Currently, it is unknown how the COVID-19 pandemic affects individuals with DS. Herein, we report an analysis of individuals with DS who were hospitalized with COVID-19 in New York, New York, USA. METHODS: In this retrospective, dual-center study of 7246 patients hospitalized with COVID-19, we analyzed all patients with DS admitted in the Mount Sinai Health System and Columbia University Irving Medical Center. We assessed hospitalization rates, clinical characteristics, and outcomes. RESULTS: We identified 12 patients with DS. Hospitalized individuals with DS are on average ten years younger than patients without DS. Patients with DS have more severe disease than controls, particularly an increased incidence of sepsis and mechanical ventilation. CONCLUSION: We demonstrate that individuals with DS who are hospitalized with COVID-19 are younger than their non-DS counterparts, and that they have more severe disease than age-matched controls. We conclude that particular care should be considered for both the prevention and treatment of COVID-19 in these patients.

Trends in the Academic Credentials of Matched Dermatology Residency Applicants.

Introduction Research can be used to enhance the competitiveness of an application and is associated with a successful match. However, current reports regarding the publication record among prospective dermatology residents may be inaccurate. We sought to accurately assess the research credentials of matched dermatology residency candidates at the time of application. Methods We performed a bibliometric analysis to identify published articles of 1152 matched dermatology candidates and calculated the h-index of each applicant at the time of application. Details on article type, first authorship, and dermatology-relatedness of articles were collected. Results The median number of publications was two and the median h-index was 0. At the time of residency application, one-quarter of matched dermatology candidates (24%, n=278) possessed no publications. Over time, the median number of publications (R 0.10, p<0.001) and h-index (R 0.07, p=0.014) of matched applicants increased. The proportion of first-authored articles, dermatology-related papers, and each article type remained constant across application cycles (p>0.0500). An additional graduate degree, completion of a research fellowship, and graduation from a non-US medical school were independently associated with greater research credentials (p<0.0500). Conclusions Each year, applicants are publishing more articles and have a greater scholarly impact than in previous application cycles. However, the verified publication volume of matched dermatology applicants is strikingly lower than the values reported in national statistics.

Investigating the Gender Pay Gap in Industry Contributions to Academic Neurosurgeons.

BACKGROUND: A large gender gap exists in industry funding for academic neurosurgeons. Selection criteria for funding distribution remain unclear. However, academic rank, scholarly productivity, and experience have been suggested as determining factors. METHODS: We conducted a cross-sectional study of industry payments to US academic neurosurgeons. We used online faculty listings to determine academic rank and gender, then used the Center for Medicare and Medicaid Services Open Payment Database to identify industry contributions. Details were collected on H-index and length of time in practice was used as a proxy for experience. RESULTS: Of the 1481 academic neurosurgeons included, men were in the majority (91% vs. 9%, P = 0.0001). Relative to their male colleagues, female assistant and associate professors received fewer payments (4 vs. 8, P = 0.0040; 2 vs. 7, P = 0.0067) at lower median values ($409 vs. $437, P = 0.0490; $163 vs. $260, P = 0.0089). H-index was more strongly associated with general payment receipt for women academic neurosurgeons (r = 0.20, P = 0.0201) than men academic neurosurgeons (r = 0.06, P = 0.0301). Experience trended toward a significant association with industry funding in men (r = 0.05, P = 0.0601). After adjustment for scholarly productivity and experience, gender-based funding inequalities became insignificant. CONCLUSIONS: In academic neurosurgery, substantial gender disparities exist in industry payments and metrics of academic success. There may be an industry selection bias toward recruitment of key opinion and thought leaders, as identified by scholarly productivity and experience. Despite the objective gender inequalities, industry funding to academic neurosurgeons appears to be equitable when metrics of academic success are considered.

Genome-wide association study of multiplex schizophrenia pedigrees.

OBJECTIVE: The authors used a genome-wide association study (GWAS) of multiply affected families to investigate the association of schizophrenia to common single-nucleotide polymorphisms (SNPs) and rare copy number variants (CNVs). METHOD: The family sample included 2,461 individuals from 631 pedigrees (581 in the primary European-ancestry analyses). Association was tested for single SNPs and genetic pathways. Polygenic scores based on family study results were used to predict case-control status in the Schizophrenia Psychiatric GWAS Consortium (PGC) data set, and consistency of direction of effect with the family study was determined for top SNPs in the PGC GWAS analysis. Within-family segregation was examined for schizophrenia-associated rare CNVs. RESULTS: No genome-wide significant associations were observed for single SNPs or for pathways. PGC case and control subjects had significantly different genome-wide polygenic scores (computed by weighting their genotypes by log-odds ratios from the family study) (best p=10(-17), explaining 0.4% of the variance). Family study and PGC analyses had consistent directions for 37 of the 58 independent best PGC SNPs (p=0.024). The overall frequency of CNVs in regions with reported associations with schizophrenia (chromosomes 1q21.1, 15q13.3, 16p11.2, and 22q11.2 and the neurexin-1 gene [NRXN1]) was similar to previous case-control studies. NRXN1 deletions and 16p11.2 duplications (both of which were transmitted from parents) and 22q11.2 deletions (de novo in four cases) did not segregate with schizophrenia in families. CONCLUSIONS: Many common SNPs are likely to contribute to schizophrenia risk, with substantial overlap in genetic risk factors between multiply affected families and cases in large case-control studies. Our findings are consistent with a role for specific CNVs in disease pathogenesis, but the partial segregation of some CNVs with schizophrenia suggests that researchers should exercise caution in using them for predictive genetic testing until their effects in diverse populations have been fully studied.