Search

Objectives: Social support (SS) and social isolation (SI) are social determinants of health (SDOH) associated with psychiatric outcomes. In electronic health records (EHRs), individual-level SS/SI is typically documented in narrative clinical notes rather than as structured coded data. Natural language processing (NLP) algorithms can automate the otherwise labor-intensive process of extraction of such information., Materials and Methods: Psychiatric encounter notes from Mount Sinai Health System (MSHS, n = 300) and Weill Cornell Medicine (WCM, n = 225) were annotated to create a gold-standard corpus. A rule-based system (RBS) involving lexicons and a large language model (LLM) using FLAN-T5-XL were developed to identify mentions of SS and SI and their subcategories (eg, social network, instrumental support, and loneliness)., Results: For extracting SS/SI, the RBS obtained higher macroaveraged F1-scores than the LLM at both MSHS (0.89 versus 0.65) and WCM (0.85 versus 0.82). For extracting the subcategories, the RBS also outperformed the LLM at both MSHS (0.90 versus 0.62) and WCM (0.82 versus 0.81)., Discussion and Conclusion: Unexpectedly, the RBS outperformed the LLMs across all metrics. An intensive review demonstrates that this finding is due to the divergent approach taken by the RBS and LLM. The RBS was designed and refined to follow the same specific rules as the gold-standard annotations. Conversely, the LLM was more inclusive with categorization and conformed to common English-language understanding. Both approaches offer advantages, although additional replication studies are warranted., (© The Author(s) 2024. Published by Oxford University Press on behalf of the American Medical Informatics Association. All rights reserved. For permissions, please email: journals.permissions@oup.com.)

Homologous recombination (HR) requires bidirectional end resection initiated by a nick formed close to a DNA double-strand break (DSB), dysregulation favoring error-prone DNA end-joining pathways. Here we investigate the role of the ATAD5, a PCNA unloading protein, in short-range end resection, long-range resection not being affected by ATAD5 deficiency. Rapid PCNA loading onto DNA at DSB sites depends on the RFC PCNA loader complex and MRE11-RAD50-NBS1 nuclease complexes bound to CtIP. Based on our cytological analyses and on an in vitro system for short-range end resection, we propose that PCNA unloading by ATAD5 is required for the completion of short-range resection. Hampering PCNA unloading also leads to failure to remove the KU70/80 complex from the termini of DSBs hindering DNA repair synthesis and the completion of HR. In line with this model, ATAD5-depleted cells are defective for HR, show increased sensitivity to camptothecin, a drug forming protein-DNA adducts, and an augmented dependency on end-joining pathways. Our study highlights the importance of PCNA regulation at DSB for proper end resection and HR., (© The Author(s) 2023. Published by Oxford University Press on behalf of Nucleic Acids Research.)

Context: Metformin is the first-line drug for treating diabetes but has a high failure rate., Objective: To identify demographic and clinical factors available in the electronic health record (EHR) that predict metformin failure., Methods: A cohort of patients with at least 1 abnormal diabetes screening test that initiated metformin was identified at 3 sites (Arizona, Mississippi, and Minnesota). We identified 22 047 metformin initiators (48% female, mean age of 57 ± 14 years) including 2141 African Americans, 440 Asians, 962 Other/Multiracial, 1539 Hispanics, and 16 764 non-Hispanic White people. We defined metformin failure as either the lack of a target glycated hemoglobin (HbA1c) (<7%) within 18 months of index or the start of dual therapy. We used tree-based extreme gradient boosting (XGBoost) models to assess overall risk prediction performance and relative contribution of individual factors when using EHR data for risk of metformin failure., Results: In this large diverse population, we observed a high rate of metformin failure (43%). The XGBoost model that included baseline HbA1c, age, sex, and race/ethnicity corresponded to high discrimination performance (C-index of 0.731; 95% CI 0.722, 0.740) for risk of metformin failure. Baseline HbA1c corresponded to the largest feature performance with higher levels associated with metformin failure. The addition of other clinical factors improved model performance (0.745; 95% CI 0.737, 0.754, P < .0001)., Conclusion: Baseline HbA1c was the strongest predictor of metformin failure and additional factors substantially improved performance suggesting that routinely available clinical data could be used to identify patients at high risk of metformin failure who might benefit from closer monitoring and earlier treatment intensification., (© The Author(s) 2023. Published by Oxford University Press on behalf of the Endocrine Society. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.)

Objective: Artificial intelligence (AI) models may propagate harmful biases in performance and hence negatively affect the underserved. We aimed to assess the degree to which data quality of electronic health records (EHRs) affected by inequities related to low socioeconomic status (SES), results in differential performance of AI models across SES., Materials and Methods: This study utilized existing machine learning models for predicting asthma exacerbation in children with asthma. We compared balanced error rate (BER) against different SES levels measured by HOUsing-based SocioEconomic Status measure (HOUSES) index. As a possible mechanism for differential performance, we also compared incompleteness of EHR information relevant to asthma care by SES., Results: Asthmatic children with lower SES had larger BER than those with higher SES (eg, ratio = 1.35 for HOUSES Q1 vs Q2-Q4) and had a higher proportion of missing information relevant to asthma care (eg, 41% vs 24% for missing asthma severity and 12% vs 9.8% for undiagnosed asthma despite meeting asthma criteria)., Discussion: Our study suggests that lower SES is associated with worse predictive model performance. It also highlights the potential role of incomplete EHR data in this differential performance and suggests a way to mitigate this bias., Conclusion: The HOUSES index allows AI researchers to assess bias in predictive model performance by SES. Although our case study was based on a small sample size and a single-site study, the study results highlight a potential strategy for identifying bias by using an innovative SES measure., (© The Author(s) 2022. Published by Oxford University Press on behalf of the American Medical Informatics Association. All rights reserved. For permissions, please email: journals.permissions@oup.com.)

Objective: Social determinants of health (SDoH) are nonclinical dispositions that impact patient health risks and clinical outcomes. Leveraging SDoH in clinical decision-making can potentially improve diagnosis, treatment planning, and patient outcomes. Despite increased interest in capturing SDoH in electronic health records (EHRs), such information is typically locked in unstructured clinical notes. Natural language processing (NLP) is the key technology to extract SDoH information from clinical text and expand its utility in patient care and research. This article presents a systematic review of the state-of-the-art NLP approaches and tools that focus on identifying and extracting SDoH data from unstructured clinical text in EHRs., Materials and Methods: A broad literature search was conducted in February 2021 using 3 scholarly databases (ACL Anthology, PubMed, and Scopus) following Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. A total of 6402 publications were initially identified, and after applying the study inclusion criteria, 82 publications were selected for the final review., Results: Smoking status (n = 27), substance use (n = 21), homelessness (n = 20), and alcohol use (n = 15) are the most frequently studied SDoH categories. Homelessness (n = 7) and other less-studied SDoH (eg, education, financial problems, social isolation and support, family problems) are mostly identified using rule-based approaches. In contrast, machine learning approaches are popular for identifying smoking status (n = 13), substance use (n = 9), and alcohol use (n = 9)., Conclusion: NLP offers significant potential to extract SDoH data from narrative clinical notes, which in turn can aid in the development of screening tools, risk prediction models, and clinical decision support systems., (© The Author(s) 2021. Published by Oxford University Press on behalf of the American Medical Informatics Association.)

Pseudoxanthoma elasticum (PXE) is a multisystem disorder characterized by ectopic mineralization of connective tissues with primary manifestations in the skin, eyes and the cardiovascular system. The classic forms of PXE are caused by mutations in the ABCC6 gene encoding the ABCC6 protein, expressed primarily in the liver. Cutis laxa (CL) manifests with loose and sagging skin with loss of recoil. In 2009 we investigated a 19-year-old patient with overlapping cutaneous features of PXE and CL, together with alpha thalassaemia. Genetic analysis failed to identify pathogenic mutations in ABCC6. More recently we developed a gene-targeted panel of next-generation sequencing technology. This panel has 29 genes, 22 of which, including ABCC6 and GGCX, are associated with ectopic mineralization phenotypes. Mutation analysis in the patient identified two heterozygous GGCX mutations: c.200&#95;201delTT in exon 2 and c.763G>A, p.V255M in exon 7. The GGCX gene encodes a γ-glutamyl carboxylase necessary for activation of blood coagulation factors in the liver. The p.V255M mutation was previously reported to result in reduced γ-glutamyl carboxylase activity in vitro, while the c.200&#95;201delTT mutation is novel. Previous studies reported that mutations in GGCX cause overlapping PXE/CL skin phenotypes in association with or without multiple vitamin K-dependent coagulation factor deficiency. Our patient had loose redundant skin, moderate-to-severe angioid streaks and characteristic calcification of elastic structures in the mid dermis, consistent with PXE/CL overlap, but no coagulation abnormalities. Our studies expand the GGCX mutation landscape in patients with PXE-like phenotypes., (© 2020 British Association of Dermatologists.)

R-loops are formed when replicative forks collide with the transcriptional machinery and can cause genomic instability. However, it is unclear how R-loops are regulated at transcription-replication conflict (TRC) sites and how replisome proteins are regulated to prevent R-loop formation or mediate R-loop tolerance. Here, we report that ATAD5, a PCNA unloader, plays dual functions to reduce R-loops both under normal and replication stress conditions. ATAD5 interacts with RNA helicases such as DDX1, DDX5, DDX21 and DHX9 and increases the abundance of these helicases at replication forks to facilitate R-loop resolution. Depletion of ATAD5 or ATAD5-interacting RNA helicases consistently increases R-loops during the S phase and reduces the replication rate, both of which are enhanced by replication stress. In addition to R-loop resolution, ATAD5 prevents the generation of new R-loops behind the replication forks by unloading PCNA which, otherwise, accumulates and persists on DNA, causing a collision with the transcription machinery. Depletion of ATAD5 reduces transcription rates due to PCNA accumulation. Consistent with the role of ATAD5 and RNA helicases in maintaining genomic integrity by regulating R-loops, the corresponding genes were mutated or downregulated in several human tumors., (© The Author(s) 2020. Published by Oxford University Press on behalf of Nucleic Acids Research.)

Background: We assessed glioma incidence and disparities in postglioma survival rate in the Olmsted County, Minnesota, population., Methods: This population-based study assessed the incidence of pathologically confirmed primary gliomas between January 1, 1995, and December 31, 2014. Age- and sex-adjusted incidence rates per 100 000 person-years were calculated and standardized to the US white 2010 population. We compared incidence trends of glioma during our study period with previously published Olmsted County data from 1950 to 1990. We assessed postglioma survival rates among individuals with different socioeconomic status (SES), which was measured by a validated individual HOUsing-based SES index (HOUSES)., Results: We identified 135 incident glioma cases (93% white) with 20 pediatric (50% female) and 115 adult cases (44% female). Overall incidence rate during our study period, 5.51 per 100 000 person-years (95% CI: 4.56-6.46), showed no significant changes and was similar to that seen in 1950 to 1990, 5.5 per 100 000 person-years. The incidence of pediatric (age < 20 years) glioma was 2.49 (95% CI: 1.40-3.58), whereas adult glioma incidence was 6.47 (95% CI: 5.26-7.67). Among those with grade II to IV gliomas, individuals with lower SES (< median HOUSES) had significantly lower 5-year survival rates compared to those with higher SES, adjusted hazard ratio 1.61 (95% CI: 1.01-2.85)., Conclusion: In a well-defined North American population, long-term glioma incidence appears stable since 1950. Significant socioeconomic disparities exist for postglioma survival., (© The Author(s) 2019. Published by Oxford University Press on behalf of the Society for Neuro-Oncology and the European Association of Neuro-Oncology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.)

Objective: To assess clinical documentation variations across health care institutions using different electronic medical record systems and investigate how they affect natural language processing (NLP) system portability., Materials and Methods: Birth cohorts from Mayo Clinic and Sanford Children's Hospital (SCH) were used in this study (n = 298 for each). Documentation variations regarding asthma between the 2 cohorts were examined in various aspects: (1) overall corpus at the word level (ie, lexical variation), (2) topics and asthma-related concepts (ie, semantic variation), and (3) clinical note types (ie, process variation). We compared those statistics and explored NLP system portability for asthma ascertainment in 2 stages: prototype and refinement., Results: There exist notable lexical variations (word-level similarity = 0.669) and process variations (differences in major note types containing asthma-related concepts). However, semantic-level corpora were relatively homogeneous (topic similarity = 0.944, asthma-related concept similarity = 0.971). The NLP system for asthma ascertainment had an F-score of 0.937 at Mayo, and produced 0.813 (prototype) and 0.908 (refinement) when applied at SCH., Discussion: The criteria for asthma ascertainment are largely dependent on asthma-related concepts. Therefore, we believe that semantic similarity is important to estimate NLP system portability. As the Mayo Clinic and SCH corpora were relatively homogeneous at a semantic level, the NLP system, developed at Mayo Clinic, was imported to SCH successfully with proper adjustments to deal with the intrinsic corpus heterogeneity., (© The Author 2017. Published by Oxford University Press on behalf of the American Medical Informatics Association. All rights reserved. For Permissions, please email: journals.permissions@oup.com.)

Context: Thiazide diuretics, the antihypertensive agent prescribed most frequently worldwide, are commonly associated with hypercalcemia. However, the epidemiology and clinical features are poorly understood., Objective: To update the incidence of thiazide-associated hypercalcemia and clarify its clinical features., Patients and Methods: In a population-based descriptive study, Olmsted County, Minnesota, residents with thiazide-associated hypercalcemia were identified through the Rochester Epidemiology Project and the Mayo Clinic Laboratory Information System from 2002-2010 and were added to the historical cohort beginning in 1992., Main Outcome: Incidence rates were adjusted to the 2010 United States white population., Results: Overall, 221 Olmsted County residents were identified with thiazide-associated hypercalcemia an average of 5.2 years after initiation of treatment. Subjects were older (mean age, 67 years) and primarily women (86.4%). The incidence of thiazide-associated hypercalcemia increased after 1997 and peaked in 2006 with an annual incidence of 20 per 100,000, compared to an overall rate of 12 per 100,000 in 1992-2010. Severe hypercalcemia was not observed in the cohort despite continuation of thiazide treatment in 62.4%. Of patients discontinuing thiazides, 71% continued to have hypercalcemia. Primary hyperparathyroidism was diagnosed in 53 patients (24%), including five patients who underwent parathyroidectomy without thiazide discontinuation., Conclusions: Many patients with thiazide-associated hypercalcemia have underlying primary hyperparathyroidism. Additionally, a sharp rise in thiazide-associated hypercalcemia incidence began in 1998, paralleling the increase observed in primary hyperparathyroidism in this community. Case ascertainment bias from targeted osteoporosis screening is the most likely explanation.

Background and Aim: Celiac disease (CD) is a lifelong immune-mediated disease with excess mortality. Early diagnosis is important to minimize disease symptoms, complications, and consumption of healthcare resources. Most patients remain undiagnosed. We developed two electronic medical record (EMR)-based algorithms to identify patients at high risk of CD and in need of CD screening., Methods: (I) Using natural language processing (NLP), we searched EMRs for 16 free text (and related) terms in 216 CD patients and 280 controls. (II) EMRs were also searched for ICD9 (International Classification of Disease) codes suggesting an increased risk of CD in 202 patients with CD and 524 controls. For each approach, we determined the optimal number of hits to be assigned as CD cases. To assess performance of these algorithms, sensitivity and specificity were calculated., Results: Using two hits as the cut-off, the NLP algorithm identified 72.9% of all celiac patients (sensitivity), and ruled out CD in 89.9% of the controls (specificity). In a representative US population of individuals without a prior celiac diagnosis (assuming that 0.6% had undiagnosed CD), this NLP algorithm could identify a group of individuals where 4.2% would have CD (positive predictive value). ICD9 code search using three hits as the cut-off had a sensitivity of 17.1% and a specificity of 88.5% (positive predictive value was 0.9%)., Discussion and Conclusions: This study shows that computerized EMR-based algorithms can help identify patients at high risk of CD. NLP-based techniques demonstrate higher sensitivity and positive predictive values than algorithms based on ICD9 code searches.

The authors performed a systematic review of the association of complement component 2(C2)/complement factor B (CFB) gene polymorphisms with age-related macular degeneration (AMD). In total, data from 19 studies published between 2006 and 2011 were pooled for 4 polymorphisms: rs9332739 and rs547154 in the C2 gene and rs4151667 and rs641153 in the CFB gene. Data extraction and assessments for risk of bias were independently performed by 2 reviewers. Allele frequencies and allele and genotypic effects were pooled. Heterogeneity and publication bias were explored. Pooled minor allele frequencies for all 4 SNPs were between 4.7% and 9.6% for all polymorphisms, except for an Indian population in which the C allele at rs9332739 was the major allele. For the C2 polymorphisms, the minor C allele at rs9332739 and the minor T allele at rs547154 carried estimated relative risks (odds ratios) of 0.55 (95% confidence interval (CI): 0.46, 0.65) and 0.47 (95% CI: 0.39, 0.57), respectively. For the CFB polymorphisms, the minor A alleles at rs4151667 and rs614153 carried estimated risks of 0.54 (95% CI: 0.45, 0.64) and 0.41 (95% CI: 0.34, 0.51), respectively. These allele effects contributed to an absolute lowering of the risk of all AMD in Caucasian populations by 2.0%-6.0%. This meta-analysis provides a robust estimate of the protective association of C2/CFB with AMD.

Background: Many health measures (e.g., blood pressure, quality of life) have meaningful fluctuation over time around a relatively stable mean level for each person., Purpose: This didactic paper describes two closely related statistical models for examining between-person and within-person relationships between two or more sets of measures collected over time: the latent intercept model with correlated residuals (LI) in structural equation modeling framework and the multivariate multilevel model (MVML) in multilevel modeling framework., Results: We illustrated that the basic LI model and the MVML model are equivalent. We presented an illustrative example using a national arthritis data resource to examine between-person and within-person relationships of symptom status, functional health, and quality of life in arthritis patients., Discussion: Additional design and modeling issues for the treatment of missing data are considered. We discuss contexts in which one of the two models may be preferred. Mplus and SAS syntax are available.

Context: Patients with primary hyperparathyroidism (PHP) often report nonspecific symptoms including mood disturbances., Objective: The objective of the study was to determine the frequency of depression in PHP and assess its response to parathyroidectomy., Methods: A case-control study at a referral center in Rochester, MN, performed Patient Health Questionnaire-9 (PHQ-9) assessments in observed (n = 81) and surgical (n = 88) PHP and benign nontoxic surgical thyroid disease (n = 85) at baseline and 1, 3, 6, and 12 months after surgery or the initial questionnaire in observed PHP. Baseline PHQ-9 scores and their response to surgery were evaluated., Results: The groups were similar in gender and depression history, but PHP patients were older. Baseline PHQ-9 scores were 1.71 points higher in PHP than controls after adjusting for age and gender (P = .004). Clinically significant PHQ-9 scores (≥10) were twice as common in PHP (31.4%) compared with thyroid subjects (15.3%). Parathyroidectomy resulted in significant and sustained reductions in PHQ-9 scores, which were greater than observed PHP at all time points (P < .001). PHP patients with clinically significant PHQ-9 scores dropped to 7.4% (P < .001) and 7.6% (P < .001) at 1 month and 1 yr after parathyroidectomy. There were greater declines in PHQ-9 scores after parathyroidectomy at 1, 3, and 6 months (P < .001) and 1 yr (P = .061) compared with thyroid surgery., Conclusions: Depression is common in patients with PHP. Parathyroidectomy results in greater improvement in PHQ-9 scores compared with thyroid surgery or observation of PHP.

Background: Psoriasis has been considered as a T-helper 1 cell-mediated autoimmune disease driven by collaboration with multiple components of innate and acquired immune cells. Natural killer (NK) cells have been shown to bridge innate and acquired immunity, and thus could potentially contribute to the pathophysiology of psoriasis., Objectives: To investigate the phenotypic changes of circulating NK cells in patients with new-onset psoriasis., Methods: Fifteen patients with plaque psoriasis (eight women and seven men) who visited our clinic after their first episode of psoriasis and did not have a history of previous systemic therapy or phototherapy participated in this study. Peripheral blood mononuclear cells were isolated and stained with a panel of antibodies against cell surface receptors expressed on T and/or NK cells and analysed by flow cytometry., Results: As compared with normal healthy volunteers, patients with new-onset psoriasis showed no significant changes in numbers of peripheral NK, NK-T or T cells. NK activating receptors 2B4, CD48, NKG2D, CD16 and CD56 were found to be unchanged in new-onset psoriasis. However, the expression of Fas (activation-induced death receptor) was upregulated, whereas the expression of the NK inhibitory receptors CD94 and NKG2A was dramatically reduced on NK cells of new-onset psoriasis. These changes occurred at the level of mean fluorescent intensity, but minimally affected percentages of cells expressing Fas, CD94 and NKG2A., Conclusions: Our findings demonstrate that changes in the expression of Fas and CD94/NKG2A receptors on NK cells may occur during new-onset psoriasis, and are likely to contribute to the pathogenesis of psoriasis.

Background: Previous comparisons of monoclonal protein quantification identified a bias between serum protein electrophoresis (SPEP) and immunonephelometry (NEPH)., Methods: We reviewed data from 2845 patients in whom a single sample provided a gamma fraction M-spike by SPEP, a heavy chain isotype by immunofixation electrophoresis (IFE), and an Ig quantification by NEPH. We examined the relationship between SPEP and NEPH. Selected sera with high monoclonal protein concentrations were diluted and reassessed., Results: For all isotypes, the relationship between SPEP and NEPH was best fitted with cubic curves. We determined the concentrations of each isotype that fitted a linear relationship. IgA had the best correspondence (slope 0.92, 95% CI 0.87-1.02), whereas IgM demonstrated a systematic bias of higher values by NEPH (slope 1.80, 95% CI 1.68-1.92). IgG demonstrated a nonlinear relationship between SPEP and NEPH, with a linear region <19.2 g/L having a slope of 0.83 (95% CI 0.79-0.89) and a second linear region having a slope of 1.47 (95% CI 1.39-1.53) at higher concentrations. Dilutions of high-concentration IgG monoclonal proteins were linear by NEPH and nonlinear by SPEP., Conclusions: There are systematic differences in the quantification of monoclonal IgM and IgG by SPEP and NEPH. The bias in IgM is from NEPH overestimation. The nonlinearity of SPEP at high monoclonal IgG concentrations may obscure changes in plasma cell populations. Clinicians should be made aware of the biases and nonlinearity in these tests to make proper conclusions regarding treatment response.

Context: The association between acute pancreatitis and primary hyperparathyroidism (PHPT) is controversial., Objective: The aim of the study was to address the incidence and disease characteristics of acute pancreatitis in PHPT from a large inception cohort of community residents., Design and Setting: Patients with acute pancreatitis were identified in an Olmsted County, Minnesota, cohort of PHPT subjects diagnosed from 1965-2001 and compared to matched control subjects., Main Outcome Measures: The estimated rate of developing acute pancreatitis was calculated by person-years method. Cox models assessed the effect of PHPT disease status on the development of acute pancreatitis., Results: Of 684 patients with PHPT, 10 patients (1.5%) developed acute pancreatitis, compared to 32 of 1364 control patients (2.3%). The estimated rate of development of acute pancreatitis in PHPT was 114 per 100,000 person-years, compared to 140 per 100,000 person-years in control subjects (P = 0.56). The estimated hazard ratio of acute pancreatitis for PHPT relative to the control subjects was 0.84 (P = 0.89). The majority of subjects with PHPT and acute pancreatitis were women (n = 7), and the mean age was 70.6 yr at the time of acute pancreatitis. PHPT was present an average of 137 months before the development of acute pancreatitis, and contributing causes for acute pancreatitis were in identified in four patients. Maximal serum calcium levels in PHPT were not significantly associated with the development of acute pancreatitis., Conclusions: Acute pancreatitis was not increased in community patients with PHPT, and therefore, there does not appear to be a causal relationship between PHPT and acute pancreatitis.