Search results
Found 41196 matches for
The Nuffield Department of Medicine (NDM) at the University of Oxford has a global reach and significant breadth in terms of capabilities and capacity.
Transformers and large language models are efficient feature extractors for electronic health record studies
Abstract Background Free-text data is abundant in electronic health records, but challenges in accurate and scalable information extraction mean less specific clinical codes are often used instead. Methods We evaluated the efficacy of feature extraction using modern natural language processing methods (NLP) and large language models (LLMs) on 938,150 hospital antibiotic prescriptions from Oxfordshire, UK. Specifically, we investigated inferring the type(s) of infection from a free-text “indication” field, where clinicians state the reason for prescribing antibiotics. Clinical researchers labelled a subset of the 4000 most frequent unique indications (representing 692,310 prescriptions) into 11 categories describing the infection source or clinical syndrome. Various models were then trained to determine the binary presence/absence of these infection types and also any uncertainty expressed by clinicians. Results We show on separate internal (n = 2000 prescriptions) and external test datasets (n = 2000 prescriptions), a fine-tuned domain-specific Bio+Clinical BERT model performs best across the 11 categories (average F1 score 0.97 and 0.98 respectively) and outperforms traditional regular expression (F1 = 0.71 and 0.74) and n-grams/XGBoost (F1 = 0.86 and 0.84) models. A zero-shot OpenAI GPT4 model matches the performance of traditional NLP models without the need for labelled training data (F1 = 0.71 and 0.86) and a fine-tuned GPT3.5 model achieves similar performance to the fine-tuned BERT-based model (F1 = 0.95 and 0.97). Infection sources obtained from free-text indications reveal specific infection sources 31% more often than ICD-10 codes. Conclusions Modern transformer-based models have the potential to be used widely throughout medicine to extract information from structured free-text records, to facilitate better research and patient care.
Tezepelumab can Restore Normal Lung Function in Patients with Severe, Uncontrolled Asthma: Pooled Results from the PATHWAY and NAVIGATOR Studies.
IntroductionThis post hoc analysis assessed the ability of tezepelumab treatment to restore normal lung function in patients with severe, uncontrolled asthma with abnormal lung function at baseline pooled from the PATHWAY and NAVIGATOR studies.MethodsPATHWAY and NAVIGATOR were multicentre, randomized, double-blind, placebo-controlled studies. Patients (12-80 years old) included in this analysis received tezepelumab 210 mg subcutaneously every 4 weeks or matched placebo for 52 weeks. Patients had a percent predicted pre-bronchodilator (BD) forced expiratory volume in 1 s (FEV1) of 1 was assessed by baseline percent predicted pre-BD FEV1 subgroup [abnormal (ResultsOf the 665 and 669 patients who received tezepelumab or placebo, respectively, 564 and 569 had abnormal lung function at baseline. Tezepelumab improved the pre-BD FEV1 from baseline to week 52 versus placebo by 0.14 L [95% confidence interval (CI) 0.09-0.19] and 0.13 L (95% CI 0.01-0.24) in patients with abnormal and normal lung function at baseline, respectively. A higher proportion of tezepelumab than placebo recipients with abnormal lung function at baseline achieved normal lung function at week 52 (17.2% vs. 9.9%, respectively). Among tezepelumab recipients, those with higher levels of type 2 inflammatory biomarkers and a shorter duration of disease at baseline were more likely to achieve normal lung function at week 52.ConclusionIn patients with severe, uncontrolled asthma, a greater proportion of tezepelumab than placebo recipients with abnormal lung function at baseline achieved normal lung function at week 52.Trial registrationPATHWAY: NCT02054130; NAVIGATOR: NCT03347279.
Treatable Traits in Patients with Obstructive Lung Diseases in a Well-Established Asthma/COPD Service for Primary Care.
PurposeThe primary objective of this study was to assess the prevalence of treatable traits (TTs) in patients with obstructive lung diseases in a primary care setting and how these TTs co-occur. The secondary objective was to assess the stability of TTs and the effect of management advice on changes in traits and health outcomes.Patients and methodsData from the Dutch asthma/COPD service (2007-2023) were studied retrospectively. Patients ≥18 years with asthma, COPD, or Asthma-COPD overlap (ACO) were included. The prevalence of eight TTs were assessed: 1) insufficient inhaler technique, 2) poor medication adherence, 3) blood eosinophilia, 4) smoking, 5) obesity, 6) physical inactivity, 7) reversible airflow limitation, and 8) anxiety and/or depression. The effect of management advice on TTs was evaluated for patients with a follow-up visit scheduled within 1-2 years.ResultsIn total, 15246 patients (COPD n=4822; ACO n=1761, asthma n=8663) were included. The highest proportions of TTs were insufficient inhaler technique: 43.6% (95% CI: 42.9-44.4), followed by poor medication adherence: 40.3% (95% CI: 39.2-41.4) and blood eosinophilia: 36.9% (95% CI: 35.8-38.1). Overall, 83.3% of patients had ≥ 1 TTs, and 48.9% of patients ≥ 2 TTs. Among patients with blood eosinophilia, a significant reduction of the trait at follow-up (OR: 0.61, 95% CI: 0.39; 0.96) and improved health status were observed when the pulmonologist advised the general practitioner to initiate or increase the dose of ICS. No significant association was found between management advice and the exacerbation rate at follow-up.ConclusionThe TTs assessed in this study are common in primary care patients, with nearly half of the patients showing a combination of at least two TTs. These TTs coexist in many different combinations. A personalized approach targeting these traits may be effective in achieving better control of these heterogeneous diseases.
The latency time of SARS-CoV- 2 Delta variant in infection- and vaccine-naive individuals from Vietnam
Abstract Background The latency time (from infection to infectiousness) guides the choice of measures required to control an infectious disease. Estimates of the SARS-CoV- 2 latency time are sparse due to lack of appropriate and representative data. Infection time is rarely known exactly and exposure information may be subject to several biases. Information on the endpoint requires repeated testing. Moreover, estimation is challenging because both the starting point and endpoint are typically interval censored and data may be subject to length-biased sampling (truncation). Methods We collected detailed information on exposure from public health reports produced during an outbreak with the SARS-CoV- 2 Delta variant in Ho Chi Minh City, Vietnam, in May-July 2021. Using a custom digital form and application facilitated reliable choices on exposure window. This comprehensive data set on exposure and test results from 1951 individuals, collected in the absence of large-scale vaccination or earlier infection, is the first of its kind outside of China. We accounted for the doubly interval censored nature of the observations and went beyond the standard assumption of a constant infection risk over calendar time (exponential growth) and allowed for flexibility regarding the latency time (generalized gamma distribution). We addressed right truncation due to a cutoff in data collection and a finite quarantine length. Employing a Bayesian approach, using the program , made the analyses relatively straightforward. Results Assuming exponential growth, our estimate of SARS-CoV- 2 Delta variant’s mean latency time was 3.22 (95% Credible Interval 2.89 - 3.55) days; the median was 1.81 (95% CrI 1.44- 2.16) days; the 95 th percentile was 10.98 (95% CrI 9.91 - 12.41) days. These values were much larger if a uniform infection risk was assumed. Conclusions Using a Bayesian approach with the program, we were able to estimate the SARS-CoV- 2 latency time distribution of the Delta variant in infection-naive and vaccine-naive individuals. Estimates were sensitive to the assumptions made regarding the risk of infection within the exposure window. Compared to earlier studies, the median latency time was shorter, while the 95 th percentile was larger. Our results stress the importance of thoughtful data collection and analysis for evidence-based control of an infectious disease.
Peptide-specific natural killer cell receptors.
Class I and II human leukocyte antigens (HLA-I and HLA-II) present peptide antigens for immunosurveillance by T cells. HLA molecules also form ligands for a plethora of innate, germline-encoded receptors. Many of these receptors engage HLA molecules in a peptide sequence independent manner, with binding sites outside the peptide binding groove. However, some receptors, typically expressed on natural killer (NK) cells, engage the HLA presented peptide directly. Remarkably, some of these receptors display exquisite specificity for peptide sequences, with the capacity to detect sequences conserved in pathogens. Here, we review evidence for peptide-specific NK cell receptors (PSNKRs) and discuss their potential roles in immunity.
Hepatitis B virus resistance to nucleos(t)ide analogue therapy: WHO consultation on questions, challenges, and a roadmap for the field.
In this Review, we summarise outputs from a multidisciplinary consultation convened by WHO between July 11 and 13, 2023, to discuss hepatitis B virus (HBV) drug resistance (HBVDR). Treatment of chronic HBV infection with highly effective nucleos(t)ide analogue agents, tenofovir and entecavir, is a crucial intervention that supports the global goal of elimination of HBV infection as a public health threat. The risk of HBVDR as a threat to treatment outcomes is currently considered low from a public health perspective; however, drug resistance can influence individual outcomes, particularly among those who are treatment-experienced. We highlight the need to develop appropriate prevention, monitoring, and surveillance approaches for HBVDR, to support investment in the global scale-up of HBV diagnosis and treatment. Recommendations for the HBVDR field will ultimately be incorporated into a WHO integrated Global Action Plan for drug-resistant HIV, viral hepatitis, and priority sexually transmitted infections.
Risk analysis for outpatient experimental infection as a pathway for affordable RSV vaccine development.
Controlled human infection models (CHIMs) are an important tool for accelerating clinical development of vaccines. CHIM costs are driven by quarantine facilities but may be reduced by performing CHIM in the outpatient setting. Furthermore, outpatient CHIMs offer benefits beyond costs, such as a participant-friendly approach and increased real-world aspect. We analyze safety, logistic and ethical risks of respiratory syncytial virus (RSV) CHIM in the outpatient setting. A review of the literature identified outpatient CHIMs involving respiratory pathogens. RSV transmission risk was assessed using data from our inpatient and outpatient RSV CHIMs (EudraCT 020-004137-21). Fifty-nine outpatient CHIMs using RSV, Streptococcus pneumoniae, rhinovirus, and an ongoing Bordetella Pertussis outpatient CHIM were included. One transmission event was recorded. In an inpatient RSV CHIM, standard droplet and isolation measures were sufficient to limit RSV transmission and no symptomatic third-party transmission was measured in the first outpatient RSV CHIM. Logistic and ethical advantages support outpatient CHIM adoption. We propose a framework for outpatient RSV CHIM with risk mitigation strategies to enhance affordable vaccine development.
Autoimmunity in inflammatory bowel disease: a holobiont perspective
Adaptive immunity towards self-antigens (autoimmunity) and intestinal commensal microbiota is a key feature of inflammatory bowel disease (IBD). Considering mucosal adaptive immunity from a holobiont perspective, where the host and its microbiome form a single physiological unit, emphasises the challenge of avoiding damaging responses to self-antigen and symbiotic microbial communities in the gut while protecting against potential pathogens. Intestinal tolerance mechanisms prevent maladaptive T and B cell responses to microbial, environmental, and self-antigens, which drive inflammation. We discuss the spectrum of antimicrobial and autoantibody responses and highlight mechanisms by which common IBD-associated adaptive immune responses contribute to disease.
Mapping TB incidence across districts in Uganda to inform health program activities
<sec><title>BACKGROUND</title>Identifying spatial variation in TB burden can help national TB programs effectively allocate resources to reach and treat all people with TB. However, data limitations pose challenges for subnational TB burden estimation.</sec><sec><title>METHODS</title>We developed a small-area modeling approach using geo-positioned prevalence survey data, case notifications, and geospatial covariates to simultaneously estimate spatial variation in TB incidence and case notification completeness across districts in Uganda from 2016–2019. TB incidence was estimated using 1) cluster-level data from the national 2014–2015 TB prevalence survey transformed to incidence, and 2) case notifications adjusted for geospatial covariates of health system access. The case notification completeness surface was fit jointly using observed case notifications and estimated incidence.</sec><sec><title>RESULTS</title>Estimated pulmonary TB incidence among adults varied >10-fold across Ugandan districts in 2019. Case detection increased nationwide from 2016 to 2019, and the number of districts with case detection rates >70% quadrupled. District-level estimates of TB incidence were five times more precise than a model using TB prevalence survey data alone.</sec><sec><title>CONCLUSION</title>A joint spatial modeling approach provides useful insights for TB program operation, outlining areas where TB incidence estimates are highest and health programs should concentrate their efforts. This approach can be applied in many countries with high TB burden.</sec>