OBJECTIVES: We report on findings from a meeting convened by the American College of Medical Informatics (ACMI) to characterize aspects of the patient experience that could be improved using informatics. MATERIALS AND METHODS: The American College of Medical Informatics fellows were invited to share their experiences as patients and suggest informatics approaches that may improve the patient experience. RESULTS: We identified 4 themes: (1) getting the right care, (2) data sharing and data interoperability, (3) guiding low-cost evaluations, and (4) predictive analytics. DISCUSSION: Despite widespread adoption of health IT, patient experiences remain far from optimal. CONCLUSION: The American College of Medical Informatics fellows identified informatics approaches, applications, and research areas that have the potential to improve patient experiences with health care systems.

Prostate cancer is a complex disease which advances in stages. While clinical failure (including metastasis) is a significant endpoint following a radical prostatectomy, it can often take years to manifest, usually too late to be optimistically treated. Instead the earlier endpoint of PSA Recurrence is frequently used as a surrogate in prognostic modeling. The central issue in these models is managing censored observations which challenge traditional regression techniques. The true target times of a majority of instances are unknown, what is known is a censored target representing some earlier indeterminate time. In this work we apply a novel transduction approach for semi-supervised survival analysis which has previously been shown to be powerful in medical prognosis. The approach considers censored samples as semi-supervised regression targets leveraging the partial nature of unsupervised information. In this work, the approach leads to a significant increase in performance for predicting advanced prostate cancer from earlier endpoints and may also be useful in other diseases for predicting advanced endpoints from earlier stages.

BACKGROUND: Biomedical information and knowledge, structural and non-structural, stored in different repositories can be semantically connected to form a hybrid knowledge network. How to compute relatedness between concepts and discover valuable but implicit information or knowledge from it effectively and efficiently is of paramount importance for precision medicine, and a major challenge facing the biomedical research community. RESULTS: In this study, a hybrid biomedical knowledge network is constructed by linking concepts across multiple biomedical ontologies as well as non-structural biomedical knowledge sources. To discover implicit relatedness between concepts in ontologies for which potentially valuable relationships (implicit knowledge) may exist, we developed a Multi-Ontology Relatedness Model (MORM) within the knowledge network, for which a relatedness network (RN) is defined and computed across multiple ontologies using a formal inference mechanism of set-theoretic operations. Semantic constraints are designed and implemented to prune the search space of the relatedness network. CONCLUSIONS: Experiments to test examples of several biomedical applications have been carried out, and the evaluation of the results showed an encouraging potential of the proposed approach to biomedical knowledge discovery.

Ontologies, seen as effective representations for sharing and reusing knowledge, have become increasingly important in biomedicine, usually focusing on taxonomic knowledge specific to a subject. Efforts have been made to uncover implicit knowledge within large biomedical ontologies by exploring semantic similarity and relatedness between concepts. However, much less attention has been paid to another potentially helpful approach: discovering implicit knowledge across multiple ontologies of different types, such as disease ontologies, symptom ontologies, and gene ontologies. In this paper, we propose a unified approach to the problem of ontology based implicit knowledge discovery - a Multi-Ontology Relatedness Model (MORM), which includes the formation of multiple related ontologies, a relatedness network and a formal inference mechanism based on set-theoretic operations. Experiments for biomedical applications have been carried out, and preliminary results show the potential value of the proposed approach for biomedical knowledge discovery.

The central challenge in predictive modeling for survival analysis in medical prognostics is managing censored observations. Traditional regression techniques are challenged by these censored samples. In such problems the true target times of a majority of instances are unknown; what is known is a censored target representing some earlier indeterminate time. Such censored samples can be considered as semi-supervised targets; however most efforts in semi-supervised regression do not take into account the partial nature of unsupervised information; samples are treated as either fully labelled or unlabeled. In this work we extend a novel transduction approach for semi-supervised survival analysis to neural networks. The true target times are approximated from the censored times through transduction. For prostate and breast cancer applications, this semi-supervised regression framework yields a significant improvement in performance.

Online learned tracking is widely used for its adaptive ability to handle appearance changes. However, it introduces potential drifting problems due to the accumulation of errors during the self-updating, especially for occluded scenarios. The recent literature demonstrates that appropriate combinations of trackers can help balance the stability and flexibility requirements. We have developed a robust tracking algorithm using a local sparse appearance model (SPT) and K-Selection. A static sparse dictionary and a dynamically updated online dictionary basis distribution are used to model the target appearance. A novel sparse representation-based voting map and a sparse constraint regularized mean shift are proposed to track the object robustly. Besides these contributions, we also introduce a new selection-based dictionary learning algorithm with a locally constrained sparse representation, called K-Selection. Based on a set of comprehensive experiments, our algorithm has demonstrated better performance than alternatives reported in the recent literature.

Over the next 10 years, more information and communication technology (ICT) will be deployed in the health system than in its entire previous history. Systems will be larger in scope, more complex, and move from regional to national and supranational scale. Yet we are at roughly the same place the aviation industry was in the 1950s with respect to system safety. Even if ICT harm rates do not increase, increased ICT use will increase the absolute number of ICT related harms. Factors that could diminish ICT harm include adoption of common standards, technology maturity, better system development, testing, implementation and end user training. Factors that will increase harm rates include complexity and heterogeneity of systems and their interfaces, rapid implementation and poor training of users. Mitigating these harms will not be easy, as organizational inertia is likely to generate a hysteresis-like lag, where the paths to increase and decrease harm are not identical.

The AMIA biomedical informatics (BMI) core competencies have been designed to support and guide graduate education in BMI, the core scientific discipline underlying the breadth of the field's research, practice, and education. The core definition of BMI adopted by AMIA specifies that BMI is 'the interdisciplinary field that studies and pursues the effective uses of biomedical data, information, and knowledge for scientific inquiry, problem solving and decision making, motivated by efforts to improve human health.' Application areas range from bioinformatics to clinical and public health informatics and span the spectrum from the molecular to population levels of health and biomedicine. The shared core informatics competencies of BMI draw on the practical experience of many specific informatics sub-disciplines. The AMIA BMI analysis highlights the central shared set of competencies that should guide curriculum design and that graduate students should be expected to master.

Online learned tracking is widely used for it's adaptive ability to handle appearance changes. However, it introduces potential drifting problems due to the accumulation of errors during the self-updating, especially for occluded scenarios. The recent literature demonstrates that appropriate combinations of trackers can help balance stability and flexibility requirements. We have developed a robust tracking algorithm using a local sparse appearance model (SPT). A static sparse dictionary and a dynamically online updated basis distribution model the target appearance. A novel sparse representation-based voting map and sparse constraint regularized mean-shift support the robust object tracking. Besides these contributions, we also introduce a new dictionary learning algorithm with a locally constrained sparse representation, called K-Selection. Based on a set of comprehensive experiments, our algorithm has demonstrated better performance than alternatives reported in the recent literature.

We present results from a comparative empirical study of two methods for constructing support vector machines (SVMs). The first method is the conventional one based on the quadratic programming approach, which builds the optimal separating hyperplane maximizing the margin between two classes (SVM-Q). The second method is based on the linear programming approach suggested by Vapnik to build a separating hyperplane with the minimum number of support vectors (SVM-L). Using synthetic data from two classes, we compare the classification performance of these SVMs, with a geometrical comparison of their separating hyperplanes and support vectors. We show that both classifiers achieve practically identical classification accuracy and generalization performance. However, SVM-L has many fewer support vectors than SVM-Q. We also prove that, in contrast to SVM-Q, which selects support vectors from the margin between two classes, support vectors of SVM-L lie on the furthermost borders of the classes, at the maximum distance from the opposite class.

Graph clustering partitions a graph into subgraphs with strongly interconnected nodes, while nodes belonging to different subgraphs are weakly connected. In this paper, we propose a new clustering method applicable to either weighted or unweighted graphs in which each cluster consists of a highly dense core region surrounded by a region with lower density. We have developed a highly efficient and robust method to identify nodes belonging to dense cores of clusters. The set of the nodes is then divided into groups, each of which is the representative of one cluster. These groups are finally expanded into complete clusters covering all the nodes of the graph. Experiments with both synthetic and real datasets for gene expression analysis and image segmentation yield very encouraging results.

Medical imaging has made great technological breakthroughs in multimodal acquisition, visualization, and analysis with many complementary image modalities to non-invasively capture human anatomy, physiology and pathology. New functional imaging techniques help elucidate the dynamics of human health and disease at much higher throughput, enabled by growth in affordable computing power, memory capacity, processor speed, and communication bandwidth. Recently, exceedingly powerful computer hardware and optimized image processing software may, for the first time, allow high volume image data processing and manipulation (like multimodal registration) to become clinically feasible on a routine basis in real-time. In reviewing recent medical imaging advances, we discuss some major open issues and directions.

We present a method for automatically extracting groups of orthologous genes from a large set of genomes by a new clustering algorithm on a weighted multipartite graph. The method assigns a score to an arbitrary subset of genes from multiple genomes to assess the orthologous relationships between genes in the subset. This score is computed using sequence similarities between the member genes and the phylogenetic relationship between the corresponding genomes. An ortholog cluster is found as the subset with the highest score, so ortholog clustering is formulated as a combinatorial optimization problem. The algorithm for finding an ortholog cluster runs in time O(absolute value(E) + absolute value(V) log absolute value(V)), where V and E are the sets of vertices and edges, respectively, in the graph. However, if we discretize the similarity scores into a constant number of bins, the runtime improves to O(absolute value(E) + absolute value(V)). The proposed method was applied to seven complete eukaryote genomes on which the manually curated database of eukaryotic ortholog clusters, KOG, is constructed. A comparison of our results with the manually curated ortholog clusters shows that our clusters are well correlated with the existing clusters.

The fourteen papers in this special issue are devoted to biomedical informatics, focusing on research and major applications for its deployment.

We have developed a method for distinguishing between correctly labeled and mislabeled data sampled from video sequences and used in the construction of a facial expression recognition classifier. The novelty of our approach lies in training a single, optimal classifier type (a support vector machine, or SVM) on multiple representations of the data, involving different "discriminating" subspaces. Results of a preliminary study on the discrimination of "high stress" vs. "low stress" facial expression data by this method confirms that our novel approach is able to distinguish subproblems where labeling is highly reliable from those where mislabeling can lead to high error rates. In helping detect data subsamples which yield misleading classification results, the method is also a rapid, highly efficient cross-validated approach for eliminating outliers.

Peer-reviewed publication of scientific research results represents the most important means of their communication. The authors have annually reviewed a large heterogeneous set of papers to produce the International Medical Informatics Association (IMIA) Yearbook of Medical Informatics. To support an objective and high-quality review process, the authors attempted to provide reviewers with a set of refined quality criteria, comprised of 80 general criteria and an additional 60 criteria for specific types of manuscripts. Authors conducted a randomized controlled trial, with 18 reviewers, to evaluate application of the refined criteria on review outcomes. Whereas the trial found that reviewers applying the criteria graded papers more strictly (lower overall scores), and that junior reviewers appreciated the availability of the criteria, there was no overall change in the interrater variability in reviewing the manuscripts. The authors describe their experience as a "case report" and provide a reference to the refined quality review criteria without claiming that the criteria represent a validated instrument for quantitative quality measurement.

In this report, the authors compare and contrast medical informatics (MI) and bioinformatics (BI) and provide a viewpoint on their complementarities and potential for collaboration in various subfields. The authors compare MI and BI along several dimensions, including: (1) historical development of the disciplines, (2) their scientific foundations, (3) data quality and analysis, (4) integration of knowledge and databases, (5) informatics tools to support practice, (6) informatics methods to support research (signal processing, imaging and vision, and computational modeling, (7) professional and patient continuing education, and (8) education and training. It is pointed out that, while the two disciplines differ in their histories, scientific foundations, and methodologic approaches to research in various areas, they nevertheless share methods and tools, which provides a basis for exchange of experience in their different applications. MI expertise in developing health care applications and the strength of BI in biological "discovery science" complement each other well. The new field of biomedical informatics (BMI) holds great promise for developing informatics methods that will be crucial in the development of genomic medicine. The future of BMI will be influenced strongly by whether significant advances in clinical practice and biomedical research come about from separate efforts in MI and BI, or from emerging, hybrid informatics subdisciplines at their interface.

Most human commonsense problem solving is done within a context, which constrains the solution space, whether it involves perception and image interpretation or not. Yet most research in image analysis still assumes context is defined a-priori by the investigator and is external to the computational image analysis system. Where explicit focus of attention and spatial contexts are used, as in active vision systems, these are problem-specific. We report on a new approach to image analysis, which includes a general framework for defining spatial context in terms of reference objects and their spatial relationships to other objects in a scene. Image analysis problems are decomposed into a sequence of sub-problems corresponding to determining a sequence of spatial contexts based on the set of dynamically chosen reference objects. The experimental results with medical image analysis and interpretation have demonstrated that using reference objects to define spatial contexts is a very effective strategy for computer image analysis systems that can purposely focus the image analysis effort on the most promising part of the entire image, such that the target object can be quickly and accurately localized by eliminating most potential false positives.

We introduce a new algorithmic method for identifying the geometrical core of proteins that does not require the usual superposition of structures.A geometrical core is defined as the set of residues such that the C,(I) -C,(J) atom distances are identicsl iu all structures of the protein family under study, where I and J are secondary structure positions in the structural units -strands, loops, or parts of them [7].The result of applying the algorithm to 53 Ig structures lead to the identification of two geometrical core sets of C, atom positions for the VL and VI domains.Applications of the core sets are described 1 IntroductionIt is well understood that in order to find any reasonable clsssi6cation of proteins into femilies one needs to find invariaut charscteristics that are shared by all members of the family-There are a number of such characteristics, depending on functional and structural properties.The main invariant of a protein fold is its geometrical core.There are several definitions of a core and of Pcore atoms".In this paper we use a core structure concept similar to those developed by Chothia and Lesk [3,4] and Gerstein and Altman ]2,i].A core is a set of residues whose Co atoms occupy the same relative positions in space.A variety of methods and core-finding algorithms have been developed.Most of them are based on finding of the optimal superposition of structures in order to determine the structurel variations of their residues ]5,13].Another way to define a geometrical core was recently suggested 181, where an invariant sytem of coordinates was introduced and a core defined as the subset of positions with approximately the same coordinates relative to this coordinate system.The main idea is to introduce a system of coordinates that is based on the geometrical properties common to all molecules of a protein family.This approach was

The authors present the case study of a 35-year informatics-based single subspecialty practice for the management of patients with chronic thyroid disease. This extensive experience provides a paradigm for the organization of longitudinal medical information by integrating individual patient care with clinical research and education. The kernel of the process is a set of worksheets easily completed by the physician during the patient encounter. It is a structured medical record that has been computerized since 1972, enabling analysis of different groups of patients to answer questions about chronic conditions and the effects of therapeutic interventions. The recording process and resulting studies serve as an important vehicle for medical education about the nuances of clinical practice. The authors suggest ways in which computerized medical records can become an integral part of medical practice, rather than a luxury or novelty.

Drawing on biophysics and engineering for primary methodology, medical imaging traditionally has had little overlap with medical informatics. This, however, is beginning to change. As new imaging modalities have proliferated, general-purpose image analysis software has been developed to incorporate not only low-level processing but also increasingly sophisticated functions that extract, transform, and combine the medical content from sets of related images. All require expert physician interaction to determine and interpret the content that is of interest for a specific patient. However, with the rapid and seemingly endless rise in computational power coinciding with unprecedented constraints on health costs, the potential and pressure for automating all but the most routine functions increase apace. This raises important issues for medical informatics, since any automation, however gradual, requires us to improve our understanding of the underlying image data and knowledge and its explicit definition, representation, and use on the computer. This special issue of JAMIA contains a section of three papers addressing some of the major problems and opportunities in medical informatics that arise in imaging. The paper by Brinkley and Rosse1 discusses the problems of developing knowledge-based image representations and methods in the context of their Digital Anatomist framework. They emphasize the need for symbolic anatomical reasoning so that whatever the image annotations made by the expert, a computer model can be developed to capture their medical content or semantics. They also highlight the importance of introducing knowledge-bases for supporting image segmentation and the technological value of network-based integration of multimodality imaging. Examples of research in functional brain imaging and mapping demonstrate the power of these approaches and techniques. Increases in computational power and storage capacity have also spurred dramatic improvements in the science and technology of graphics and databases. These in turn have enabled rapid visualization of large amounts of graphic information (generated by models that may well be parametrized from image data) and growing research on the unique characteristics and requirements of image databases. The paper by Tagare, Jaffe and Duncan2 addresses some of these issues from the perspective of the medical user of such databases, emphasizing the problems of defining the semantics of image knowledge bases and the implications of the imprecision inherent in medical datasets. They discuss the problems of indexing by feature content and supporting spatial queries that will distinguish changes in shape, size, and topological arrangement of anatomical structures, emphasizing the importance of developments in object-oriented schemas for defining anatomical reference structures. There is increasing urgency to standardize the exchange of medical image data because of the needs of practice, research, and education. These needs include shared access to and utilization of medical image databases by multiple interacting specialists working on different computer platforms and with different software systems, the development of massive reference digital image datasets like that of the Visible Human Project, and sharing data in large clinical studies. The paper by Bidgood3 provides a comprehensive introduction to the evolving DICOM (formerly ACRNEMA) standard of image data description and interchange. It discusses the many difficult issues involved in standardizing the definition of image information content for an electronic medical record in a way that preserves the flexibility of reporting structure and content. The interpretation data interchange (IDI) model of DICOM incorporates an object-oriented analysis of image interpretation reporting, semantic meta-analysis of the message components of prevailing health care data interchange standards, informal analysis of real-world scenarios, and trial encodings with iterative rounds of expert review. It has led to the specification of a Structured Interpretation (SI) Information Object Description (IOD), which provides the basis for an evolving description standard for medical image data interchange. Once it has reached a suitable level of stability, the next step would involve proposing equivalent extensions to the Health Level Seven (HL7) Standard, in order to integrate medical imaging information with other patient health care data in a platform- and software-transparent manner. As can be seen from the above, it is still early days for medical imaging informatics. The structuring of medical information and knowledge described in the above articles is only the beginning of a challenging scientific, technological, and empirical process that will have to draw on a multitude of disciplines. Besides the central computer vision, graphics and visualization, database, software engineering, artificial intelligence, distributed processing, networking, and interface research, basic work is needed in the underlying psychophysics, cognitive science, and mathematical/statistical modeling of the imaging modalities and their interpretation by synergistic human-machine methods. Considerable efforts will be needed to connect existing text-based semantic models with visual/graphical models of medical information and knowledge so that practitioners, researchers, educators, students, and patients can not only navigate but also understand the results of querying and interacting with our evolving multimedia databases. The medical informatics community has many challenges ahead if it is to seize the opportunity for serving as the integrative catalyst in this process.

The potential of lossy image compression is investigated in a well defined task domain, specifically, clinical assessment of chromatic, surgical and hematopathology specimens. Two criteria were employed, just noticeable difference and largest tolerable distortion. Compression tolerances differed among pathologists, but conformed to well defined upper and lower limits. The level of tolerable compression was significantly larger for the second criterion. It is concluded that lossy image compression is feasible for diagnostic pathology and may hold promise for telepathology applications.

Casimir A. Kulikowski, PhD; AIM: Quo Vadis?, Journal of the American Medical Informatics Association, Volume 3, Issue 6, 1 November 1996, Pages 432–433, https:/

This paper presents a new approach to the knowledge-based composition of processes for image interpretation and analysis. Its computer implementation in the VISIPLAN (Vision Planner) system provides a way of modeling the composition of image analysis processes within a unified, object-centered hierarchical planning framework. The approach has been tested and shown to be flexible in handling a reasonably broad class of multi-modality biomedical image analysis and interpretation problems. It provides a relatively general design or planning framework, within which problem specific image analysis and recognition processes can be generated more efficiently and effectively. In this way, generality is gained at the design and planning stages, even though the final implementation stage of interpretation processes is almost invariably problem- and domain-specific.>

This paper presents a new knowledge-based approach to the composition of biomedical image analysis processes. A prototype system has been developed that provides the imaging domain expert with the tools needed to formulate, specify and experiment with the design of analysis strategies. The system has been tested in the area of MR image analysis with promising results.>

Truncated optimal entropy-minimizing expansions can serve to characterize classes of multivariate data. A method is presented here by which the level of truncation and the corresponding dimensionalities of the class subspaces can be chosen to ensure adequate discrimination. The subspaces are chosen to maximize the average margin of correct classification of the paradigms of one class subject to constraints on the other margins.