Search

The emergency department (ED) is a fast-paced environment responsible for large volumes of patients with varied disease acuity. Operational pressures on EDs are increasing, which creates the imperative to efficiently identify patients at imminent risk of acute deterioration. The aim of this study is to systematically compare the performance of machine learning algorithms based on logistic regression, gradient boosted decision trees, and support vector machines for predicting imminent clinical deterioration for patients based on cross-sectional patient data extracted from electronic patient records (EPR) at the point of entry to the hospital. We apply state-of-the-art machine learning methods to predict early patient deterioration, based on their first recorded vital signs, observations, laboratory results, and other predictors documented in the EPR. Clinical deterioration in this study is measured by in-hospital mortality and/or admission to critical care. We build on prior work by incorporating interpretable machine learning and fairness-aware modelling, and use a dataset comprising 118, 886 unplanned admissions to Salford Royal Hospital, UK, to systematically compare model variations for predicting mortality and critical care utilisation within 24 hours of admission. We compare model performance to the National Early Warning Score 2 (NEWS2) and yield up to a 0.366 increase in average precision, up to a 21.16 % reduction in daily alert rate, and a median 0.599 reduction in differential bias amplification across the protected demographics of age and sex. We use Shapely Additive exPlanations to justify the models' outputs, verify that the captured data associations align with domain knowledge, and pair predictions with the causal context of each patient's most influential characteristics. Introducing our modelling to clinical practice has the potential to reduce alert fatigue and identify high-risk patients with a lower NEWS2 that might be missed currently, but further work is needed to trial the models in clinical practice. We encourage future research to follow a systematised approach to data-driven risk modelling to obtain clinically applicable support tools. [ABSTRACT FROM AUTHOR]

The emergency department (ED) is a fast-paced environment responsible for large volumes of patients with varied disease acuity. Operational pressures on EDs are increasing, which creates the imperative to efficiently identify patients at imminent risk of acute deterioration. The aim of this study is to systematically compare the performance of machine learning algorithms based on logistic regression, gradient boosted decision trees, and support vector machines for predicting imminent clinical deterioration for patients based on cross-sectional patient data extracted from electronic patient records (EPR) at the point of entry to the hospital. We apply state-of-the-art machine learning methods to predict early patient deterioration, based on their first recorded vital signs, observations, laboratory results, and other predictors documented in the EPR. Clinical deterioration in this study is measured by in-hospital mortality and/or admission to critical care. We build on prior work by incorporating interpretable machine learning and fairness-aware modelling, and use a dataset comprising 118, 886 unplanned admissions to Salford Royal Hospital, UK, to systematically compare model variations for predicting mortality and critical care utilisation within 24 hours of admission. We compare model performance to the National Early Warning Score 2 (NEWS2) and yield up to a 0.366 increase in average precision, up to a 21.16 % reduction in daily alert rate, and a median 0.599 reduction in differential bias amplification across the protected demographics of age and sex. We use Shapely Additive exPlanations to justify the models' outputs, verify that the captured data associations align with domain knowledge, and pair predictions with the causal context of each patient's most influential characteristics. Introducing our modelling to clinical practice has the potential to reduce alert fatigue and identify high-risk patients with a lower NEWS2 that might be missed currently, but further work is needed to trial the models in clinical practice. We encourage future research to follow a systematised approach to data-driven risk modelling to obtain clinically applicable support tools. [ABSTRACT FROM AUTHOR]

Right from the beginning, attendance has played an important role in the education systems, not only in student success but in the overall interest of the matter. Although all schools try to accentuate good attendance, still some schools find it hard to achieve the required level (96% in UK) of average attendance. The most productive way of increasing the pupils′ attendance rate is to predict when it is going to go down, understand the reasons—why it happened—and act on the affecting factors so as to prevent it. Artificial intelligence (AI) is an automated machine learning solution for different types of problems. Several machine learning (ML) models like logistic regression, decision trees, etc. are easy to understand; however, complicated (Neural Network, BART etc.) ML models are not transparent but are black-boxes for humans. It is not always evident how machine intelligence arrived at a decision. However, not always, but in critical applications it is important that humans can understand the reasons for such decisions. In this paper, we present a methodology on the application example of pupil attendance for constructing explanations for AI classification algorithms. The methodology includes building a model of the application in the Isabelle Insider and Infrastructure framework (IIIf) and an algorithm (PCR) that helps us to obtain a detailed logical rule to specify the performance of the black-box algorithm, hence allowing us to explain it. The explanation is provided within the logical model of the IIIf, thus is suitable for human audiences. It has been shown that the RR-cycle of IIIf can be adapted to provide a method for iteratively extracting an explanation by interleaving attack tree analysis with precondition refinement, which finally yields a general rule that describes the decision taken by a black-box algorithm produced by Artificial intelligence. [ABSTRACT FROM AUTHOR]

Arterial stiffness, measured by arterial stiffness index (ASI), could be considered a main denominator in target organ damage among hypertensive subjects. Currently, no reported ASI normal references have been reported. The index of arterial stiffness is evaluated by calculation of a stiffness index. Predicted ASI can be estimated regardless to age, sex, mean blood pressure, and heart rate, to compose an individual stiffness index [(measured ASI–predicted ASI)/predicted ASI]. A stiffness index greater than zero defines arterial stiffness. Thus, the purpose of this study was 1) to determine determinants of stiffness index 2) to perform threshold values to discriminate stiffness index and then 3) to determine hierarchical associations of the determinants by performing a decision tree model among hypertensive participants without CV diseases. A study was conducted from 53,363 healthy participants in the UK Biobank survey to determine predicted ASI. Stiffness index was applied on 49,452 hypertensives without CV diseases to discriminate determinants of positive stiffness index (N = 22,453) from negative index (N = 26,999). The input variables for the models were clinical and biological parameters. The independent classifiers were ranked from the most sensitives: HDL cholesterol≤1.425 mmol/L, smoking pack years≥9.2pack-years, Phosphate≥1.172 mmol/L, to the most specifics: Cystatin c≤0.901 mg/L, Triglycerides≥1.487 mmol/L, Urate≥291.9 μmol/L, ALT≥22.13 U/L, AST≤32.5 U/L, Albumin≤45.92 g/L, Testosterone≥5.181 nmol/L. A decision tree model was performed to determine rules to highlight the different hierarchization and interactions between these classifiers with a higher performance than multiple logistic regression (p<0.001). The stiffness index could be an integrator of CV risk factors and participate in future CV risk management evaluations for preventive strategies. Decision trees can provide accurate and useful classification for clinicians. [ABSTRACT FROM AUTHOR]

• Postpartum haemorrhage (PPH) is a leading cause of morbidity and mortality worldwide. • Preventing blood loss where possible is an important public health concern. • There is a high cost of the PPH treatment pathway. • PPH Butterfly is highly likely to offer value for money in the UK NHS setting. • Extrapolation on an international scale to lower and income countries could prevent maternal mortality. To explore the cost-effectiveness of a novel PPH device as compared with usual care. A decision analytical model was used to explore the cost-effectiveness of the PPH Butterfly device compared with usual care. This was part of a United Kingdom, UK, clinical trial ISRCTN15452399 using a matched historical cohort who had standard PPH management without the use of the PPH Butterfly device. The economic evaluation was conducted from a UK National Health Service (NHS) perspective. Liverpool Women's Hospital, UK. 57 women with 113 matched controls. The PPH Butterfly is a novel device that has been invented and developed in the UK to facilitate bimanual compression of the uterus in the treatment of PPH. Main outcome measures included healthcare costs, blood loss, and maternal morbidity events. Mean treatment costs in the Butterfly cohort were £3,459.66 as compared with standard care £3,223.93. Treatment with the Butterfly device resulted in decreased total blood loss in comparison with standard care. The Butterfly device had an incremental cost-effectiveness ratio of £3,795.78 per PPH progression avoided (defined as ≤ 1000 ml additional blood loss from device insertion point). If the NHS is prepared to pay £8,500 per PPH progression avoided, then the Butterfly device is cost-effective with a probability of 87 percent. In the PPH Butterfly treatment arm there were 9% fewer cases of massive obstetric haemorrhage (severe PPH of more than 2000mls or more than 4 units of blood transfusion required) recorded as compared with the standard care historical cohort. As a low-cost device, the PPH Butterfly device is cost-effective but can be cost-saving to the NHS. The PPH pathway can result in high-cost resource use such as blood transfusion or high dependence unit hospital stays. The Butterfly device is a relative low-cost device in a UK NHS setting with a high probability of being cost-effective. The National Institute for Health and Care Excellence (NICE) can use this evidence in considering the adoption of innovative technologies such as the Butterfly device in the NHS. Extrapolation on an international scale to lower and middle-income countries could prevent mortality associated with PPH. [ABSTRACT FROM AUTHOR]

Previous studies have demonstrated hysterosalpingography (HSG) in general, and specifically with an oil-soluble contrast medium, directly increases pregnancy rates. Decision modelling was performed to compare fertility management using three HSG diagnostic strategies: (i) water-soluble contrast medium (WSCM)-HSG; (ii) Lipiodol® Ultra Fluid (LUF)-HSG; and (iii) No HSG, for women aged ≤39 years with unexplained infertility. Four reimbursement scenarios were modelled to reflect the various funding arrangements across the regions of the United Kingdom. Compared with WSCM-HSG, the live birth rates after 24 months increased by 3.4% with LUF-HSG and decreased by 2.7% with no HSG. From a patient perspective, fertility management with LUF-HSG is the most cost-effective strategy with cost-savings ranging from £299 to £857 per patient depending on the funding arrangement for IVF. From an NHS perspective, fertility management with LUF-HSG is cost-effective when 2 or more IVF cycles are NHS-funded. If none of the IVF cycles are NHS-funded, fertility management with LUF-HSG can be considered cost-effective if society is willing to pay £8,353 for an additional live birth. The findings from this analysis suggest that fertility management with WSCM-HSG is cost-effective compared to no HSG and LUF-HSG is the most cost-effective with increased live birth rates after 24 months. [ABSTRACT FROM AUTHOR]

Simple Summary: Oesophago-gastric cancer is one of the commonest cancers worldwide, yet it can be particularly difficult to diagnose given that initial symptoms are often non-specific and routine screening is not available. Cancer risk-assessment tools, which calculate cancer risk based on symptoms and other risk factors present in the primary care record, can aid decisions on referrals for cancer investigations, facilitating earlier diagnosis. Diagnosing common cancers earlier could help improve survival rates. Using UK primary care electronic health record data, we compared five different machine learning techniques for probabilistic classification of cancer patients against a current widely used UK primary care cancer risk-assessment tool. The machine learning algorithms outperformed the current risk-assessment tool, with a higher overall accuracy and an ability to reasonably identify 11–25% more cancer patients. We conclude that machine-learning-based risk-assessment tools could help better identify suitable patients for further investigation and support earlier diagnosis. Oesophago-gastric cancer is difficult to diagnose in the early stages given its typical non-specific initial manifestation. We hypothesise that machine learning can improve upon the diagnostic performance of current primary care risk-assessment tools by using advanced analytical techniques to exploit the wealth of evidence available in the electronic health record. We used a primary care electronic health record dataset derived from the UK General Practice Research Database (7471 cases; 32,877 controls) and developed five probabilistic machine learning classifiers: Support Vector Machine, Random Forest, Logistic Regression, Naïve Bayes, and Extreme Gradient Boosted Decision Trees. Features included basic demographics, symptoms, and lab test results. The Logistic Regression, Support Vector Machine, and Extreme Gradient Boosted Decision Tree models achieved the highest performance in terms of accuracy and AUROC (0.89 accuracy, 0.87 AUROC), outperforming a current UK oesophago-gastric cancer risk-assessment tool (ogRAT). Machine learning also identified more cancer patients than the ogRAT: 11.0% more with little to no effect on false positives, or up to 25.0% more with a slight increase in false positives (for Logistic Regression, results threshold-dependent). Feature contribution estimates and individual prediction explanations indicated clinical relevance. We conclude that machine learning could improve primary care cancer risk-assessment tools, potentially helping clinicians to identify additional cancer cases earlier. This could, in turn, improve survival outcomes. [ABSTRACT FROM AUTHOR]

Objective: To assess whether specific non-communicable diseases are associated with hearing aid adoption in older adults with hearing loss. Design: A cross-sectional, observational study. Study sample: Data was obtained from one of the largest pharmacy-led health and beauty retailers in the United Kingdom. In total, 17,172 older adults were included. Results: Greater odds of adopting hearing aids were associated with being older (Odds ratio [OR] = 1.03; 95% confidence interval [CI] = 1.02, 1.03), having fewer self-reported hearing difficulties (OR = 0.61; 95% CI = 0.58, 0.64), and greater audiometric hearing loss (OR = 1.68; 95% CI = 1.54, 1.83). The odds of adopting hearing aids were lower if individuals self-reported hypertension (OR = 0.87; 95% CI = 0.79, 0.97) and diabetes (OR = 0.83; 95% CI = 0.72, 0.95). Using a decision tree model, self-reported hearing difficulties, audiometric hearing loss, and age were the best combination of variables to differentiate between individuals that did and did not adopt hearing aids. Conclusions: This study demonstrates that hearing aid adoption is lower in older adults with cardiovascular disease risk factors. Thus, there is a need for healthcare professionals to identify older adults living with hearing loss and cardiovascular ill-health, ensuring that they receive appropriate patient-centred support to manage their health. [ABSTRACT FROM AUTHOR]

We present a simple and efficient hypothesis-free machine learning pipeline for risk factor discovery that accounts for non-linearity and interaction in large biomedical databases with minimal variable pre-processing. In this study, mortality models were built using gradient boosting decision trees (GBDT) and important predictors were identified using a Shapley values-based feature attribution method, SHAP values. Cox models controlled for false discovery rate were used for confounder adjustment, interpretability, and further validation. The pipeline was tested using information from 502,506 UK Biobank participants, aged 37–73 years at recruitment and followed over seven years for mortality registrations. From the 11,639 predictors included in GBDT, 193 potential risk factors had SHAP values ≥ 0.05, passed the correlation test, and were selected for further modelling. Of the total variable importance summed up, 60% was directly health related, and baseline characteristics, sociodemographics, and lifestyle factors each contributed about 10%. Cox models adjusted for baseline characteristics, showed evidence for an association with mortality for 166 out of the 193 predictors. These included mostly well-known risk factors (e.g., age, sex, ethnicity, education, material deprivation, smoking, physical activity, self-rated health, BMI, and many disease outcomes). For 19 predictors we saw evidence for an association in the unadjusted but not adjusted analyses, suggesting bias by confounding. Our GBDT-SHAP pipeline was able to identify relevant predictors 'hidden' within thousands of variables, providing an efficient and pragmatic solution for the first stage of hypothesis free risk factor identification. [ABSTRACT FROM AUTHOR]

Purpose: Food bolus and oesophageal foreign bodies are a common presentation that may be managed by otolaryngologists, gastroenterologists, acute medicine physicians and accident and emergency. The condition is highly variable with presentations ranging from well patients whose obstruction spontaneously passes to peri-arrest with severe aspiration or impending airway compromise. Management of this condition is heterogeneous and often depends on the specialty the patient is originally admitted under. There exist European and American guidelines from the perspective of gastroenterology, but there are no UK-based guidelines and limited consideration of the role of the otolaryngologists and rigid oesophagoscopy. Methods: An extensive literature search was carried out to generate conclusions on key management questions for food bolus and oesophageal foreign bodies. This was then summarised into both a written summary of the evidence and a graphical decision tree. Results: This paper is a review article and presents conclusions regarding management options for food bolus and oesophageal foreign bodies. Conclusion: This article considers the current evidence surrounding investigation and management of oesophageal food bolus and foreign body. It draws conclusions regarding presentation, investigation and subsequent operative treatment. As part of this process, we propose a graphical decision tree to assist in management decisions. [ABSTRACT FROM AUTHOR]

Brain–computer interface (BCI)-controlled assistive robotic systems have been developed with increasing success with the aim to rehabilitation of patients after brain injury to increase independence and quality of life. While such systems may use surgically implanted invasive sensors, non-invasive alternatives can be better suited due to the ease of use, reduced cost, improvements in accuracy and reliability with the advancement of the technology and practicality of use. The consumer-grade BCI devices are often capable of integrating multiple types of signals, including Electroencephalogram (EEG) and Electromyogram (EMG) signals. This paper summarizes the development of a portable and cost-efficient BCI-controlled assistive technology using a non-invasive BCI headset "OpenBCI" and an open source robotic arm, U-Arm, to accomplish tasks related to rehabilitation, such as access to resources, adaptability or home use. The resulting system used a combination of EEG and EMG sensor readings to control the arm. To avoid risks of injury while the device is being used in clinical settings, appropriate measures were incorporated into the software control of the arm. A short survey was used following the system usability scale (SUS), to measure the usability of the technology to be trialed in clinical settings. From the experimental results, it was found that EMG is a very reliable method for assistive technology control, provided that the user specific EMG calibration is done. With the EEG, even though the results were promising, due to insufficient detection of the signal, the controller was not adequate to be used within a neurorehabilitation environment. The survey indicated that the usability of the system is not a barrier for moving the system into clinical trials. For the rehabilitation of patients suffering from neurological disabilities (particularly those suffering from varying degrees of paralysis), it is necessary to develop technology that bypasses the limitations of their condition. For example, if a patient is unable to walk due to the unresponsiveness in their motor neurons, technology can be developed that used an alternate input to move an exoskeleton, which enables the patient to walk again with the assistance of the exoskeleton. This research focuses on neuro-rehabilitation within the framework of the NHS at the Kent and Canterbury Hospital in UK. The hospital currently does not have any system in place for self-driven rehabilitation and instead relies on traditional rehabilitation methods through assistance from physicians and exercise regimens to maintain muscle movement. This paper summarises the development of a portable and cost-efficient BCI controlled assistive technology using a non-invasive BCI headset "OpenBCI" and an open source robotic arm, U-Arm, to accomplish tasks related to rehabilitation, such as access to resources, adaptability or home use. The resulting system used a combination of EEG and EMG sensor readings to control the arm, which could perform a number of different tasks such as picking/placing objects or assist users in eating. [ABSTRACT FROM AUTHOR]

Objective: Decision-tree methods are machine-learning methods which provide results that are relatively easy to interpret and apply by human decision makers. The resulting decision trees show how baseline patient characteristics can be combined to predict treatment outcomes for individual patients, for example. This paper introduces GLMM trees, a decision-tree method for multilevel and longitudinal data. Method: To illustrate, we apply GLMM trees to a dataset of 3,256 young people (mean age 11.33, 48% girls) receiving treatment at one of several mental-health service providers in the UK. Two treatment outcomes (mental-health difficulties scores corrected for baseline) were regressed on 18 demographic, case and severity characteristics at baseline. We compared the performance of GLMM trees with that of traditional GLMMs and random forests. Results: GLMM trees yielded modest predictive accuracy, with cross-validated multiple R values of.18 and.25. Predictive accuracy did not differ significantly from that of traditional GLMMs and random forests, while GLMM trees required evaluation of a lower number of variables. Conclusion: GLMM trees provide a useful data-analytic tool for clinical prediction problems. The supplemental material provides a tutorial for replicating the GLMM tree analyses in R. [ABSTRACT FROM AUTHOR]

Background: Alcohol consumption is influenced by the characteristics of drinking occasions, for example, location, timing, or the composition of the drinking group. However, the relative importance of occasion characteristics is not yet well understood. This study aims to identify which characteristics, and combinations of characteristics, are associated with units consumed within drinking occasions. It also tests whether accounting for occasion characteristics improves the prediction of consumption compared to using demographic information only. Methods: The data come from a cross‐sectional, nationally representative, online market research survey. Our sample includes 18,409 British drinkers aged 18 + who recorded the characteristics of 46,072 drinking occasions using 7‐day retrospective drinking diaries in 2018. We used decision tree modeling and nested linear regression to predict units consumed in occasions using information on drinking location/venue, occasion timing, company, occasion type (e.g., a quiet night in), occasion motivation, drink type and packaging, food eaten and entertainment/ other activities during the occasion. We estimated models separately for 6 age‐sex groups and controlled for usual drinking frequency, and social grade in nested linear regression models. Open Science Framework preregistration: https://osf.io/42epd. Results: Our 6 final models accounted for between 55% and 71% of the variance in drinking occasion alcohol consumption. Beyond demographic characteristics (1 to 9%) and occasion duration (24 to 60%), further occasion characteristics and combinations of characteristics accounted for 31 to 70% of the total explained variance. The characteristics most strongly associated with heavy alcohol consumption were long occasion duration, drinking spirits as doubles, and drinking wine. Spirits were also consumed in light occasions, but as singles. This suggests that the serving size is an important differentiator of light and heavy occasions. Conclusions: Combinations of occasion duration and drink type are strongly predictive of alcohol consumption in adults' drinking occasions. Accounting for characteristics of drinking occasions, both individually and in combination, substantially improves the prediction of alcohol consumption. [ABSTRACT FROM AUTHOR]

Traffic safety has always been an important issue in sustainable transportation development, and the prediction of traffic accident severity remains a crucial challenging issue in the domain of traffic safety. A huge variety of forecasting models have been proposed to meet this challenge. These models gradually evolved from linear to nonlinear forms and from traditional statistical regression models to current popular machine learning models. Recently, a machine learning algorithm called Deep Forests based on the decision tree ensemble has aroused widespread concern, which was proposed for the first time by a research team of Nanjing University. This algorithm was proved to be more accurate and robust in comparison with other machine learning algorithms. Motivated by this benefit, this study employs the UK road safety dataset to propose a novel method for predicting the severity of traffic accidents based on the Deep Forests algorithm. To verify the superiority of our proposed method, several other machine learning algorithm-based perdition models were implemented to predict traffic accident severity with the same dataset, and the prediction results show that the Deep Forests algorithm present good stability, fewer hyper-parameters, and the highest accuracy under different level of training data volume. It is expected that the findings from this study would be helpful for the establishment or improvement of effective traffic safety system within a sustainable transportation system, which is of great significance for helping government managers to establish timely proactive strategies in traffic accident prevention and effectively improve road traffic safety. [ABSTRACT FROM AUTHOR]

Major depressive disorder (MDD) has been the subject of many neuroimaging case–control classification studies. Although some studies report accuracies ≥80%, most have investigated relatively small samples of clinically‐ascertained, currently symptomatic cases, and did not attempt replication in larger samples. We here first aimed to replicate previously reported classification accuracies in a small, well‐phenotyped community‐based group of current MDD cases with clinical interview‐based diagnoses (from STratifying Resilience and Depression Longitudinally cohort, 'STRADL'). We performed a set of exploratory predictive classification analyses with measures related to brain morphometry and white matter integrity. We applied three classifier types—SVM, penalised logistic regression or decision tree—either with or without optimisation, and with or without feature selection. We then determined whether similar accuracies could be replicated in a larger independent population‐based sample with self‐reported current depression (UK Biobank cohort). Additional analyses extended to lifetime MDD diagnoses—remitted MDD in STRADL, and lifetime‐experienced MDD in UK Biobank. The highest cross‐validation accuracy (75%) was achieved in the initial current MDD sample with a decision tree classifier and cortical surface area features. The most frequently selected decision tree split variables included surface areas of bilateral caudal anterior cingulate, left lingual gyrus, left superior frontal, right precentral and paracentral regions. High accuracy was not achieved in the larger samples with self‐reported current depression (53.73%), with remitted MDD (57.48%), or with lifetime‐experienced MDD (52.68–60.29%). Our results indicate that high predictive classification accuracies may not immediately translate to larger samples with broader criteria for depression, and may not be robust across different classification approaches. [ABSTRACT FROM AUTHOR]

The clinical utility of the 70-gene signature (MammaPrint®) to guide chemotherapy use in T1-3N0-1M0 breast cancer was demonstrated in the Microarray in Node-Negative and 1 to 3 Positive Lymph Node Disease May Avoid Chemotherapy (MINDACT) study. One thousand four ninety seven of 3356 (46.2%) enrolled patients with high clinical risk (in accordance with the modified Adjuvant! Online clinical-pathological assessment) had a low-risk 70-gene signature. Using patient-level data from the MINDACT trial, the cost-effectiveness of using the 70-gene signature to guide adjuvant chemotherapy selection for clinical high risk, estrogen receptor positive (ER+), human epidermal growth factor 2 negative (HER2-) patients was analysed. A hybrid decision tree-Markov model simulated treatment strategies in accordance with the 70-gene signature with clinical assessment versus clinical assessment alone, over a 10-year time horizon. Primary outcomes were quality-adjusted life years (QALYs), country-specific costs and incremental cost-effectiveness ratios (ICERs) for six countries: Belgium, France, Germany, Netherlands, UK and the US. Treatment strategies guided by the 70-gene signature result in more QALYs compared with clinical assessment alone. Costs of the 70-gene signature strategy were lower in five of six countries. This led to dominance of the 70-gene signature in Belgium, France, Germany, Netherlands and the US and to a cost-effective situation in the UK (ICER £22,910/QALY). Annual national cost savings were €4.2M (Belgium), €24.7M (France), €45.1M (Germany), €12.7M (Netherlands) and $244M (US). UK budget increase was £8.4M. Using the 70-gene signature to safely guide chemotherapy de-escalation in clinical high risk patients with ER+/HER2- tumours is cost-effective compared with using clinical assessment alone. Long-term follow-up and outcomes from the MINDACT trial are necessary to address uncertainties in model inputs. • The 70-gene signature is cost-effective compared with clinical assessment alone • Minimal survival differences are balanced against quality of life gains • Results are consistent amongst the five studied countries [ABSTRACT FROM AUTHOR]

Objective: The M6 risk-prediction model was published as part of a two-step protocol using an initial progesterone level of ≤ 2 nmol/L to identify probable failing pregnancies (Step 1) followed by the M6 model (Step 2). The M6 model has been shown to have good triage performance for stratifying women with a pregnancy of unknown location (PUL) as being at low or high risk of harboring an ectopic pregnancy (EP). This study validated the triage performance of the two-step protocol in clinical practice by evaluating the number of protocol-related adverse events and how effectively patients were triaged.Methods: This was a prospective multicenter interventional study of 3272 women with a PUL, carried out between January 2015 and January 2017 in four district general hospitals and four university teaching hospitals in the UK. The final pregnancy outcome was defined as: a failed PUL (FPUL), an intrauterine pregnancy (IUP) or an EP (including persistent PUL (PPUL)). FPUL and IUP were grouped as low-risk and EP/PPUL as high-risk PUL. Serum progesterone and human chorionic gonadotropin (hCG) levels were measured at presentation in all patients. If the initial progesterone level was ≤ 2 nmol/L, patients were discharged and were asked to have a follow-up urine pregnancy test in 2 weeks to confirm a negative result. If the progesterone level was > 2 nmol/L or a measurement had not been taken, hCG level was measured again at 48 h and results were entered into the M6 model. Patients were managed according to the outcome predicted by the protocol. Those classified as 'low risk, probable FPUL' were advised to perform a urine pregnancy test in 2 weeks and those classified as 'low risk, probable IUP' were invited for a scan a week later. When a woman with a PUL was classified as high risk (i.e. risk of EP ≥ 5%) she was reviewed clinically within 48 h. One center used a progesterone cut-off of ≤ 10 nmol/L and its data were analyzed separately. If the recommended management protocol was not adhered to, this was recorded as a protocol deviation and classified as: unscheduled visit for clinician reason, unscheduled visit for patient reason or incorrect timing of blood test or ultrasound scan. The classifications outlined in the UK Good Clinical Practice (GCP) guidelines were used to evaluate the incidence of adverse events. Data were analyzed using descriptive statistics.Results: Of the 3272 women with a PUL, 2625 were included in the final analysis (317 met the exclusion criteria or were lost to follow-up, while 330 were evaluated using a progesterone cut-off of ≤ 10 nmol/L). Initial progesterone results were available for 2392 (91.1%) patients. In Step 1, 407 (15.5%) patients were classified as low risk (progesterone ≤ 2 nmol/L), of whom seven (1.7%) were ultimately diagnosed with an EP. In 279 of the remaining 2218 women with a PUL, the M6 model was not applied owing to protocol deviation or because the outcome was already known (usually on the basis of an ultrasound scan) before a second hCG reading was taken; of these patients, 30 were diagnosed with an EP. In Step 2, 1038 women with a PUL were classified as low risk, of whom eight (0.8%) had a final outcome of EP. Of 901 women classified as high risk at Step 2, 275 (30.5%) had an EP. Therefore, 275/320 (85.9%) EPs were correctly classified as high risk. Overall, 1445/2625 PUL (55.0%) were classified as low risk, of which 15 (1.0%) were EP. None of these cases resulted in a ruptured EP or significant clinical harm. Sixty-two women participating in the study had an adverse event, but no woman had a serious adverse event as defined in the UK GCP guidelines.Conclusions: This study has shown that the two-step protocol incorporating the M6 model effectively triaged the majority of women with a PUL as being at low risk of an EP, minimizing the follow-up required for these patients after just two visits. There were few misclassified EPs and none of these women came to significant clinical harm or suffered a serious adverse clinical event. The two-step protocol incorporating the M6 model is an effective and clinically safe way of rationalizing the management of women with a PUL. Copyright © 2019 ISUOG. Published by John Wiley & Sons Ltd. [ABSTRACT FROM AUTHOR]

Critical care admission may be necessary for surgical patients requiring organ support or invasive monitoring in the peri-operative period. Unplanned critical care admission poses a potential risk to patients and pressure on services. Existing guidelines base admission criteria on predicted risk of 30-day mortality; however, this may not provide the best predictor of which patients would benefit from this service, and how unplanned admission might be avoided. A systematic review of MEDLINE, Embase, CINAHL, Web of Science, the Cochrane database and the grey literature identified 44 studies assessing risk factors for unplanned critical care admission in adult populations undergoing non-cardiac, non-thoracic and non-neurological surgery. Comparative, quantitative analysis of the admission criteria was not feasible due to heterogeneity in study design. Age, anaemia, ASA physical status, body mass index, comorbidity burden, emergency surgery, high-risk surgery, male sex, obstructive sleep apnoea, increased blood loss and operative duration were all independent risk factors for unplanned critical care admission. Age, body mass index, comorbidity extent and emergency surgery were the most common independent risk factors identified in the USA, UK, Asia and Australia. These risk factors could be used in the development of a risk tool or decision tree for determining which patients might benefit from planned critical care admission. Future work should involve testing the sensitivity and specificity of these measures, either alone or in combination, to guide planned critical care admission, reduce patient deterioration and unplanned admissions. [ABSTRACT FROM AUTHOR]

This paper treats with a decision tree consisting of RBF neural networks in the nodes to decrease the classification time and to improve accuracy as well as generalization. We propose two knowledge transferring mechanisms between nodes to reduce the duplicate computations in training process. The resulted classifier is titled as ensemble of RBF neural networks in decision tree structure with knowledge transferring or shortly ERDK. We accelerate ERDK by applying a cut-point mechanism to prune its tree structure. The results on a great number of benchmark datasets show that ERDK provides promising results. Particularly, AUCarea, F-measure, and G-mean for ERDK are 95, 90, and 90%, respectively. As a case study, we finally apply ERDK on Britain incident dataset. [ABSTRACT FROM AUTHOR]

Research has shown online learners' performance to have a strong association with their demographic characteristics, such as regional belonging, socio-economic standing, education level, age, gender, and disability status. Despite a growing number of studies exploring factors for successful online learning outcomes, most researchers have utilised one or a combination of very few learner characteristics. Moreover, a limited number of studies scrutinised the impact of individual characteristics on learning outcomes as learners progress in a course. The current research aims to explore the dynamic impact of demographic characteristics on academic outcomes in the online learning environment. We investigated and compared the dynamic influence of six demographic characteristics on online learning outcomes using a sample of 8581 UK based learners across four Open University online courses from four different disciplines. We found region, neighborhood poverty level, and prior education respectively, to be strong predictors of overall learning outcomes. However, at a fine-grain level, such influence varied temporally as the course progressed, as well as between different courses. To conclude with, we discussed the implications for institutional support on adopting a tailored approach towards a more personalised student support system. • In Decision trees, some predictive variables were more contributing than others. • Region, multiple deprivation, and education mostly contributed in tree induction. • Linkage between learning outcome and gender or disability was trivial. • Precision and recall for each class calculated separately (accuracy upto 83.14%). • Evidently, predictive models were biased towards majority class (Pass in this case). [ABSTRACT FROM AUTHOR]

Objectives: To calculate fractional inspired oxygen concentration (FiO2) thresholds in ward patients and add these to the National Early Warning Score (NEWS). To evaluate the performance of NEWS-FiO2 against NEWS when predicting in-hospital death and unplanned intensive care unit (ICU) admission.Methods: A multi-centre, retrospective, observational cohort study was carried out in five hospitals from two UK NHS Trusts. Adult admissions with at least one complete set of vital sign observations recorded electronically were eligible. The primary outcome measure was an 'adverse event' which comprised either in-hospital death or unplanned ICU admission. Discrimination was assessed using the Area Under the Receiver Operating Characteristic curve (AUROC).Results: A cohort of 83,304 patients from a total of 271,363 adult admissions were prescribed oxygen. In this cohort, NEWS-FiO2 (AUROC 0.823, 95% CI 0.819-0.824) outperformed NEWS (AUORC 0.811, 95% CI 0.809-0.814) when predicting in-hospital death or unplanned ICU admission within 24 h of a complete set of vital sign observations.Conclusions: NEWS-FiO2 generates a performance gain over NEWS when studied in ward patients requiring oxygen. This warrants further study, particularly in patients with respiratory disorders. [ABSTRACT FROM AUTHOR]

The article discusses a study that examined the attributes of fathers who receive social worker contact using a sample of individuals from a British cohort study, the Avon Longitudinal Study of Parents and Children. Study highlights include the use of classification tree models to predict whether fathers or mothers are more likely to receive social work contact, events associated with social work contact for mothers and fathers, and interaction between age and sex captured by the decision tree.

A key focus of the racing industry is to reduce the number of race-day events where horses die suddenly or are euthanased due to catastrophic injury. The objective of this study was therefore to determine risk factors for race-day fatalities in Thoroughbred racehorses, using a cohort of all horses participating in flat racing in Great Britain between 2000 and 2013. Horse-, race- and course-level data were collected and combined with all race-day fatalities, recorded by racecourse veterinarians in a central database. Associations between exposure variables and fatality were assessed using logistic regression analyses for (1) all starts in the dataset and (2) starts made on turf surfaces only. There were 806,764 starts in total, of which 548,571 were on turf surfaces. A total of 610 fatalities were recorded; 377 (61.8%) on turf. In both regression models, increased firmness of the going, increasing racing distance, increasing average horse performance, first year of racing and wearing eye cover for the first time all increased the odds of fatality. Generally, the odds of fatality also increased with increasing horse age whereas increasing number of previous starts reduced fatality odds. In the ‘all starts’ model, horses racing in an auction race were at 1.46 (95% confidence interval (CI) 1.06–2.01) times the odds of fatality compared with horses not racing in this race type. In the turf starts model, horses racing in Group 1 races were at 3.19 (95% CI 1.71–5.93) times the odds of fatality compared with horses not racing in this race type. Identification of novel risk factors including wearing eye cover and race type will help to inform strategies to further reduce the rate of fatality in flat racing horses, enhancing horse and jockey welfare and safety. [ABSTRACT FROM AUTHOR]

Objective The aim of the present study was to develop a toolkit combining various risk factors to predict the risk of developing a postpartum hemorrhage (PPH) during a cesarean delivery. Study Design A retrospective cohort study of 24,230 women who had cesarean delivery between January 2003 and December 2013 at a tertiary care teaching hospital within the United Kingdom serving a multiethnic population. Data were extracted from hospital databases, and risk factors for PPH were identified. Hothorn et al recursive partitioning algorithm was used to infer a conditional decision tree. For each of the identified combinations of risk factors, two probabilities were calculated: the probability of a patient producing ≥1,000 and ≥ 2,000 mL blood loss. Results The Leicester PPH predict score was then tested on the randomly selected remaining 25% (n = 6,095) of the data for internal validity. Reliability testing showed an intraclass correlation of 0.98 and mean absolute error of 239.8 mL with the actual outcome. Conclusion The proposed toolkit enables clinicians to predict the risk of postpartum hemorrhage. As a result, preventative measures for postpartum hemorrhage could be undertaken. Further external validation of the current toolkit is required. [ABSTRACT FROM AUTHOR]

Objective: To determine the economic impact of the introduction of carbetocin for the prevention of postpartum haemorrhage (PPH) at caesarean section, compared to oxytocin.Study Design: The model is a decision tree conducted from a UK National Health Service perspective. 1500 caesarean sections (both elective and emergency) were modelled over a 12 month period. Efficacy data was taken from a published Cochrane meta-analysis, and costs from NHS Reference costs, the British National Formulary and the NHS electronic Medicines Information Tool. A combination of hospital audit data and expert input from an advisory board of clinicians was used to inform resource use estimates. The main outcome measures were the incidence of PPH and total cost over a one year time horizon, as a result of using carbetocin compared to oxytocin for prevention of PPH at caesarean section.Results: The use of carbetocin compared to oxytocin for prevention of PPH at caesarean section was associated with a reduction of 30 (88 vs 58) PPH events (>500ml blood loss), and a cost saving of £27,518. In probabilistic sensitivity analysis, carbetocin had a 91.5% probability of producing better outcomes, and a 69.4% chance of being dominant (both cheaper and more effective) compared to oxytocin.Conclusion: At list price, the introduction of carbetocin appears to provide improved clinical outcomes along with cost savings, though this is subject to uncertainty regarding the underlying data in efficacy, resource use, and cost. [ABSTRACT FROM AUTHOR]

Objectives. To identify and quantify resource required and associated costs for implementing TNF-α inhibitor (TNFi) drug level and anti-drug antibody (ADAb) tests in UK rheumatology practice. Methods. A microcosting study, assuming the UK National Health Service perspective, identified the direct medical costs associated with providing TNFi drug level and ADAb testing in clinical practice. Resource use and costs per patient were identified via four stages: identification of a patient pathway with resource implications; estimation of the resources required; identification of the cost per unit of resource (2015 prices); and calculation of the total costs per patient. Univariate and multiway sensitivity analyses were performed using the variation in resource use and unit costs. Results. Total costs for TNFi drug level and concurrent ADAb testing, assessed using ELISAs on trough serum levels, were £152.52/patient (range: £147.68-159.24) if 40 patient samples were tested simultaneously. For the base-case analysis, the pre-testing phase incurred the highest costs, which included booking an additional appointment to acquire trough blood samples. The additional appointment was the key driver of costs per patient (67% of the total cost), and labour accounted for 10% and consumables 23% of the total costs. Performing ELISAs once per patient (rather than in duplicate) reduced the total costs to £133.78/patient. Conclusion. This microcosting study is the first assessing the cost of TNFi drug level and ADAb testing. The results could be used in subsequent cost-effectiveness analyses of TNFi pharmacological tests to target treatments and inform future policy recommendations. [ABSTRACT FROM AUTHOR]

Objective: Many dietary assessment methods attempt to estimate total food and nutrient intake. If the intention is simply to determine whether participants achieve dietary recommendations, this leads to much redundant data. We used data mining techniques to explore the number of foods that intake information was required on to accurately predict achievement, or not, of key dietary recommendations. Design: We built decision trees for achievement of recommendations for fruit and vegetables, sodium, fat, saturated fat and free sugars using data from a national dietary surveillance data set. Decision trees describe complex relationships between potential predictor variables (age, sex and all foods listed in the database) and outcome variables (achievement of each of the recommendations). Setting: UK National Diet and Nutrition Survey (NDNS, 2008-12). Subjects: The analysis included 4156 individuals. Results: Information on consumption of 113 out of 3911 (3%) foods, plus age and sex was required to accurately categorize individuals according to all five recommendations. The best trade-off between decision tree accuracy and number of foods included occurred at between eleven (for fruit and vegetables) and thirty-two (for fat, plus age) foods, achieving an accuracy of 72% (for fat) to 83% (for fruit and vegetables), with similar values for sensitivity and specificity. Conclusions: Using information on intake of 113 foods, it is possible to predict with 72-83% accuracy whether individuals achieve key dietary recommendations. Substantial further research is required to make use of these findings for dietary assessment. [ABSTRACT FROM AUTHOR]

With a growing awareness that hoarding can be found in many psychiatric conditions, clinicians may recognise an increase in referrals for people presenting with hoarding. Here, the authors present three cases of hoarding behaviour stemming from different psychiatric conditions and discuss how hoarding can be managed. [ABSTRACT FROM AUTHOR]

Adults with autism face high rates of unemployment. Supported employment enables individuals with autism to secure and maintain a paid job in a regular work environment. The objective of this study was to assess the cost-effectiveness of supported employment compared with standard care (day services) for adults with autism in the United Kingdom. Thus, a decision-analytic economic model was developed, which used outcome data from the only trial that has evaluated supported employment for adults with autism in the United Kingdom. The main analysis considered intervention costs, while cost-savings associated with changes in accommodation status and National Health Service and personal social service resource use were examined in secondary analyses. Two outcome measures were used: the number of weeks in employment and the quality-adjusted life year. Supported employment resulted in better outcomes compared with standard care, at an extra cost of £18 per additional week in employment or £5600 per quality-adjusted life year. In secondary analyses that incorporated potential cost-savings, supported employment dominated standard care (i.e. it produced better outcomes at a lower total cost). The analysis suggests that supported employment schemes for adults with autism in the United Kingdom are cost-effective compared with standard care. Further research needs to confirm these findings. [ABSTRACT FROM PUBLISHER]

The administration of sedative drugs at the end of life raises both clinical and ethical challenges. This article details the evolution of sedation decision-making practices at a 14-bedded UK hospice over the course of 5 years through an initial review and two follow-up audits. Key areas, such as documented consideration of hydration status and discussion with family members, have been improved following the implementation of a checklist of decision-making prompts, demonstrating how practice can be improved and sustained over time. [ABSTRACT FROM AUTHOR]

Hens MJ, Alonso-Ferreira V, Villaverde-Hueso A, Abaitua I, Posada de la Paz M. Cost-effectiveness analysis of burning mouth syndrome therapy. Community Dent Oral Epidemiol 2012; 40: 185-192. © 2011 John Wiley & Sons A/S Abstract - Objective: To study the cost-effectiveness of four alternative treatments for burning mouth syndrome (BMS). Methods: A cost-effectiveness analysis was conducted from a healthcare payer perspective of four therapy strategies (amisulpride, paroxetine, sertraline and topical clonazepam), using a decision-tree model that incorporated direct healthcare costs and probabilities associated with the possible events and outcomes. Average cost-effectiveness and incremental cost-effectiveness ratios were calculated. Sensitivity analyses included the costs of brand name and generic drugs in five European countries (France, Italy, the Netherlands, Spain and UK), as well as two scenarios with different treatment length. Results: Of the drugs analysed, topical clonazepam proved to be the most cost-effective therapy. Although generic proved more efficient than brand name drugs, they displayed no advantage over brand name topical clonazepam. The Netherlands was the country with the highest overall drug efficiency. Sensitivity analyses highlighted the robustness of the model, because topical clonazepam proved to be the most efficient therapy under all the different scenarios. Conclusions: Topical clonazepam, which previous analyses of clinical evidence have shown to be the drug of choice for BMS, also proved to be the most cost-effective of the drugs analysed for this condition. [ABSTRACT FROM AUTHOR]

Map of Medicine is an evidence-based online clinical knowledge resource. Procured at significant cost by healthcare providers in the UK, Sweden and Denmark, it is important to establish the beneficial impact that investment has had on healthcare practise and, ultimately, on patient outcomes. The objective of this study is to review the evidence for the impact of the Map of Medicine on clinical practice. A systematic review of peer-reviewed and grey literature was conducted. Nine healthcare databases, Google Scholar and Google were searched for articles containing the terms 'map of medicine'. The search identified 133 articles. Eleven of the articles identified met the inclusion criteria. The main finding of the study is the paucity of evidence available on the impact of the Map of Medicine and the variable quality of that which does exist. There are some encouraging early indications for the Map of Medicine as a tool within service redesign, leading to an increase in appropriate referrals to secondary care, decreased patient waiting times and considerable cost savings. Further research with study designs that can generate high quality evidence for the impact of Map of Medicine is essential in order to support policy decisions. [ABSTRACT FROM AUTHOR]

AIM OF THE PAPER: To examine the application of the decision tree approach to collaborative clinical decision-making in mental health care in the United Kingdom (UK). BACKGROUND: While this approach to decision-making has been examined in the acute care setting, there is little published evidence of its use in clinical decision-making within the mental health setting. The complexities of dual diagnosis (schizophrenia and substance misuse in this case example) and the varied viewpoints of different professionals often hamper the decision-making process. This paper highlights how the approach was used successfully as a multiprofessional collaborative approach to decision-making in the context of British community mental health care. DESIGN: A selective review of the relevant literature and a case study application of the decision tree framework. CONCLUSIONS: The process of applying the decision tree framework to clinical decision-making in mental health practice can be time consuming and client inclusion within the process is not always appropriate. The approach offers a method of assigning numerical values to support complex multiprofessional decision-making as well as considering underpinning literature to inform the final decision. Use of the decision tree offers a common framework that can assist professionals to examine the options available to them in depth, while considering the complex variables that influence decision-making in collaborative mental health practice. Use of the decision tree warrants further consideration in mental health care in terms of practice and education. [ABSTRACT FROM AUTHOR]

Objective: To examine the relative cost effectiveness of topical calcipotriol and short-contact dithranol in the treatment of mild to moderate plaque psoriasis.Design and Setting: This was a modelling study from the perspective of the UK National Health Service as payer.Methods: The interventions were compared using 2 decision-analytic models: one using a short term horizon (12 weeks) and the other using a longer term horizon (up to 1 year). Clinical success of treatment and relapse rates were obtained from the results of randomised controlled trials.Main Outcome Measures and Results: The outcome measure used was the degree of improvement in psoriasis as judged by the patient. In the short term comparison, calcipotriol was the most effective treatment (60.8% success), but it was also the most expensive (96.03 Pounds; 2000 values). The incremental cost per success was 577.50 Pounds using a 12-week course of calcipotriol compared with short-contact dithranol. Over the long term horizon, first-line treatment with calcipotriol still had the highest expected cost per successful treatment (164.91 Pounds), but the incremental cost using this strategy was 38.66 Pounds compared with short-contact dithranol. In terms of cost per successful day's treatment (i.e. the cost for a day in which the patient reported a marked improvement or cleared lesions), the cost of each additional successful day's treatment was 19.93 Pounds when using calcipotriol as first-line treatment rather than short-contact dithranol.Conclusion: The combination of differences in drug acquisition costs and relapse rates can lead to large differences in the comparative cost effectiveness of topical treatments for plaque psoriasis. From the perspective of the prescriber, the results of this analysis suggest that selecting short-contact dithranol as first-line treatment was the most cost-effective strategy. [ABSTRACT FROM AUTHOR]

Nicotine consumption is considered a major health problem, where many of those who wish to quit smoking relapse. The problem is that overtime smoking as behaviour is changing into a habit, in which it is connected to internal (e.g., nicotine level, craving) and external (action, time, location) triggers. Smoking cessation apps have proved their efficiency to support smoking who wish to quit smoking. However, still, these applications suffer from several drawbacks, where they are highly relying on the user to initiate the intervention by submitting the factor the causes the urge to smoke. This research describes the creation of a combined Control Theory and deep learning model that can learn the smoker's daily routine and predict smoking events. The model's structure combines a Control Theory model of smoking with a 1D-CNN classifier to adapt to individual differences between smokers and predict smoking events based on motion and geolocation values collected using a mobile device. Data were collected from 5 participants in the UK, and analysed and tested on 3 different machine learning model (SVM, Decision tree, and 1D-CNN), 1D-CNN has proved it's efficiency over the three methods with average overall accuracy 86.6%. The average MSE of forecasting the nicotine level was (0.04) in the weekdays, and (0.03) in the weekends. The model has proved its ability to predict the smoking event accurately when the participant is well engaged with the app. [ABSTRACT FROM AUTHOR]

The article offers information on the Incident Decision Tree developed by the National Patient Safety Agency to help National Health Service managers in England and Wales determine fair course of action towards staff involved in patient safety incidents. Several tests involved in the model include deliberate harm test and foresight test.

In recent decades there have been many assessments of machine learning models for short-term flood prediction, but few comparisons using water level forecasts. Here, we use high-resolution water level and rainfall data from the UK Environment Agency to compare and assess the usefulness of multiple models including deep artificial neural networks to accurately predict water level at different points of the upper Soar catchment area (UK) with lead-times of 6 to 48 h. Six types of machine learning algorithms in addition to 3 neural network models are developed for every lead time, with increasing complexities and data-pre-processing techniques. The models are then tested for their generalization properties using the complete test set times series and focusing on the rising limb of the hydrographs. Preliminary results indicate that the inclusion of water level from multiple upstream stations with different lag times leads to better performance in most of the models. Overall, the ensemble of decision trees showed good performance, but regularised neural network approach has the highest accuracy. [ABSTRACT FROM AUTHOR]

Reveals a technique that is helping decision makers cost out the litigation process. Functions of a decision-making tree; Components of modern dispute resolution; Important drives of a decision; Effectiveness of decision trees.