Federated Learning Survival Model and Potential Radiotherapy Decision Support Impact Assessment for Non-small Cell Lung Cancer Using Real-World Data

AIMS

The objective of this study was to develop a two-year overall survival model for inoperable stage I-III non-small cell lung cancer (NSCLC) patients using routine radiation oncology data over a federated (distributed) learning network and evaluate the potential of decision support for curative versus palliative radiotherapy.

METHODS

A federated infrastructure of data extraction, de-identification, standardisation, image analysis, and modelling was installed for seven clinics to obtain clinical and imaging features and survival information for patients treated in 2011-2019. A logistic regression model was trained for the 2011-2016 curative patient cohort and validated for the 2017-2019 cohort. Features were selected with univariate and model-based analysis and optimised using bootstrapping. System performance was assessed by the receiver operating characteristic (ROC) and corresponding area under curve (AUC), C-index, calibration metrics and Kaplan-Meier survival curves, with risk groups defined by model probability quartiles. Decision support was evaluated using a case-control analysis using propensity matching between treatment groups.

RESULTS

1655 patient datasets were included. The overall model AUC was 0.68. Fifty-eight percent of patients treated with palliative radiotherapy had a low-to-moderate risk prediction according to the model, with survival times not significantly different (p = 0.87 and 0.061) from patients treated with curative radiotherapy classified as high-risk by the model. When survival was simulated by risk group and model-indicated treatment, there was an estimated 11% increase in survival rate at two years (p < 0.01).

CONCLUSION

Federated learning over multiple institution data can be used to develop and validate decision support systems for lung cancer while quantifying the potential impact of their use in practice. This paves the way for personalised medicine, where decisions can be based more closely on individual patient details from routine care.

Noninvasive drug adherence monitoring of antipsychotic patients via finger sweat testing

Collection of finger sweat is explored here as a rapid and convenient way of monitoring patient adherence to antipsychotic drugs. Finger sweat samples (n = 426) collected from patients receiving treatment with clozapine, quetiapine and olanzapine were analysed by liquid chromatography mass spectrometry, including a subgroup of patients with paired plasma samples. Finger sweat samples were also analysed from a negative control group and patients who had handled antipsychotic medication only. The finger sweat test (based on the detection of parent drug in one donated sample) was 100% effective in monitoring adherence within commonly prescribed dosing ranges. In comparison to participants who handled the medication only, the test could distinguish between contact and administration through monitoring of the drug metabolite, or the level of parent drug. Additionally, in a subgroup of patients prescribed clozapine, a statistically significant correlation was observed between the mass of parent drug in finger sweat and plasma concentration. The finger sweat technology shows promise as a dignified, noninvasive method to monitor treatment adherence in patients taking antipsychotics.

The Creation of LIFE-M: The Longitudinal, Intergenerational Family Electronic Micro-Database Project

This paper describes the creation of the Longitudinal, Intergenerational Family Electronic Micro-Database (LIFE-M), a new data resource linking vital records and decennial censuses for millions of individuals and families living in the late 19th and 20th centuries in the United States. This combination of records provides a life-course and intergenerational perspective on the evolution of health and economic outcomes. Vital records also enable the linkage of women, because they contain a crosswalk between women's birth (i.e., "maiden") and married names. We describe (1) the data sources, coverage, and linking sequence; (2) the process and supervised machine-learning methods to linking records longitudinally and across generations; and (3) the resulting linked samples, including linking rates, representativeness, and weights.

Comparison of suspected and confirmed internal EVD-related infections: a prospective multi-centre U.K. observational study

Background

Diagnosis of internal external ventricular drain (EVD)-related infections (iERI) is an area of diagnostic difficulty. Empiric treatment is often initiated on clinical suspicion. There is limited guidance around antimicrobial management of confirmed versus suspected iERI.

Methods

Data on patients requiring EVD insertion were collected from 21 neurosurgical units in the United Kingdom from 2014 to 2015. Confirmed iERI was defined as clinical suspicion of infection with positive cerebrospinal fluid (CSF) culture and/or Gram stain. Cerebrospinal fluid, blood, and clinical parameters and antimicrobial management were compared between the 2 groups. Mortality and Modified Rankin Scores were compared at 30 days post-EVD insertion.

Results

Internal EVD-related infection was suspected after 46 of 495 EVD insertions (9.3%), more common after an emergency insertion. Twenty-six of 46 were confirmed iERIs, mostly due to Staphylococci (16 of 26). When confirmed and suspected infections were compared, there were no differences in CSF white cell counts or glucose concentrations, nor peripheral blood white cell counts or C-reactive protein concentrations. The incidence of fever, meningism, and seizures was also similar, although altered consciousness was more common in people with confirmed iERI. Broad-spectrum antimicrobial usage was prevalent in both groups with no difference in median duration of therapy (10 days [interquartile range {IQR}, 7–24.5] for confirmed cases and 9.5 days [IQR, 5.75–14] for suspected, P = 0.3). Despite comparable baseline characteristics, suspected iERI was associated with lower mortality and better neurological outcomes.

Conclusions

Suspected iERI could represent sterile inflammation or lower bacterial load leading to false-negative cultures. There is a need for improved microbiology diagnostics and biomarkers of bacterial infection to permit accurate discrimination and improve antimicrobial stewardship.

547 Postoperative Outcomes Are Significantly Worse Among Patients Undergoing Repair of Ruptured Versus Unruptured Iliac Artery Aneurysms – a 10-Year Longitudinal Cohort Study

Aim

The European Society for Vascular Surgery defines iliac artery aneurysms (IAA) as beyond 1.5 times its normal diameter. Common iliac arteries (CIA) beyond 1.8cm in men and 1.5cm in women are considered aneurysmal. This study aimed to assess outcomes following IAA rupture as their natural history is poorly understood and treatment recommendations based on low-level evidence.

Method

Patients with IAAs at a single vascular centre between 1st January 2010 and 31st August 2021 were identified from a prospectively collected departmental database and Caldicott-approved data collection performed. The primary outcomes included aneurysm rupture, rupture diameter, post-operative complications, 30-day, 1-year and 5-year mortality rates. Statistical analysis with SPSS® was performed using chi-squared tests.

Results

Of 203 patients included, 90.6% were men and median(IQR) age at detection was 77 (71–83). Co-morbidities included hypertension (54.2%), hyperlipidaemia (42.9%) and ischaemic heart disease (35.5%). IAA were in the CIA (85.2%), IIA (21.7%) and EIA (2.0%), mostly asymptomatic (78.8%). Overall IAA rupture rate was 7.9% with CIA (81.2%) and EIA (18.8%). Mean (SD) diameters at rupture were 4.6 (2.4)cm for CIA and 4.6 (3.0)cm for IIA. Post-operative major adverse cardiovascular events (MACE) more frequently occurred following repair of ruptured compared to unruptured IAA (33.3% vs 3.5%, p=.011). Mortality at 30-days, 1-year and 5-years postoperatively were higher following repair of ruptured vs unruptured aneurysms (88.9%, 88.9%, 100% vs 1.2%, 10.6%, 36.1% respectively).

Conclusions

Early detection and elective treatment of IAA aneurysms before they approach 4.6cm may reduce rupture risk, morbidity and mortality associated with emergency repair following rupture.

548 A 10-Year Longitudinal Cohort Study Assessing Growth Rates and Surveillance Intervals for Common Iliac Artery Aneurysms

Aim

The European Society for Vascular Surgery defines common iliac artery (CIA) aneurysms as greater than 1.8cm in men and 1.5cm in women. Their reported growth rate is 1–4mm/year depending on their diameter. This study aimed to assess the natural history and surveillance intervals for CIA aneurysms as intervention threshold is based on low-quality evidence.

Method

Patients diagnosed with an IAA at a single vascular centre between 1st January 2010 and 31st August 2021 were identified from a prospectively collected departmental database and Caldicott-approved data collection performed. The primary outcomes included diameter-based mean aneurysm growth rates and median surveillance intervals. Statistical analysis with SPSS® was performed using chi-squared tests.

Results

Of the 203 patients included, 90.6% were men and median (IQR) age at detection was 77 (71–83). IAA were located in the CIA (85.2%), IIA (21.7%) and EIA (2.0%) with the majority being asymptomatic (78.8%). CT was most frequently used as the imaging modality for IAA surveillance (66.3%), followed by ultrasound scan (29.8%) and MRA (3.9%). Growth rate for CIA aneurysms measuring 1.0–1.9cm were -2.1mm/year, 2.0–2.9cm were 0.8mm/year, 3.0–3.9cm were 3.5mm/year, 4.0–4.9cm were 9.4mm/year, 5.0–5.9cm were 2.9mm/year and &gt;6.0cm were 13.8mm/year. Median surveillance intervals for CIA aneurysms at 1.0–1.9cm were 12-monthly, 2.0–2.9cm were 11-monthly, 3.0–3.9cm were 5-monthly, 4.0–4.9cm were 5-monthly, 5.0–5.9cm were 5.5-monthly and &gt;6.0cm were 14.5-monthly. Mean(SD) CIA diameter at rupture was 4.6 (2.4)cm.

Conclusions

CIA aneurysms demonstrate faster growth rates as they enlarge and may require more frequent clinical assessments, surveillance, and consideration for repair prior to rupture.

O069 Infra-inguinal bypass graft surveillance is an opportunity to optimise statin and antiplatelet therapy to reduce 12-month major amputation and mortality rates

Introduction

Infra-inguinal bypass graft failure within the first two postoperative years can result from stenotic lesions within the conduit and near anastomosis. This study assesses graft surveillance uptake, patency, amputation and death rates among infra-inguinal bypass surgery patients at 1, 6 and 12-months postoperatively.

Methods

Patients undergoing infra-inguinal bypass at a single vascular centre between 1st January 2018 and 31st December 2019 were identified from the prospectively collected database. Primary outcomes at 1, 6 and 12-months postoperatively included: duplex-ultrasound (DUS) surveillance uptake, patency, major amputations and death rates. Statistical analysis with SPSS® was performed using chi-squared tests and paired sample t-tests.

Results

Of the 91 patients included, 79.1% were men and median (IQR) age was 71(62–76). At 1, 6 and 12 months, DUS uptake was 74.2%, 77.5% and 73.3% respectively, primary-assisted patency rates were 88.2%, 78.8% and 65.3% respectively and secondary patency rates were 97.6%, 96.3% and 96.3% respectively. Major amputation rates at 1, 6 and 12 months were 5.4%, 8.6%, 8.6% respectively. Death rates at 1, 6 and 12 months were 3.2%, 5.4% and 10.8% respectively. Patients on statin therapy post-operatively had lower 12-month mortality than those not on statin therapy (7.7% vs 30.0%, p=.028). Major amputation rate at 12 months was lower among patients prescribed antiplatelet therapy (6.3% vs 40%, p=.008).

Conclusion

DUS graft surveillance is a vital opportunity to adequately optimise statin and antiplatelet therapy post-revascularisation to reduce 12-month major amputation and mortality rates.

Take-home message

Infra-inguinal bypass graft surveillance is an opportunity to optimise statin and antiplatelet therapy to reduce 12-month major amputation and mortality rates.

Radiotherapy Quality Assurance in the PORTEC-3 (TROG 08.04) Trial

AIMS

Quality assurance in radiotherapy (QART) is essential to ensure the scientific integrity of a clinical trial. This paper reports the findings of the retrospective QART assessment for all centres that participated in PORTEC-3; a randomised controlled trial that compared pelvic radiotherapy with concurrent chemoradiotherapy to the pelvis followed by adjuvant chemotherapy. The trial showed an overall survival benefit for the addition of the chemotherapy in the management of women with high-risk endometrial cancer.

MATERIALS AND METHODS

Clinicians were invited to upload a randomly selected case/s treated at each of the participating sites. Panel reviewers analysed the contours to certify that the target volumes and organ at risk structures were contoured according to guidelines. The results were categorised into acceptable, minor variation, major variation or unevaluable. The radiotherapy plans were dosimetrically evaluated using the well-established Trans-Tasman Radiation Oncology Group (TROG) protocol.

RESULTS

Between August 2010 and January 2018, data from 146 patients of 686 consecutively treated patients were retrospectively reviewed. All 16 Australia and New Zealand and 71 of 77 international centres uploaded data for evaluation. In total, 3514 dosimetric and contour variables were reviewed. Of these, 3136 variables were deemed acceptable (89.2%), with 335 minor (9.6%) and 43 major variations (1.2%). Major contour variations included the clinical target volume vaginal vault, clinical target volume parametria and differential planning target volume vault expansion.

CONCLUSION

The results of the QART assessment confirmed high uniformity and low rates of both minor and major deviations in contouring and dosimetry in all sites. This supports the safe introduction of the PORTEC-3 treatment protocol into routine clinical practice.

Comparison of Intensive Care and Trauma-specific Scoring Systems in Critically Ill Patients

INTRODUCTION

Amongst critically ill trauma patients admitted to ICU and still alive and in ICU after 24 hours, it is unclear which trauma scoring system offers the best performance in predicting in-hospital mortality.

METHODS

The Australia and New Zealand Intensive Care Society Adult Patient Database and Victorian State Trauma Registry were linked using a unique patient identification number. Six scoring systems were evaluated: the Australian and New Zealand Risk of Death (ANZROD), Acute Physiology and Chronic Health Evaluation III (APACHE III) score and associated APACHE III Risk of Death (ROD), Trauma and Injury Severity Score (TRISS), Injury Severity Score (ISS), New Injury Severity Score (NISS) and the Revised Trauma Score (RTS). Patients who were admitted to ICU for longer than 24 hours were analysed. Performance of each scoring system was assessed primarily by examining the area under the receiver operating characteristic curve (AUROC) and in addition using standardised mortality ratios, Brier score and Hosmer-Lemeshow C statistics where appropriate. Subgroup assessments were made for patients aged 65 years and older, patients between 18 and 40 years of age, major trauma centre and head injury.

RESULTS

Overall, 5,237 major trauma patients who were still alive and in ICU after 24 hours were studied from 25 ICUs in Victoria, Australia between July 2008 and January 2018. Hospital mortality was 10.7%. ANZROD (AUROC 0.91; 95% CI 0.90-0.92), APACHE III ROD (AUROC 0.88; 95% CI 0.87-0.90), and APACHE III (AUROC 0.88; 95% CI 0.87-0.89) were the best performing tools for predicting hospital mortality. TRISS had acceptable overall performance (AUROC 0.78; 95% CI 0.76-0.80) while ISS (AUROC 0.61; 95% CI 0.59-0.64), NISS (AUROC 0.68; 95% CI 0.65-0.70) and RTS (AUROC 0.69; 95% CI 0.67-0.72) performed poorly. The performance of each scoring system was highest in younger adults and poorest in older adults.

CONCLUSION

In ICU patients admitted with a trauma diagnosis and still alive and in ICU after 24 hours, ANZROD and APACHE III had a superior performance when compared with traditional trauma-specific scoring systems in predicting hospital mortality. This was observed both overall and in each of the subgroup analyses. The anatomical scoring systems all performed poorly in the ICU population of Victoria, Australia.

Cognitive impairment reversed by cinacalcet administration in primary hyperparathyroidism

Primary hyperparathyroidism (pHPT) is a common endocrine disorder. Often serum calcium is minimally elevated with few symptoms. In elderly patients with multiple co-morbidities, the decision to "watch and wait" is often most clinically appropriate as operative intervention is associated with high peri-operative risk. We present an elderly patient with mild hypercalcemia secondary to primary hyperparathyroidism. The clinical decision was initially to watch and wait. The patient subsequently developed cognitive impairment and was diagnosed with mixed Alzheimer's disease/vascular dementia. She became dependent for all care and housebound. A therapeutic trial of cinacalcet was commenced following a further acute rise in serum calcium. Significant reversal of her functional and cognitive deficit occurred. She was no longer fully dependent. Mini mental state examination (MMSE) improved from 8/30 to 21/30. In vulnerable neural systems, even mild elevation in serum calcium may have a profound effect on cognition and function. We propose a therapeutic trial of cinacalcet in such patients.

Predictors of death and new disability after critical illness: a multicentre prospective cohort study

PURPOSE

This study aimed to determine the prevalence and predictors of death or new disability following critical illness.

METHODS

Prospective, multicentre cohort study conducted in six metropolitan intensive care units (ICU). Participants were adults admitted to the ICU who received more than 24 h of mechanical ventilation. The primary outcome was death or new disability at 6 months, with new disability defined by a 10% increase in the WHODAS 2.0.

RESULTS

Of 628 patients with the primary outcome available (median age of 62 [49-71] years, 379 [61.0%] had a medical admission and 370 (58.9%) died or developed new disability by 6 months. Independent predictors of death or new disability included age [OR 1.02 (1.01-1.03), P = 0.001], higher severity of illness (APACHE III) [OR 1.02 (1.01-1.03), P < 0.001] and admission diagnosis. Compared to patients with a surgical admission diagnosis, patients with a cardiac arrest [OR (95% CI) 4.06 (1.89-8.68), P < 0.001], sepsis [OR (95% CI) 2.43 (1.32-4.47), P = 0.004], or trauma [OR (95% CI) 6.24 (3.07-12.71), P < 0.001] diagnosis had higher odds of death or new disability, while patients with a lung transplant [OR (95% CI) 0.21 (0.07-0.58), P = 0.003] diagnosis had lower odds. A model including these three variables had good calibration (Brier score 0.20) and acceptable discriminative power with an area under the receiver operating characteristic curve of 0.76 (95% CI 0.72-0.80).

CONCLUSION

Less than half of all patients mechanically ventilated for more than 24 h were alive and free of new disability at 6 months after admission to ICU. A model including age, illness severity and admission diagnosis has acceptable discriminative ability to predict death or new disability at 6 months.

918 COVID-19: A Detailed Analysis on Fit-Testing for Respiratory Protective Equipment in the UK

Introduction

There is limited data in the literature regarding the adequacy of generic FFP3 masks and their facial fit to ensure adequate protection. Mask fit-testing is therefore essential to protect healthcare workers.

Method

Using the Freedom of Information Act, 137 acute NHS trusts in the UK were approached on the 26/3/2020 by an independent researcher to provide data on the outcome of fit testing at each site.

Results

85 Trusts responded to the FOI with 51 trusts providing pertinent data relevant to the FOI request. There was a total of 72 mask types used across 51 trusts. The commonest of which was the FFP3M1863 (used by 47/51 trusts, 92.16%). A positive correlation was found between staff members and number of mask types used (r = 0.75, P = &lt;0.05).

Overall fit-testing pass rates were provided by 32 trusts. The mean percentage pass rate was 80.74%.

Gender specific failure rates were provided by seven trusts. 4386 males underwent fit-testing in comparison to 16305 females. Across all seven trusts 20.08% of men tested failed the fit-test while only 19.89% of women failed the fit-test.

Conclusions

Our results may be utilised in choosing respirators for fit testing programme in healthcare-workers during the COVID-19 pandemic.

Fit-testing of respiratory protective equipment in the UK during the initial response to the COVID-19 pandemic

Background

Public Health England guidance stipulates the use of filtering facepiece (FFP3) masks for healthcare workers engaged in aerosol-generating procedures. Mask fit-testing of respiratory protective equipment is essential to protect healthcare workers from aerosolized particles.

Aim

To analyse the outcome of mask fit-testing across National Health Service (NHS) hospitals in the UK during the first wave of the coronavirus disease 2019 (COVID-19) pandemic.

Methods

Using the Freedom of Information Act, 137 NHS hospitals were approached on May 26^th, 2020 by an independent researcher to provide data on the outcome of fit-testing at each site.

Findings

Ninety-six hospitals responded to the request between May 26^th, 2020 to October 29^th, 2020. There was a total of 86 mask types used across 56 hospitals, 13 of which were used in at least 10% of these hospitals; the most frequently used was the FFP3M1863, used by 92.86% of hospitals. Overall fit-testing pass rates were provided by 32 hospitals with mean pass rate of 80.74%. The most successful masks, in terms of fit-test failure rates, were the Alpha Solway 3030V and the Alpha Solway S–3V (both reporting mean fit-test failures of 2%). Male- and female-specific pass and failure rates were provided by seven hospitals. Across the seven hospitals, 20.1% of men tested failed the fit-test for all masks used, whereas 19.9% of women tested failed the fit-test for all masks used. Failure rates were significantly higher in staff from Black, Asian, and Minority Ethnic (BAME) backgrounds 644/2507 (25.69%) across four hospitals.

Conclusion

Twenty percent of healthcare workers tested during the first response to the pandemic failed fit-testing for masks. A small sample revealed that this was most prominent in staff from BAME backgrounds.

16 Pharmacological Interventions for Sialorrhoea in Parkinson’s Disease: A Systematic Review

Introduction

We looked at patients with Idiopathic Parkinson’s disease and reviewed articles that used any pharmacological therapy to attempt to reduce the volume or burden of sialorrhoea. The control was patients on placebo or receiving usual best care. The primary outcome was symptom burden of sialorrhoea.

Method

The review was registered on Prospero (CRD42016042470.) 7 electronic databases (MEDLINE, EMBASE, CINAHL, PsycINFO, Cochrane Central Register of Controlled Trials (CENTRAL), LILACS) were searched until April 2019 using search terms for sialorrhoea and Parkinson’s disease. We additional searched through the last 6 years of abstracts from the International Movement Disorder Conference, Movement Disorder Society International Congress and World Parkinson’s Congress. Hand searching was performed of published journals from the Movement Disorder Journal of the Movement Disorder Society and Neurology. Inclusion criteria included patients aged over 18; patients with a diagnosis of Idiopathic Parkinson’s disease; patients with the complication of sialorrhoea and any pharmacological therapy aimed at reducing sialorrhoea. All articles were assessed for risk of bias independently by two assessors using the Cochrane risk of bias tool.

Results

7 articles were included from the 1,015 screened citations. 3 studies used Botox B as an intervention with clear objective measures and low risk of bias. These studies did show a significant reduction in sialorrhoea however the largest study showed some reporting bias and the others had small numbers and a short follow up period. 1 study looked at Botox A which showed significant improvement in sialorrhoea but had short follow-up.

Conclusion

Most studies have very small numbers and were followed up for only 1 month. There was significant heterogeneity in outcome measures but little evidence of significant harm from from the treatments studied. Overall, more robust evidence is required in order to achieve a gold standard in the treatment of sialorrhoea.

African American Activism: The Predictive Role of Race Related Stress, Racial Identity, and Social Justice Beliefs

Historically, African American activism has played a pivotal role in advancing social change in the United States. As such, there is an interest in examining possible factors that may engender activism among African Americans. Therefore, the purpose of this study was to extend research by Szymanski and Lewis (2015), which explored potential predictors of activism among African Americans. With a sample of 458 African American undergraduates, race-related stress, racial identity dimensions, and social justice variables were examined. A four-stage, multiple linear hierarchical regression model and two multiple mediation bootstrap analyses were employed. Race-related stress and racial identity attitudes significantly and uniquely predicted involvement in African American activism, complementing existing literature. Beyond that, social justice beliefs predicted African American activism over and above racial identity and race-related stress. More specifically, social justice subjective norms (i.e., social influence) was the most important predictor of activism for African American undergraduates. Implications for social justice development within institutions of higher education are discussed.

How Well Do Automated Linking Methods Perform? Lessons from U.S. Historical Data

This paper reviews the literature in historical record linkage in the U.S. and examines the performance of widely-used record linking algorithms and common variations in their assumptions. We use two high-quality, hand-linked datasets and one synthetic ground truth to examine the direct effects of linking algorithms on data quality. We find that (1) no algorithm (including hand-linking) consistently produces representative samples; (2) 15 to 37 percent of links chosen by widely-used algorithms are classified as errors by trained human reviewers; and (3) false links are systematically related to baseline sample characteristics, showing that some algorithms may induce systematic measurement error into analyses. A case study shows that the combined effects of (1)-(3) attenuate estimates of the intergenerational income elasticity by up to 20 percent, and common variations in algorithm assumptions result in greater attenuation. As current practice moves to automate linking and increase link rates, these results highlight the important potential consequences of linking errors on inferences with linked data. We conclude with constructive suggestions for reducing linking errors and directions for future research.

Functional Outcomes in Patients Admitted to the Intensive Care Unit with Traumatic Brain Injury and Exposed to Hyperoxia: A Retrospective Multicentre Cohort Study

Background

Supplemental oxygen administration to critically ill patients is ubiquitous in the intensive care unit (ICU). Uncertainty persists as to whether hyperoxia is benign in patients with traumatic brain injury (TBI), particularly in regard to their long-term functional neurological outcomes.

Methods

We conducted a retrospective multicenter cohort study of invasively ventilated patients with TBI admitted to the ICU. A database linkage between the Australian and New Zealand Intensive Care Society Adult Patient Database (ANZICS-APD) and the Victorian State Trauma Registry (VSTR) was utilized. The primary exposure variable was minimum acute physiology and chronic health evaluation (APACHE) III P_aO₂ in the first 24 h of ICU. We defined hypoxia as P_aO₂ < 60 mmHg, normoxia as 60–299 mmHg, and hyperoxia as ≥ 300 mmHg. The primary outcome was a Glasgow Outcome Scale-Extended (GOSE) < 5 at 6 months while secondary outcomes included 12 and 24 months GOSE and mortality at each of these timepoints. Additional sensitivity analyses were undertaken in the following subgroups: isolated head injury, patients with operative intervention, head injury severity, and P_aO₂ either subcategorized by increments of 60 mmHg or treated as a continuous variable.

Results

A total of 3699 patients met the inclusion criteria. The mean age was 42.8 years, 77.7% were male and the mean acute physiology and chronic health evaluation (APACHE) III score was 60.1 (26.3). 2842 patients experienced normoxia, and 783 hyperoxia. The primary outcome occurred in 1470 (47.1%) of patients overall with 1123 (47.1%) from the normoxia group and 312 (45.9%) from the hyperoxia group—odds ratio 0.99 (0.78–1.25). No significant differences in outcomes between groups at 6, 12, and 24 months were observed. Sensitivity analyses did not identify subgroups that were adversely affected by exposure to hyperoxia.

Conclusions

No associations were observed between hyperoxia in ICU during the first 24 h and adverse neurological outcome at 6 months in ventilated TBI patients.

Electronic supplementary material

The online version of this article (10.1007/s12028-020-01033-y) contains supplementary material, which is available to authorized users.

45 Implementing A Frailty at the Front Door Service in the Emergency Department

Background

University Hospital Hairmyres is a small District General Hospital in Lanarkshire Scotland. We have an active Care of the Elderly Department with a well-established Acute Care of the Elderly (ACE) team of Advanced Nurse Practitioners, supported by Consultants. This team delivers Comprehensive Geriatric Assessment (CGA) to frail older people in acute medical receiving as well as offering liaison to medical, surgical and orthopaedic wards.

Local problem and intervention: Our patients were not always being managed by the correct professionals in a timely manner, leading to delays especially in the Emergency Department (ED). We set up a Frailty at the Front Door (FAFD) service to address this, commencing July 2018. Supported by additional consultant sessions, we re-focused the ACE team on assessing and managing frail patients in the ED. The aim was to get the right patient to the right place at the right time and to manage acutely ill people at home where this was safe. Where admission was required we aimed to admit directly to a specialty bed, bypassing acute receiving wards.

Methods

We routinely collect important data including number of frail patients, %patients receiving CGA within 24 hours, number of discharges. To assess the impact of our change we analysed the data by plotting on run charts and statistical process control charts. In addition we assessed the effect on referrals from medical specialties and the number of direct-to-specialty admissions.

Results

After the 22 July 2018 we noticed a significant increase in patients screened for frailty, and a significant increase in discharges. We were able to reliably sustain over 95% of frail patients getting CGA within 24 hrs. There was an increased use of hospital at home. There was a reduction in referrals from medical wards (median = 10/week before, 5/week after intervention). Between August 2018 and May 2019 we were able to admit 163 patients directly to specialty beds. There was no change in re-admission rate.

Conclusions

We successfully changed our service to have consultant delivered Frailty at the Front Door, assessing more frail patients. Most importantly, we have an improved patient pathway, both managing more people at home but also reducing ward moves by achieving direct to specialty admissions. CGA can be safely delivered in the ED.

Simple Strategies for Improving Inference with Linked Data: A Case Study of the 1850-1930 IPUMS Linked Representative Historical Samples

New large-scale linked data are revolutionizing quantitative history and demography. This paper proposes two complementary strategies for improving inference with linked historical data: the use of validation variables to identify higher quality links and a simple, regression-based weighting procedure to increase the representativeness of custom research samples. We demonstrate the potential value of these strategies using the 1850-1930 Integrated Public Use Microdata Series Linked Representative Samples (IPUMS-LRS)-a high quality, publicly available linked historical dataset. We show that, while incorrect linking rates appear low in the IPUMS-LRS, researchers can reduce error rates further using validation variables. We also show how researchers can reweight linked samples to balance observed characteristics in the linked sample with those in a reference population using a simple regression-based procedure.

The safety and efficacy of neuromodulation using percutaneous electrical nerve stimulation for the management of trigeminal-mediated headshaking in 168 horses

Background

Early results from the use of neuromodulation by percutaneous electrical nerve stimulation for the management of trigeminal‐mediated headshaking in horses were promising but lacked sufficient case numbers and long‐term follow‐up. The neuromodulatory procedure has since been established as EquiPENS™.

Objectives

The aim of this study was to report long‐term results from a larger number of cases and to investigate for predictors of outcome.

Study design

Prospective case series using international, multi‐centre data.

Methods

Eligible cases were horses with a veterinary diagnosis of trigeminal‐mediated headshaking, which received EquiPENS™ neuromodulation at trained centres between August 2013 and November 2017. The standard protocol was an initial three‐procedure course, with additional procedures should a horse go into remission but then relapse. Data collected included signalment, history, diagnostic tests performed, details of any complications, whether horses had gone into remission and the length of remission.

Results

Results were obtained from 168 horses, with 530 procedures. The complication rate was 8.8% of procedures. In all but one case, complications were mild and transient, without self‐trauma. Remission of headshaking following the initial course occurred in 53% (72/136) of horses. Median length of time recorded in remission was 9.5 weeks (range 2 days to 156 weeks ongoing). Where signs recurred, most horses went back into remission following additional procedures, usually for longer than from the previous procedure. No predictors for outcome were determined.

Main limitations

No placebo or control group, owner‐assessed results.

Conclusions

EquiPENS™ neuromodulation can be an effective and safe treatment for the management of trigeminal‐mediated headshaking in some horses. An increased understanding of neuromodulation could help optimise the technique. Advances in treatment for trigeminal‐mediated headshaking will remain limited until there is a greater understanding of the aetiopathogenesis of the condition.

Insights into Pasteurellaceae carriage dynamics in the nasal passages of healthy beef calves

We investigated three bovine respiratory pathobionts in healthy cattle using qPCR optimised and validated to quantify Histophilus somni, Mannheimia haemolytica and Pasteurella multocida over a wide dynamic range. A longitudinal study was conducted to investigate the carriage and density of these bacteria in the nasal passages of healthy beef calves (N = 60) housed over winter in an experimental farm setting. The three pathobiont species exhibited remarkably different carriage rates and density profiles. At housing, high carriage rates were observed for P. multocida (95%), and H. somni (75%), while fewer calves were positive for M. haemolytica (13%). Carriage rates for all three bacterial species declined over the 75-day study, but not all individuals became colonised despite sharing of environment and airspace. Colonisation patterns ranged from continuous to intermittent and were different among pathobiont species. Interval-censored exponential survival models estimated the median duration of H. somni and P. multocida carriage at 14.8 (CI_95%: 10.6–20.9) and 55.5 (CI_95%: 43.3–71.3) days respectively, and found higher density P. multocida carriage was associated with slower clearance (p = 0.036). This work offers insights into the dynamics of pathobiont carriage and provides a potential platform for further data collection and modelling studies.

The Impact of Bispectral Index Monitoring on Sedation Administration in Mechanically Ventilated Patients

The aim of this prospective randomised controlled trial was to assess the effectiveness of the Bispectral Index (BIS) monitor in supporting clinical sedation management decisions in mechanically ventilated intensive care unit patients. Fifty adult mechanically ventilated surgical and general intensive care unit patients receiving sedative infusions of morphine and midazolam were randomly allocated to receive BIS monitoring (n=25) or standard sedation management (n=25). In the BIS group, sedation was titrated to maintain a BIS value of greater than 70. In the standard management group, sedative needs were titrated based on subjective assessment and clinical signs.

There was no statistically significant difference in the amount of sedation administered (morphine P =0.67 and midazolam P =0.85). However, there was a statistically significant difference in sedation administration over time. Patients in the BIS group received increasing amounts of sedation over time whilst those in the control group received decreasing amounts of sedation over time. The same inverse relationship existed for both sedative agents (morphine P=0.005, midazolam P=0.03). Duration of mechanical ventilation was comparable in the two groups. We conclude that the use of BIS monitoring did not reduce the amount of sedation used, the length of mechanical ventilation time or the length of ICU stay.

Impact of stone characteristics on cavitation in burst wave lithotripsy

Non-invasive kidney stone treatments such as shock wave lithotripsy (SWL) and burst wave lithotripsy (BWL) rely on the delivery of pressure waves through tissue to the stone. In both SWL and BWL, the potential to hinder comminution by exciting cavitation proximal to the stone has been reported. To elucidate how different stones alter prefocal cavitation in BWL, different natural and synthetic stones were treated in vitro using a therapy transducer operating at 350 kHz (peak negative pressure 7 MPa, pulse length 20 cycles, pulse repetition frequency 10 Hz). Stones were held in a confined volume of water designed to mimic the geometry of a kidney calyx, with the water filtered and degassed to maintain conditions for which the cavitation threshold (in the absence of a stone) matches that from in vivo observations. Stone targeting and cavitation monitoring were performed via ultrasound imaging using a diagnostic probe aligned coaxially with the therapy transducer. Quantitative differences in the extent and location of cavitation activity were observed for different stone types-e.g., "softer" stones (natural and synthetic) that disintegrate into "dusty" fragments produced larger prefocal cavitation clouds. Future work will focus on correlation of such cavitation metrics with stone fragmentation.

Prevalence and antimicrobial resistance patterns of Salmonella isolated from poultry farms in southeastern United States

Salmonella spp. are among the most common foodborne pathogens, and increase in the occurrence of antimicrobial drug-resistant Salmonella poses a severe risk to public health. The main objective of this study was to determine changes in Salmonella prevalence and their antimicrobial resistance on poultry farms following recommendations to changes in biosecurity practices. Four poultry farms were sampled by collecting cloacal swabs, drag swabs, and litter samples prior to recommended biosecurity changes (March-April) and post recommendations (October-November). Prevalence of Salmonella was 3 to 4% during pre-recommendations, while the prevalence was higher (P > 0.05), ranging from 5 to 14% during post recommendations. Higher Salmonella prevalence was observed for pre- and post-recommendation phases by sample type in cloacal and drag samples -5% for farm 1, drag swab -6% on farm 2, cloacal swab -6% for farm 3, and drag swab -17% on farm 4. The PCR confirmed Salmonella were serotyped and tested for antimicrobial resistance. Six serotypes of Salmonella were identified with S. Enteritidis (52%) being the most prevalent, followed by S. Berta (38%), S. Mbandaka (7%), S. Typhimurium (2%), S. Kentucky (0.4%), and S. Tennessee (0.4%). A total of 7% isolates exhibited resistance to at least one of the 8 antimicrobials. Higher resistance was observed for tetracycline, streptomycin, and nalidixic acid. A single isolate of S. Mbandaka exhibited multidrug resistance to tetracycline, amoxicillin/clavulanic acid, and ampicillin. Based on these prevalence results, it can be inferred that, irrespective of implementation of improved biosecurity practices, seasonal variation can cause changes in the prevalence of Salmonella on the farms. Resistance to clinically important antimicrobials used to treat salmonellosis is a concern to public health.

Predicting medical emergency team calls, cardiac arrest calls and re-admission after intensive care discharge: creation of a tool to identify at-risk patients

We aimed to develop a predictive model for intensive care unit (ICU)-discharged patients at risk of post-ICU deterioration. We performed a retrospective, single-centre cohort observational study by linking the hospital admission, patient pathology, ICU, and medical emergency team (MET) databases. All patients discharged from the Alfred Hospital ICU to wards between July 2012 and June 2014 were included. The primary outcome was a composite endpoint of any MET call, cardiac arrest call or ICU re-admission. Multivariable logistic regression analysis was used to identify predictors of outcome and develop a risk-stratification model. Four thousand, six hundred and thirty-two patients were included in the study. Of these, 878 (19%) patients had a MET call, 51 (1.1%) patients had cardiac arrest calls, 304 (6.5%) were re-admitted to ICU during the same hospital stay, and 964 (21%) had MET calls, cardiac arrest calls or ICU re-admission. A discriminatory predictive model was developed (area under the receiver operating characteristic curve 0.72 [95% confidence intervals {CI} 0.70 to 0.73]) which identified the following factors: increasing age (odds ratio [OR] 1.012 [95% CI 1.007 to 1.017] <i>P</i> <0.001), ICU admission with subarachnoid haemorrhage (OR 2.26 [95% CI 1.22 to 4.16] <i>P</i>=0.009), admission to ICU from a ward (OR 1.67 [95% CI 1.31 to 2.13] <i>P</i> <0.001), Acute Physiology and Chronic Health Evaluation (APACHE) III score without the age component (OR 1.005 [95% CI 1.001 to 1.010] <i>P</i>=0.025), tracheostomy on ICU discharge (OR 4.32 [95% CI 2.9 to 6.42] <i>P</i> <0.001) and discharge to cardiothoracic (OR 2.43 [95%CI 1.49 to 3.96] <i>P</i> <0.001) or oncology wards (OR 2.27 [95% CI 1.05 to 4.89] <i>P</i>=0.036). Over the two-year period, 361 patients were identified as having a greater than 50% chance of having post-ICU deterioration. Factors are identifiable to predict patients at risk of post-ICU deterioration. This knowledge could be used to guide patient follow-up after ICU discharge, optimise healthcare resources, and improve patient outcomes and service delivery.

Effects of Tendon Splitting on Experimentally-Induced Acute Equine Tendinitis

The objective of this study was to compare the early phases of healing between surgically split and unsplit (control) tendons in a superficial digital flexor tendinitis model. Tendinitis was induced in both forelimbs of six horses by injection with collagenase. One tendon from each horse was randomly chosen to undergo tendon splitting. In six horses, the tendons were evaluated by ultrasonography before (week 0) and after (week 1) the onset of tendinitis and at weeks two, three, and four following tendon splitting. Three horses were euthanatized at the end of week four. Three horses were maintained for an additional month and evaluated at week eight before euthanasia. Following euthanasia, microangiographic and histologic evaluations were performed. Ultrasonographic evaluation revealed a significant (p <0.05) decrease in mean tendon lesion area in the split tendons when compared to the controls at weeks three, four, and eight. By week eight, all lesions had resolved in the split but not the control tendons. Mean lesion grade was lower in the split tendons at weeks two, three, four, and eight. Tendon area was similar (p >0.05) between groups at all periods after surgery. There was a more mature vascular pattern and greater peritendinous vascular response in the split tendons compared to control tendons. Histological examination under plain and polarized light revealed more normal collagen orientation and wave formation in the repair tissue in the split tendons at weeks four and eight. The fibroblast nuclei density was significantly (p <0.05) lower in the split tendons at week eight. These findings suggest that during the time period of this study, tendon splitting results in a more rapid decrease in lesion size and superior repair tissue organization when compared to controls.

In a collagenase-induced tendinitis model, tendon splitting was shown to result in a greater reduction in tendon lesion size and superior repair tissue organization when compared to control tendons.

Day or overnight transfusion in critically ill patients: does it matter?

BACKGROUND AND OBJECTIVES

The timing of blood administration in critically ill patients is first driven by patients' needs. This study aimed to define the epidemiology and significance of overnight transfusion in critically ill patients.

MATERIALS AND METHODS

This is a post hoc analysis of a prospective multicentre observational study including 874 critically ill patients receiving red blood cells, platelets, fresh frozen plasma (FFP) or cryoprecipitate. Characteristics of patients receiving blood only during the day (8 am up until 8 pm) were compared to those receiving blood only overnight (8 pm up until 8 am). Characteristics of transfusion were compared, and factors independently associated with major bleeding were analysed.

RESULTS

The 287 patients transfused during the day only had similar severity and mortality to the 258 receiving blood products overnight only. Although bleeding-related admission diagnoses were similar, major bleeding was the indication for transfusion in 12% of patients transfused in daytime only versus 30% of patients transfused at night only (P < 0·001). Similar total amount of blood products were transfused at day and night (2856 versus 2927); however, patients were more likely to receive FFP and cryoprecipitate at night compared with daytime. Overnight transfusion was independently associated with increased odds of major bleeding (odds ratio, 3·16, 95% confidence interval, 2·00-5·01).

CONCLUSION

Transfusion occurs evenly across day and night in ICU; nonetheless, there are differences in type of blood products administered that reflect differences in indication. Critically ill patients were more likely to receive blood for major bleeding at night irrespective of admission diagnosis.

Changes in mucosal and serum immunoglobulin levels of California sea lions during early development

To date, most studies on pinniped immunoglobulins have focused on circulating antibodies. However, systemic and local immune activities differ in terms of maturation, intensity, and types of effectors that participate. Here, we examined levels of three immunoglobulin isotypes, IgG, IgM and IgA, in the blood and mucosal membranes of free-living California sea lion pups. We investigated whether age, body condition and sex influenced their concentration. Isotype levels varied among tissues, with age-related patterns that could be indicative of differential regulation along development. Serum IgG and IgA increased linearly with age, reaching adult levels at five months of age, while IgM remained unchanged. Mucosal isotypes tended to be low in newborns and remained so until five months of age. Regardless of age, pups with better condition tended to have higher anal IgG levels and higher genital IgA levels, suggesting that their synthesis and transport to the mucosal membranes is costly. Intersex differences were only observed in the genital mucosa, where all isotypes differed between male and female pups, regardless of age, presumably due to histological and anatomical differences.

Assessing contemporary intensive care unit outcome: development and validation of the Australian and New Zealand Risk of Death admission model

The Australian and New Zealand Risk of Death (ANZROD) model currently used for benchmarking intensive care units (ICUs) in Australia and New Zealand utilises physiological data collected up to 24 hours after ICU admission to estimate the risk of hospital mortality. This study aimed to develop the Australian and New Zealand Risk of Death admission (ANZROD₀) model to predict hospital mortality using data available at presentation to ICU and compare its performance with the ANZROD in Australian and New Zealand hospitals. Data pertaining to all ICU admissions between 1 January 2006 and 31 December 2015 were extracted from the Australian and New Zealand Intensive Care Society Adult Patient Database. Hospital mortality was modelled using logistic regression with development (two-thirds) and validation (one-third) datasets. All predictor variables available at ICU admission were considered for inclusion in the ANZROD₀ model. Model performance was assessed using Brier score, standardised mortality ratio and area under the receiver operating characteristic curve. The relationship between ANZROD₀ and ANZROD predicted risk of death was assessed using linear regression. After standard exclusions, 1,097,416 patients were available for model development and validation. Observed mortality was 9.5%. Model performance measures (Brier score, standardised mortality ratio and area under the receiver operating characteristic curve) for the ANZROD₀ and ANZROD in the validation dataset were 0.069, 1.0 and 0.853; 0.057, 1.0 and 0.909, respectively. There was a strong positive correlation between the mortality predictions with an overall R² of 0.73. We found that the ANZROD₀ model had acceptable calibration and discrimination. Predictions from the models had high correlations in all major diagnostic groups, with the exception of cardiac surgery and possibly trauma and sepsis.

How do probiotics and prebiotics function at distant sites?

The realisation that microbes regarded as beneficial to the host can impart effects at sites distant from their habitat, has raised many possibilities for treatment of diseases. The objective of a workshop hosted in Turku, Finland, by the International Scientific Association for Probiotics and Prebiotics, was to assess the evidence for these effects and the extent to which early life microbiome programming influences how the gut microbiota communicates with distant sites. In addition, we examined how probiotics and prebiotics might affect the skin, airways, heart, brain and metabolism. The growing levels of scientific and clinical evidence showing how microbes influence the physiology of many body sites, leads us to call for more funding to advance a potentially exciting avenue for novel therapies for many chronic diseases.