Clustering of 27,525,663 death records from the United States Based on health conditions associated with death: an example of big health data exploration
Following a population-based approach to palliative care, this article analyzes more than 27 million US death certificates from 2006 to 2016 with respect to common comorbidities, socio-demographic background and circumstances of death. Viscovery SOMine's Big Data capabilities are used to cluster the death records in the 346-dimensional space of health conditions, to identify population-based patterns, and to derive implications for palliative care.
Disease-specific comorbidity clusters in COPD and accelerated aging
This article analyzes the connection between disease-specific comorbidity clusters in COPD patients and accelerated aging. Viscovery SOMine is used to cluster COPD patients and a comparable control group to evaluate the decrease in telomere length in COPD-specific clusters.
Comprehensive lung function assessment does not allow to infer response to pulmonary rehabilitation in patients with COPD
This article examines whether the extent of improvement in lung function, exercise performance, daily activities, mood, and disease-specific health status of COPD patients undergoing pulmonary rehabilitation can be predicted from the lung function profile prior to rehabilitation. The lung function data is clustered with Viscovery SOMine to obtain and statistically analyze groups of patients with similar profiles.
The respiratory physiome: clustering based on a comprehensive lung function assessment in patients with COPD
The aim of this paper is to provide a comprehensive description of lung function for patients with chronic obstructive pulmonary disease (COPD) and study its connection to functional performance and health status. The Viscovery SOMine cluster model shows that lung-function impairment is a multidimensional problem, which cannot be characterized by a single measurement. In addition, it is shown that the lung-function profile alone is a poor predictor for functional performance, highlighting the necessity of additional performance tests and questionnaires to provide optimal treatment for patients.
HIV/Human herpesvirus co-infections: impact on tryptophan-kynurenine pathway and immune reconstitution
This article attempts to clarify the impact of human herpesvirus co-infections of patients infected with human immunodeficiency virus (HIV) on the kynurenine/tryptophan (KT) ratio and long-term CD4 T-cell recovery in antiretroviral treatment. Viscovery SOMine clustering shows that different types and combinations of herpes infections have varying effect on K/T ratio.
Objectively identified comorbidities in COPD: impact on pulmonary rehabilitation outcomes
Based on the Viscovery SOMine cluster model from "Clusters of comorbidities based on validated objective measurements and systemic inflammation in patients with chronic obstructive pulmonary disease" by Vanfleteren et al., rehabilitation outcomes of patients with chronic obstructive pulmonary disease (COPD) are analyzed. This analysis shows that the existence of comorbidities has only a small influence on the achievable outcomes.
Differential response to pulmonary rehabilitation in COPD: multidimensional profiling
Multidimensional response to pulmonary rehabilitation in patients with chronic obstructive pulmonary disease (COPD) is studied. Viscovery SOMine identified distinct response profiles; moreover, response outcome depends only slightly on the baseline status of patients.
Cluster analyses in a sample of COPD patients
To optimize pharmacological treatment and recognize mortality risk factors, a segmentation of chronic obstructive pulmonary disease (COPD) patient groups is obtained. Viscovery Profiler is used on metered data (physical examination, laboratory tests, spirometry), medical history (smoking history, Charlson comorbidity index), and survey data (St. George's Respiratory Questionnaire) and identifies six distinct clusters of patients.
Integrative understanding of macular morphologic patterns in diabetic retinopathy based on self-organizing map
This article analyzes the morphologic patterns in patients with diabetic retinopathy. Several parameters obtained via OCT scans are clustered via Viscovery SOMine, identifying distinct morphological profiles.
Classification of age-related changes in lumbar spine with the help of MRI scores
The variation of normal age-related changes in the lumbar spine is characterized and distinguished from pathological deformities. Viscovery SOMine is used on geometric measurements obtained by MRI scans of the lumbar spine and shows that vertebral height and fat signals do not significantly change during normal ageing, whereas changes in disc height, para-spinal-muscle signal intensity and psoas muscle are distinctly correlated with ageing.
HMGA2 expression in white adipose tissue linking cellular senescence with diabetes
This study analyzes the connection between obesity, gain of white adipose tissue, type 2 diabetes and HMGA2 expression. The Viscovery SOMine model suggests a higher risk for type 2 diabetes in patients with a high HMGA2 expression.
Clusters of comorbidities based on validated objective measurements and systemic inflammation in patients with chronic obstructive pulmonary disease
The co-occurrence of clinically important comorbidities in patients with chronic obstructive pulmonary disease (COPD) and the prevalence of multimorbidity in the pathophysiology of COPD are analyzed. Viscovery Profiler is used to identify comorbidity clusters in a data set of COPD patients and to characterize the clusters in terms of health status, clinical outcomes, and systemic inflammation.
A psycho-cognitive segmentation of organ donors in Egypt using Kohonen’s self-organizing maps
The psychographic background of organ donors is the focus of this paper. Viscovery SOMine is used to cluster panel data about Egyptian citizens’ attitudes towards organ donation.
47 glioblastoma gene expression profile diagnostics by the artificial neural networks
The connection between gene expression profiles and glioblastoma incidences is analyzed. Viscovery SOMine and a feed-forward neural network are used to classify the data and both produce very good results.
Psychographic clustering of blood donors in Egypt using Kohonen's self-organizing maps
This paper aims to find the motives behind donating (or not donating) blood. Viscovery SOMine is used on a data set consisting of panel data about Egyptian citizens’ view on blood donation.
Combining data mining and case-based reasoning for intelligent decision support for pathology ordering by general practitioners
In this paper, a novel methodology for integrating data mining and case-based reasoning for decision support for pathology ordering is proposed. It is demonstrated how this methodology can facilitate intelligent decision support that is both patient-oriented and deeply rooted in practical peer-group evidence. Knowledge extracted through data mining with Kohonen’s self-organizing maps constitutes the base that, with further assistance of the modern data visualization tool Viscovery SOMine and online processing interfaces, can facilitate more informed evidential decision making by doctors in the area of pathology ordering.
Using supervised and unsupervised techniques to determine groups of patients with different doctor–patient stability
Similarities between any groupings found with unsupervised classification using Viscovery SOMine and supervised classification using classification and regression trees are compared and used to identify insights into factors associated with doctor–patient stability. Both methods result in many similar groupings, indicating that self-perceived health and age are important indicators of stability. Profiles of patients that are at risk are identified.
Using self organising feature maps to unravel process complexity in a hospital emergency department: a decision support perspective
This book chapter describes how to use self-organizing maps to create a decision support system for hospital management practice. Viscovery SOMine is used to cluster emergency patients with respect to treatments performed and thus to reveal actual work processes.
Towards process-of-care aware emergency department information systems: a clustering approach to activity views elicitation
The focus of this article is on building an emergency department IT system and an analytical model to predict the patient workflow. The system is based on the treatment cluster model introduced by Ceglowski et. al. (2017) in their article "Combining data mining and discrete event simulation for a value-added view of a hospital emergency department" to predict likely treatment paths and resources needed for new patients.
Combining data mining and discrete event simulation for a value-added view of a hospital emergency department
Data from treatments given to emergency patients are clustered using Viscovery SOMine. Analysis reveals clusters of treatments related to injury (e.g., tetanus injections, dressings, sutures) and clusters of treatments related to illness (e.g., arterial blood gases, echocardiograms, and intravenous drug infusion). The results provide insight into the complex relationship between urgency, patient treatment and discharge, and the occurrences of queues for treatment.
An investigation of emergency department overcrowding using data mining and simulation: a patient treatment type perspective
To analyze the problem of overcrowding in emergency departments, homogenous clusters of patient treatment with similar activities are identified. Techniques from the dissociated methods of data mining and management science are combined within the hypothesis and experimentation framework of the scientific method. Viscovery SOMine is used for discovery of patient treatment patterns. The clusters are combined with patient urgency and disposition to create “patient treatment types” that are tracked through the emergency department.
On visual exploration of breast cancer data using the self-organizing map
This study describes a showcase analysis of 497 incidences of breast cancer to confirm well-known facts and demonstrate the use of self-organizing maps in cancer research. Viscovery SOMine is used for exploratory data analysis with respect to patient age, estrogen-receptor status, lymph-node status, size and histological grade of the tumor. A positive correlation between tumor size and histological grade and tumor size and likelihood of metastases as well as an inverse correlation between estrogen level and histological grade are found.
Histological heterogeneity of human glioblastomas investigated with an unsupervised neural network (SOM)
This paper aims at validating the World Health Organization classification of human glioblastomas and identifying new sub-groups. Viscovery SOMine is used to cluster 1489 glioblastomas with respect to 50 histological features, age and sex of the patients, providing new interesting insights into this form of brain tumor.
Analysis of hippocampal atrophy in alcoholic patients by a Kohonen feature map
The correlation of hippocampal volume with homocysteine, folate, vitamin B12 and B6 content in alcoholic patients and healthy controls is examined by applying Viscovery SOMine and conventional statistics. Viscovery SOMine proves to be a sensitive tool for visualization of statistical correlations in data sets, even when no further statistical information is available.
A neural clustering approach to iso-resource grouping for acute healthcare in Australia
The Case Mix funding formula is the most widely used approach for classifying patients according to diagnostic related groups (DRGs). Although it is clinically meaningful, experience suggests that DRG groupings do not necessarily present a sound basis for relevant knowledge generation. An alternative grouping of patients based on a neural clustering approach is proposed, generating homogeneous groups of patients with similar resource-utilization characteristics. Features of the data and the dependencies between the variables are identified and evaluated from the Viscovery SOMine map.
Clinical–pathological classification of glioblastomas investigated by a non-supervised neural network
Using a variant of unsupervised neural networks, the ability to reproduce a clinical–pathological classification of patients with glioblastomas is examined. This resulting self-organizing map provides a powerful means to visualize and analyze complex data sets without prior statistical knowledge and allows a specific visual evaluation of new treatments and a more effective comparison with established tumor management.
Light microscope heterogeneity of human glioblastomas investigated with an unsupervised neural network (SOM)
As an alternative to statistical evaluation of histological variability of glioblastomas, 1266 human glioblastomas are investigated to discover whether they can be correctly classified using self-organizing maps generated with Viscovery SOMine. Five clusters of glioblastomas with a maximum significance are found. A useful classification, comparable to the classification suggested by the World Health Organization, as well as the visualization of multidimensional histological features of human glioblastomas, is achieved. The data can be used to improve patient management.