Bayesian emulation and history matching of JUNE

We analyze JUNE: a detailed model of COVID-19 transmission with high spatial and demographic resolution, developed as part of the RAMP initiative. JUNE requires substantial computational resources to evaluate, making model calibration and general uncertainty analysis extremely challenging. We describe and employ the uncertainty quantification approaches of Bayes linear emulation and history matching to mimic JUNE and to perform a global parameter search, hence identifying regions of parameter space that produce acceptable matches to observed data, and demonstrating the capability of such methods. This article is part of the theme issue 'Technical challenges of modelling real-life epidemics and examples of overcoming these'.

Bayesian emulation and history matching of JUNE

We analyze JUNE: a detailed model of COVID-19 transmission with high spatial and demographic resolution, developed as part of the RAMP initiative. JUNE requires substantial computational resources to evaluate, making model calibration and general uncertainty analysis extremely challenging. We describe and employ the uncertainty quantification approaches of Bayes linear emulation and history matching to mimic JUNE and to perform a global parameter search, hence identifying regions of parameter space that produce acceptable matches to observed data, and demonstrating the capability of such methods. This article is part of the theme issue 'Technical challenges of modelling real-life epidemics and examples of overcoming these'.

Outbreak of SARS-CoV-2 Delta variant on a single liquified natural gas (LNG) vessel, with estimates of vaccine effectiveness

In August 2021, there was an outbreak of severe acute respiratory syndrome corona virus-2 (SARS-CoV-2) Delta variant on an international liquified natural gas (LNG) vessel offshore to Gladstone, Queensland. Fourteen of the 26 crew members aboard the vessel tested positive for SARS-COV-2 on PCR during the outbreak. Vaccine effectiveness (VE) was 52% for all lab-confirmed SARS-CoV-2 infections, 65% for symptomatic SARS-CoV-2 infection and 100% for severe SARS-CoV-2. The attack rate (AR) of SARS-CoV-2 Delta variant was 54% (14/26). With heightened public health measures and infection control practices, we were able to declare the outbreak over in 26 days.

P-280 Potential for improvement and current limitations of Artificial Intelligence (AI) for embryo selection: analysis of external validation data

Study question

What are the prospects of improvement and the limitations of an AI system for embryo selection?

Summary answer

The predictive performance of AI can be enhanced by including additional factors, on top of embryo images, and by assessing images with centered blastocysts.

What is known already

We previously reported the external validation of IVFvision.ai, an AI algorithm that differentiates between Day-5 blastocysts with a positive or negative implantation outcome. IVFvision.ai had higher AUC and overall accuracy in predicting implantation compared to KIDScoreD5 and senior embryologists. Here we report a secondary analysis of external validation data, focusing on a) the improvement of the predictive ability of IVFvision.ai by incorporating data from additional sources, and b) the impact of the blastocyst image quality on the performance of IVFvision.ai.

Study design, size, duration

This is a secondary analysis of external validation data. External validation of IVFvision.ai was performed at a University IVF Clinic using 113 anonymised Embryoscope images of single D5 blastocyst transfers with known implantation outcome.

Participants/materials, setting, methods

The performance of IVFvision.ai and three senior Embryologists to correctly classify blastocysts according to implantation outcome were compared in images in which the whole blastocyst was visible (centred blastocysts, n = 62) vs images in which part of the blastocyst was not visible (off-centred blastocysts, n = 51). Logistic regression models were created: a) IVFvision alone, b) IVFvision+age, c) IVFvision+fertilisation_method, d) IVFvision+KIDScoreD5, e) IVFvision+age+Fertilisation_method+KIDScoreD5. The AUC of each model in predicting implantation was estimated using ROC curve analysis.

Main results and the role of chance

The AUC of IVFVision.ai (0.675 vs 0.432), Embryologist 1 (0.570 vs 0.390), Embryologist 2 (0.663 vs 0.448) and Embryologist 3 (0.628 vs 0.485) were higher for images with centered blastocysts compared to non-centered blastocysts, respectively. There was a progressive increase of AUC with the addition of more factors in the predictive models. a) IVFvision alone: AUC=0.675, b) IVFvision+age: AUC=0.675 c) IVFvision+KIDScoreD5: AUC=0.721 d) IVFvision+fertilisation_method=0.740, e) IVFvision+age+Fertilisation_method+KIDScoreD5=0.768.

Limitations, reasons for caution

The retrospective nature of the study and the small sample of the study raise the need for further prospective studies with a larger number of embryos.

Wider implications of the findings

The highest performance of IVFvision.ai is achieved in images with centred blastocysts, suggesting that implantation cannot be predicted accurately in images with non-centred blastocysts. In addition, we provide provide proof of concept that training AI systems using data from different sources, in addition to embryo images, may increase overall accuracy.

Trial registration number

not applicable

Protocol on establishing a prospective enhanced surveillance of vaccine preventable diseases in residential aged care facilities in Central Queensland, Australia: an observational study

INTRODUCTION

Infectious diseases are a major cause of mortality and morbidity among the highly vulnerable occupants of residential aged care facilities (RACFs). The burden of vaccine preventable diseases (VPDs) among RACFs residents is mostly unknown and there is a lack of quality data from population-based prospective VPD surveillance in RACFs. The increasing burden of emerging and existing VPDs (eg, COVID-19, influenza, pneumococcal, pertussis and varicella-zoster) necessitates the establishment of an active enhanced surveillance system to provide real-time evidence to devise strategies to reduce the burden of VPDs in RACFs.

METHOD AND ANALYSIS

This study proposes a prospective active enhanced surveillance that will be implemented in RACFs across the Central Queensland (CQ) region. The study aims to measure the burden, identify aetiologies, risk factors, predictors of severe outcomes (eg, hospitalisations, mortality) and impact of the existing National Immunization Program (NIP) funded vaccines in preventing VPDs in this vulnerable population. CQ Public Health Unit (CQPHU) will implement the active surveillance by collecting demographic, clinical, pathological, diagnostic, therapeutic and clinical outcome data from the RACFs based on predefined selection criteria and case report forms as per routine public health practices. Descriptive statistics, univariate and multivariate regression analysis will be conducted to identify the predictors of morbidity and clinical outcomes following infection.

ETHICS AND DISSEMINATION

The study has been approved by the CQHHS Human Research Ethics Committee (HREC) (reference number HREC/2021/QCQ/74305). This study involves data that is routinely collected as part of the surveillance of notifiable conditions under the Public Health Act 2005. The CQHHS HREC approved a request to waive consent requirements of study participants as researchers will be provided non-identifiable data. The findings from the study will be actively disseminated through publication in peer-reviewed journals, conference presentations, social and print media, federal, state, and local authorities to reflect on the results that may facilitate revision of policy and highlight the stakeholders, funding bodies both locally and internationally.

Social determinants of health influence disease activity and functional disability in Polyarticular Juvenile Idiopathic Arthritis

BACKGROUND

Social determinants of health (SDH) greatly influence outcomes during the first year of treatment in rheumatoid arthritis, a disease similar to polyarticular juvenile idiopathic arthritis (pJIA). We investigated the correlation of community poverty level and other SDH with the persistence of moderate to severe disease activity and functional disability over the first year of treatment in pJIA patients enrolled in the Childhood Arthritis and Rheumatology Research Alliance Registry.

METHODS

In this cohort study, unadjusted and adjusted generalized linear mixed effects models analyzed the effect of community poverty and other SDH on disease activity, using the clinical Juvenile Arthritis Disease Activity Score-10, and disability, using the Child Health Assessment Questionnaire, measured at baseline, 6, and 12 months.

RESULTS

One thousand six hundred eighty-four patients were identified. High community poverty (≥20% living below the federal poverty level) was associated with increased odds of functional disability (OR 1.82, 95% CI 1.28-2.60) but was not statistically significant after adjustment (aOR 1.23, 95% CI 0.81-1.86) and was not associated with increased disease activity. Non-white race/ethnicity was associated with higher disease activity (aOR 2.48, 95% CI: 1.41-4.36). Lower self-reported household income was associated with higher disease activity and persistent functional disability. Public insurance (aOR 1.56, 95% CI 1.06-2.29) and low family education (aOR 1.89, 95% CI 1.14-3.12) was associated with persistent functional disability.

CONCLUSION

High community poverty level was associated with persistent functional disability in unadjusted analysis but not with persistent moderate to high disease activity. Race/ethnicity and other SDH were associated with persistent disease activity and functional disability.

Epidemiology of mumps outbreaks and the impact of an additional dose of MMR vaccine for outbreak control in regional Queensland, Australia, 2017-2018

Background

In recent years, there have been ongoing outbreaks of mumps reported in Northern and North-Western Queensland, Western Australia and the Northern Territory, Australia. We aimed to define the epidemiology of mumps outbreaks in Central Queensland, Australia between October 2017 and October 2018 and evaluate the effectiveness of an additional dose of measles, mumps, rubella (MMR) vaccine.

Methods

A retrospective case control study was conducted, including outbreak investigations with laboratory-confirmed cases of mumps and subsequent comparison with matched controls. We analysed mandatory notifications from the Queensland Health Notifiable Conditions System database and immunisation information from the Queensland Health Vaccination Information and Admin System (VIVAS) and the Australian Immunisation Register.

Results

Between October 2017 and October 2018, there were 93 cases of mumps reported in Central Queensland with three distinct outbreaks: a discrete Indigenous community; a correctional facility; and a boarding school. Among all cases, 74 (79.6%) were fully vaccinated and 14 (15.1%) were partially vaccinated with MMR vaccine. Eighty-six cases (92.5%) were reported among Aboriginal and Torres Strait Islander people. In all outbreaks, an additional dose of MMR vaccine was offered with 35.4%, 73.6% and 35.8% of the target population being immunised in the discrete Indigenous community, the correctional facility and the boarding school, respectively. Prior to this additional dose of MMR, the mumps attack rate was 31.0 (95% confidence interval [95% CI]: 24.2-39.0) per 1000 population, compared to the post-additional dose MMR attack rate of 10.6 (95% CI: 6.7-15.9) per 1000 population.

Conclusion

An additional or booster dose of MMR should be included as an effective public health intervention strategy, particularly in communal or high-density living conditions to control mumps outbreaks in highly vaccinated populations.

Oral vaccination of dogs with a monovalent live-avirulent vaccine confers 1 year of immunity against Bordetella bronchiseptica challenge

This experimental challenge study assessed immune protection 1 year after a single dose of live-attenuated oral Bordetella bronchiseptica (Bb) vaccine in dogs. Forty Bb-seronegative 7-9-week-old puppies were randomly assigned at Day 0 to receive a single oral dose of either Bb vaccine (n = 20; vaccinated group) or sterile water (n = 20; control group). Groups were housed separately until comingling 1 day pre-challenge (Day 365). Challenge with virulent aerosolized Bb occurred at Day 366. Clinical scores were obtained at Days 1-7, and 366-380. Bb microagglutination test (MAT) titers were obtained at Days -7, 0, monthly post-vaccination, and Days 358, 365, and 380. Nasal swabs were collected for microbiological assessment at Days -7, 0, 365, and 367-380. Oral Bb vaccination was not associated with side effects. Pre-challenge, vaccinated dogs developed persistent Bb MAT titers and control dogs remained seronegative. Post-challenge, duration of cough was longer in control dogs (least square means [LSM], 8.6 days) than vaccinated dogs (LSM, 1.5 days; P < 0.0001), with more control dogs having cough on 2 or more consecutive days (control group, n = 17/19, 89.5%; vaccinated group, n = 3/19, 15.8%; P = 0.0011). Post-challenge, Bb shedding occurred in all control dogs and 5/19 (26%) vaccinated dogs. Average duration of Bb shedding was longer in the control group (11.9 days vs. 0.6 days; P < 0.0001) and nasal Bb loads were higher in the control group (P < 0.00001). Orally administered Bb vaccine stimulated immunity that was still protective against virulent Bb challenge after 1 year.

A norovirus gastroenteritis outbreak in an Australian child-care center: A household-level analysis

There is a large burden of norovirus disease in child-care centers in Australia and around the world. Despite the ubiquity of norovirus outbreaks in child-care centers, little is known about the extent of this burden within the child-care center and the surrounding household clusters. Therefore, we performed an in-depth analysis of a gastroenteritis outbreak to examine the patterns of transmissions, household attack rates and the basic reproduction number (R₀) for Norovirus in a child-care facility. We used data from parental interviews of suspected cases sent home with gastroenteritis at a child-care center between 24^th of August and 18^th of September 2020. A total of 52 persons in 19 household clusters were symptomatic in this outbreak investigation. Of all transmissions, 23 (46.9%) occurred in the child-care center, the rest occurring in households. We found a household attack rate of 36.5% (95% CI 27.3, 47.1%). Serial intervals were estimated as mean 2.5 ± SD1.45 days. The R₀, using time-dependent methods during the growth phase of the outbreak (days 2 to 8) was 2.4 (95% CI 1.50, 3.50). The count of affected persons of a child-care center norovirus outbreak is approximately double the count of the total symptomatic staff and attending children. In the study setting, each symptomatic child-care attendee likely infected one other child-care attendee or staff and just over one household contact on average.

Bone turnover markers as determinants of bone density and fracture in men with distal forearm fractures: the pathogenesis examined in the Mr F study

The pathogenesis for low-trauma wrist fractures in men is not fully understood. This study found that these men had evidence of significantly higher bone turnover compared with control subjects. Bone turnover markers were negative predictors of bone mineral density and were a predictor of fracture.

INTRODUCTION

Men with distal forearm fractures have reduced bone density, an increased risk of osteoporosis and of further fractures. The aim of this study was to investigate whether or not men with distal forearm fractures had evidence of altered bone turnover activity.

METHODS

Fifty eight men with low-trauma distal forearm fracture and 58 age-matched healthy control subjects were recruited. All subjects underwent a DXA scan of the forearm, both hips, and lumbar spine, biochemical investigations, and health questionnaires. Measurements of beta crosslaps (βCTX), procollagen type I N-terminal propeptide (PINP), sclerostin, Dickkopf-1 (Dkk1), and fibroblast growth factor 23 (FGF 23) were made.

RESULTS

Men with fracture had significantly higher PINP than controls at 39.2 ng/ml (SD 19.5) versus 33.4 ng/ml (SD13.1) (p<0.001). They also had significantly higher βCTX at 0.45 ng/ml (SD 0.21) versus 0.37 ng/ml (SD 0.17) (p= 0.037). Fracture subjects had significantly lower aBMD and PINP was a negative predictor of aBMD at the total hip and βCTX a negative predictor of forearm aBMD. Sclerostin was a positive predictor of aBMD at the lumbar spine and hip sites. Sex hormone binding globulin (SHBG) at 37nmol/L (SD 15.0) was lower in fracture cohort compared to 47.9 nmol/L (SD 19.2) (p=0.001) in control. Multiple regression revealed that the best model for prediction of fracture included SHBG, P1NP, and ultra-distal forearm aBMD. The likelihood of distal forearm fracture was decreased by 5.1% for each nmol/L increase in SHBH and by 1.4% for every mg/cm² increase in ultra-distal forearm aBMD, but increased by 6.1 % for every ng/ml increase in P1NP. Men in the highest quartile of PINP had a significantly greater likelihood of distal forearm fracture than those in the lowest quartile.

CONCLUSION

The fracture group had significantly higher PINP and βCTX compared with the control group, and these markers were negative predictors of aBMD at the total hip and forearm sites, respectively. Sclerostin was a positive predictor of the variance of spinal and hip aBMD. Likelihood of forearm fracture was best predicted by a combination of SHBG, PINP, and ultra-distal forearm aBMD. Findings of such cross-sectional data should be treated with caution, as longitudinal studies would be required to confirm or refute them.

1037 Early Detection and Management of Preoperative Anaemia in An Elective Vascular Surgery Cohort

Introduction

Iron deficiency anaemia is the most common cause of anaemia in the surgical population. Patients receiving transfusions during major surgery encounter more adverse outcomes compared to those that do not, indicating the early management of pre-operative anemia is likely to reduce major surgical patient morbidity and mortality. Previous studies have demonstrated that 33% of patients undergoing major vascular surgery operations are anaemic (haemaglobin, Hb, &lt;130g/L).

Aim

We aim to reduce the proportion of anaemic patients undergoing major vascular surgery by 50%.

Method

A point-of-care HemoCue machine was used to measure Hb in patients who were referred for elective major surgery in 11 vascular clinics. Patients with Hb &lt; 130g/L were prescribed a 1-month course of Ferrous Sulfate (200mg TDS), and a letter was sent to their GP requesting iron function tests. Post-intervention Hb levels were rechecked following the preoperative anaesthetics review.

Results

11 patients were referred for major surgery, of which 4 were identified as anaemic (36%). The mean Hb concentration was 121.3g/L, which increased to 137.3g/L following oral iron therapy. The mean duration of follow up was 36 days (range 0-94 days).

Conclusions

The implementation of a HemoCue machine to identify and manage preoperative anaemia was successful in our pilot study. Further work should include full integration of our pathway into current vascular clinics without student support. This will enable evaluation of the impact of our intervention on a wider scale.

423 Patient Perspective on the Use of the Independent Sector to Maintain Elective NHS Services During the COVID-19 Pandemic

Introduction

COVID-19 resulted in the suspension of planned treatments for patients worldwide leaving millions suffering the physical and mental effects of delay. Reports indicate that when services have been re-established, patients have been afraid to take up appointments. Hospitals put processes in place to counter this, notably separating emergency and elective patient cohorts. Most notably seen by the co-operation between the NHS and independent private healthcare providers in March 2020 at the height of the pandemic. We undertook a study to ascertain the perceptions of NHS patients who underwent elective treatment at independent ‘cold’ sites during the COVID-19 pandemic.

Method

A cross-sectional study with structured telephone interviews of patients who had planned elective treatments at ‘cold’ independent hospitals between March and September 2020. 1150 patients were identified, and a 20% sample formed a 230 patient study group, with 158 (70%) agreeing to participate.

Results

30% of patients delayed their treatment due to COVID related concerns, with 76% of these only accepting treatment because this was at a ‘cold’ site. 46% of patients perceived treatment at a ‘cold’ site as the most important factor contributing to their safety. 153 patients (97%) supported the paid arrangement between the NHS and the independent sector to provide separate ‘cold’ sites for elective treatments.

Discussion

Safely restarting elective services to allow important planned treatments to take place, as was the pandemic continues, is a priority. Our study indicates that physical separation of patient pathways impacted most on patient confidence, and that the use of ‘cold’ sites is a viable option.