One promising solution to address physician data entry needs is through the development of so-called "digital scribes," or tools which aim to automate clinical documentation via automatic speech recognition (ASR) of patient-clinician conversations. Evaluation of specialized ASR models in this domain, useful for understanding feasibility and development opportunities, has been difficult because most models have been under development. Following the commercial release of such models, we report an independent evaluation of four models, two general-purpose, and two for medical conversation with a corpus of 36 primary care conversations. We identify word error rates (WER) of 8.8%-10.5% and word-level diarization error rates (WDER) ranging from 1.8%-13.9%, which are generally lower than previous reports. The findings indicate that, while there is room for improvement, the performance of these specialized models, at least under ideal recording conditions, may be amenable to the development of downstream applications which rely on ASR of patient-clinician conversations.

We present the requirements, design, and evaluation of the cryogenic continuously rotating half-wave plate (CHWP) for the Simons Observatory (SO). SO is a cosmic microwave background polarization experiment at Parque Astronómico de Atacama in northern Chile that covers a wide range of angular scales using both small (⌀0.42 m) and large (⌀6 m) aperture telescopes. In particular, the small aperture telescopes (SATs) focus on large angular scales for primordial B-mode polarization. To this end, the SATs employ a CHWP to modulate the polarization of the incident light at 8 Hz, suppressing atmospheric 1/f noise and mitigating systematic uncertainties that would otherwise arise due to the differential response of detectors sensitive to orthogonal polarizations. The CHWP consists of a 505 mm diameter achromatic sapphire HWP and a cryogenic rotation mechanism, both of which are cooled down to ∼50 K to reduce detector thermal loading. Under normal operation, the HWP is suspended by a superconducting magnetic bearing and rotates with a constant 2 Hz frequency, controlled by an electromagnetic synchronous motor. We find that the number of superconductors and the number of magnets that make up the superconducting magnetic bearing are important design parameters, especially for the rotation mechanism's vibration performance. The rotation angle is detected through an angular encoder with a noise level of 0.07 μrad s. During a cooldown process, the rotor is held in place by a grip-and-release mechanism that serves as both an alignment device and a thermal path. In this paper, we provide an overview of the SO SAT CHWP: its requirements, hardware design, and laboratory performance.

Background

Xpert MTB/RIF, the most widely used automated nucleic acid amplification test for tuberculosis, is available in more than 130 countries. Although diagnostic accuracy is well documented, anticipated improvements in patient outcomes have not been clearly identified. We performed an individual patient data meta-analysis to examine improvements in patient outcomes associated with Xpert MTB/RIF.

Methods

We searched PubMed, Embase, ClinicalTrials.gov, and the Pan African Clinical Trials Registry from inception to Feb 1, 2018, for randomised controlled trials (RCTs) comparing the use of Xpert MTB/RIF with sputum smear microscopy as tests for tuberculosis diagnosis in adults (aged 18 years or older). We excluded studies of patients with extrapulmonary tuberculosis, and studies in which mortality was not assessed. We used a two-stage approach for our primary analysis and a one-stage approach for the sensitivity analysis. To assess the primary outcome of cumulative 6-month all-cause mortality, we first performed logistic regression models (random effects for cluster randomised trials, with robust SEs for multicentre studies) for each trial, and then pooled the odds ratio (OR) estimates by a fixed-effects (inverse variance) or random-effects (Der Simonian Laird) meta-analysis. We adjusted for age and gender, and stratified by HIV status and previous tuberculosis-treatment history. The study protocol has been registered with PROSPERO, number CRD42014013394.

Findings

Our search identified 387 studies, of which five RCTs were eligible for analysis. 8567 adult clinic attendees (4490 [63·5%] of 7074 participants for whom data were available were HIV-positive) were tested for tuberculosis with Xpert MTB/RIF (Xpert group) versus sputum smear microscopy (sputum smear group), across five low-income and middle-income countries (South Africa, Brazil, Zimbabwe, Zambia, and Tanzania). The primary outcome (reported in three studies) occurred in 182 (4·5%) of 4050 patients in the Xpert group and 217 (5·3%) of 4093 patients in the smear group (pooled adjusted OR 0·88, 95% CI 0·68-1·14 [p=0·34]; for HIV-positive individuals OR 0·83, 0·65-1·05 [p=0·12]). Kaplan-Meier estimates showed a lower rate of death (12·73 per 100 person-years in the Xpert group vs 16·38 per 100 person-years in the sputum smear group) for HIV-positive patients (hazard ratio 0·76, 95% CI 0·60-0·97; p=0·03). The risk of bias was assessed as reasonable and the statistical heterogeneity across studies was low (I²<20% for the primary outcome).

Interpretation

Despite individual patient data analysis from five RCTs, we were unable to confidently rule in nor rule out an Xpert MTB/RIF-associated reduction in mortality among outpatients tested for tuberculosis. Reduction in mortality among HIV-positive patients in a secondary analysis suggests the possibility of population-level impact.

Funding

US National Institutes of Health.

Enteropathogenic infections cause pathophysiological changes in the host but their effects beyond the gastrointestinal tract are undefined. Here, using Citrobacter rodentium infection in mouse, which mimics human diarrheal enteropathogenic Escherichia coli, we show that gastrointestinal infection negatively affects bone remodeling, leading to compromised bone architecture. Transmission of infection through fecal-oral route from Citrobacter rodentium-infected to non-infected mice caused bone loss in non-infected cage mates. Mice with B cell deficiency (Igh6−/− mice) failed to clear C. rodentium infection and exhibited more severe and long-term bone loss compared to WT mice. Unbiased cytokine profiling showed an increase in circulating tumor necrosis factor α (TNFα) levels following Citrobacter rodentium infection, and immunoneutralization of TNFα prevented infection-induced bone loss completely in WT and immunocompromised mice. These findings reveal rapid, relaying, and modifiable effects of enteropathogenic infections on an extraintestinal organ—bone, and provide insights into the mechanism(s) through which these infections affect extraintestinal organ homeostasis.

We present a measurement of the B-mode polarization power spectrum of the cosmic microwave background (CMB) using data taken from 2014 July to 2016 December with the Polarbear experiment. The CMB power spectra are measured using observations at 150 GHz with an instantaneous array sensitivity of NETarray=23μ K√s on a 670 square degree patch of sky centered at (R.A., decl.) = (+0h12m0s, -59°18′). A continuously rotating half-wave plate is used to modulate polarization and to suppress low-frequency noise. We achieve 32 μK arcmin effective polarization map noise with a knee in sensitivity of ℓ = 90, where the inflationary gravitational-wave signal is expected to peak. The measured B-mode power spectrum is consistent with a ΛCDM lensing and single dust component foreground model over a range of multipoles 50 ≤ ℓ ≤ 600. The data disfavor zero CℓBB at 2.2σ using this ℓ range of Polarbear data alone. We cross-correlate our data with Planck full mission 143, 217, and 353 GHz frequency maps and find the low-ℓ B-mode power in the combined data set to be consistent with thermal dust emission. We place an upper limit on the tensor-to-scalar ratio r < 0.90 at the 95% confidence level after marginalizing over foregrounds.

We present the first measurement of cross-correlation between the lensing potential, reconstructed from cosmic microwave background (CMB) polarization data, and the cosmic shear field from galaxy shapes. This measurement is made using data from the Polarbear CMB experiment and the Subaru Hyper Suprime-Cam (HSC) survey. By analyzing an 11 deg2 overlapping region, we reject the null hypothesis at 3.5σ and constrain the amplitude of the cross-spectrum to , where is the amplitude normalized with respect to the Planck 2018 prediction, based on the flat Λ cold dark matter cosmology. The first measurement of this cross-spectrum without relying on CMB temperature measurements is possible owing to the deep Polarbear map with a noise level of ∼6 μK arcmin, as well as the deep HSC data with a high galaxy number density of . We present a detailed study of the systematics budget to show that residual systematics in our results are negligibly small, which demonstrates the future potential of this cross-correlation technique.

Machine learning is an important applied research area in particle physics, beginning with applications to high-level physics analysis in the 1990s and 2000s, followed by an explosion of applications in particle and event identification and reconstruction in the 2010s. In this document we discuss promising future research and development areas in machine learning in particle physics with a roadmap for their implementation, software and hardware resource requirements, collaborative initiatives with the data science community, academia and industry, and training the particle physics community in data science. The main objective of the document is to connect and motivate these areas of research and development with the physics drivers of the High-Luminosity Large Hadron Collider and future neutrino experiments and identify the resource needs for their implementation. Additionally we identify areas where collaboration with external communities will be of great benefit.

Background

While vaccines have established utility against COVID-19, phase 3 efficacy studies have generally not comprehensively evaluated protection provided by previous infection or hybrid immunity (previous infection plus vaccination). Individual patient data from US government-supported harmonized vaccine trials provide an unprecedented sample population to address this issue. We characterized the protective efficacy of previous SARS-CoV-2 infection and hybrid immunity against COVID-19 early in the pandemic over three-to six-month follow-up and compared with vaccine-associated protection.

Methods

In this post-hoc cross-protocol analysis of the Moderna, AstraZeneca, Janssen, and Novavax COVID-19 vaccine clinical trials, we allocated participants into four groups based on previous-infection status at enrolment and treatment: no previous infection/placebo; previous infection/placebo; no previous infection/vaccine; and previous infection/vaccine. The main outcome was RT-PCR-confirmed COVID-19 >7-15 days (per original protocols) after final study injection. We calculated crude and adjusted efficacy measures.

Findings

Previous infection/placebo participants had a 92% decreased risk of future COVID-19 compared to no previous infection/placebo participants (overall hazard ratio [HR] ratio: 0.08; 95% CI: 0.05-0.13). Among single-dose Janssen participants, hybrid immunity conferred greater protection than vaccine alone (HR: 0.03; 95% CI: 0.01-0.10). Too few infections were observed to draw statistical inferences comparing hybrid immunity to vaccine alone for other trials. Vaccination, previous infection, and hybrid immunity all provided near-complete protection against severe disease.

Interpretation

Previous infection, any hybrid immunity, and two-dose vaccination all provided substantial protection against symptomatic and severe COVID-19 through the early Delta period. Thus, as a surrogate for natural infection, vaccination remains the safest approach to protection.

Funding

National Institutes of Health.