Data analysis and data mining: current issues in biomedical informatics

Bioactive Molecules from the Innate Immunity of Ascidians and Innovative Methods of Drug Discovery: A Computational Approach Based on Artificial Intelligence

The study of bioactive molecules of marine origin has created an important bridge between biological knowledge and its applications in biotechnology and biomedicine. Current studies in different research fields, such as biomedicine, aim to discover marine molecules characterized by biological activities that can be used to produce potential drugs for human use. In recent decades, increasing attention has been paid to a particular group of marine invertebrates, the Ascidians, as they are a source of bioactive products. We describe omics data and computational methods relevant to identifying the mechanisms and processes of innate immunity underlying the biosynthesis of bioactive molecules, focusing on innovative computational approaches based on Artificial Intelligence. Since there is increasing attention on finding new solutions for a sustainable supply of bioactive compounds, we propose that a possible improvement in the biodiscovery pipeline might also come from the study and utilization of marine invertebrates' innate immunity.

Detection of atypical response trajectories in biomedical longitudinal databases

Many health care professionals and institutions manage longitudinal databases, involving follow-ups for different patients over time. Longitudinal data frequently manifest additional complexities such as high variability, correlated measurements and missing data. Mixed effects models have been widely used to overcome these difficulties. This work proposes the use of linear mixed effects models as a tool that allows to search conceptually different types of anomalies in the data simultaneously.

Drug repositioning in drug discovery of T2DM and repositioning potential of antidiabetic agents

Repositioning or repurposing drugs account for a substantial part of entering approval pipeline drugs, which indicates that drug repositioning has huge market potential and value. Computational technologies such as machine learning methods have accelerated the process of drug repositioning in the last few decades years. The repositioning potential of type 2 diabetes mellitus (T2DM) drugs for various diseases such as cancer, neurodegenerative diseases, and cardiovascular diseases have been widely studied. Hence, the related summary about repurposing antidiabetic drugs is of great significance. In this review, we focus on the machine learning methods for the development of new T2DM drugs and give an overview of the repurposing potential of the existing antidiabetic agents.

Correlates of physical activity behavior in adults: a data mining approach

PURPOSE

A data mining approach was applied to establish a multilevel hierarchy predicting physical activity (PA) behavior, and to methodologically identify the correlates of PA behavior.

METHODS

Cross-sectional data from the population-based Northern Finland Birth Cohort 1966 study, collected in the most recent follow-up at age 46, were used to create a hierarchy using the chi-square automatic interaction detection (CHAID) decision tree technique for predicting PA behavior. PA behavior is defined as active or inactive based on machine-learned activity profiles, which were previously created through a multidimensional (clustering) approach on continuous accelerometer-measured activity intensities in one week. The input variables (predictors) used for decision tree fitting consisted of individual, demographical, psychological, behavioral, environmental, and physical factors. Using generalized linear mixed models, we also analyzed how factors emerging from the model were associated with three PA metrics, including daily time (minutes per day) in sedentary (SED), light PA (LPA), and moderate-to-vigorous PA (MVPA), to assure the relative importance of methodologically identified factors.

RESULTS

Of the 4582 participants with valid accelerometer data at the latest follow-up, 2701 and 1881 had active and inactive profiles, respectively. We used a total of 168 factors as input variables to classify these two PA behaviors. Out of these 168 factors, the decision tree selected 36 factors of different domains from which 54 subgroups of participants were formed. The emerging factors from the model explained minutes per day in SED, LPA, and/or MVPA, including body fat percentage (SED: B = 26.5, LPA: B = - 16.1, and MVPA: B = - 11.7), normalized heart rate recovery 60 s after exercise (SED: B = -16.1, LPA: B = 9.9, and MVPA: B = 9.6), average weekday total sitting time (SED: B = 34.1, LPA: B = -25.3, and MVPA: B = -5.8), and extravagance score (SED: B = 6.3 and LPA: B = - 3.7).

CONCLUSIONS

Using data mining, we established a data-driven model composed of 36 different factors of relative importance from empirical data. This model may be used to identify subgroups for multilevel intervention allocation and design. Additionally, this study methodologically discovered an extensive set of factors that can be a basis for additional hypothesis testing in PA correlates research.

Predictors of User Engagement With Facebook Posts Generated by a National Sample of Lesbian, Gay, Bisexual, Transgender, and Queer Community Centers in the United States: Content Analysis

BACKGROUND

Lesbian, gay, bisexual, transgender, and queer (LGBTQ) community centers remain important venues for reaching and providing crucial health and social services to LGBTQ individuals in the United States. These organizations commonly use Facebook to reach their target audiences, but little is known about factors associated with user engagement with their social media presence.

OBJECTIVE

This study aimed to identify factors associated with engagement with Facebook content generated by LGBTQ community centers in the United States.

METHODS

Content generated by LGBTQ community centers in 2017 was downloaded using Facebook's application programming interface. Posts were classified by their content and sentiment. Correlates of user engagement were identified using negative binomial regression.

RESULTS

A total of 32,014 posts from 175 community centers were collected. Posts with photos (incidence rate ratio, [IRR] 1.07; 95% CI 1.06-1.09) and videos (IRR 1.54; 95% CI 1.52-1.56) that contained a direct invitation for engagement (IRR 1.03; 95% CI 1.02-1.04), that expressed a positive sentiment (IRR 1.11; 95% CI 1.10-1.12), and that contained content related to stigma (IRR 1.16; 95% CI 1.14-1.17), mental health (IRR 1.33; 95% CI 1.31-1.35), and politics (IRR 1.28; 95% CI 1.27-1.29) received higher levels of engagement.

CONCLUSIONS

The results of this study provide support for the use of Facebook to extend the reach of LGBTQ community centers and highlight multiple factors that can be leveraged to optimize engagement.

Linear discriminant analysis and principal component analysis to predict coronary artery disease

Coronary artery disease is one of the most prevalent chronic pathologies in the modern world, leading to the deaths of thousands of people, both in the United States and in Europe. This article reports the use of data mining techniques to analyse a population of 10,265 people who were evaluated by the Department of Advanced Biomedical Sciences for myocardial ischaemia. Overall, 22 features are extracted, and linear discriminant analysis is implemented twice through both the Knime analytics platform and R statistical programming language to classify patients as either normal or pathological. The former of these analyses includes only classification, while the latter method includes principal component analysis before classification to create new features. The classification accuracies obtained for these methods were 84.5 and 86.0 per cent, respectively, with a specificity over 97 per cent and a sensitivity between 62 and 66 per cent. This article presents a practical implementation of traditional data mining techniques that can be used to help clinicians in decision-making; moreover, principal component analysis is used as an algorithm for feature reduction.

Concepts of Artificial Intelligence for Computer-Assisted Drug Discovery

Artificial intelligence (AI), and, in particular, deep learning as a subcategory of AI, provides opportunities for the discovery and development of innovative drugs. Various machine learning approaches have recently (re)emerged, some of which may be considered instances of domain-specific AI which have been successfully employed for drug discovery and design. This review provides a comprehensive portrayal of these machine learning techniques and of their applications in medicinal chemistry. After introducing the basic principles, alongside some application notes, of the various machine learning algorithms, the current state-of-the art of AI-assisted pharmaceutical discovery is discussed, including applications in structure- and ligand-based virtual screening, de novo drug design, physicochemical and pharmacokinetic property prediction, drug repurposing, and related aspects. Finally, several challenges and limitations of the current methods are summarized, with a view to potential future directions for AI-assisted drug discovery and design.

An Enhanced Symptom Clustering with Profile Based Prescription Suggestion in Biomedical application

The application of data mining has been increasing day to day whereas the data base is also enhancing simultaneously. Hence retrieving required content from a huge data base is a critical task. This paper focus on biomedical engineering field, it concentrates on initial stage of database such as data preprocessing and cleansing to deal with noise and missing data in large biomedical data sets. The database of biomedical is huge and enhancing nature retrieving of specific content will be a critical task. Suggesting prescription with respect to identified disease based on profile analysis of specific patient is not available in current system. This paper proposes a recommendation system of prescription based on disease identification is done by combining user and professional suggestion with profile based analysis. Hence this focuses on profile based suggestions and report will be generated. The retrieving of specific suggestion from a huge database is done by hybrid feature selection algorithm. This approach focuses on enabling recommendation based on user profile and implementing Hybrid feature selection algorithm to retrieve specific content from a huge database. Hence it attains better retrieval of required content from a huge database compared to other existing approaches and suggests better recommendation with respect to user profile.

Regularization and grouping -omics data by GCA method: A transcriptomic case

The paper presents the application of Grade Correspondence Analysis (GCA) and Grade Correspondence Cluster Analysis (GCCA) for ordering and grouping -omics datasets, using transcriptomic data as an example. Based on gene expression data describing 256 patients with Multiple Myeloma it was shown that the GCA method could be used to find regularities in the analyzed collections and to create characteristic gene expression profiles for individual groups of patients. GCA iteratively permutes rows and columns to maximize the tau-Kendall or rho-Spearman coefficients, which makes it possible to arrange rows and columns in such a way that the most similar ones remain in each other’s neighbourhood. In this way, the GCA algorithm highlights regularities in the data matrix. The ranked data can then be grouped using the GCCA method, and after that aggregated in clusters, providing a representation that is easier to analyze–especially in the case of large sets of gene expression profiles. Regularization of transcriptomic data, which is presented in this manuscript, has enabled division of the data set into column clusters (representing genes) and row clusters (representing patients). Subsequently, rows were aggregated (based on medians) to visualise the gene expression profiles for patients with Multiple Myeloma in each collection. The presented analysis became the starting point for characterisation of differentiated genes and biochemical processes in which they are involved. GCA analysis may provide an alternative analytical method to support differentiation and analysis of gene expression profiles characterising individual groups of patients.

Thirty biologically interpretable clusters of transcription factors distinguish cancer type

Background

Transcription factors are essential regulators of gene expression and play critical roles in development, differentiation, and in many cancers. To carry out their regulatory programs, they must cooperate in networks and bind simultaneously to sites in promoter or enhancer regions of genes. We hypothesize that the mRNA co-expression patterns of transcription factors can be used both to learn how they cooperate in networks and to distinguish between cancer types.

Results

We recently developed a new algorithm, Thresher, that combines principal component analysis, outlier filtering, and von Mises-Fisher mixture models to cluster genes (in this case, transcription factors) based on expression, determining the optimal number of clusters in the process. We applied Thresher to the RNA-Seq expression data of 486 transcription factors from more than 10,000 samples of 33 kinds of cancer studied in The Cancer Genome Atlas (TCGA). We found that 30 clusters of transcription factors from a 29-dimensional principal component space were able to distinguish between most cancer types, and could separate tumor samples from normal controls. Moreover, each cluster of transcription factors could be either (i) linked to a tissue-specific expression pattern or (ii) associated with a fundamental biological process such as cell cycle, angiogenesis, apoptosis, or cytoskeleton. Clusters of the second type were more likely also to be associated with embryonically lethal mouse phenotypes.

Conclusions

Using our approach, we have shown that the mRNA expression patterns of transcription factors contain most of the information needed to distinguish different cancer types. The Thresher method is capable of discovering biologically interpretable clusters of genes. It can potentially be applied to other gene sets, such as signaling pathways, to decompose them into simpler, yet biologically meaningful, components.

Electronic supplementary material

The online version of this article (10.1186/s12864-018-5093-z) contains supplementary material, which is available to authorized users.

The Preterm Clinical Network (PCN) Database: a web-based systematic method of collecting data on the care of women at risk of preterm birth

Background

Despite much research effort, there is a paucity of conclusive evidence in the field of preterm birth prediction and prevention. The methods of monitoring and prevention strategies offered to women at risk vary considerably around the UK and depend on local maternity care provision. It is becoming increasingly recognised that this experience and knowledge, if captured on a larger scale, could be a utilized as a valuable source of evidence for others. The UK Preterm Clinical Network (UKPCN) was established with the aim of improving care and outcomes for women at risk of preterm birth through the sharing of a wealth of experience and knowledge, as well as the building of clinical and research collaboration. The design and development of a bespoke internet-based database was fundamental to achieving this aim.

Method

Following consultation with UKPCN members and agreement on a minimal dataset, the Preterm Clinical Network (PCN) Database was constructed to collect data from women at risk of preterm birth and their children. Information Governance and research ethics committee approval was given for the storage of historical as well as prospectively collected data. Collaborating centres have instant access to their own records, while use of pooled data is governed by the PCN Database Access Committee. Applications are welcomed from UKPCN members and other established research groups. The results of investigations using the data are expected to provide insights into the effectiveness of current surveillance practices and preterm birth interventions on a national and international scale, as well as the generation of ideas for innovation and research. To date, 31 sites are registered as Data Collection Centres, four of which are outside the UK.

Conclusion

This paper outlines the aims of the PCN Database along with the development process undertaken from the initial idea to live launch.

Predictors of Facebook User Engagement With Health-Related Content for Gay, Bisexual, and Other Men Who Have Sex With Men: Content Analysis

Background

Social media is used by community-based organizations (CBOs) to promote the well-being of gay and bisexual men (GBM). However, few studies have quantified which factors facilitate the diffusion of health content tailored for sexual minorities.

Objective

The aim of this study was to identify post characteristics that can be leveraged to optimize the health promotion efforts of CBOs on Facebook.

Methods

The Facebook application programming interface was used to collect 5 years’ of posts shared across 10 Facebook pages administered by Vancouver-based CBOs promoting GBM health. Network analysis assessed basic indicators of network structure. Content analyses were conducted using informatics-based approaches. Hierarchical negative binomial regression of post engagement data was used to identify meaningful covariates of engagement.

Results

In total, 14,071 posts were shared and 21,537 users engaged with these posts. Most users (n=13,315) engaged only once. There was moderate correlation between the number of posts and the number of CBOs users engaged with (r=.53, P<.001). Higher user engagement was positively associated with positive sentiment, sharing multimedia, and posting about pre-exposure prophylaxis, stigma, and mental health. Engagement was negatively associated with asking questions, posting about dating, and sharing posts during or after work (versus before).

Conclusions

Results highlight the existence of a core group of Facebook users who facilitate diffusion. Factors associated with greater user engagement present CBOs with a number of strategies for improving the diffusion of health content.

Emerging Diagnostic and Therapeutic Potentials of Human Hair Proteomics

The use of noninvasive human substrates to interrogate pathophysiological conditions has become essential in the post- Human Genome Project era. Due to its high turnover rate, and its long term capability to incorporate exogenous and endogenous substances from the circulation, hair testing is emerging as a key player in monitoring long term drug compliance, chronic alcohol abuse, forensic toxicology, and biomarker discovery, among other things. Novel high-throughput 'omics based approaches like proteomics have been underutilized globally in comprehending human hair morphology and its evolving use as a diagnostic testing substrate in the era of precision medicine. There is paucity of scientific evidence that evaluates the difference in drug incorporation into hair based on lipid content, and very few studies have addressed hair growth rates, hair forms, and the biological consequences of hair grooming or bleaching. It is apparent that protein-based identification using the human hair proteome would play a major role in understanding these parameters akin to DNA single nucleotide polymorphism profiling, up to single amino acid polymorphism resolution. Hence, this work seeks to identify and discuss the progress made thus far in the field of molecular hair testing using proteomic approaches, and identify ways in which proteomics would improve the field of hair research, considering that the human hair is mostly composed of proteins. Gaps in hair proteomics research are identified and the potential of hair proteomics in establishing a historic medical repository of normal and disease-specific proteome is also discussed.

A Shared Opportunity for Improving Electronic Medical Record Data

With the recent rapid adoption of electronic medical records (EMRs), studies reporting results based on EMR data have become increasingly common. While analyzing data extracted from our EMR for a retrospective study, we identified various types of erroneous data entries. This report investigates the root causes of the incompleteness, inconsistency, and inaccuracy of the medical records analyzed in our study. While experienced health information management professionals are well aware of the many shortcomings with EMR data, the aims of this case study are to highlight the significance of the negative impact of erroneous EMR data, to provide fundamental principles for managing EMRs, and to provide recommendations to help facilitate the successful use of electronic health data, whether to inform clinical decisions or for clinical research.

New approach to generating insights for aging research based on literature mining and knowledge integration

The proportion of the elderly population in most countries worldwide is increasing dramatically. Therefore, social interest in the fields of health, longevity, and anti-aging has been increasing as well. However, the basic research results obtained from a reductionist approach in biology and a bioinformatic approach in genome science have limited usefulness for generating insights on future health, longevity, and anti-aging-related research on a case by case basis. We propose a new approach that uses our literature mining technique and bioinformatics, which lead to a better perspective on research trends by providing an expanded knowledge base to work from. We demonstrate that our approach provides useful information that deepens insights on future trends which differs from data obtained conventionally, and this methodology is already paving the way for a new field in aging-related research based on literature mining. One compelling example of this is how our new approach can be a useful tool in drug repositioning.

Perioperative and ICU Healthcare Analytics within a Veterans Integrated System Network: a Qualitative Gap Analysis

Health care facilities are implementing analytics platforms as a way to document quality of care. However, few gap analyses exist on platforms specifically designed for patients treated in the Operating Room, Post-Anesthesia Care Unit, and Intensive Care Unit (ICU). As part of a quality improvement effort, we undertook a gap analysis of an existing analytics platform within the Veterans Healthcare Administration. The objectives were to identify themes associated with 1) current clinical use cases and stakeholder needs; 2) information flow and pain points; and 3) recommendations for future analytics development. Methods consisted of semi-structured interviews in 2 phases with a diverse set (n = 9) of support personnel and end users from five facilities across a Veterans Integrated Service Network. Phase 1 identified underlying needs and previous experiences with the analytics platform across various roles and operational responsibilities. Phase 2 validated preliminary feedback, lessons learned, and recommendations for improvement. Emerging themes suggested that the existing system met a small pool of national reporting requirements. However, pain points were identified with accessing data in several information system silos and performing multiple manual validation steps of data content. Notable recommendations included enhancing systems integration to create "one-stop shopping" for data, and developing a capability to perform trends analysis. Our gap analysis suggests that analytics platforms designed for surgical and ICU patients should employ approaches similar to those being used for primary care patients.

Fibroblast growth factor 23 weakens chemotaxis of human blood neutrophils in microfluidic devices

Neutrophil trafficking in tissues critically regulates the body’s immune response. Neutrophil migration can either play a protective role in host defense or cause health problems. Fibroblast growth factor 23 (FGF23) is a known biomarker for chronic kidney disease (CKD) and was recently shown to impair neutrophil arrest on endothelium and transendothelial migration. In the present study, we further examined the effect of FGF23 on human blood neutrophil chemotaxis using two new microfluidic devices. Our results showed that chemotaxis of FGF23 pre-treated neutrophils to a fMLP gradient, in the presence or absence of a uniform FGF23 background, is quantitatively lower compared to the control cells. This effect is accompanied with a stronger drifting of FGF23 pre-treated cells along the flow. However, without the FGF23 pre-treatment, the FGF23 background only reduces chemotaxis of transmigrated cells through the thin barrier channel to the fMLP gradient. The effect of FGF23 on neutrophil migration and the correlation between multiple cell migration parameters are further revealed by chemotactic entropy and principle component analysis. Collectively, these results revealed the effect of FGF23 on weakening neutrophil chemotaxis, which shed light on FGF23 mediated neutrophil migration with direct disease relevance such as CKD.

Optimizing R with SparkR on a commodity cluster for biomedical research

BACKGROUND AND OBJECTIVES

Medical researchers are challenged today by the enormous amount of data collected in healthcare. Analysis methods such as genome-wide association studies (GWAS) are often computationally intensive and thus require enormous resources to be performed in a reasonable amount of time. While dedicated clusters and public clouds may deliver the desired performance, their use requires upfront financial efforts or anonymous data, which is often not possible for preliminary or occasional tasks. We explored the possibilities to build a private, flexible cluster for processing scripts in R based on commodity, non-dedicated hardware of our department.

METHODS

For this, a GWAS-calculation in R on a single desktop computer, a Message Passing Interface (MPI)-cluster, and a SparkR-cluster were compared with regards to the performance, scalability, quality, and simplicity.

RESULTS

The original script had a projected runtime of three years on a single desktop computer. Optimizing the script in R already yielded a significant reduction in computing time (2 weeks). By using R-MPI and SparkR, we were able to parallelize the computation and reduce the time to less than three hours (2.6 h) on already available, standard office computers. While MPI is a proven approach in high-performance clusters, it requires rather static, dedicated nodes. SparkR and its Hadoop siblings allow for a dynamic, elastic environment with automated failure handling. SparkR also scales better with the number of nodes in the cluster than MPI due to optimized data communication.

CONCLUSION

R is a popular environment for clinical data analysis. The new SparkR solution offers elastic resources and allows supporting big data analysis using R even on non-dedicated resources with minimal change to the original code. To unleash the full potential, additional efforts should be invested to customize and improve the algorithms, especially with regards to data distribution.

Clinical Decision Support: a 25 Year Retrospective and a 25 Year Vision

OBJECTIVE

The objective of this review is to summarize the state of the art of clinical decision support (CDS) circa 1990, review progress in the 25 year interval from that time, and provide a vision of what CDS might look like 25 years hence, or circa 2040.

METHOD

Informal review of the medical literature with iterative review and discussion among the authors to arrive at six axes (data, knowledge, inference, architecture and technology, implementation and integration, and users) to frame the review and discussion of selected barriers and facilitators to the effective use of CDS.

RESULT

In each of the six axes, significant progress has been made. Key advances in structuring and encoding standardized data with an increased availability of data, development of knowledge bases for CDS, and improvement of capabilities to share knowledge artifacts, explosion of methods analyzing and inferring from clinical data, evolution of information technologies and architectures to facilitate the broad application of CDS, improvement of methods to implement CDS and integrate CDS into the clinical workflow, and increasing sophistication of the end-user, all have played a role in improving the effective use of CDS in healthcare delivery.

CONCLUSION

CDS has evolved dramatically over the past 25 years and will likely evolve just as dramatically or more so over the next 25 years. Increasingly, the clinical encounter between a clinician and a patient will be supported by a wide variety of cognitive aides to support diagnosis, treatment, care-coordination, surveillance and prevention, and health maintenance or wellness.

In silico methods for drug repurposing and pharmacology

Data in the biological, chemical, and clinical domains are accumulating at ever-increasing rates and have the potential to accelerate and inform drug development in new ways. Challenges and opportunities now lie in developing analytic tools to transform these often complex and heterogeneous data into testable hypotheses and actionable insights. This is the aim of computational pharmacology, which uses in silico techniques to better understand and predict how drugs affect biological systems, which can in turn improve clinical use, avoid unwanted side effects, and guide selection and development of better treatments. One exciting application of computational pharmacology is drug repurposing-finding new uses for existing drugs. Already yielding many promising candidates, this strategy has the potential to improve the efficiency of the drug development process and reach patient populations with previously unmet needs such as those with rare diseases. While current techniques in computational pharmacology and drug repurposing often focus on just a single data modality such as gene expression or drug-target interactions, we argue that methods such as matrix factorization that can integrate data within and across diverse data types have the potential to improve predictive performance and provide a fuller picture of a drug's pharmacological action. WIREs Syst Biol Med 2016, 8:186-210. doi: 10.1002/wsbm.1337 For further resources related to this article, please visit the WIREs website.

Multimodality monitoring: informatics, integration data display and analysis

The goal of multimodality neuromonitoring is to provide continuous, real-time assessment of brain physiology to prevent, detect, and attenuate secondary brain injury. Clinical informatics deals with biomedical data, information, and knowledge including their acquisition, storage, retrieval, and optimal use for clinical decision-making. An electronic literature search was conducted for English language articles describing the use of informatics in the intensive care unit setting from January 1990 to August 2013. A total of 64 studies were included in this review. Clinical informatics infrastructure should be adopted that enables a wide range of linear and nonlinear analytical methods be applied to patient data. Specific time epochs of clinical interest should be reviewable. Analysis strategies of monitor alarms may help address alarm fatigue. Ergonomic data display that present results from analyses with clinical information in a sensible uncomplicated manner improve clinical decision-making. Collecting and archiving the highest resolution physiologic and phenotypic data in a comprehensive open format data warehouse is a crucial first step toward information management and two-way translational research for multimodality monitoring. The infrastructure required is largely the same as that needed for telemedicine intensive care applications, which under the right circumstances improves care quality while reducing cost.

Success/Failure Prediction of Noninvasive Mechanical Ventilation in Intensive Care Units. Using Multiclassifiers and Feature Selection Methods

OBJECTIVES

This paper addresses the problem of decision-making in relation to the administration of noninvasive mechanical ventilation (NIMV) in intensive care units.

METHODS

Data mining methods were employed to find out the factors influencing the success/failure of NIMV and to predict its results in future patients. These artificial intelligence-based methods have not been applied in this field in spite of the good results obtained in other medical areas.

RESULTS

Feature selection methods provided the most influential variables in the success/failure of NIMV, such as NIMV hours, PaCO2 at the start, PaO2 / FiO2 ratio at the start, hematocrit at the start or PaO2 / FiO2 ratio after two hours. These methods were also used in the preprocessing step with the aim of improving the results of the classifiers. The algorithms provided the best results when the dataset used as input was the one containing the attributes selected with the CFS method.

CONCLUSIONS

Data mining methods can be successfully applied to determine the most influential factors in the success/failure of NIMV and also to predict NIMV results in future patients. The results provided by classifiers can be improved by preprocessing the data with feature selection techniques.

Prediction of breast cancer survival through knowledge discovery in databases

The collection of large volumes of medical data has offered an opportunity to develop prediction models for survival by the medical research community. Medical researchers who seek to discover and extract hidden patterns and relationships among large number of variables use knowledge discovery in databases (KDD) to predict the outcome of a disease. The study was conducted to develop predictive models and discover relationships between certain predictor variables and survival in the context of breast cancer. This study is Cross sectional. After data preparation, data of 22,763 female patients, mean age 59.4 years, stored in the Surveillance Epidemiology and End Results (SEER) breast cancer dataset were analyzed anonymously. IBM SPSS Statistics 16, Access 2003 and Excel 2003 were used in the data preparation and IBM SPSS Modeler 14.2 was used in the model design. Support Vector Machine (SVM) model outperformed other models in the prediction of breast cancer survival. Analysis showed SVM model detected ten important predictor variables contributing mostly to prediction of breast cancer survival. Among important variables, behavior of tumor as the most important variable and stage of malignancy as the least important variable were identified. In current study, applying of the knowledge discovery method in the breast cancer dataset predicted the survival condition of breast cancer patients with high confidence and identified the most important variables participating in breast cancer survival.

Optimizing data collection for public health decisions: a data mining approach

Background

Collecting data can be cumbersome and expensive. Lack of relevant, accurate and timely data for research to inform policy may negatively impact public health. The aim of this study was to test if the careful removal of items from two community nutrition surveys guided by a data mining technique called feature selection, can (a) identify a reduced dataset, while (b) not damaging the signal inside that data.

Methods

The Nutrition Environment Measures Surveys for stores (NEMS-S) and restaurants (NEMS-R) were completed on 885 retail food outlets in two counties in West Virginia between May and November of 2011. A reduced dataset was identified for each outlet type using feature selection. Coefficients from linear regression modeling were used to weight items in the reduced datasets. Weighted item values were summed with the error term to compute reduced item survey scores. Scores produced by the full survey were compared to the reduced item scores using a Wilcoxon rank-sum test.

Results

Feature selection identified 9 store and 16 restaurant survey items as significant predictors of the score produced from the full survey. The linear regression models built from the reduced feature sets had R² values of 92% and 94% for restaurant and grocery store data, respectively.

Conclusions

While there are many potentially important variables in any domain, the most useful set may only be a small subset. The use of feature selection in the initial phase of data collection to identify the most influential variables may be a useful tool to greatly reduce the amount of data needed thereby reducing cost.

On determining factors for good research in biomedical and health informatics. Some lessons learned

OBJECTIVE

What are the determining factors for good research in medical informatics or, from a broader perspective, in biomedical and health informatics?

METHOD

From the many lessons learned during my professional career, I tried to identify a fair sampling of such factors. On the occasion of giving the IMIA Award of Excellence lecture during MedInfo 2013, they were presented for discussion.

RESULTS

Sixteen determining factors (df) have been identified: early identification and promotion (df1), appropriate education (df2), stimulating persons and environments (df3), sufficient time and backtracking opportunities (df4), breadth of medical informatics competencies (df5), considering the necessary preconditions for good medical informatics research (df6), easy access to high-quality knowledge (df7), sufficient scientific career opportunities (df8), appropriate conditions for sustainable research (df9), ability to communicate and to solve problems (df10), as well as to convey research results (df11) in a highly inter- and multidisciplinary environment, ability to think for all and, when needed, taking the lead (df12), always staying unbiased (df13), always keeping doubt (df14), but also always trying to provide solutions (df15), and, finally, being aware that life is more (df16).

CONCLUSIONS

Medical Informatics is an inter- and multidisciplinary discipline "avant la lettre". Compared to monodisciplinary research, inter- and multidisciplinary research does not only provide significant opportunities for solving major problems in science and in society. It also faces considerable additional challenges for medical informatics as a scientific field. The determining factors, presented here, are in my opinion crucial for conducting successful research and for developing a research career. Since medical informatics as a field has today become an important driving force for research progress, especially in biomedicine and health care, but also in fields like computer science, it may be helpful to consider such factors in relation with research and education in our discipline.

Black box integration of computer-aided diagnosis into PACS deserves a second chance: results of a usability study concerning bone age assessment

Usability aspects of different integration concepts for picture archiving and communication systems (PACS) and computer-aided diagnosis (CAD) were inquired on the example of BoneXpert, a program determining the skeletal age from a left hand's radiograph. CAD-PACS integration was assessed according to its levels: data, function, presentation, and context integration focusing on usability aspects. A user-based study design was selected. Statements of seven experienced radiologists using two alternative types of integration provided by BoneXpert were acquired and analyzed using a mixed-methods approach based on think-aloud records and a questionnaire. In both variants, the CAD module (BoneXpert) was easily integrated in the workflow, found comprehensible and fitting in the conceptual framework of the radiologists. Weak points of the software integration referred to data and context integration. Surprisingly, visualization of intermediate image processing states (presentation integration) was found less important as compared to efficient handling and fast computation. Seamlessly integrating CAD into the PACS without additional work steps or unnecessary interrupts and without visualizing intermediate images may considerably improve software performance and user acceptance with efforts in time.

Improvement of adequate use of warfarin for the elderly using decision tree-based approaches

OBJECTIVES

Due to the narrow therapeutic range and high drug-to-drug interactions (DDIs), improving the adequate use of warfarin for the elderly is crucial in clinical practice. This study examines whether the effectiveness of using warfarin among elderly inpatients can be improved when machine learning techniques and data from the laboratory information system are incorporated.

METHODS

Having employed 288 validated clinical cases in the DDI group and 89 cases in the non-DDI group, we evaluate the prediction performance of seven classification techniques, with and without an Adaptive Boosting (AdaBoost) algorithm. Measures including accuracy, sensitivity, specificity and area under the curve are used to evaluate model performance.

RESULTS

Decision tree-based classifiers outperform other investigated classifiers in all evaluation measures. The classifiers supplemented with AdaBoost can generally improve the performance. In addition, weight, congestive heart failure, and gender are among the top three critical variables affecting prediction accuracy for the non-DDI group, while age, ALT, and warfarin doses are the most influential factors for the DDI group.

CONCLUSION

Medical decision support systems incorporating decision tree-based approaches improve predicting performance and thus may serve as a supplementary tool in clinical practice. Information from laboratory tests and inpatients' history should not be ignored because related variables are shown to be decisive in our prediction models, especially when the DDIs exist.

Generating reference models for structurally complex data. Application to the stabilometry medical domain

OBJECTIVES

We present a framework specially designed to deal with structurally complex data, where all individuals have the same structure, as is the case in many medical domains. A structurally complex individual may be composed of any type of single-valued or multivalued attributes, including time series, for example. These attributes are structured according to domain-dependent hierarchies. Our aim is to generate reference models of population groups. These models represent the population archetype and are very useful for supporting such important tasks as diagnosis, detecting fraud, analyzing patient evolution, identifying control groups, etc.

METHODS

We have developed a conceptual model to represent structurally complex data hierarchically. Additionally, we have devised a method that uses the similarity tree concept to measure how similar two structurally complex individuals are, plus an outlier detection and filtering method. These methods provide the groundwork for the method that we have designed for generating reference models of a set of structurally complex individuals. A key idea of this method is to use event-based analysis for modeling time series.

RESULTS

The proposed framework has been applied to the medical field of stabilometry. To validate the outlier detection method we used 142 individuals, and there was a match between the outlier ratings by the experts and by the system for 139 individuals (97.8%). To validate the reference model generation method, we applied k-fold cross validation (k = 5) with 60 athletes (basketball players and ice-skaters), and the system correctly classified 55 (91.7%). We then added 30 non-athletes as a control group, and the method output the correct result in a very high percentage of cases (96.6%).

CONCLUSIONS

We have achieved very satisfactory results for the tests on data from such a complex domain as stabilometry and for the comparison of the reference model generation method with other methods. This supports the validity of this framework.

Disease-based modeling to predict fluid response in intensive care units

OBJECTIVE

To compare general and disease-based modeling for fluid resuscitation and vasopressor use in intensive care units.

METHODS

Retrospective cohort study involving 2944 adult medical and surgical intensive care unit (ICU) patients receiving fluid resuscitation. Within this cohort there were two disease-based groups, 802 patients with a diagnosis of pneumonia, and 143 patients with a diagnosis of pancreatitis. Fluid resuscitation either progressing to subsequent vasopressor administration or not was used as the primary outcome variable to compare general and disease-based modeling.

RESULTS

Patients with pancreatitis, pneumonia and the general group all shared three common predictive features as core variables, arterial base excess, lactic acid and platelets. Patients with pneumonia also had non-invasive systolic blood pressure and white blood cells added to the core model, and pancreatitis patients additionally had temperature. Disease-based models had significantly higher values of AUC (p < 0.05) than the general group (0.82 ± 0.02 for pneumonia and 0.83 ± 0.03 for pancreatitis vs. 0.79 ± 0.02 for general patients).

CONCLUSIONS

Disease-based predictive modeling reveals a different set of predictive variables compared to general modeling and improved performance. Our findings add support to the growing body of evidence advantaging disease specific predictive modeling.

Intelligent data analysis to model and understand live cell time-lapse sequences

BACKGROUND

One important aspect of cellular function, which is at the basis of tissue homeostasis, is the delivery of proteins to their correct destinations. Significant advances in live cell microscopy have allowed tracking of these pathways by following the dynamics of fluorescently labelled proteins in living cells.

OBJECTIVES

This paper explores intelligent data analysis techniques to model the dynamic behavior of proteins in living cells as well as to classify different experimental conditions.

METHODS

We use a combination of decision tree classification and hidden Markov models. In particular, we introduce a novel approach to "align" hidden Markov models so that hidden states from different models can be cross-compared.

RESULTS

Our models capture the dynamics of two experimental conditions accurately with a stable hidden state for control data and multiple (less stable) states for the experimental data recapitulating the behaviour of particle trajectories within live cell time-lapse data.

CONCLUSIONS

In addition to having successfully developed an automated framework for the classification of protein transport dynamics from live cell time-lapse data our model allows us to understand the dynamics of a complex trafficking pathway in living cells in culture.

Statistical process control for monitoring standardized mortality ratios of a classification tree model

OBJECTIVES

The ratio of observed to expected mortality (standardized mortality ratio, SMR), is a key indicator of quality of care. We use PreControl Charts to investigate SMR behavior over time of an existing tree-model for predicting mortality in intensive care units (ICUs) and its implications for hospital ranking. We compare the results to those of a logistic regression model.

METHODS

We calculated SMRs of 30 equally-sized consecutive subsets from a total of 12,143 ICU patients aged 80 years or older and plotted them on a PreControl Chart. We calculated individual hospital SMRs in 2009, with and without repeated recalibration of the models on earlier data.

RESULTS

The overall SMR of the tree-model was stable over time, in contrast to logistic regression. Both models were stable after repeated recalibration. The overall SMR of the tree on the whole validation set was statistically significantly different (SMR 1.00 ± 0.012 vs. 0.94 ± 0.01) and worse in performance than the logistic regression model (AUC 0.76 ± 0.005 vs. 0.79 ± 0.004; Brier score 0.17 ± 0.012 vs. 0.16 ± 0.010). The individual SMRs' range in 2009 was 0.53-1.31 for the tree and 0.64-1.27 for logistic regression. The proportion of individual hospitals with SMR >1, hinting at poor quality of care, reduced from 38% to 29% after recalibration for the tree, and increased from 15% to 35% for logistic regression.

CONCLUSIONS

Although the tree-model has seemingly a longer shelf life than the logistic regression model, its SMR may be less useful for quality of care assessment as it insufficiently responds to changes in the population over time.