Using machine learning to improve bovine tuberculosis control in herd level outbreaks.

Bovine tuberculosis (bTB), a chronic disease of cattle, is caused by the Mycobacterium bovis infection. Despite having a serious social and economic impact in the United Kingdom and Ireland, there is no antemortem gold standard diagnostic test. Tuberculin skin tests (CICT) are commonly used as a control measure with the interferon gamma (IFN-γ) assay being applied in certain circumstances. This paper utilizes data gathered describing tuberculin regression in reactors (test positive cattle) following the CICT at 72 ± 4 h post injection in herds with large bTB outbreaks. The work then applies machine learning techniques (Decision Trees, Bagging Trees and Random Forests, alongside several balancing approaches) to predict which cattle were likely to be truly infected with tuberculosis, enabling identification of atypical breakdowns that require extra investigation and providing a mechanism for quality assurance of the existing CICT bTB surveillance scheme. The analysis showed that Random Forests (RF) trained using SMOTE balancing had the joint best performance and accuracy (0.90). The importance of the two components of the interferon gamma assay within the RF model also indicated that varying the assay threshold for large outbreaks would be beneficial. Furthermore, the combined use of the RF and IFN- γ models could lead to the improved detection of infection within breakdown herds, reducing the scale and duration of outbreaks. An additional use of these models would be for quality assuring the current bTB surveillance based on CICT and post mortem inspection. Quality control is well recognized as an essential component of a disease surveillance/eradication programme.Clinical Relevance- Bovine tuberculosis remains a disease that is hard to control on a national level. The use of the machine learning model could lead to significant improved detection of infection within breakdown herds, reducing the scale and duration of outbreaks. Advanced modelling, such as this, has the potential to strengthen the efficacy of disease surveillance and the eradication strategy and can meaningfully contribute to animal disease national control plans.

The CP-ABM approach for modelling COVID-19 infection dynamics and quantifying the effects of non-pharmaceutical interventions.

The motivation for this research is to develop an approach that reliably captures the disease dynamics of COVID-19 for an entire population in order to identify the key events driving change in the epidemic through accurate estimation of daily COVID-19 cases. This has been achieved through the new CP-ABM approach which uniquely incorporates Change Point detection into an Agent Based Model taking advantage of genetic algorithms for calibration and an efficient infection centric procedure for computational efficiency. The CP-ABM is applied to the Northern Ireland population where it successfully captures patterns in COVID-19 infection dynamics over both waves of the pandemic and quantifies the significant effects of non-pharmaceutical interventions (NPI) on a national level for lockdowns and mask wearing. To our knowledge, there is no other approach to date that has captured NPI effectiveness and infection spreading dynamics for both waves of the COVID-19 pandemic for an entire country population.

Identification of Bridge Key Performance Indicators Using Survival Analysis for Future Network-Wide Structural Health Monitoring.

Machine learning and statistical approaches have transformed the management of infrastructure systems such as water, energy and modern transport networks. Artificial Intelligence-based solutions allow asset owners to predict future performance and optimize maintenance routines through the use of historic performance and real-time sensor data. The industrial adoption of such methods has been limited in the management of bridges within aging transport networks. Predictive maintenance at bridge network level is particularly complex due to the considerable level of heterogeneity encompassed across various bridge types and functions. This paper reviews some of the main approaches in bridge predictive maintenance modeling and outlines the challenges in their adaptation to the future network-wide management of bridges. Survival analysis techniques have been successfully applied to predict outcomes from a homogenous data set, such as bridge deck condition. This paper considers the complexities of European road networks in terms of bridge type, function and age to present a novel application of survival analysis based on sparse data obtained from visual inspections. This research is focused on analyzing existing inspection information to establish data foundations, which will pave the way for big data utilization, and inform on key performance indicators for future network-wide structural health monitoring.

Cattle movements in Northern Ireland form a robust network: implications for disease management.

The movements of undetected infected animals can facilitate long-distance pathogen spread, making control and eradication difficult by (re)infecting disease-free populations. Characterising movement patterns is essential in understanding pathogen spread and how potential interventions, particularly animal movement restrictions, could help as a control mechanism. In Northern Ireland (NI), cattle movements are important contributors to a significant portion of agricultural trade. They can be disrupted due to statutory interventions, for example, during bovine tuberculosis (bTB) control. Identifying populations at risk of becoming infected would allow for improved resource allocation. This could be through targeting herds with an above-average risk of becoming infected or spreading (amplifying) infection, and restricting their movement to manage future outbreaks. In this study, cattle movements were investigated using social network analysis (SNA) at the monthly temporal scale across NI during 2010-2015. Targeted and random herd restrictions were compared and their impact on the structure and connectivity of the networks' was assessed (e.g. connected component subgraphs). This work was contextualised in relation to bTB, the most persistent infectious disease currently impacting agriculture in NI, where reduced connectivity would represent potential reduced vulnerability from infection introduction. There was seasonal variation in network size and level of connectivity with spring and autumn being the largest and most connected due to common farming practices in NI. Across the study period, there was limited inter-annual variation in global network metrics. On average there were 6.28 movements between each pair of nodes each month, low reciprocity (mean of 0.155) and the networks were moderately accessible with an average path length of 4.28. Movements were not confined to within each disease management area but frequently occurred between these areas (mean assortativity of -0.0731) and herds with high degree interacted with herds of low degree (mean assortativity of -0.351). The Giant Weakly Connected Component (GWCC) spanned most of the networks (between 75% and 100% of nodes); however the Giant Strongly Connected Component (GSCC) included, at most, 23% of the network. There was heterogeneous contributions across NI with little participation in the GSCC from some disease management areas, and the GSCC was comprised predominantly of 'beef breeders', 'beef rearers', and 'other/mixed' type herds. Targeted restrictions were more effective at fragmenting the network than randomly restricting movements when 25% of nodes or more were removed. Cattle networks in NI are extremely interconnected and robust to movement restrictions, suggesting potential vulnerability to movement-facilitated pathogen spread, such as bTB.

Expediting assessments of database performance for streams of respiratory parameters.

A new methodology is proposed to compare database performance for streams of patient respiratory data from patients in an intensive care unit. New metrics are proposed through which databases may be compared both for this and similar streaming applications in the domain of the Internet of Things. Studies are reported using simulated patient data for four freely available databases. The statistical technique of non-parametric bootstrapping is used to minimise the total running time of the tests. We report mean values and bias corrected and accelerated confidence intervals for each metric and use these to compare the databases. We find that, among the four databases tested, ScaleDB is an optimum database technology when handling between 200 and 800 patients in this application, while PostgreSQL performs best outside of this range. Comparing the non-parametric bootstrapping method to a complete set of tests shows that the two approaches give results differing by a few percent.

A two-stage approach to the joint analysis of longitudinal and survival data utilising the Coxian phase-type distribution.

The Coxian phase-type distribution is a special type of Markov model which can be utilised both to uncover underlying stages of a survival process and to make inferences regarding the rates of flow of individuals through these latent stages before an event of interest occurs. Such models can be utilised, for example, to identify individuals who are likely to deteriorate faster through a series of disease states and thus require more aggressive medical intervention. Within this paper, a two-stage approach to the analysis of longitudinal and survival data is presented. In Stage 1, a linear mixed effects model is first used to represent how some longitudinal response of interest changes through time. Within this linear mixed effects model, the individuals' random effects can be considered as a proxy measure for the effect of the individuals' genetic profiles on the response of interest. In Stage 2, the Coxian phase-type distribution is employed to represent the survival process. The individuals' random effects, estimated in Stage 1, are incorporated as covariates within the Coxian phase-type distribution so as to evaluate their effect on the individuals' rates of flow through the system represented by the Coxian. The approach is illustrated using data collected on individuals suffering from chronic kidney disease, where focus is given to an emerging longitudinal biomarker of interest - an individual's haemoglobin level.

Predicting elderly patient length of stay in hospital and community care using a series of conditional Coxian phase-type distributions, further conditioned on a survival tree.

Increasing demand on hospital resources by an ageing population is impacting significantly on the number of beds available and, in turn, the length of time that elderly patients must wait for a bed before being admitted to hospital. This research presents a new methodology that models patient pathways and allows the accurate prediction of patient length of stay in hospital, using a phase-type survival tree to cluster patients based on their covariates and length of stay in hospital. A type of Markov model, called the conditional Coxian phase-type distribution is then implemented, with the probability density function for the time spent at a particular stage of care, for example, the first community discharge, conditioned on the length of stay experienced at the previous stage, namely the initial hospital admission. This component of the methodology is subsequently applied to each cohort of patients over a number of hospital and community stages in order to build up the profile of patient readmissions and associated timescales for each cohort. It is then possible to invert the methodology, so that the length of stay for an observation representing a new patient admission may be estimated at each stage of care, based on the assigned cohort at the initial hospital stage. This approach provides hospital managers with an accurate understanding of the rates with which different groups of patients move between hospital and community care, which may be used to reduce the negative effects of bed-blocking and the premature discharge of patients without a required period of convalescence. This has the benefit of assisting hospital managers with the effective allocation of vital healthcare resources. The approach presented is different to previous research in that it allows the inclusion of patient covariate information into the methodology describing patient transitions between hospital and community care stages in an aggregate Markov process. A data set containing hospital readmission data for elderly patients from the Abruzzo region of Italy is used as a case study in the application of the presented methodology.

Education achievement and type 2 diabetes-what mediates the relationship in older adults? Data from the ESTHER study: a population-based cohort study.

OBJECTIVE: The study aims to identify the mediating factors of the relationship between education achievement and incident type 2 diabetes mellitus (T2DM) in older adults. DESIGN: Population-based cohort study. SETTING: Participants were recruited from the German federal state of Saarland. PARTICIPANTS: Participants were excluded if they had prevalent T2DM or missing data on prevalent T2DM, missing or zero follow-up time for incident T2DM or were under 50 years of age. The total sample consisted of 7462 individuals aged 50-75 years (42.8% men, mean age 61.7 years) at baseline (2000-02). The median follow-up time was 8.0 years. METHODS: Cox proportional hazards regression was initially used to determine the direct association between education achievement and incident T2DM. Using the Baron and Kenny approach, we then investigated the associations between education achievement and incident T2DM with the potential mediators. The contribution of each of the putative mediating variables was then calculated. RESULTS: A clear socioeconomic gradient was observed with regard to T2DM incidence with the lowest educated individuals at a greater risk of developing the disease during the follow-up period: HR (95% CI) high education: 0.52 (0.34 to 0.80); medium education: 0.80 (0.66 to 0.96). Seven of the variables considered explained a proportion of the education-T2DM relationship (body mass index, alcohol consumption, hypertension, fasting triglycerides, high-density lipoprotein (HDL) cholesterol, physical activity and smoking status), where the contribution of the variables ranged from 1.0% to 17.7%. Overall, the mediators explained 31.7% of the relationship. CONCLUSION: By identifying the possible mediating factors of the relationship between education achievement and incident T2DM in older adults, the results of this study can be used to assist with the development of public health strategies that aim to reduce socioeconomic inequalities in T2DM.

A conditional approach for modelling patient readmissions to hospital using a mixture of Coxian phase-type distributions incorporating Bayes' theorem.

The number of elderly patients requiring hospitalisation in Europe is rising. With a greater proportion of elderly people in the population comes a greater demand for health services and, in particular, hospital care. Thus, with a growing number of elderly patients requiring hospitalisation competing with non-elderly patients for a fixed (and in some cases, decreasing) number of hospital beds, this results in much longer waiting times for patients, often with a less satisfactory hospital experience. However, if a better understanding of the recurring nature of elderly patient movements between the community and hospital can be developed, then it may be possible for alternative provisions of care in the community to be put in place and thus prevent readmission to hospital. The research in this paper aims to model the multiple patient transitions between hospital and community by utilising a mixture of conditional Coxian phase-type distributions that incorporates Bayes' theorem. For the purpose of demonstration, the results of a simulation study are presented and the model is applied to hospital readmission data from the Lombardy region of Italy. Copyright © 2016 John Wiley & Sons, Ltd.

Treatment of retinopathy of prematurity in Northern Ireland, 2000-2011: a population-based study.

PURPOSE: To evaluate the incidence of treatment-requiring retinopathy of prematurity (ROP) over a 12-year-period in Northern Ireland. METHODS: The medical records of all infants treated for ROP from January 2000 to December 2011 were retrospectively reviewed and cross-referenced with the Neonatal Intensive Care Outcomes Research and Evaluation (NICORE) database. RESULTS: The Northern Ireland population data showed an increase in the number of live births from 2000 to 2011. The proportion of babies born with a birth weight <1501 g and/or <32 weeks' gestational age remained constant (χ(2) trend = 3.220, P = 0.0727), although the proportion of these babies who died prior to 42 weeks' gestation decreased from 2000 to 2011 (P = 0.0196 using χ(2) for trend = 5.445; P = 0.0354 using χ(2) = 20.809). The prevalence of treatment-requiring ROP in these infants increased from 1.05% in 2000 to 5.78% in 2011 (P < 0.001 using χ(2) trend = 16.309; P < 0.001 using χ(2) = 31.378). CONCLUSIONS: The present population-based study highlights that the incidence of treatment- requiring ROP is increasing in Northern Ireland. The increasing number of infants requiring treatment will need to be taken into consideration in the commissioning process for ROP services in Northern Ireland.

Hotspots identification and ranking for road safety improvement: an alternative approach.

During the last decade, the concept of composite performance index, brought from economic and business statistics, has become a popular practice in the field of road safety, namely for the identification and classification of worst performing areas or time slots also known as hotspots. The overall quality of a composite index depends upon the complexity of phenomena of interest as well as the relevance of the methodological approach used to aggregate the various indicators into a single composite index. However, current aggregation methods used to estimate the composite road safety performance index suffer from various deficiencies at both the theoretical and operational level; these include the correlation and compensability between indicators, the weighting of the indicators as well as their high "degree of freedom" which enables one to readily manipulate them to produce desired outcomes (Munda and Nardo, 2003, 2005, 2009). The objective of this study is to contribute to the ongoing research effort on the estimation of road safety composite index for hotspots' identification and ranking. The aggregation method for constructing the composite road safety performance index introduced in this paper, strives to minimize the aforementioned deficiencies of the current approaches. Furthermore, this new method can be viewed as an intelligent decision support system for road safety performance evaluation, in order to prioritize interventions for road safety improvement.

The practical implications of using standardized estimation equations in calculating the prevalence of chronic kidney disease.

BACKGROUND: Kidney Disease Outcomes Quality Initiative (KDOQI) chronic kidney disease (CKD) guidelines have focused on the utility of using the modified four-variable MDRD equation (now traceable by isotope dilution mass spectrometry IDMS) in calculating estimated glomerular filtration rates (eGFRs). This study assesses the practical implications of eGFR correction equations on the range of creatinine assays currently used in the UK and further investigates the effect of these equations on the calculated prevalence of CKD in one UK region METHODS: Using simulation, a range of creatinine data (30-300 micromol/l) was generated for male and female patients aged 20-100 years. The maximum differences between the IDMS and MDRD equations for all 14 UK laboratory techniques for serum creatinine measurement were explored with an average of individual eGFRs calculated according to MDRD and IDMS < 60 ml/min/1.73 m(2) and 30 ml/min/1.73 m(2). Similar procedures were applied to 712,540 samples from patients > or = 18 years (reflecting the five methods for serum creatinine measurement utilized in Northern Ireland) to explore, graphically, maximum differences in assays. CKD prevalence using both estimation equations was compared using an existing cohort of observed data. RESULTS: Simulated data indicates that the majority of laboratories in the UK have small differences between the IDMS and MDRD methods of eGFR measurement for stages 4 and 5 CKD (where the averaged maximum difference for all laboratory methods was 1.27 ml/min/1.73 m(2) for females and 1.59 ml/min/1.73 m(2) for males). MDRD deviated furthest from the IDMS results for the Endpoint Jaffe method: the maximum difference of 9.93 ml/min/1.73 m(2) for females and 5.42 ml/min/1.73 m(2) for males occurred at extreme ages and in those with eGFR > 30 ml/min/1.73 m(2). Observed data for 93,870 patients yielded a first MDRD eGFR < 60 ml/min/1.73 m(2) in 2001. 66,429 (71%) had a second test > 3 months later of which 47,093 (71%) continued to have an eGFR < 60 ml/min/1.73 m(2). Estimated crude prevalence was 3.97% for laboratory detected CKD in adults using the MDRD equation which fell to 3.69% when applying the IDMS equation. Over 95% of this difference in prevalence was explained by older females with stage 3 CKD (eGFR 30-59 ml/min/1.73 m(2)) close to the stage 2 CKD (eGFR 60-90 ml/min/1.73 m(2)) interface. CONCLUSIONS: Improved accuracy of eGFR is obtainable by using IDMS correction especially in the earlier stages of CKD 1-3. Our data indicates that this improved accuracy could lead to reduced prevalence estimates and potentially a decreased likelihood of onward referral to nephrology services particularly in older females.

Rasch analysis of the daily living tasks dependent on vision (DLTV).

PURPOSE: To examine internal consistency, refine the response scale, and obtain a linear scoring system for the visual function instrument, the Daily Living Tasks Dependent on Vision (DLTV). METHODS: Data were available from 186 participants with a clinical diagnosis of AMD who completed the 22-item DLTV (DLTV-22) according to four-point ordinal response scale. An independent group of 386 participants with AMD were administered a reduced version of the DLTV with 11 items (DLTV-11), according to a five-point response scale. Rasch analysis was performed on both datasets and used to generate item statistics for measure order, response odds ratios per item and per person, and infit and outfit mean square statistics. The Rasch output from the DLTV-22 was examined to identify redundant items and for factorial validity and person item measure separation reliabilities. RESULTS: The average rating for the DLTV-22 changed monotonically with the magnitude of the latent person trait. The expected versus observed average measures were extremely close, with step calibrations evenly separated for the four-point ordinal scale. In the case of the DLTV-11, step calibrations were not as evenly separated, suggesting that the five-point scale should be reduced to either a four- or three-point scale. Five items in the DLTV-22 were removed, and all 17 remaining items had good infit and outfit mean squares. PCA with residuals from Rasch analysis identified two domains containing 7 and 10 items each. The domains had high person separation reliabilities (0.86 and 0.77 for domains 1 and 2, respectively) and item measure reliabilities (0.99 and 0.98 for domains 1 and 2, respectively). CONCLUSIONS: With the improved internal consistency, establishment of the accuracy and precision of the rating scale for the DLTV and the establishment of a valid domain structure we believe that it constitutes a useful instrument for assessing visual function in older adults with age-related macular degeneration.

Estimating the costs for a group of geriatric patients using the Coxian phase-type distribution.

The length of stay in hospital of geriatric patients may be modelled using the Coxian phase-type distribution. This paper examines previous methods which have been used to model health-care costs and presents a new methodology to estimate the costs for a cohort of patients for their duration of stay in hospital, assuming there are no further admissions. The model, applied to 1392 patients admitted into the geriatric ward of a local hospital in Northern Ireland, between 2002 and 2003, should be beneficial to hospital managers, as future decisions and policy changes could be tested on the model to investigate their influence on costs before the decisions were carried out on a real ward.

Modeling the health care costs of geriatric inpatients.

This paper extends a method for modeling the survival of patients in hospitals to allow the expected cost to be estimated for the patients' accumulated duration of time in care. An extension of Bayesian network (BN) theory has previously been developed to model patients' survival time in hospitals with respect to the graphical and probabilistic representation of the interrelationships between the patients' clinical variables. Unlike previous BN techniques, this extended model can accommodate continuous times that are skewed in nature. This paper presents the theory behind such an approach and extends it by attaching a cost variable to the survival times, enabling the costing and efficient management of groups of patients in hospitals. An application of the model is illustrated by considering a group of 4260 patients admitted into the geriatric department of a U.K. hospital between 1994-1997. Results are derived for the distribution for their length of stay in the hospital and associated costs. The model's practical use is highlighted by illustrating how hospital managers could benefit using such a method for investigating the influence of future decisions and policy changes on the hospital's expenditure.

Addressing bed costs for the elderly: a new methodology for modelling patient outcomes and length of stay.

The proportion of elderly in the population has dramatically increased and will continue to do so for at least the next 50 years. Medical resources throughout the world are feeling the added strain of the increasing proportion of elderly in the population. The effective care of elderly patients in hospitals may be enhanced by accurately modelling the length of stay of the patients in hospital and the associated costs involved. This paper examines previously developed models for patient length of stay in hospital and describes the recently developed conditional phase-type distribution (C-Ph) to model patient duration of stay in relation to explanatory patient variables. The Clinics data set was used to demonstrate the C-Ph methodology. The resulting model highlighted a strong relationship between Barthel grade, patient outcome and length of stay showing various groups of patient behaviour. The patients who stay in hospital for a very long time are usually those that consume the largest amount of hospital resources. These have been identified as the patients whose resulting outcome is transfer. Overall, the majority of transfer patients spend a considerably longer period of time in hospital compared to patients who die or are discharged home. The C-Ph model has the potential for considering costs where different costs are attached to the various phases or subgroups of patients and the anticipated cost of care estimated in advance. It is hoped that such a method will lead to the successful identification of the most cost effective case-mix management of the hospital ward.

Using Coxian phase-type distributions to identify patient characteristics for duration of stay in hospital.

Coxian phase-type distributions are a special type of Markov model that describes duration until an event occurs in terms of a process consisting of a sequence of latent phases. This paper considers the use of Coxian phase-type distributions for modelling patient duration of stay for the elderly in hospital and investigates the potential for using the resulting distribution as a classifying variable to identify common characteristics between different groups of patients according to their (anticipated) length of stay in hospital. The identification of common characteristics for patient length of stay groups would offer hospital managers and clinicians possible insights into the overall management and bed allocation of the hospital wards.