Managing extreme AI risks amid rapid progress

Preparation requires technical research and development, as well as adaptive, proactive governance.

Alcohol-related emergency department presentations and hospital admissions around the time of minimum unit pricing in Ireland

BACKGROUND

Minimum unit pricing (MUP) was recently introduced in Ireland to reduce alcohol-related harms. The size of the impact of alcohol on hospital emergency departments (EDs) in Ireland is poorly understood due to inconsistent alcohol screening and documentation.

AIMS

We sought to systematically characterise the volume, timing, and nature of alcohol-related presentations and admissions to a busy urban ED in Dublin, Ireland.

METHOD

Patients presenting to the ED were assessed by a dedicated clinician during selected time periods before (Nov-Dec 2021) and after (Feb-Apr 2022) the introduction of MUP. A total of 725 interviews were conducted over 168 h in the ED.

FINDINGS

Alcohol consumption was a factor in 19.4% of ED presentations and in 17.3% of hospital admissions across the entire study period. A reduction in overall alcohol-related ED presentations was noted in the period following MUP, although it is not possible to conclude a direct effect.

CONCLUSION

Alcohol-related harm places a significant strain on EDs and hospitals, and the impact of MUP on hospital burden in Ireland merits further evaluation. Effective measures at local and population levels are urgently required to address this burden.

Proactive Contact Tracing

The COVID-19 pandemic has spurred an unprecedented demand for interventions that can reduce disease spread without excessively restricting daily activity, given negative impacts on mental health and economic outcomes. Digital contact tracing (DCT) apps have emerged as a component of the epidemic management toolkit. Existing DCT apps typically recommend quarantine to all digitally-recorded contacts of test-confirmed cases. Over-reliance on testing may, however, impede the effectiveness of such apps, since by the time cases are confirmed through testing, onward transmissions are likely to have occurred. Furthermore, most cases are infectious over a short period; only a subset of their contacts are likely to become infected. These apps do not fully utilize data sources to base their predictions of transmission risk during an encounter, leading to recommendations of quarantine to many uninfected people and associated slowdowns in economic activity. This phenomenon, commonly termed as "pingdemic," may additionally contribute to reduced compliance to public health measures. In this work, we propose a novel DCT framework, Proactive Contact Tracing (PCT), which uses multiple sources of information (e.g. self-reported symptoms, received messages from contacts) to estimate app users' infectiousness histories and provide behavioral recommendations. PCT methods are by design proactive, predicting spread before it occurs. We present an interpretable instance of this framework, the Rule-based PCT algorithm, designed via a multi-disciplinary collaboration among epidemiologists, computer scientists, and behavior experts. Finally, we develop an agent-based model that allows us to compare different DCT methods and evaluate their performance in negotiating the trade-off between epidemic control and restricting population mobility. Performing extensive sensitivity analysis across user behavior, public health policy, and virological parameters, we compare Rule-based PCT to i) binary contact tracing (BCT), which exclusively relies on test results and recommends a fixed-duration quarantine, and ii) household quarantine (HQ). Our results suggest that both BCT and Rule-based PCT improve upon HQ, however, Rule-based PCT is more efficient at controlling spread of disease than BCT across a range of scenarios. In terms of cost-effectiveness, we show that Rule-based PCT pareto-dominates BCT, as demonstrated by a decrease in Disability Adjusted Life Years, as well as Temporary Productivity Loss. Overall, we find that Rule-based PCT outperforms existing approaches across a varying range of parameters. By leveraging anonymized infectiousness estimates received from digitally-recorded contacts, PCT is able to notify potentially infected users earlier than BCT methods and prevent onward transmissions. Our results suggest that PCT-based applications could be a useful tool in managing future epidemics.

Self-reported beta-lactam allergy in government and private hospitals in Cape Town, South Africa

BACKGROUND

Up to a quarter of inpatients in high-income countries (HICs) self-report beta-lactam allergy (BLA), which if incorrect,increases the use of alternative antibiotics, worsening individual health outcomes and driving bacterial resistance. In HICs, up to 95% ofself-reported BLAs are incorrect. The epidemiology of BLA in low- and middle-income African countries is unknown.

OBJECTIVES

To describe the epidemiology and de-labelling outcomes of self-reported BLA in hospitalised South African (SA) patients.

METHODS

Point-prevalence surveys were conducted at seven hospitals (adult, paediatric, government and privately funded, district andtertiary level) in Cape Town, SA, between April 2019 and June 2021. Ward prescription records and in-person interviews were conductedto identify and risk-stratify BLA patients using the validated PEN-FAST tool. De-labelling was attempted at the tertiary allergy clinic atGroote Schuur Hospital.

RESULTS

A total of 1 486 hospital inpatients were surveyed (1 166 adults and 320 children). Only 48 patients (3.2%) self-reported a BLA,with a higher rate in private than in government-funded hospitals (6.3% v. 2.8%; p=0.014). Using the PEN-FAST tool, only 10.4% (n=5/48)of self-reported BLA patients were classified as high risk for true penicillin hypersensitivity. Antibiotics were prescribed to 70.8% (n=34/48)of self-reported BLA patients, with 64.7% (n=22/34) receiving a beta-lactam. Despite three attempts to contact patients for de-labelling atthe allergy clinic, only 3/36 underwent in vivo testing, with no positive results, and 1 patient proceeded to a negative oral challenge.

CONCLUSION

Unlike HICs, self-reported BLA is low among inpatients in SA. The majority of those who self-reported BLA were low risk fortype 1 hypersensitivity, but outpatient de-labelling efforts were largely unsuccessful.

Filling gaps in trustworthy development of AI

[Figure: see text].

Transmural electrophysiological heterogeneities in action potential duration increase the upper limit of vulnerability

Transmural dispersion in action potential duration (APD) has been shown to contribute to arrhythmia induction in the heart. However, its role in termination of lethal arrhythmias by defibrillation shocks has never been examined. The goal of this study is to investigate how transmural dispersion in APD affects cardiac vulnerability to electric shocks, in an attempt to better understand the mechanisms behind defibrillation failure. This study used a three- dimensional, geometrically accurate finite element bidomain rabbit ventricular model. Transmural heterogeneities in ionic currents were incorporated based on experimental data to generate the transmural APD profile recorded in adult rabbits during pacing. Results show that the incorporation of transmural APD heterogeneities in the model causes an increase in the upper limit of vulnerability from 26.7 V/cm in the homogeneous APD ventricles to 30.5 V/cm in the ventricles with heterogeneous transmural APD profile. Examination of shock-end virtual electrode polarisation and postshock electrical activity reveals that the higher ULV in the heterogeneous model is caused by increased dispersion in postshock repolarisation within the LV wall, which increases the likelihood of the establishment of intramural re-entrant circuits.