Development of a clinical prediction model for recurrence and mortality outcomes after Clostridioides difficile infection using a machine learning approach.

OBJECTIVES: Clostridioides difficile infection (CDI) is associated with a large burden of morbidity and mortality worldwide. Previous studies have developed models for predicting recurrence and mortality following CDI, but no machine learning predictive models have been developed specifically using data from Japanese patients. METHODS: Using a database of records from acute care hospitals in Japan, we extracted records from January 2012 to September 2016 (plus a 60-day lookback window). A total of 19,159 patients were included. We used a machine learning approach, XGBoost, and compared it to a traditional unregularized logistic regression model. The first 80% of the dataset (by patient index date) was used to optimize model hyperparameters and train the final models, and evaluation was performed on the remaining 20%. We measured model performance by the area under the receiver operator curve and assessed feature importance using Shapley additive explanations. RESULTS: Performance was similar between the machine learning approach and the classical logistic regression model. Logistic regression performed slightly better than XGBoost for predicting mortality. CONCLUSION: XGBoost performed slightly better than logistic regression for predicting recurrence, but it was not competitive with existing published models. Despite this, a future machine learning-based application provided in a bedside setting at low cost might be a clinically useful tool.

An early warning system to predict and mitigate wheat rust diseases in Ethiopia.

Wheat rust diseases pose one of the greatest threats to global food security, including subsistence farmers in Ethiopia. The fungal spores transmitting wheat rust are dispersed by wind and can remain infectious after dispersal over long distances. The emergence of new strains of wheat rust has exacerbated the risks of severe crop loss. We describe the construction and deployment of a near realtime early warning system (EWS) for two major wind-dispersed diseases of wheat crops in Ethiopia that combines existing environmental research infrastructures, newly developed tools and scientific expertise across multiple organisations in Ethiopia and the UK. The EWS encompasses a sophisticated framework that integrates field and mobile phone surveillance data, spore dispersal and disease environmental suitability forecasting, as well as communication to policy-makers, advisors and smallholder farmers. The system involves daily automated data flow between two continents during the wheat season in Ethiopia. The framework utilises expertise and environmental research infrastructures from within the cross-disciplinary spectrum of biology, agronomy, meteorology, computer science and telecommunications. The EWS successfully provided timely information to assist policy makers formulate decisions about allocation of limited stock of fungicide during the 2017 and 2018 wheat seasons. Wheat rust alerts and advisories were sent by short message service and reports to 10 000 development agents and approximately 275 000 smallholder farmers in Ethiopia who rely on wheat for subsistence and livelihood security. The framework represents one of the first advanced crop disease EWSs implemented in a developing country. It provides policy-makers, extension agents and farmers with timely, actionable information on priority diseases affecting a staple food crop. The framework together with the underpinning technologies are transferable to forecast wheat rusts in other regions and can be readily adapted for other wind-dispersed pests and disease of major agricultural crops.

THE PARADOX OF EUTHANIZING KOALAS (PHASCOLARCTOS CINEREUS) TO SAVE POPULATIONS FROM ELIMINATION.

Koala (Phascolarctos cinereus) populations in the wild are in sharp decline in Australia due to deforestation, road accidents, dog attacks, and disease from infection with sexually transmitted Chlamydia spp. Severely diseased koalas that are captured are euthanized for humane reasons because antibiotics are not effective. Paradoxically, we propose that euthanizing more koalas could help to increase koala population numbers. We investigated the potential impact of systematically euthanizing diseased koalas. Using data from a well-studied koala population, and an individual-based computer simulation model, we predict that such a program would result in a larger population of koalas after 7 yr than would exist without the program. If terminally diseased and sterile koalas are euthanized and other infected captured koalas are given antibiotics, chlamydial infection could be eliminated and population growth observed after 4 yr. The practical implementation of such a program would be facilitated with further development of tools to diagnose infection and internal disease in the field.

The potential impact of vaccination on the prevalence of gonorrhea.

Gonorrhea, one of the most common sexually transmitted infections worldwide, can lead to serious sequelae, including infertility and increased HIV transmission. Recently, untreatable, multidrug-resistant Neisseria gonorrhoeae strains have been reported. In the absence of new antibiotics, and given the speed with which resistance has emerged to all previously used antibiotics, development of a vaccine would be the ideal solution to this public health emergency. Understanding the desired characteristics, target population, and expected impact of an anti-gonococcal vaccine is essential to facilitate vaccine design, assessment and implementation. The modeling presented herein aims to fill these conceptual gaps, and inform future gonococcal vaccine development. Using an individual-based, epidemiological simulation model, gonococcal prevalence was simulated in a heterosexual population of 100,000 individuals after the introduction of vaccines with varied efficacy (10-100%) and duration of protection (2.5-20 years). Model simulations predict that gonococcal prevalence could be reduced by at least 90% after 20 years, if all 13-year-olds were given a non-waning vaccine with 50% efficacy, or a vaccine with 100% efficacy that wanes after 7.5 years. A 40% reduction in prevalence is achievable with a non-waning vaccine of only 20% efficacy. We conclude that a vaccine of moderate efficacy and duration could have a substantive impact on gonococcal prevalence, and disease sequelae, if coverage is high and protection lasts over the highest risk period (i.e., most sexual partner change) among young people.

Is it time to switch to doxycycline from azithromycin for treating genital chlamydial infections in women? Modelling the impact of autoinoculation from the gastrointestinal tract to the genital tract.

BACKGROUND: Single-dose azithromycin is recommended over multi-dose doxycycline as treatment for chlamydial infection. However, even with imperfect adherence, doxycycline is more effective in treating genital and rectal infection. Recently, it has been suggested that autoinoculation from the rectum to the genitals may be a source of persistent chlamydial infection in women. We estimated the impact autoinoculation may have on azithromycin and doxycycline effectiveness. METHODS: We estimate treatment effectiveness using a simple mathematical model, incorporating data on azithromycin and doxycycline efficacy from recent meta-analyses, and data on prevalence of rectal infection in women with genital chlamydial infection. RESULTS: When the possibility of autoinoculation is taken into account, we calculate that doxycycline effectiveness may be 97% compared to just 82% for azithromycin. CONCLUSIONS: Consideration should be given to re-evaluating azithromycin as the standard treatment for genital chlamydia in women.

Target cell limitation constrains chlamydial load in persistent infections: results from mathematical modelling applied to mouse genital tract infection data.

The interactions between chlamydial pathogens and their host contribute to the outcome of infection. Nonresolving infections in immunodeficient mice can provide insights into these mechanisms by allowing observation of a form of persistent infection. Using a mathematical model, we predict that in a nonresolving infection, the number of chlamydiae in the host will attain a stable equilibrium and that this equilibrium will be independent of the inoculum size. We test this hypothesis by infecting RAG(-/-) mice with 10(4)-10(7) inclusion-forming units (IFU) of Chlamydia muridarum and comparing the IFU levels at equilibrium. There were no statistically significant differences in equilibrium IFU levels between the reference group and other inoculation groups, supporting the hypothesis. Using the mathematical model, we estimated that at equilibrium just 3% of the chlamydiae infect a target cell. We predict that the equilibrium IFU level is highly sensitive to the rate of replenishment of healthy cells. The limitation of target cells is a key driver of infection dynamics, affecting both the peak of infection and the equilibrium level of persistent infections. Target cell limitation likely plays an important role in the dynamics of human infections as well.

A 5-year Chlamydia vaccination programme could reverse disease-related koala population decline: predictions from a mathematical model using field data.

BACKGROUND: Many koala populations around Australia are in serious decline, with a substantial component of this decline in some Southeast Queensland populations attributed to the impact of Chlamydia. A Chlamydia vaccine for koalas is in development and has shown promise in early trials. This study contributes to implementation preparedness by simulating vaccination strategies designed to reverse population decline and by identifying which age and sex category it would be most effective to target. METHODS: We used field data to inform the development and parameterisation of an individual-based stochastic simulation model of a koala population endemic with Chlamydia. The model took into account transmission, morbidity and mortality caused by Chlamydia infections. We calibrated the model to characteristics of typical Southeast Queensland koala populations. As there is uncertainty about the effectiveness of the vaccine in real-world settings, a variety of potential vaccine efficacies, half-lives and dosing schedules were simulated. RESULTS: Assuming other threats remain constant, it is expected that current population declines could be reversed in around 5-6 years if female koalas aged 1-2 years are targeted, average vaccine protective efficacy is 75%, and vaccine coverage is around 10% per year. At lower vaccine efficacies the immunological effects of boosting become important: at 45% vaccine efficacy population decline is predicted to reverse in 6 years under optimistic boosting assumptions but in 9 years under pessimistic boosting assumptions. Terminating a successful vaccination programme at 5 years would lead to a rise in Chlamydia prevalence towards pre-vaccination levels. CONCLUSION: For a range of vaccine efficacy levels it is projected that population decline due to endemic Chlamydia can be reversed under realistic dosing schedules, potentially in just 5 years. However, a vaccination programme might need to continue indefinitely in order to maintain Chlamydia prevalence at a sufficiently low level for population growth to continue.

Spending of HIV resources in Asia and Eastern Europe: systematic review reveals the need to shift funding allocations towards priority populations.

INTRODUCTION: It is increasingly important to prioritize the most cost-effective HIV interventions. We sought to summarize the evidence on which types of interventions provide the best value for money in regions with concentrated HIV epidemics. METHODS: We conducted a systematic review of peer-reviewed and grey literature reporting measurements of cost-effectiveness or cost-benefit for HIV/AIDS interventions in Asia and Eastern Europe. We also collated HIV/AIDS spending assessment data from case-study countries in the region. RESULTS: We identified 91 studies for inclusion, 47 of which were from peer-reviewed journals. Generally, in concentrated settings, prevention of mother-to-child transmission programmes and prevention programmes targeting people who inject drugs and sex workers had lower incremental cost-effectiveness ratios than programmes aimed at the general population. The few studies evaluating programmes targeting men who have sex with men indicate moderate cost-effectiveness. Collation of prevention programme spending data from 12 countries in the region (none of which had generalized epidemics) indicated that resources for the general population/non-targeted was greater than 30% for eight countries and greater than 50% for five countries. CONCLUSIONS: There is a misalignment between national spending on HIV/AIDS responses and the most affected populations across the region. In concentrated epidemics, scarce funding should be directed more towards most-at-risk populations. Reaching consensus on general principles of cost-effectiveness of programmes by epidemic settings is difficult due to inconsistent evaluation approaches. Adopting a standard costing, impact evaluation, benefits calculation, analysis and reporting framework would enable cross comparisons and improve HIV resource prioritization and allocation.