137 research outputs found
Applications of phase type survival trees in HIV disease progression modelling
No full textIt is important to model progression of a disease to understanding if the patient's condition is improving or getting worse. In the case of HIV disease, the change in the patient's CD4+ T cell count is used to calculate the progression of HIV disease i.e. if the CD4 count goes down it represent the progression of the patient's HIV disease. Due to the lack of an effective cure for HIV disease, it is crucial to monitor the disease progression to managing HIV disease effectively. Therefore, this study is aimed to model HIV disease progression by using phase type survival trees to cluster patients into homogenous groups based on their disease progression to understand the effect of different factors of prognostic significance and their interactions affecting the disease progression. The proposed methods are evaluated using an empirical data of 1,838 HIV-infected patients. The methods developed in this study can also be used for modelling the progression of other chronic conditions or diseases
Markov Chain Modelling for Geriatric Patient Care
No full textSummary
Objectives:
To show that Markov chain modelling can be applied to data on geriatric patients and use these models to assess the effects of covariates.
Methods:
Phase-type distributions were fitted by maximum likelihood to data on times spent by the patients in hospital and in community-based care. Data on the different events that ended the patients’ periods of care were used to estimate the dependence of the probabilities of these events on the phase from which the time in care ended. The age of the patients at admission to care and the year of admission were also included as covariates.
Results:
Differential effects of these covariates were shown on the various parameters of the fitted model, and interpretations of these effects made.
Conclusions:
Models based on phase-type distributions were appropriate for describing times spent in care, as the ordered phases had an interpretable structure corresponding to increasing amounts of care being given.</jats:p
Using Markov Models to Characterize and Predict Process Target Compliance
No full textProcesses are everywhere, covering disparate fields such as business, industry, telecommunications, and healthcare. They have previously been analyzed and modelled with the aim of improving understanding and efficiency as well as predicting future events and outcomes. In recent years, process mining has appeared with the aim of uncovering, observing, and improving processes, often based on data obtained from logs. This typically requires task identification, predicting future pathways, or identifying anomalies. We here concentrate on using Markov processes to assess compliance with completion targets or, inversely, we can determine appropriate targets for satisfactory performance. Previous work is extended to processes where there are a number of possible exit options, with potentially different target completion times. In particular, we look at distributions of the number of patients failing to meet targets, through time. The formulae are illustrated using data from a stroke patient unit, where there are multiple discharge destinations for patients, namely death, private nursing home, or the patient’s own home, where different discharge destinations may require disparate targets. Key performance indicators (KPIs) of this sort are commonplace in healthcare, business, and industrial processes. Markov models, or their extensions, have an important role to play in this work where the approach can be extended to include more expressive assumptions, with the aim of assessing compliance in complex scenarios
When Harry left Sally: A New Estimate of Marital Disruption in the U.S., 1860 - 1948
No full textDivorce rate is a poor indicator of marital instability because many marital disruptions never become divorces. This paper provides the first estimate of the rate of marital disruption in the U.S. in 1860 - 1948. Marital disruption rate was similar to divorce rate after the Civil War but the two rates wildly diverged in the early 20th century. In 1900 - 1930, the disruption rate was as much as double the divorce rate, implying that perhaps half of all disruptions never reached the court. In the long run, the cohort rate of marital disruption increased from about 10% in the mid-1860s to about 30% in the 1940s.divorce, marital disruption, separation
Length of stay-based clustering methods for patient grouping
No full textLength of stay (LOS) is often used as a proxy measure of a patient’ resource consumption because of the practical difficulties of directly measuring resource consumption and the easiness of calculating LOS. Grouping patient spells according to their LOS has proved to be a challenge in health care applications due to the inherent variability in the LOS distribution. Sound methods for LOS-based patient grouping should certainly lead to a better planning of bed allocation, and patient admission and discharge. Grouping patient spells according to their LOS in a computational efficient manner is still a research issue that has not been fully addressed. For instance, grouping patient spells according to LOS intervals (e.g. 0-3 days, 4-9 days, 10-21 days etc.), has previously been defined by non-algorithmic approaches using clinical judgement, visual inspection of the LOS distribution or according to the perceived casemix. The aim of this paper is to present a novel methodology of grouping patients according to their length of stay based on fitting Gaussian mixture models to LOS observations. This method was developed as part of an innovative prediction tool that helps identify groups of patients exhibiting similar resource consumption levels as these are approximated by patient LOS. As part of evaluating the approach, we also compare it to two alternative clustering approaches, K-means and the two-step algorithm. Computational results show the superiority of this method compared to alternative clustering approaches in terms of its ability to extract clinically meaningful patient groups as applied to a skewed LOS dataset
Models for extracting information on patient pathways
No full textIn this paper, we present a random effects approach to modelling of patient flow. Individual patient experience in care as represented by their pathways through the system is modelled. An application to the University College of London Hospital (UCLH) neonatal unit is presented. Using the multinomial logit random effects model, we demonstrate a methodology to extract useful information on patient pathways. This modelling technique is useful for identifying interesting pathways such as those resulting in high probabilities of death/survival, and those resulting in short or long length of stay. Patient-specific discharge probabilities may also be predicted as a function of the frailties; which are modelled as random effects. In the current climate of healthcare cost concerns these will assist healthcare managers in their task of allocating resources to different departments or units of healthcare institution. Two classes of models are presented; one based on patient pathways in which different random effects distribution assumptions are made and the other in which the random effects are regressed on patient characteristics. Intuitively, with the introduction of individual patient frailties, we can argue that both clinical and operational patient flows are being captured in this modelling framework
A Grid implementation for profiling hospitals based on patient readmissions
No full textGenerally, high level of readmission is associated with poor patient care, hence its relation to the quality of care is plausible. Frequent patient readmissions have personal, financial and organisational consequences. This has motivated healthcare commissioners in England to use emergency readmission as an indicator in the performance rating framework. A statistical model, known as the multilevel transition model was previously developed, where individual hospitals propensity for first readmission, second readmission, third (and so on) were considered to be measures of performance. Using these measures, we defined a new performance index. During the period 1997 and 2004, the national (England) hospital episodes statistics dataset comprise more than 5 million patient readmissions. Implementing a statistical model using the complete population dataset could possibly take weeks to estimate the parameters. Moreover, it is not statistically sound to utilise the full population dataset. To resolve the problem, we extract 1000 random samples from the original data, where each random sample is likely to lead to differing hospital performance measures. For computational efficiency a Grid implementation of the model is developed. Using a stand-alone computer, it would take approximately 500 hours to estimate 1000 samples, whereas in the Grid implementation, the full 1000 samples were analysed in less than 24 hours. Analysing the output from the full 1000 sample, we noticed that 4 out of the 5 worst performing hospitals treating cancer patients were in London
Dual contextual module for neural machine translation
No full textSelf-attention-based encoder-decoder frameworks have drawn increasing attention in recent years. The self-attention mechanism generates contextual representations by attending to all tokens in the sentence. Despite improvements in performance, recent research argues that the self-attention mechanism tends to concentrate more on the global context with less emphasis on the contextual information available within the local neighbourhood of tokens. This work presents the Dual Contextual (DC) module, an extension of the conventional self-attention unit, to effectively leverage both the local and global contextual information. The goal is to further improve the sentence representation ability of the encoder and decoder subnetworks, thus enhancing the overall performance of the translation model. Experimental results on WMT’14 English-German (En→De) and eight IWSLT translation tasks show that the DC module can further improve the translation performance of the Transformer model
- …
