A patient's electronic medical record contains a large amount of unstructured textual information. As patient records become increasingly dense owing to an aging population and increased occurrence of chronic diseases, a tool is needed to help organize and navigate patient data in a way that facilitates a clinician's ability to understand this information and that improves efficiency. A system has been developed for physicians that summarizes clinical information from a patient record. This system provides a gestalt view of the patient's record by organizing information about each disease along four dimensions (axes): time (eg, disease progression over time), space (eg, tumor in left frontal lobe), existence (eg, certainty of existence of a finding), and causality (eg, response to treatment). A display is generated from information provided by radiology reports and discharge summaries. Natural language processing is used to identify clinical abnormalities (problems, symptoms, findings) from these reports as well as associated properties and relationships. This information is presented in an integrated format that organizes extracted findings into a problem list, depicts the information on a timeline grid, and provides direct access to relevant reports and images. The goal of this system is to improve the structure of clinical information and its presentation to the physician, thereby simplifying the information retrieval and knowledge discovery necessary to bridge the gap between acquiring raw data and making an informed diagnosis.

Dynamic Bayesian Belief networks (DBNs) have been commonly used to represent temporal data in several domains; however, an ideal representation requires a near perfect mapping between the process being modeled and the DBN. Furthermore, DBNs assume a full set of observations collected at a fixed frequency. Bayesian model selection has arisen to address biased inference and underlying assumptions about the data (e.g., distribution, representativeness) to choose a model that best fits the given observations. Per patient case, a Bayesian model is generated to maximize specificity, and the collective set of models is averaged to fit all examples. This paper demonstrates the advantages of patient-specific modeling over a DBN-driven approach. Results evaluating this approach are presented based on models for two longitudinal clinical datasets (neuro-oncology, knee osteoarthritis). Largely, the patient-specific models show improved performance in prediction relative to the DBNs.

valuation is a cornerstone of informatics, allowing us to objectively assess the strengths and weaknesses of a given tool. These insights ultimately provide insight and feedback for the improvement of a system and its approach in the future. Thus, this final chapter aims to provide an overview of the fundamental techniques that are used in informatics evaluations. The basis upon which any quantitative evaluation starts is with statistics and formal study design. A review of inferential statistical concepts is provided from the perspective of biostatistics (confidence intervals; hypothesis testing; error assessment including sensitivity/ specificity and receiver operating characteristics). Under study design, differences between observational investigations and controlled experiments are covered. Issues pertaining to population selection and study errors are briefly introduced. With these general tools, we then look to more specific informatics evaluations, using information retrieval (IR) systems and usability studies as examples to motivate further discussion. Methods for designing both types of evaluations and endpoint metrics are described in detail. Biostatistics and Study Design: A PrimerCentral to any evaluation is an understanding of statistics and the systematic methods used to design experiments that are unbiased and that will correctly answer questions of efficacy and impact. The focus of statistical analysis is the interpretation of a collection of data describing some phenomena. Descriptive statistics (e.g., mean, median, mode) provide a summary of the collection, whereas inferential statistics aim to draw inferences about a population from a (random) sample. We start this chapter with a brief review of biostatistical concepts common to evaluation in biomedical informatics, leading into a discussion of study design and decision-making methods. Note that this section is not intended to be an instructional resource for statistics, but rather assumes some basic statistical knowledge on the part of the reader. For more detailed coverage of foundational concepts, the reader is referred to [15]. Statistical ConceptsInferential statistics is concerned with the estimation of parameters that describe a population. Common tasks include: point estimates from a distribution (e.g., calculating the mean from a random sample); interval estimates (e.g., confidence intervals); hypothesis testing; and prediction (or, in the context of biostatistics, medical decision making). Interval estimates and hypothesis testing are covered in the sections immediately below; and medical decision making is covered in a separate section. Confidence IntervalsWhen inferring values about a population, there is an inherent question of how "good" the estimate might be. Confidence intervals indicate the reliability of an estimate, providing an upper and lower bound around an estimated parameter. For instance, assume that a drug test shows that 40% of subjects experience improvement; a 95% confidence interval on this statistic would mean th...