Detecting changes in population trends in infection surveillance using community SARS-CoV-2 prevalence as an exemplar.

Detecting and quantifying changes in growth rates of infectious diseases is vital to informing public health strategy and can inform policymakers' rationale for implementing or continuing interventions aimed at reducing impact. Substantial changes in SARS-CoV-2 prevalence with emergence of variants provides opportunity to investigate different methods to do this. We included PCR results from all participants in the UK's COVID-19 Infection Survey between August 2020-June 2022. Change-points for growth rates were identified using iterative sequential regression (ISR) and second derivatives of generalised additive models (GAMs). Consistency between methods and timeliness of detection were compared. Of 8,799,079 visits, 147,278 (1.7%) were PCR-positive. Change-points associated with emergence of major variants were estimated to occur a median 4 days earlier (IQR 0-8) in GAMs versus ISR. When estimating recent change-points using successive data periods, four change-points (4/96) identified by GAMs were not found when adding later data or by ISR. Change-points were detected 3-5 weeks after they occurred in both methods but could be detected earlier within specific subgroups. Change-points in growth rates of SARS-CoV-2 can be detected in near real-time using ISR and second derivatives of GAMs. To increase certainty about changes in epidemic trajectories both methods could be run in parallel.

Immunogenicity of COVID-19 vaccines in patients with follicular lymphoma receiving frontline chemoimmunotherapy.

Immune responses to primary COVID-19 vaccination were investigated in 58 patients with follicular lymphoma (FL) as part of the PETReA trial of frontline therapy (EudraCT 2016-004010-10). COVID-19 vaccines (BNT162b2 or ChAdOx1) were administered before, during or after cytoreductive treatment comprising rituximab (depletes B cells) and either bendamustine (depletes CD4+ T cells) or cyclophosphamide-based chemotherapy. Blood samples obtained after vaccine doses 1 and 2 (V1, V2) were analysed for antibodies and T cells reactive to the SARS-CoV-2 spike protein using the Abbott Architect and interferon-gamma ELISpot assays respectively. Compared to 149 healthy controls, patients with FL exhibited lower antibody but preserved T-cell responses. Within the FL cohort, multivariable analysis identified low pre-treatment serum IgA levels and V2 administration during induction or maintenance treatment as independent determinants of lower antibody and higher T-cell responses, and bendamustine and high/intermediate FLIPI-2 score as additional determinants of a lower antibody response. Several clinical scenarios were identified where dichotomous immune responses were estimated with >95% confidence based on combinations of predictive variables. In conclusion, the immunogenicity of COVID-19 vaccines in FL patients is influenced by multiple disease- and treatment-related factors, among which B-cell depletion showed differential effects on antibody and T-cell responses.

The novel 2024 WHO Neisseria gonorrhoeae reference strains for global quality assurance of laboratory investigations and superseded WHO N. gonorrhoeae reference strains-phenotypic, genetic and reference genome characterization.

OBJECTIVES: MDR and XDR Neisseria gonorrhoeae strains remain major public health concerns internationally, and quality-assured global gonococcal antimicrobial resistance (AMR) surveillance is imperative. The WHO global Gonococcal Antimicrobial Surveillance Programme (GASP) and WHO Enhanced GASP (EGASP), including metadata and WGS, are expanding internationally. We present the phenotypic, genetic and reference genome characteristics of the 2024 WHO gonococcal reference strains (n = 15) for quality assurance worldwide. All superseded WHO gonococcal reference strains (n = 14) were identically characterized. MATERIAL AND METHODS: The 2024 WHO reference strains include 11 of the 2016 WHO reference strains, which were further characterized, and four novel strains. The superseded WHO reference strains include 11 WHO reference strains previously unpublished. All strains were characterized phenotypically and genomically (single-molecule PacBio or Oxford Nanopore and Illumina sequencing). RESULTS: The 2024 WHO reference strains represent all available susceptible and resistant phenotypes and genotypes for antimicrobials currently and previously used (n = 22), or considered for future use (n = 3) in gonorrhoea treatment. The novel WHO strains include internationally spreading ceftriaxone resistance, ceftriaxone resistance due to new penA mutations, ceftriaxone plus high-level azithromycin resistance and azithromycin resistance due to mosaic MtrRCDE efflux pump. AMR, serogroup, prolyliminopeptidase, genetic AMR determinants, plasmid types, molecular epidemiological types and reference genome characteristics are presented for all strains. CONCLUSIONS: The 2024 WHO gonococcal reference strains are recommended for internal and external quality assurance in laboratory examinations, especially in the WHO GASP, EGASP and other GASPs, but also in phenotypic and molecular diagnostics, AMR prediction, pharmacodynamics, epidemiology, research and as complete reference genomes in WGS analysis.

Retrospective analysis of hospital electronic health records reveals unseen cases of acute hepatitis with unknown aetiology in adults in Oxfordshire.

BACKGROUND: An outbreak of acute severe hepatitis of unknown aetiology (AS-Hep-UA) in children during 2022 was subsequently linked to infections with adenovirus-associated virus 2 and other 'helper viruses', including human adenovirus. It is possible that evidence of such an outbreak could be identified at a population level based on routine data captured by electronic health records (EHR). METHODS: We used anonymised EHR to collate retrospective data for all emergency presentations to Oxford University Hospitals NHS Foundation Trust in the UK, between 2016-2022, for all ages from 18 months and older. We investigated clinical characteristics and temporal distribution of presentations of acute hepatitis and of adenovirus infections based on laboratory data and clinical coding. We relaxed the stringent case definition adopted during the AS-Hep-UA to identify all cases of acute hepatitis with unknown aetiology (termed AHUA). We compared events within the outbreak period (defined as 1st Oct 2021-31 Aug 2022) to the rest of our study period. RESULTS: Over the study period, there were 903,433 acute presentations overall, of which 391 (0.04%) were classified as AHUA. AHUA episodes had significantly higher critical care admission rates (p < 0.0001, OR = 41.7, 95% CI:26.3-65.0) and longer inpatient admissions (p < 0.0001) compared with the rest of the patient population. During the outbreak period, significantly more adults (≥ 16 years) were diagnosed with AHUA (p < 0.0001, OR = 3.01, 95% CI: 2.20-4.12), and there were significantly more human adenovirus (HadV) infections in children (p < 0.001, OR = 1.78, 95% CI:1.27-2.47). There were also more HAdV tests performed during the outbreak (p < 0.0001, OR = 1.27, 95% CI:1.17-1.37). Among 3,707 individuals who were tested for HAdV, 179 (4.8%) were positive. However, there was no evidence of more acute hepatitis or increased severity of illness in HadV-positive compared to negative cases. CONCLUSIONS: Our results highlight an increase in AHUA in adults coinciding with the period of the outbreak in children, but not linked to documented HAdV infection. Tracking changes in routinely collected clinical data through EHR could be used to support outbreak surveillance.

Target enrichment improves culture-independent detection of Neisseria gonorrhoeae and antimicrobial resistance determinants direct from clinical samples with Nanopore sequencing.

Multi-drug-resistant Neisseria gonorrhoeae infection is a significant public health risk. Rapidly detecting N. gonorrhoeae and antimicrobial-resistant (AMR) determinants by metagenomic sequencing of urine is possible, although high levels of host DNA and overgrowth of contaminating species hamper sequencing and limit N. gonorrhoeae genome coverage. We performed Nanopore sequencing of nucleic acid amplification test-positive urine samples and culture-positive urethral swabs with and without probe-based target enrichment, using a custom SureSelect panel, to investigate whether selective enrichment of N. gonorrhoeae DNA improves detection of both species and AMR determinants. Probes were designed to cover the entire N. gonorrhoeae genome, with tenfold enrichment of probes covering selected AMR determinants. Multiplexing was tested in a subset of samples. The proportion of sequence bases classified as N. gonorrhoeae increased in all samples after enrichment, from a median (IQR) of 0.05 % (0.01-0.1 %) to 76 % (42-82 %), giving a corresponding median improvement in fold genome coverage of 365 times (112-720). Over 20-fold coverage, required for robust AMR determinant detection, was achieved in 13/15(87 %) samples, compared to 2/15(13 %) without enrichment. The four samples multiplexed together also achieved >20-fold genome coverage. Coverage of AMR determinants was sufficient to predict resistance conferred by changes in chromosomal genes, where present, and genome coverage also enabled phylogenetic relationships to be reconstructed. Probe-based target enrichment can improve N. gonorrhoeae genome coverage when sequencing DNA extracts directly from urine or urethral swabs, allowing for detection of AMR determinants. Additionally, multiplexing prior to enrichment provided enough genome coverage for AMR detection and reduces the costs associated with this method.

Deep reinforcement learning for multi-class imbalanced training: applications in healthcare.

With the rapid growth of memory and computing power, datasets are becoming increasingly complex and imbalanced. This is especially severe in the context of clinical data, where there may be one rare event for many cases in the majority class. We introduce an imbalanced classification framework, based on reinforcement learning, for training extremely imbalanced data sets, and extend it for use in multi-class settings. We combine dueling and double deep Q-learning architectures, and formulate a custom reward function and episode-training procedure, specifically with the capability of handling multi-class imbalanced training. Using real-world clinical case studies, we demonstrate that our proposed framework outperforms current state-of-the-art imbalanced learning methods, achieving more fair and balanced classification, while also significantly improving the prediction of minority classes. Supplementary Information: The online version contains supplementary material available at 10.1007/s10994-023-06481-z.

No evidence of difference in mortality with amoxicillin versus co-amoxiclav for hospital treatment of community-acquired pneumonia.

OBJECTIVES: Current guidelines recommend broad-spectrum antibiotics for high-severity community-acquired pneumonia (CAP), potentially contributing to antimicrobial resistance (AMR). We aim to compare outcomes in CAP patients treated with amoxicillin (narrow-spectrum) versus co-amoxiclav (broad-spectrum), to understand if narrow-spectrum antibiotics could be used more widely. METHODS: We analysed electronic health records from adults (≥16 y) admitted to hospital with a primary diagnosis of pneumonia between 01-January-2016 and 30-September-2023 in Oxfordshire, United Kingdom. Patients receiving baseline ([-12 h,+24 h] from admission) amoxicillin or co-amoxiclav were included. The association between 30-day all-cause mortality and baseline antibiotic was examined using propensity score (PS) matching and inverse probability treatment weighting (IPTW) to address confounding by baseline characteristics and disease severity. Subgroup analyses by disease severity and sensitivity analyses with missing covariates imputed were also conducted. RESULTS: Among 16,072 admissions with a primary diagnosis of pneumonia, 9685 received either baseline amoxicillin or co-amoxiclav. There was no evidence of a difference in 30-day mortality between patients receiving initial co-amoxiclav vs. amoxicillin (PS matching: marginal odds ratio 0.97 [0.76-1.27], p = 0.61; IPTW: 1.02 [0.78-1.33], p = 0.87). Results remained similar across stratified analyses of mild, moderate, and severe pneumonia. Results were also similar with missing data imputed. There was also no evidence of an association between 30-day mortality and use of additional macrolides or additional doxycycline. CONCLUSIONS: There was no evidence of co-amoxiclav being advantageous over amoxicillin for treatment of CAP in 30-day mortality at a population-level, regardless of disease severity. Wider use of narrow-spectrum empirical treatment of moderate/severe CAP should be considered to curb potential for AMR.

Distinct patterns of vital sign and inflammatory marker responses in adults with suspected bloodstream infection.

OBJECTIVES: To identify patterns in inflammatory marker and vital sign responses in adult with suspected bloodstream infection (BSI) and define expected trends in normal recovery. METHODS: We included patients ≥16 y from Oxford University Hospitals with a blood culture taken between 1-January-2016 and 28-June-2021. We used linear and latent class mixed models to estimate trajectories in C-reactive protein (CRP), white blood count, heart rate, respiratory rate and temperature and identify CRP response subgroups. Centile charts for expected CRP responses were constructed via the lambda-mu-sigma method. RESULTS: In 88,348 suspected BSI episodes; 6908 (7.8%) were culture-positive with a probable pathogen, 4309 (4.9%) contained potential contaminants, and 77,131(87.3%) were culture-negative. CRP levels generally peaked 1-2 days after blood culture collection, with varying responses for different pathogens and infection sources (p < 0.0001). We identified five CRP trajectory subgroups: peak on day 1 (36,091; 46.3%) or 2 (4529; 5.8%), slow recovery (10,666; 13.7%), peak on day 6 (743; 1.0%), and low response (25,928; 33.3%). Centile reference charts tracking normal responses were constructed from those peaking on day 1/2. CONCLUSIONS: CRP and other infection response markers rise and recover differently depending on clinical syndrome and pathogen involved. However, centile reference charts, that account for these differences, can be used to track if patients are recovering line as expected and to help personalise infection.

A scalable federated learning solution for secondary care using low-cost microcomputing: privacy-preserving development and evaluation of a COVID-19 screening test in UK hospitals.

BACKGROUND: Multicentre training could reduce biases in medical artificial intelligence (AI); however, ethical, legal, and technical considerations can constrain the ability of hospitals to share data. Federated learning enables institutions to participate in algorithm development while retaining custody of their data but uptake in hospitals has been limited, possibly as deployment requires specialist software and technical expertise at each site. We previously developed an artificial intelligence-driven screening test for COVID-19 in emergency departments, known as CURIAL-Lab, which uses vital signs and blood tests that are routinely available within 1 h of a patient's arrival. Here we aimed to federate our COVID-19 screening test by developing an easy-to-use embedded system-which we introduce as full-stack federated learning-to train and evaluate machine learning models across four UK hospital groups without centralising patient data. METHODS: We supplied a Raspberry Pi 4 Model B preloaded with our federated learning software pipeline to four National Health Service (NHS) hospital groups in the UK: Oxford University Hospitals NHS Foundation Trust (OUH; through the locally linked research University, University of Oxford), University Hospitals Birmingham NHS Foundation Trust (UHB), Bedfordshire Hospitals NHS Foundation Trust (BH), and Portsmouth Hospitals University NHS Trust (PUH). OUH, PUH, and UHB participated in federated training, training a deep neural network and logistic regressor over 150 rounds to form and calibrate a global model to predict COVID-19 status, using clinical data from patients admitted before the pandemic (COVID-19-negative) and testing positive for COVID-19 during the first wave of the pandemic. We conducted a federated evaluation of the global model for admissions during the second wave of the pandemic at OUH, PUH, and externally at BH. For OUH and PUH, we additionally performed local fine-tuning of the global model using the sites' individual training data, forming a site-tuned model, and evaluated the resultant model for admissions during the second wave of the pandemic. This study included data collected between Dec 1, 2018, and March 1, 2021; the exact date ranges used varied by site. The primary outcome was overall model performance, measured as the area under the receiver operating characteristic curve (AUROC). Removable micro secure digital (microSD) storage was destroyed on study completion. FINDINGS: Clinical data from 130 941 patients (1772 COVID-19-positive), routinely collected across three hospital groups (OUH, PUH, and UHB), were included in federated training. The evaluation step included data from 32 986 patients (3549 COVID-19-positive) attending OUH, PUH, or BH during the second wave of the pandemic. Federated training of a global deep neural network classifier improved upon performance of models trained locally in terms of AUROC by a mean of 27·6% (SD 2·2): AUROC increased from 0·574 (95% CI 0·560-0·589) at OUH and 0·622 (0·608-0·637) at PUH using the locally trained models to 0·872 (0·862-0·882) at OUH and 0·876 (0·865-0·886) at PUH using the federated global model. Performance improvement was smaller for a logistic regression model, with a mean increase in AUROC of 13·9% (0·5%). During federated external evaluation at BH, AUROC for the global deep neural network model was 0·917 (0·893-0·942), with 89·7% sensitivity (83·6-93·6) and 76·6% specificity (73·9-79·1). Site-specific tuning of the global model did not significantly improve performance (change in AUROC <0·01). INTERPRETATION: We developed an embedded system for federated learning, using microcomputing to optimise for ease of deployment. We deployed full-stack federated learning across four UK hospital groups to develop a COVID-19 screening test without centralising patient data. Federation improved model performance, and the resultant global models were generalisable. Full-stack federated learning could enable hospitals to contribute to AI development at low cost and without specialist technical expertise at each site. FUNDING: The Wellcome Trust, University of Oxford Medical and Life Sciences Translational Fund.

Risk of SARS-CoV-2 reinfection during multiple Omicron variant waves in the UK general population.

SARS-CoV-2 reinfections increased substantially after Omicron variants emerged. Large-scale community-based comparisons across multiple Omicron waves of reinfection characteristics, risk factors, and protection afforded by previous infection and vaccination, are limited. Here we studied ~45,000 reinfections from the UK's national COVID-19 Infection Survey and quantified the risk of reinfection in multiple waves, including those driven by BA.1, BA.2, BA.4/5, and BQ.1/CH.1.1/XBB.1.5 variants. Reinfections were associated with lower viral load and lower percentages of self-reporting symptoms compared with first infections. Across multiple Omicron waves, estimated protection against reinfection was significantly higher in those previously infected with more recent than earlier variants, even at the same time from previous infection. Estimated protection against Omicron reinfections decreased over time from the most recent infection if this was the previous or penultimate variant (generally within the preceding year). Those 14-180 days after receiving their most recent vaccination had a lower risk of reinfection than those >180 days from their most recent vaccination. Reinfection risk was independently higher in those aged 30-45 years, and with either low or high viral load in their most recent previous infection. Overall, the risk of Omicron reinfection is high, but with lower severity than first infections; both viral evolution and waning immunity are independently associated with reinfection.

Nanopore sequencing of influenza A and B in Oxfordshire and the United Kingdom, 2022-23.

OBJECTIVES: We evaluated Nanopore sequencing for influenza surveillance. METHODS: Influenza A and B PCR-positive samples from hospital patients in Oxfordshire, UK, and a UK-wide population survey from winter 2022-23 underwent Nanopore sequencing following targeted rt-PCR amplification. RESULTS: From 941 infections, successful sequencing was achieved in 292/388 (75 %) available Oxfordshire samples: 231 (79 %) A/H3N2, 53 (18 %) A/H1N1, and 8 (3 %) B/Victoria and in 53/113 (47 %) UK-wide samples. Sequencing was more successful at lower Ct values. Most same-sample replicate sequences had identical haemagglutinin segments (124/141, 88 %); 36/39 (92 %) Illumina vs. Nanopore comparisons were identical, and 3 (8 %) differed by 1 variant. Comparison of Oxfordshire and UK-wide sequences showed frequent inter-regional transmission. Infections were closely-related to 2022-23 vaccine strains. Only one sample had a neuraminidase inhibitor resistance mutation. 849/941 (90 %) Oxfordshire infections were community-acquired. 63/88 (72 %) potentially healthcare-associated cases shared a hospital ward with ≥ 1 known infectious case. 33 epidemiologically-plausible transmission links had sequencing data for both source and recipient: 8 were within ≤ 5 SNPs, of these, 5 (63 %) involved potential sources that were also hospital-acquired. CONCLUSIONS: Nanopore influenza sequencing was reproducible and antiviral resistance rare. Inter-regional transmission was common; most infections were genomically similar. Hospital-acquired infections are likely an important source of nosocomial transmission and should be prioritised for infection prevention and control.

Improving the representativeness of UK's national COVID-19 Infection Survey through spatio-temporal regression and post-stratification.

Population-representative estimates of SARS-CoV-2 infection prevalence and antibody levels in specific geographic areas at different time points are needed to optimise policy responses. However, even population-wide surveys are potentially impacted by biases arising from differences in participation rates across key groups. Here, we used spatio-temporal regression and post-stratification models to UK's national COVID-19 Infection Survey (CIS) to obtain representative estimates of PCR positivity (6,496,052 tests) and antibody prevalence (1,941,333 tests) for different regions, ages and ethnicities (7-December-2020 to 4-May-2022). Not accounting for vaccination status through post-stratification led to small underestimation of PCR positivity, but more substantial overestimations of antibody levels in the population (up to 21 percentage points), particularly in groups with low vaccine uptake in the general population. There was marked variation in the relative contribution of different areas and age-groups to each wave. Future analyses of infectious disease surveys should take into account major drivers of outcomes of interest that may also influence participation, with vaccination being an important factor to consider.

Changes in the investigation and management of suspected myocardial infarction and injury during COVID-19: a multi-centre study using routinely collected healthcare data.

Objective: The COVID-19 pandemic was associated with a reduction in the incidence of myocardial infarction (MI) diagnosis, in part because patients were less likely to present to hospital. Whether changes in clinical decision making with respect to the investigation and management of patients with suspected MI also contributed to this phenomenon is unknown. Methods: Multicentre retrospective cohort study in three UK centres contributing data to the National Institute for Health Research Health Informatics Collaborative. Patients presenting to the Emergency Department (ED) of these centres between 1st January 2020 and 1st September 2020 were included. Three time epochs within this period were defined based on the course of the first wave of the COVID-19 pandemic: pre-pandemic (epoch 1), lockdown (epoch 2), post-lockdown (epoch 3). Results: During the study period, 10,670 unique patients attended the ED with chest pain or dyspnoea, of whom 6,928 were admitted. Despite fewer total ED attendances in epoch 2, patient presentations with dyspnoea were increased (p < 0.001), with greater likelihood of troponin testing in both chest pain (p = 0.001) and dyspnoea (p < 0.001). There was a dramatic reduction in elective and emergency cardiac procedures (both p < 0.001), and greater overall mortality of patients (p < 0.001), compared to the pre-pandemic period. Positive COVID-19 and/or troponin test results were associated with increased mortality (p < 0.001), though the temporal risk profile differed. Conclusions: The first wave of the COVID-19 pandemic was associated with significant changes not just in presentation, but also the investigation, management, and outcomes of patients presenting with suspected myocardial injury or MI.

Interpretable machine learning-based decision support for prediction of antibiotic resistance for complicated urinary tract infections.

Urinary tract infections are one of the most common bacterial infections worldwide; however, increasing antimicrobial resistance in bacterial pathogens is making it challenging for clinicians to correctly prescribe patients appropriate antibiotics. In this study, we present four interpretable machine learning-based decision support algorithms for predicting antimicrobial resistance. Using electronic health record data from a large cohort of patients diagnosed with potentially complicated UTIs, we demonstrate high predictability of antibiotic resistance across four antibiotics - nitrofurantoin, co-trimoxazole, ciprofloxacin, and levofloxacin. We additionally demonstrate the generalizability of our methods on a separate cohort of patients with uncomplicated UTIs, demonstrating that machine learning-driven approaches can help alleviate the potential of administering non-susceptible treatments, facilitate rapid effective clinical interventions, and enable personalized treatment suggestions. Additionally, these techniques present the benefit of providing model interpretability, explaining the basis for generated predictions.