An event model of medical information representation

Predicting individual physiologically acceptable states at discharge from a pediatric intensive care unit

Objective

Quantify physiologically acceptable PICU-discharge vital signs and develop machine learning models to predict these values for individual patients throughout their PICU episode.

Methods

EMR data from 7256 survivor PICU episodes (5632 patients) collected between 2009 and 2017 at Children's Hospital Los Angeles was analyzed. Each episode contained 375 variables representing physiology, labs, interventions, and drugs. Between medical and physical discharge, when clinicians determined the patient was ready for ICU discharge, they were assumed to be in a physiologically acceptable state space (PASS) for discharge. Each patient's heart rate, systolic blood pressure, diastolic blood pressure in the PASS window were measured and compared to age-normal values, regression-quantified PASS predictions, and recurrent neural network (RNN) PASS predictions made 12 hours after PICU admission.

Results

Mean absolute errors (MAEs) between individual PASS values and age-normal values (HR: 21.0 bpm; SBP: 10.8 mm Hg; DBP: 10.6 mm Hg) were greater (p < .05) than regression prediction MAEs (HR: 15.4 bpm; SBP: 9.9 mm Hg; DBP: 8.6 mm Hg). The RNN models best approximated individual PASS values (HR: 12.3 bpm; SBP: 7.6 mm Hg; DBP: 7.0 mm Hg).

Conclusions

The RNN model predictions better approximate patient-specific PASS values than regression and age-normal values.

A BPMN Based Notation for the Representation of Workflows in Hospital Protocols

Flowcharts used for hospital protocols have a series of ambiguities and limitations in order to express some types of information. In this article, a notation proposal for flowcharts that partially avoids these problems is presented. This new notation is an adaptation of BPMNE2, an extension of the Business Process Model and Notation (BPMN), which allows direct modelling of procedures that follow the Hazard Analysis and Critical Control Points (HACCP) model. The new notation has been validated in the hospital context, specifically in the field of hazardous drugs (HDs). To measure usability from the perspective of the health staff and auditors, the System Usability Scale (SUS) was used. A total of 47 experts took part in the assessment, resulting in a SUS score of 71, that corresponds to an acceptable level of usability. The feedback provided by these participants allows us to discover benefits and drawbacks of the proposal. Also, it is noteworthy that 76.6% of professionals prefer to migrate to the new notation from the ISO 5807:1985 notation, the most commonly used model. In addition to the direct benefits of this notation from the human point of view, its machine-understandable nature provides the required support for its integration into software tools for intelligent monitoring and auditing.

An information model for computable cancer phenotypes

BACKGROUND

Standards, methods, and tools supporting the integration of clinical data and genomic information are an area of significant need and rapid growth in biomedical informatics. Integration of cancer clinical data and cancer genomic information poses unique challenges, because of the high volume and complexity of clinical data, as well as the heterogeneity and instability of cancer genome data when compared with germline data. Current information models of clinical and genomic data are not sufficiently expressive to represent individual observations and to aggregate those observations into longitudinal summaries over the course of cancer care. These models are acutely needed to support the development of systems and tools for generating the so called clinical "deep phenotype" of individual cancer patients, a process which remains almost entirely manual in cancer research and precision medicine.

METHODS

Reviews of existing ontologies and interviews with cancer researchers were used to inform iterative development of a cancer phenotype information model. We translated a subset of the Fast Healthcare Interoperability Resources (FHIR) models into the OWL 2 Description Logic (DL) representation, and added extensions as needed for modeling cancer phenotypes with terms derived from the NCI Thesaurus. Models were validated with domain experts and evaluated against competency questions.

RESULTS

The DeepPhe Information model represents cancer phenotype data at increasing levels of abstraction from mention level in clinical documents to summaries of key events and findings. We describe the model using breast cancer as an example, depicting methods to represent phenotypic features of cancers, tumors, treatment regimens, and specific biologic behaviors that span the entire course of a patient's disease.

CONCLUSIONS

We present a multi-scale information model for representing individual document mentions, document level classifications, episodes along a disease course, and phenotype summarization, linking individual observations to high-level summaries in support of subsequent integration and analysis.

Lessons learned in detailed clinical modeling at Intermountain Healthcare

BACKGROUND AND OBJECTIVE

Intermountain Healthcare has a long history of using coded terminology and detailed clinical models (DCMs) to govern storage of clinical data to facilitate decision support and semantic interoperability. The latest iteration of DCMs at Intermountain is called the clinical element model (CEM). We describe the lessons learned from our CEM efforts with regard to subjective decisions a modeler frequently needs to make in creating a CEM. We present insights and guidelines, but also describe situations in which use cases conflict with the guidelines. We propose strategies that can help reconcile the conflicts. The hope is that these lessons will be helpful to others who are developing and maintaining DCMs in order to promote sharing and interoperability.

METHODS

We have used the Clinical Element Modeling Language (CEML) to author approximately 5000 CEMs.

RESULTS

Based on our experience, we have formulated guidelines to lead our modelers through the subjective decisions they need to make when authoring models. Reported here are guidelines regarding precoordination/postcoordination, dividing content between the model and the terminology, modeling logical attributes, and creating iso-semantic models. We place our lessons in context, exploring the potential benefits of an implementation layer, an iso-semantic modeling framework, and ontologic technologies.

CONCLUSIONS

We assert that detailed clinical models can advance interoperability and sharing, and that our guidelines, an implementation layer, and an iso-semantic framework will support our progress toward that goal.

The National Institutes of Health's Biomedical Translational Research Information System (BTRIS): design, contents, functionality and experience to date

The US National Institutes of Health (NIH) has developed the Biomedical Translational Research Information System (BTRIS) to support researchers' access to translational and clinical data. BTRIS includes a data repository, a set of programs for loading data from NIH electronic health records and research data management systems, an ontology for coding the disparate data with a single terminology, and a set of user interface tools that provide access to identified data from individual research studies and data across all studies from which individually identifiable data have been removed. This paper reports on unique design elements of the system, progress to date and user experience after five years of development and operation.

Desiderata for healthcare integrated data repositories based on architectural comparison of three public repositories

Integrated data repositories (IDRs) are indispensable tools for numerous biomedical research studies. We compare three large IDRs (Informatics for Integrating Biology and the Bedside (i2b2), HMO Research Network's Virtual Data Warehouse (VDW) and Observational Medical Outcomes Partnership (OMOP) repository) in order to identify common architectural features that enable efficient storage and organization of large amounts of clinical data. We define three high-level classes of underlying data storage models and we analyze each repository using this classification. We look at how a set of sample facts is represented in each repository and conclude with a list of desiderata for IDRs that deal with the information storage model, terminology model, data integration and value-sets management.

Development of clinical contents model markup language for electronic health records

OBJECTIVES

To develop dedicated markup language for clinical contents models (CCM) to facilitate the active use of CCM in electronic health record systems.

METHODS

Based on analysis of the structure and characteristics of CCM in the clinical domain, we designed extensible markup language (XML) based CCM markup language (CCML) schema manually.

RESULTS

CCML faithfully reflects CCM in both the syntactic and semantic aspects. As this language is based on XML, it can be expressed and processed in computer systems and can be used in a technology-neutral way.

CONCLUSIONS

CCML HAS THE FOLLOWING STRENGTHS: it is machine-readable and highly human-readable, it does not require a dedicated parser, and it can be applied for existing electronic health record systems.

Building a robust, scalable and standards-driven infrastructure for secondary use of EHR data: the SHARPn project

The Strategic Health IT Advanced Research Projects (SHARP) Program, established by the Office of the National Coordinator for Health Information Technology in 2010 supports research findings that remove barriers for increased adoption of health IT. The improvements envisioned by the SHARP Area 4 Consortium (SHARPn) will enable the use of the electronic health record (EHR) for secondary purposes, such as care process and outcomes improvement, biomedical research and epidemiologic monitoring of the nation's health. One of the primary informatics problem areas in this endeavor is the standardization of disparate health data from the nation's many health care organizations and providers. The SHARPn team is developing open source services and components to support the ubiquitous exchange, sharing and reuse or 'liquidity' of operational clinical data stored in electronic health records. One year into the design and development of the SHARPn framework, we demonstrated end to end data flow and a prototype SHARPn platform, using thousands of patient electronic records sourced from two large healthcare organizations: Mayo Clinic and Intermountain Healthcare. The platform was deployed to (1) receive source EHR data in several formats, (2) generate structured data from EHR narrative text, and (3) normalize the EHR data using common detailed clinical models and Consolidated Health Informatics standard terminologies, which were (4) accessed by a phenotyping service using normalized data specifications. The architecture of this prototype SHARPn platform is presented. The EHR data throughput demonstration showed success in normalizing native EHR data, both structured and narrative, from two independent organizations and EHR systems. Based on the demonstration, observed challenges for standardization of EHR data for interoperable secondary use are discussed.

Implementation of workflow engine technology to deliver basic clinical decision support functionality

Background

Workflow engine technology represents a new class of software with the ability to graphically model step-based knowledge. We present application of this novel technology to the domain of clinical decision support. Successful implementation of decision support within an electronic health record (EHR) remains an unsolved research challenge. Previous research efforts were mostly based on healthcare-specific representation standards and execution engines and did not reach wide adoption. We focus on two challenges in decision support systems: the ability to test decision logic on retrospective data prior prospective deployment and the challenge of user-friendly representation of clinical logic.

Results

We present our implementation of a workflow engine technology that addresses the two above-described challenges in delivering clinical decision support. Our system is based on a cross-industry standard of XML (extensible markup language) process definition language (XPDL). The core components of the system are a workflow editor for modeling clinical scenarios and a workflow engine for execution of those scenarios. We demonstrate, with an open-source and publicly available workflow suite, that clinical decision support logic can be executed on retrospective data. The same flowchart-based representation can also function in a prospective mode where the system can be integrated with an EHR system and respond to real-time clinical events. We limit the scope of our implementation to decision support content generation (which can be EHR system vendor independent). We do not focus on supporting complex decision support content delivery mechanisms due to lack of standardization of EHR systems in this area. We present results of our evaluation of the flowchart-based graphical notation as well as architectural evaluation of our implementation using an established evaluation framework for clinical decision support architecture.

Conclusions

We describe an implementation of a free workflow technology software suite (available at http://code.google.com/p/healthflow) and its application in the domain of clinical decision support. Our implementation seamlessly supports clinical logic testing on retrospective data and offers a user-friendly knowledge representation paradigm. With the presented software implementation, we demonstrate that workflow engine technology can provide a decision support platform which evaluates well against an established clinical decision support architecture evaluation framework. Due to cross-industry usage of workflow engine technology, we can expect significant future functionality enhancements that will further improve the technology's capacity to serve as a clinical decision support platform.

The evolving use of a clinical data repository: facilitating data access within an electronic medical record

We demonstrate the integration of a longitudinal, internally-developed legacy clinical information system with a vendor-based electronic health record application. The expense of developing the interface across systems was minimal, and it was successfully incorporated within the vendor EHR. Users consistently used the legacy data view while using the EHR; usage reached up to 50% of accessed data. Data showing concurrent increases in the use of the legacy data view and adoption of the new EHR suggests the method may improve adoption of the new system.

Early experiences in evolving an enterprise-wide information model for laboratory and clinical observations

As Electronic Healthcare Records become more prevalent, there is an increasing need to ensure unambiguous data capture, interpretation, and exchange within and across heterogeneous applications. To address this need, a common, uniform, and comprehensive approach for representing clinical information is essential. At Partners HealthCare System, we are investigating the development and implementation of enterprise-wide information models to specify the representation of clinical information to support semantic interoperability. This paper summarizes our early experiences in: (1) defining a process for information model development, (2) reviewing and comparing existing healthcare information models, (3) identifying requirements for representation of laboratory and clinical observations, and (4) exploring linkages to existing terminology and data standards. These initial findings provide insight to the various challenges ahead and guidance on next steps for adoption of information models at our organization.

Strategies for referent tracking in electronic health records

The goal of referent tracking is to create an ever-growing pool of data relating to the entities existing in concrete spatiotemporal reality. In the context of Electronic Healthcare Records (EHRs) the relevant concrete entities are not only particular patients but also their parts, diseases, therapies, lesions, and so forth, insofar as these are salient to diagnosis and treatment. Within a referent tracking system, all such entities are referred to directly and explicitly, something which cannot be achieved when familiar concept-based systems are used in what is called "clinical coding." In this paper, we describe the components of a referent tracking system in an informal way and we outline the procedures that would have to be followed by healthcare personnel in using such a system. We argue that the referent tracking paradigm can be introduced with only minor--though nevertheless ontologically important--technical changes to existing EHR infrastructures, but that it will require a radically different mindset on the part of those involved in clinical coding and terminology development from that which has prevailed hitherto.

Disseminating natural language processed clinical narratives

Through Natural Language Processing (NLP) techniques, information can be extracted from clinical narratives for a variety of applications (e.g., patient management). While the complex and nested output of NLP systems can be expressed in standard formats, such as the eXtensible Markup Language (XML), these representations may not be directly suitable for certain end-users or applications. The availability of a âeuro tabular' format that simplifies the content and structure of NLP output may facilitate the dissemination and use by users who are more familiar with common spreadsheet, database, or statistical tools. In this paper, we describe the knowledge-based design of a tabular representation for NLP output and development of a transformation program for the structured output of MedLEE, an NLP system at our institution. Through an evaluation, we found that the simplified tabular format is comparable to existing more complex NLP formats in effectiveness for identifying clinical conditions in narrative reports.

Automation of a problem list using natural language processing

Background

The medical problem list is an important part of the electronic medical record in development in our institution. To serve the functions it is designed for, the problem list has to be as accurate and timely as possible. However, the current problem list is usually incomplete and inaccurate, and is often totally unused. To alleviate this issue, we are building an environment where the problem list can be easily and effectively maintained.

Methods

For this project, 80 medical problems were selected for their frequency of use in our future clinical field of evaluation (cardiovascular). We have developed an Automated Problem List system composed of two main components: a background and a foreground application. The background application uses Natural Language Processing (NLP) to harvest potential problem list entries from the list of 80 targeted problems detected in the multiple free-text electronic documents available in our electronic medical record. These proposed medical problems drive the foreground application designed for management of the problem list. Within this application, the extracted problems are proposed to the physicians for addition to the official problem list.

Results

The set of 80 targeted medical problems selected for this project covered about 5% of all possible diagnoses coded in ICD-9-CM in our study population (cardiovascular adult inpatients), but about 64% of all instances of these coded diagnoses. The system contains algorithms to detect first document sections, then sentences within these sections, and finally potential problems within the sentences. The initial evaluation of the section and sentence detection algorithms demonstrated a sensitivity and positive predictive value of 100% when detecting sections, and a sensitivity of 89% and a positive predictive value of 94% when detecting sentences.

Conclusion

The global aim of our project is to automate the process of creating and maintaining a problem list for hospitalized patients and thereby help to guarantee the timeliness, accuracy and completeness of this information.

Medical problem and document model for natural language understanding

We are developing tools to help maintain a complete, accurate and timely problem list within a general purpose Electronic Medical Record system. As a part of this project, we have designed a system to automatically retrieve medical problems from free-text documents. Here we describe an information model based on XML (eXtensible Markup Language) and compliant with the CDA (Clinical Document Architecture). This model is used to ease the exchange of clinical data between the Natural Language Understanding application that retrieves potential problems from narrative document, and the problem list management application.

Derivation and evaluation of a document-naming nomenclature

OBJECTIVE

The Computerized Patient Record System is deployed at all 173 Veterans Affairs (VA) medical centers. Providers access clinical notes in the system from a note title menu. Following its implementation at the Nashville VA Medical Center, users expressed dissatisfaction with the time required find notes among hundreds of irregularly structured titles. The authors' objective was to develop a document-naming nomenclature (DNN) that creates informative, structured note titles that improve information access.

DESIGN

One thousand ninety-four unique note titles from two VA medical centers were reviewed. A note-naming nomenclature and compositional syntax were derived. Compositional order was determined by user preference survey.

MEASUREMENTS

The DNN was evaluated by modeling note titles from the Salt Lake City VA Medical Center (n=877), Vanderbilt University Medical Center (n=554), and the Mayo Clinic (n=42). A preliminary usability evaluation was conducted on a structured title display and sorting application.

RESULTS

Classes of note title components were found by inspection. Components describe characteristics of the author, the health care event, and the organizational unit providing care. Terms were taken from VA medical center information systems and national standards. The DNN model accurately described 97 to 99 percent of note titles from the test sites. The DNN term coverage varied, depending on component and site. Users found the DNN title format useful and the DNN-based title sorting and note review application easy to learn and quick to use.

CONCLUSION

The DNN accurately models note titles at five medical centers. Preliminary usability data indicate that DNN integration with title parsing and sorting software enhances information access.

Applications of telemedicine and telecommunications to disaster medicine: historical and future perspectives

Disaster management utilizes diverse technologies to accomplish a complex set of tasks. Despite a decade of experience, few published reports have reviewed application of telemedicine (clinical care at a distance enabled by telecommunication) in disaster situations. Appropriate new telemedicine applications can improve future disaster medicine outcomes, based on lessons learned from a decade of civilian and military disaster (wide-area) telemedicine deployments. This manuscript reviews the history of telemedicine activities in actual disasters and similar scenarios as well as ongoing telemedicine innovations that may be applicable to disaster situations. Emergency care providers must begin to plan effectively to utilize disaster-specific telemedicine applications to improve future outcomes.

Desiderata for controlled medical vocabularies in the twenty-first century

Builders of medical informatics applications need controlled medical vocabularies to support their applications and it is to their advantage to use available standards. In order to do so, however, these standards need to address the requirements of their intended users. Over the past decade, medical informatics researchers have begun to articulate some of these requirements. This paper brings together some of the common themes which have been described, including: vocabulary content, concept orientation, concept permanence, nonsemantic concept identifiers, polyhierarchy, formal definitions, rejection of "not elsewhere classified" terms, multiple granularities, multiple consistent views, context representation, graceful evolution, and recognized redundancy. Standards developers are beginning to recognize and address these desiderata and adapt their offerings to meet them.

Development of the Logical Observation Identifier Names and Codes (LOINC) vocabulary

The LOINC (Logical Observation Identifier Names and Codes) vocabulary is a set of more than 10,000 names and codes developed for use as observation identifiers in standardized messages exchanged between clinical computer systems. The goal of the study was to create universal names and codes for clinical observations that could be used by all clinical information systems. The LOINC names are structured to facilitate rapid matching, either automated or manual, between local vocabularies and the universal LOINC codes. If LOINC codes are used in clinical messages, each system participating in data exchange needs to match its local vocabulary to the standard vocabulary only once. This will reduce both the time and cost of implementing standardized interfaces. The history of the development of the LOINC vocabulary and the methodology used in its creation are described.

Department of Veterans Affairs, University of Utah consortium participation in the NLM/AHCPR Large Scale Vocabulary Test

The Large Scale Vocabulary Test (LSVT) was designed to evaluate how well the Metathesaurus plus planned additions to Meta covered the documentation needs of clinicians. Our consortium collected 10,538 clinical narratives from patient problem lists recorded at 65 Veterans Hospitals, internal medicine ambulatory care practices, diagnostic history and physical examination data elements from Iliad, and nursing shift notes and emergency transport patient records. The results showed 94% of submitted terms resulted in acceptable matches. 49% of submitted terms were judged to be synonymous with the match terms, 35% were judged to be more specific (usually due to modifiers), 2%, were less specific, and 6% had an associative relationship. In 8% of cases either no match was found by the LSVT interface or all proposed matches were rejected by the raters. The LSVT content was quite suitable for coding our narratives. Necessary improvements for an electronic record would include the ability to compose modifiers together with root concepts.

Progress notes model

The largest part of the medical record today consists of notes documenting the care delivered to patients and the clinical events relevant to diagnosis and treatment. These "progress notes" serve as the repository of medical facts and clinical thinking, and are intended as a concise vehicle of communication about a patient's condition to those who access the health record. They should be readable, easily understood, complete, accurate, and concise. They must also be flexible enough to logically convey to others what happened during an encounter, e.g., the chain of events during the visit, as well as guaranteeing full accountability for documented material, e.g., who recorded the information and when it was recorded. This paper describes a model for progress notes, which addresses the above needs, and outlines the rationale and principles which led to that model.

Managing information with SNOMED: understanding the model

SNOMED is an inventory of medical terms and concepts. Each concept has a unique representation, it's primary termcode. The primary termcode carries hierarchical and inheritable information about each concept it represents. Additional hierarchical and inheritable information is carried by cross references linked to each concept. The cross references provide the defining characteristics for each concept thereby creating a computer readable definition for each SNOMED entry. These are some of the attributes that define the SNOMED model.

Access to data: comparing AccessMed with Query by Review

OBJECTIVE

To evaluate the performance of tools for authoring patient database queries.

DESIGN

Query by Review, a tool that exploits the training that users have undergone to master a result review system, was compared with AccessMed, a vocabulary browser that supports lexical matching and the traversal of hierarchical and semantic links. Seven subjects (Medical Logic Module authors) were asked to use both tools to gather the vocabulary terms necessary to perform each of eight laboratory queries.

MEASUREMENTS

The proportion of queries that were correct; intersubject agreement.

RESULTS

Query by Review had better performance than AccessMed (38% correct queries versus 18%, p = 0.002), but both figures were low. Poor intersubject agreement (28% for Query by Review and 21% for AccessMed) corroborated the relatively low performance. Subjects appeared to have trouble distinguishing laboratory tests from laboratory batteries, picking terms relevant to the particular data type required, and using classes in the vocabulary's hierarchy.

CONCLUSION

Query by Review, with its more constrained user interface, performed somewhat better than AccessMed, a more general tool. Neither tool achieved adequate performance, however, which points to the difficulty of formulating a query for a clinical database and the need for further work.

Building national electronic medical record systems via the World Wide Web

Electronic medical record systems (EMRSs) currently do not lend themselves easily to cross-institutional clinical care and research. Unique system designs coupled with a lack of standards have led to this difficulty. The authors have designed a preliminary EMRS architecture (W3-EMRS) that exploits the multiplatform, multiprotocol, client-server technology of the World Wide Web. The architecture abstracts the clinical information model and the visual presentation away from the underlying EMRS. As a result, computation upon data elements of the EMRS and their presentation are no longer tied to the underlying EMRS structures. The architecture is intended to enable implementation of programs that provide uniform access to multiple, heterogeneous legacy EMRSs. The authors have implemented an initial prototype of W3-EMRS that accesses the database of the Boston Children's Hospital Clinician's Workstation.

Enroute toward a computer based patient record: the ACIS project

The clinical arm of the University of Utah operates a 400 bed general hospital and 30 specialty and subspecialty clinics and with the advent of managed care will soon be linked to a much wider network of health care facilities throughout the state. Each of these patient care facilities maintains its own patient record. To provide common access to data on any patient needed to provide continuity to the care process where ever it occurs, the institution has embarked on the development of a completely electronic patient record. In this paper we will describe the route we have taken to achieve this goal in hopes of providing some sign posts along the way for others seeking the same destination.

Representing hospital events as complex conditionals

We have developed an approach to medical knowledge representation whereby simple medical concepts are combined to yield complex statements of testable medical logic. The logic is created from a small number of generic medical concepts that are instantiated and combined to create the rules. Rule writing is done through a rule editor and requires knowledge of the system's data dictionaries, though no programming is required. We have used the approach to create a large knowledge base including panic lab alerting rules, drug-laboratory interaction alerting rules, an adverse drug event monitor, and a drug-age interaction detection program. The rules have been used as part of an alerting system and for data collection to determine the frequency of events of interest. The scheme is extensible and yields a readable form of the created knowledge. The scheme holds great promise as a durable form of medical knowledge representation.