Extracting Sexual Trauma Mentions from Electronic Medical Notes Using Natural Language Processing

Patient history of sexual trauma is of clinical relevance to healthcare providers as survivors face adverse health-related outcomes. This paper describes a method for identifying mentions of sexual trauma within the free text of electronic medical notes. A natural language processing pipeline for information extraction was developed and scaled to handle a large corpus of electronic medical notes used for this study from US Veterans Health Administration medical facilities. The tool was used to identify sexual trauma mentions and create snippets around every asserted mention based on a domain-specific lexicon developed for this purpose. All snippets were evaluated by trained human reviewers. An overall positive predictive value (PPV) of 0.90 for identifying sexual trauma mentions from the free text and a PPV of 0.71 at the patient level are reported. The metrics are superior for records from female patients.

Distance to Veterans Administration Medical Centers as a Barrier to Specialty Care for Homeless Women Veterans

Homeless women Veterans have a high prevalence of chronic mental and physical conditions that necessitate frequent healthcare visits, but travel burdens to specialty services may be overwhelming to navigate for this population, especially for those in rural settings. Access to specialty care is a key priority in the Veterans Health Administration (VHA) and understanding the geographic distribution and rural designation of this population in relation to medical centers (VAMC) can assist in care coordination. We identified 41,747 women Veterans age 18-44y with administrative evidence of homelessness in the VHA anytime during 2002-2015. We found 7% live in rural settings and 29% live >40miles from a VAMC. The mean travel distance for homeless women Veterans with a rural designation to a VAMC specialty center was 107 miles. Developing interventions to overcome this travel burden and engage vulnerable Veterans in necessary care can improve overall health outcomes for this high-risk population.

Fibromyalgia syndrome care of Iraq- and Afghanistan-deployed Veterans in Veterans Health Administration

Little is known regarding fibromyalgia syndrome (FMS) care among Operation Iraqi Freedom/Operation Enduring Freedom/Operation New Dawn (OIF/OEF/OND) Veterans. Current recommendations include interdisciplinary, team-based combined care approaches and limited opioid use. In this study of OIF/OEF/OND Veterans who accessed Veterans Health Administration services between 2002 and 2012, we hypothesized that combined care (defined as at least 4 primary care visits/yr with visits to mental health and/or rheumatology) versus <4 primary care visits/yr only would be associated with lower risk of at least 2 opioid prescriptions 12 mo following an FMS diagnosis. Using generalized linear models with a log-link, the Poisson family, and robust standard errors, we estimated risk ratios (RRs) and 95% confidence intervals (CIs). We found that 1% of Veterans had at least 2 FMS diagnoses (International Classification of Diseases-9th Revision-Clinical Modification code 729.1) or at least 1 FMS diagnosis by rheumatology. Veterans with (vs without) FMS were more likely to be female, older, Hispanic, and never/currently married. Combined primary, mental health, and rheumatology care was associated with at least 2 opioid prescriptions (RR [95% CI] for males 2.2 [1.1-4.4] and females 2.8 [0.4-18.6]). Also, combined care was associated with at least 2 nonopioid pain-related prescriptions, a practice supported by evidence-based clinical practice guidelines. In tandem, these results provide mixed evidence of benefit of combined care for FMS. Future studies of healthcare encounter characteristics, care coordination, and benefits for Veterans with FMS are needed.

Nursing Informatics Certification Worldwide: History, Pathway, Roles, and Motivation

INTRODUCTION

Official recognition and certification for informatics professionals are essential aspects of workforce development.

OBJECTIVE

To describe the history, pathways, and nuances of certification in nursing informatics across the globe; compare and contrast those with board certification in clinical informatics for physicians.

METHODS

(1) A review of the representative literature on informatics certification and related competencies for nurses and physicians, and relevant websites for nursing informatics associations and societies worldwide; (2) similarities and differences between certification processes for nurses and physicians, and (3) perspectives on roles for nursing informatics professionals in healthcare Results: The literature search for 'nursing informatics certification' yielded few results in PubMed; Google Scholar yielded a large number of citations that extended to magazines and other non-peer reviewed sources. Worldwide, there are several nursing informatics associations, societies, and workgroups dedicated to nursing informatics associated with medical/health informatics societies. A formal certification program for nursing informatics appears to be available only in the United States. This certification was established in 1992, in concert with the formation and definition of nursing informatics as a specialty practice of nursing by the American Nurses Association. Although informatics is inherently interprofessional, certification pathways for nurses and physicians have developed separately, following long-standing professional structures, training, and pathways aligned with clinical licensure and direct patient care. There is substantial similarity with regard to the skills and competencies required for nurses and physicians to obtain informatics certification in their respective fields. Nurses may apply for and complete a certification examination if they have experience in the field, regardless of formal training. Increasing numbers of informatics nurses are pursuing certification.

CONCLUSIONS

The pathway to certification is clear and wellestablished for U.S. based informatics nurses. The motivation for obtaining and maintaining nursing informatics certification appears to be stronger for nurses who do not have an advanced informatics degree. The primary difference between nursing and physician certification pathways relates to the requirement of formal training and level of informatics practice. Nurse informatics certification requires no formal education or training and verifies knowledge and skill at a more basic level. Physician informatics certification validates informatics knowledge and skill at a more advanced level; currently this requires documentation of practice and experience in clinical informatics and in the future will require successful completion of an accredited two-year fellowship in clinical informatics. For the profession of nursing, a graduate degree in nursing or biomedical informatics validates specialty knowledge at a level more comparable to the physician certification. As the field of informatics and its professional organization structures mature, a common certification pathway may be appropriate. Nurses, physicians, and other healthcare professionals with informatics training and certification are needed to contribute their expertise in clinical operations, teaching, research, and executive leadership.

The Veteran-Initiated Electronic Care Coordination: A Multisite Initiative to Promote and Evaluate Consumer-Mediated Health Information Exchange

INTRODUCTION

Information continuity is critical to person-centered care when patients receive care from multiple healthcare systems. Patients can access their electronic health record data through patient portals to facilitate information exchange. This pilot was developed to improve care continuity for rural Veterans by (1) promoting the use of the Department of Veterans Affairs (VA) patient portal to share health information with non-VA providers, and (2) evaluating the impact of health information sharing at a community appointment.

MATERIALS AND METHODS

Veterans from nine VA healthcare systems were trained to access and share their VA Continuity of Care Document (CCD) with their non-VA providers. Patients and non-VA providers completed surveys on their experiences.

RESULTS

Participants (n = 620) were primarily older, white, and Vietnam era Veterans. After training, 78% reported the CCD would help them be more involved in their healthcare and 86% planned to share it regularly with non-VA providers. Veterans (n = 256) then attended 277 community appointments. Provider responses from these appointments (n = 133) indicated they were confident in the accuracy of the information (97%) and wanted to continue to receive the CCD (96%). Ninety percent of providers reported the CCD improved their ability to have an accurate medication list and helped them make medication treatment decisions. Fifty percent reported they did not order a laboratory test or another procedure because of information available in the CCD.

CONCLUSIONS

This pilot demonstrates feasibility and value of patient access to a CCD to facilitate information sharing between VA and non-VA providers. Outreach and targeted education are needed to promote consumer-mediated health information exchange.

Military Sexual Trauma and Co-occurring Posttraumatic Stress Disorder, Depressive Disorders, and Substance Use Disorders among Returning Afghanistan and Iraq Veterans

PURPOSE

Posttraumatic stress disorder (PTSD), depressive disorders (DD), and substance use disorders (SUD) are prevalent among veterans. A positive military sexual trauma (MST) screen is associated with higher likelihood of each of these disorders. The current study examined the associations between MST, gender, and co-occurring PTSD, DD, and SUD among veterans receiving services at the Department of Veterans Affairs to inform assessment and treatment. We were specifically interested in the interactions between MST and gender on co-occurring disorders.

METHODS

The sample included 494,822 Department of Veterans Affairs service-seeking veterans (12.5% women) deployed to Iraq and Afghanistan who recently separated from the military and were screened for MST between 2004 and 2013.

MAIN FINDINGS

Veterans with positive MST screens had higher odds than those with negative screens of individual and co-occurring PTSD, DD, and SUD. The association between positive MST screens and diagnostic outcomes, including PTSD, was stronger for women than for men, and the association between positive MST screens and some diagnostic outcomes, including DD, was stronger for men than for women.

CONCLUSIONS

These results highlight the importance of assessing for and recognizing the potential MST and gender interactions in the clinical context among veterans with co-occurring PTSD, DD, and/or SUD.

v3NLP Framework: Tools to Build Applications for Extracting Concepts from Clinical Text

Introduction:

Substantial amounts of clinically significant information are contained only within the narrative of the clinical notes in electronic medical records. The v3NLP Framework is a set of “best-of-breed” functionalities developed to transform this information into structured data for use in quality improvement, research, population health surveillance, and decision support.

Background:

MetaMap, cTAKES and similar well-known natural language processing (NLP) tools do not have sufficient scalability out of the box. The v3NLP Framework evolved out of the necessity to scale-up these tools up and provide a framework to customize and tune techniques that fit a variety of tasks, including document classification, tuned concept extraction for specific conditions, patient classification, and information retrieval.

Innovation:

Beyond scalability, several v3NLP Framework-developed projects have been efficacy tested and benchmarked. While v3NLP Framework includes annotators, pipelines and applications, its functionalities enable developers to create novel annotators and to place annotators into pipelines and scaled applications.

Discussion:

The v3NLP Framework has been successfully utilized in many projects including general concept extraction, risk factors for homelessness among veterans, and identification of mentions of the presence of an indwelling urinary catheter. Projects as diverse as predicting colonization with methicillin-resistant Staphylococcus aureus and extracting references to military sexual trauma are being built using v3NLP Framework components.

Conclusion:

The v3NLP Framework is a set of functionalities and components that provide Java developers with the ability to create novel annotators and to place those annotators into pipelines and applications to extract concepts from clinical text. There are scale-up and scale-out functionalities to process large numbers of records.

Differential Risk for Homelessness Among US Male and Female Veterans With a Positive Screen for Military Sexual Trauma

IMPORTANCE

Military sexual trauma (MST) is associated with adverse physical and mental health outcomes following military separation. Recent research suggests that MST may be a determinant in several factors associated with postdeployment homelessness.

OBJECTIVE

To evaluate MST as an independent risk factor for homelessness and to determine whether risk varies by sex.

DESIGN, SETTING, AND PARTICIPANTS

A retrospective cohort study of US veterans who used Veterans Health Administration (VHA) services between fiscal years 2004 and 2013 was conducted using administrative data from the Department of Defense and VHA. Included in the study were 601 892 US veterans deployed in Iraq or Afghanistan who separated from the military between fiscal years 2001 and 2011 and subsequently used VHA services.

EXPOSURE

Positive response to screen for MST administered in VHA facilities.

MAIN OUTCOMES AND MEASURES

Administrative evidence of homelessness within 30 days, 1 year, and 5 years following the first VHA encounter after last deployment.

RESULTS

The mean (SD) age of the 601 892 participants was 38.9 (9.4) years, 527 874 (87.7%) were male, 310 854 (51.6%) were white, and 382 361 (63.5%) were enlisted in the Army. Among veterans with a positive screen for MST, rates of homelessness were 1.6% within 30 days, 4.4% within 1 year, and 9.6% within 5 years, more than double the rates of veterans with a negative MST screen (0.7%, 1.8%, and 4.3%, respectively). A positive screen for MST was significantly and independently associated with postdeployment homelessness. In regression models adjusted for demographic and military service characteristics, odds of experiencing homelessness were higher among those who screened positive for MST compared with those who screened negative (30-day: adjusted odds ratio [AOR], 1.89; 95% CI, 1.58-2.24; 1-year: AOR, 2.27; 95% CI, 2.04-2.53; and 5-year: AOR, 2.63; 95% CI, 2.36-2.93). Military sexual trauma screen status remained independently associated with homelessness after adjusting for co-occurring mental health and substance abuse diagnoses in follow-up regression models (30-day: AOR, 1.62; 95% CI, 1.36-1.93; 1-year: AOR, 1.49; 95% CI, 1.33-1.66; and 5-year: AOR, 1.39; 95% CI, 1.24-1.55). In the fully adjusted models, the interaction between MST status and sex was significant in the 30-day and 1-year cohorts (30-day: AOR, 1.54; 95% CI, 1.18-2.02; and 1-year: AOR, 1.46; 95% CI, 1.23-1.74), denoting higher risk for homelessness among males with a positive screen for MST.

CONCLUSIONS AND RELEVANCE

A positive screen for MST was independently associated with postdeployment homelessness, with male veterans at greater risk than female veterans. These results underscore the importance of the MST screen as a clinically important marker of reintegration outcomes among veterans. These findings demonstrate significant long-term negative effects and inform our understanding of the public health implications of sexual abuse and harassment.

Estimates of the risk of large or long-lasting outbreaks of Middle East respiratory syndrome after importations outside the Arabian Peninsula

•

MERS outbreak clusters outside the Arabian Peninsula ranged in size from 1 to 186.

•

Cluster data show declining transmission rate in later transmission generations.

•

Model projects tempered risk of large, long-lasting outbreaks after importations.

•

Explosive outbreaks are possible, but control measures are likely to be effective.

We quantify outbreak risk after importations of Middle East respiratory syndrome outside the Arabian Peninsula. Data from 31 importation events show strong statistical support for lower transmissibility after early transmission generations. Our model projects the risk of ≥10, 100, and 500 transmissions as 11%, 2%, and 0.02%, and ≥1, 2, 3, and 4 generations as 23%, 14%, 0.9%, and 0.05%, respectively. Our results suggest tempered risk of large, long-lasting outbreaks with appropriate control measures.

Using network projections to explore co-incidence and context in large clinical datasets: Application to homelessness among U.S. Veterans

INTRODUCTION

Network projections of data can provide an efficient format for data exploration of co-incidence in large clinical datasets. We present and explore the utility of a network projection approach to finding patterns in health care data that could be exploited to prevent homelessness among U.S. Veterans.

METHOD

We divided Veteran ICD-9-CM (ICD9) data into two time periods (0-59 and 60-364days prior to the first evidence of homelessness) and then used Pajek social network analysis software to visualize these data as three different networks. A multi-relational network simultaneously displayed the magnitude of ties between the most frequent ICD9 pairings. A new association network visualized ICD9 pairings that greatly increased or decreased. A signed, subtraction network visualized the presence, absence, and magnitude difference between ICD9 associations by time period.

RESULT

A cohort of 9468 U.S. Veterans was identified as having administrative evidence of homelessness and visits in both time periods. They were seen in 222,599 outpatient visits that generated 484,339 ICD9 codes (average of 11.4 (range 1-23) visits and 2.2 (range 1-60) ICD9 codes per visit). Using the three network projection methods, we were able to show distinct differences in the pattern of co-morbidities in the two time periods. In the more distant time period preceding homelessness, the network was dominated by routine health maintenance visits and physical ailment diagnoses. In the 59days immediately prior to the homelessness identification, alcohol related diagnoses along with economic circumstances such as unemployment, legal circumstances, along with housing instability were noted.

CONCLUSION

Network visualizations of large clinical datasets traditionally treated as tabular and difficult to manipulate reveal rich, previously hidden connections between data variables related to homelessness. A key feature is the ability to visualize changes in variables with temporality and in proximity to the event of interest. These visualizations lend support to cognitive tasks such as exploration of large clinical datasets as a prelude to hypothesis generation.

The role of heterogeneity in contact timing and duration in network models of influenza spread in schools

Influenza poses a significant health threat to children, and schools may play a critical role in community outbreaks. Mathematical outbreak models require assumptions about contact rates and patterns among students, but the level of temporal granularity required to produce reliable results is unclear. We collected objective contact data from students aged 5–14 at an elementary school and middle school in the state of Utah, USA, and paired those data with a novel, data-based model of influenza transmission in schools. Our simulations produced within-school transmission averages consistent with published estimates. We compared simulated outbreaks over the full resolution dynamic network with simulations on networks with averaged representations of contact timing and duration. For both schools, averaging the timing of contacts over one or two school days caused average outbreak sizes to increase by 1–8%. Averaging both contact timing and pairwise contact durations caused average outbreak sizes to increase by 10% at the middle school and 72% at the elementary school. Averaging contact durations separately across within-class and between-class contacts reduced the increase for the elementary school to 5%. Thus, the effect of ignoring details about contact timing and duration in school contact networks on outbreak size modelling can vary across different schools.

Estimates of Outbreak Risk from New Introductions of Ebola with Immediate and Delayed Transmission Control

Identifying incoming patients can have a larger risk-reduction effect than efforts to reduce transmissions from identified patients.

While the ongoing Ebola outbreak continues in the West Africa countries of Guinea, Sierra Leone, and Liberia, health officials elsewhere prepare for new introductions of Ebola from infected evacuees or travelers. We analyzed transmission data from patients (i.e., evacuees, international travelers, and those with locally acquired illness) in countries other than the 3 with continuing Ebola epidemics and quantitatively assessed the outbreak risk from new introductions by using different assumptions for transmission control (i.e., immediate and delayed). Results showed that, even in countries that can quickly limit expected number of transmissions per case to <1, the probability that a single introduction will lead to a substantial number of transmissions is not negligible, particularly if transmission variability is high. Identifying incoming infected travelers before symptom onset can decrease worst-case outbreak sizes more than reducing transmissions from patients with locally acquired cases, but performing both actions can have a synergistic effect.

Finding 'Evidence of Absence' in Medical Notes: Using NLP for Clinical Inferencing

Extracting evidence of the absence of a target of interest from medical text can be useful in clinical inferencing. The purpose of our study was to develop a natural language processing (NLP) pipelineto identify the presence of indwelling urinary catheters from electronic medical notes to aid in detection of catheter-associated urinary tract infections (CAUTI). Finding clear evidence that a patient does not have an indwelling urinary catheter is useful in making a determination regarding CAUTI. We developed a lexicon of seven core concepts to infer the absence of a urinary catheter. Of the 990,391 concepts extractedby NLP from a large corpus of 744,285 electronic medical notes from 5589 hospitalized patients, 63,516 were labeled as evidence of absence.Human review revealed three primary causes for false negatives. The lexicon and NLP pipeline were refined using this information, resulting in outputs with an acceptable false positive rate of 11%.

PlateRunner: A Search Engine to Identify EMR Boilerplates

Medical text contains boilerplated content, an artifact of pull-down forms from EMRs. Boilerplated content is the source of challenges for concept extraction on clinical text. This paper introduces PlateRunner, a search engine on boilerplates from the US Department of Veterans Affairs (VA) EMR. Boilerplates containing concepts should be identified and reviewed to recognize challenging formats, identify high yield document titles, and fine tune section zoning. This search engine has the capability to filter negated and asserted concepts, save and search query results. This tool can save queries, search results, and documents found for later analysis.

Improving Bridging from Informatics Practice to Theory

BACKGROUND

In 1962, Methods of Information in Medicine ( MIM ) began to publish papers on the methodology and scientific fundamentals of organizing, representing, and analyzing data, information, and knowledge in biomedicine and health care. Considered a companion journal, Applied Clinical Informatics ( ACI ) was launched in 2009 with a mission to establish a platform that allows sharing of knowledge between clinical medicine and health IT specialists as well as to bridge gaps between visionary design and successful and pragmatic deployment of clinical information systems. Both journals are official journals of the International Medical Informatics Association.

OBJECTIVES

As a follow-up to prior work, we set out to explore congruencies and interdependencies in publications of ACI and MIM. The objectives were to describe the major topics discussed in articles published in ACI in 2014 and to determine if there was evidence that theory in 2014 MIM publications was informed by practice described in ACI publications in any year. We also set out to describe lessons learned in the context of bridging informatics practice and theory and offer opinions on how ACI editorial policies could evolve to foster and improve such bridging.

METHODS

We conducted a retrospective observational study and reviewed all articles published in ACI during the calendar year 2014 (Volume 5) for their main theme, conclusions, and key words. We then reviewed the citations of all MIM papers from 2014 to determine if there were references to ACI articles from any year. Lessons learned in the context of bridging informatics practice and theory and opinions on ACI editorial policies were developed by consensus among the two authors.

RESULTS

A total of 70 articles were published in ACI in 2014. Clinical decision support, clinical documentation, usability, Meaningful Use, health information exchange, patient portals, and clinical research informatics emerged as major themes. Only one MIM article from 2014 cited an ACI article. There are several lessons learned including the possibility that there may not be direct links between MIM theory and ACI practice articles. ACI editorial policies will continue to evolve to reflect the breadth and depth of the practice of clinical informatics and articles received for publication. Efforts to encourage bridging of informatics practice and theory may be considered by the ACI editors.

CONCLUSIONS

The lack of direct links from informatics theory-based papers published in MIM in 2014 to papers published in ACI continues as was described for papers published during 2012 to 2013 in the two companion journals. Thus, there is little evidence that theory in MIM has been informed by practice in ACI.

Scaling Out and Evaluation of OBSecAn, an Automated Section Annotator for Semi-Structured Clinical Documents, on a Large VA Clinical Corpus

"Identifying and labeling" (annotating) sections improves the effectiveness of extracting information stored in the free text of clinical documents. OBSecAn, an automated ontology-based section annotator, was developed to identify and label sections of semi-structured clinical documents from the Department of Veterans Affairs (VA). In the first step, the algorithm reads and parses the document to obtain and store information regarding sections into a structure that supports the hierarchy of sections. The second stage detects and makes correction to errors in the parsed structure. The third stage produces the section annotation output using the final parsed tree. In this study, we present the OBSecAn method and its scale to a million document corpus and evaluate its performance in identifying family history sections. We identify high yield sections for this use case from note titles such as primary care and demonstrate a median rate of 99% in correctly identifying a family history section.

Exploiting the UMLS Metathesaurus for extracting and categorizing concepts representing signs and symptoms to anatomically related organ systems

OBJECTIVE

To develop a method to exploit the UMLS Metathesaurus for extracting and categorizing concepts found in clinical text representing signs and symptoms to anatomically related organ systems. The overarching goal is to classify patient reported symptoms to organ systems for population health and epidemiological analyses.

MATERIALS AND METHODS

Using the concepts' semantic types and the inter-concept relationships as guidance, a selective portion of the concepts within the UMLS Metathesaurus was traversed starting from the concepts representing the highest level organ systems. The traversed concepts were chosen, filtered, and reviewed to obtain the concepts representing clinical signs and symptoms by blocking deviations, pruning superfluous concepts, and manual review. The mapping process was applied to signs and symptoms annotated in a corpus of 750 clinical notes.

RESULTS

The mapping process yielded a total of 91,000 UMLS concepts (with approximately 300,000 descriptions) possibly representing physical and mental signs and symptoms that were extracted and categorized to the anatomically related organ systems. Of 1864 distinct descriptions of signs and symptoms found in the 750 document corpus, 1635 of these (88%) were successfully mapped to the set of concepts extracted from the UMLS. Of 668 unique concepts mapped, 603 (90%) were correctly categorized to their organ systems.

CONCLUSION

We present a process that facilitates mapping of signs and symptoms to their organ systems. By providing a smaller set of UMLS concepts to use for comparing and matching patient records, this method has the potential to increase efficiency of information extraction pipelines.

Identifying Homelessness among Veterans Using VA Administrative Data: Opportunities to Expand Detection Criteria

Researchers at the U.S. Department of Veterans Affairs (VA) have used administrative criteria to identify homelessness among U.S. Veterans. Our objective was to explore the use of these codes in VA health care facilities. We examined VA health records (2002-2012) of Veterans recently separated from the military and identified as homeless using VA conventional identification criteria (ICD-9-CM code V60.0, VA specific codes for homeless services), plus closely allied V60 codes indicating housing instability. Logistic regression analyses examined differences between Veterans who received these codes. Health care services and co-morbidities were analyzed in the 90 days post-identification of homelessness. VA conventional criteria identified 21,021 homeless Veterans from Operations Enduring Freedom, Iraqi Freedom, and New Dawn (rate 2.5%). Adding allied V60 codes increased that to 31,260 (rate 3.3%). While certain demographic differences were noted, Veterans identified as homeless using conventional or allied codes were similar with regards to utilization of homeless, mental health, and substance abuse services, as well as co-morbidities. Differences were noted in the pattern of usage of homelessness-related diagnostic codes in VA facilities nation-wide. Creating an official VA case definition for homelessness, which would include additional ICD-9-CM and other administrative codes for VA homeless services, would likely allow improved identification of homeless and at-risk Veterans. This also presents an opportunity for encouraging uniformity in applying these codes in VA facilities nationwide as well as in other large health care organizations.

Chronic multisymptom illness among female Veterans deployed to Iraq and Afghanistan

BACKGROUND

Chronic multisymptom illness (CMI) may be more prevalent among female Operation Enduring Freedom/Operation Iraqi Freedom/Operation New Dawn (OEF/OIF/OND) deployed Veterans due to deployment-related experiences.

OBJECTIVES

To investigate CMI-related diagnoses among female OEF/OIF/OND Veterans.

RESEARCH DESIGN

We estimated the prevalence of the International Classification of Disease-9th edition-Clinical Modification coded CMI-related diagnoses of chronic fatigue syndrome, fibromyalgia (FM), and irritable bowel syndrome (IBS) among female OEF/OIF/OND Veterans with Veterans Health Administration (VHA) visits, FY2002-2012 (n=78,435). We described the characteristics of female Veterans with and without CMI-related diagnoses and VHA settings of first CMI-related diagnoses.

RESULTS

The prevalence of CMI-related diagnoses among female OEF/OIF/OND Veterans was 6397 (8.2%), over twice as high as the prevalence 95,424 (3.9%) among the totality of female Veterans currently accessing VHA (P<0.01). There were statistically significant differences in age, education, marital status, military component, service branch, and proportions of those with depression and/or post-traumatic stress disorder diagnoses across females with and without CMI-related diagnoses. Diagnoses were mainly from primary care, women's health, and physical medicine and rehabilitation clinics.

CONCLUSIONS

CMI-related diagnoses were more prevalent among female OEF/OIF/OND Veterans compared with all female Veterans who currently access VHA. Future studies of the role of mental health diagnoses as confounders or mediators of the association of OEF/OIF/OND deployment and CMI are warranted. These and other factors associated with CMI may provide a basis for enhanced screening to facilitate recognition of these conditions. Further work should evaluate models of care and healthcare utilization related to CMI in female Veterans.

Fertility, pregnancies and outcomes reported by females with common variable immune deficiency and hypogammaglobulinemia: results from an internet-based survey

BACKGROUND

Issues of fertility and pregnancy place an extra burden on females with primary immunodeficiencies. Patients lack reliable information and providers lack guidelines to counsel patients on these anxiety-provoking matters.

OBJECTIVE

To collate concerns and experiences related to fertility and pregnancy from females with humoral immune deficiencies.

METHODS

We conducted an internet-based survey of female patients who self-identified as having a diagnosis of primary humoral immune deficiency.

RESULTS

Responses from 490 women with common variable immune deficiency and 100 with hypogammaglobulinemia were evaluated. The reported fertility measure (% of women who had had a birth) was statistically significantly lower as compared to the general US population (70 % vs. 85 %, p < 0.0001) whereas the rates of spontaneous pregnancy loss were comparable. This group reported a total of 966 pregnancies; 72 % resulted in a live birth. A majority of the pregnancies progressed with no incident and with continuation of their IgG replacement therapy; 23 % reported an increase in IgG dosing during pregnancy. Only 15 % of those reporting a first pregnancy indicated that they had been diagnosed with immune deficiency prior to their first pregnancy; these women expressed concern regarding the effect of immune deficiency on their fertility, pregnancy and decision to have children.

CONCLUSION

With inherent limitations of self-reported responses to surveys, females with humoral immune deficiencies reported relatively good rates of fertility and pregnancies ending in live births. Results of the survey will serve as peer support for patients and inform counseling guidelines for providers.

Clinical Informatics Board Specialty Certification for Physicians: A Global View

Clinical informatics workforce development is a high priority for medicine. Professional board certification for physicians is an important tool to demonstrating excellence. The recent recognition of clinical informatics as a subspecialty board in the U.S. has generated interest and excitement among the U.S. informatics community. To determine the extent of similar programs in countries around the world, we performed literature searches with relevant keywords and internet searches of websites of informatics societies around the world for mentions or descriptions of certifications and reviewed publicly available sources. The U.S. certification was prominent in the recent published literature. Germany and Belgium have long-standing certifications with South Korea and Sri Lanka considering similar programs. This is the first global view of clinical informatics board certification for physicians. Training and certification for non-physician informatics professionals in allied areas are widespread. Official recognition and certification for physicians and all informatics professionals represents a key component of capacity building and a means of addressing the shortage of a skilled informatics workforce. Wider adoption of certification programs may further attracting talent and accelerate growth of the field.

Taming Big Data: An Information Extraction Strategy for Large Clinical Text Corpora

Concepts of interest for clinical and research purposes are not uniformly distributed in clinical text available in electronic medical records. The purpose of our study was to identify filtering techniques to select 'high yield' documents for increased efficacy and throughput. Using two large corpora of clinical text, we demonstrate the identification of 'high yield' document sets in two unrelated domains: homelessness and indwelling urinary catheters. For homelessness, the high yield set includes homeless program and social work notes. For urinary catheters, concepts were more prevalent in notes from hospitalized patients; nursing notes accounted for a majority of the high yield set. This filtering will enable customization and refining of information extraction pipelines to facilitate extraction of relevant concepts for clinical decision support and other uses.

Nora: A Vocabulary Discovery Tool for Concept Extraction

Coverage of terms in domain-specific terminologies and ontologies is often limited in controlled medical vocabularies. Creating and augmenting such terminologies is resource intensive. We developed Nora as an interactive tool to discover terminology from text corpora; the output can then be employed to refine and enhance natural language processing-based concept extraction tasks. Nora provides a visualization of chains of words foraged from word frequency indexes from a text corpus. Domain experts direct and curate chains that contain relevant terms, which are further curated to identify lexical variants. A test of Nora demonstrated an increase of a domain lexicon in homelessness and related psychosocial factors by 38%, yielding an additional 10% extracted concepts.

Extracting Concepts Related to Homelessness from the Free Text of VA Electronic Medical Records

Mining the free text of electronic medical records (EMR) using natural language processing (NLP) is an effective method of extracting information not always captured in administrative data. We sought to determine if concepts related to homelessness, a non-medical condition, were amenable to extraction from the EMR of Veterans Affairs (VA) medical records. As there were no off-the-shelf products, a lexicon of terms related to homelessness was created. A corpus of free text documents from outpatient encounters was reviewed to create the reference standard for NLP training and testing. V3NLP Framework was used to detect instances of lexical terms and was compared to the reference standard. With a positive predictive value of 77% for extracting relevant concepts, this study demonstrates the feasibility of extracting positively asserted concepts related to homelessness from the free text of medical records.

Sophia: A Expedient UMLS Concept Extraction Annotator

An opportunity exists for meaningful concept extraction and indexing from large corpora of clinical notes in the Veterans Affairs (VA) electronic medical record. Currently available tools such as MetaMap, cTAKES and HITex do not scale up to address this big data need. Sophia, a rapid UMLS concept extraction annotator was developed to fulfill a mandate and address extraction where high throughput is needed while preserving performance. We report on the development, testing and benchmarking of Sophia against MetaMap and cTAKEs. Sophia demonstrated improved performance on recall as compared to cTAKES and MetaMap (0.71 vs 0.66 and 0.38). The overall f-score was similar to cTAKES and an improvement over MetaMap (0.53 vs 0.57 and 0.43). With regard to speed of processing records, we noted Sophia to be several fold faster than cTAKES and the scaled-out MetaMap service. Sophia offers a viable alternative for high-throughput information extraction tasks.

Exploring patterns in resource utilization prior to the formal identification of homelessness in recently returned veterans

There are limited data on resources utilized by US Veterans prior to their identification as being homeless. We performed visual analytics on longitudinal medical encounter data prior to the official recognition of homelessness in a large cohort of OEF/OIF Veterans. A statistically significant increase in numbers of several categories of visits in the immediate 30 days prior to the recognition of homelessness was noted as compared to an earlier period. This finding has the potential to inform prediction algorithms based on structured data with a view to intervention and mitigation of homelessness among Veterans.

Recognizing Questions and Answers in EMR Templates Using Natural Language Processing

Templated boilerplate structures pose challenges to natural language processing (NLP) tools used for information extraction (IE). Routine error analyses while performing an IE task using Veterans Affairs (VA) medical records identified templates as an important cause of false positives. The baseline NLP pipeline (V3NLP) was adapted to recognize negation, questions and answers (QA) in various template types by adding a negation and slot:value identification annotator. The system was trained using a corpus of 975 documents developed as a reference standard for extracting psychosocial concepts. Iterative processing using the baseline tool and baseline+negation+QA revealed loss of numbers of concepts with a modest increase in true positives in several concept categories. Similar improvement was noted when the adapted V3NLP was used to process a random sample of 318,000 notes. We demonstrate the feasibility of adapting an NLP pipeline to recognize templates.

Detecting earlier indicators of homelessness in the free text of medical records

Early warning indicators to identify US Veterans at risk of homelessness are currently only inferred from administrative data. References to indicators of risk or instances of homelessness in the free text of medical notes written by Department of Veterans Affairs (VA) providers may precede formal identification of Veterans as being homeless. This represents a potentially untapped resource for early identification. Using natural language processing (NLP), we investigated the idea that concepts related to homelessness written in the free text of the medical record precede the identification of homelessness by administrative data. We found that homeless Veterans were much higher utilizers of VA resources producing approximately 12 times as many documents as non-homeless Veterans. NLP detected mentions of either direct or indirect evidence of homelessness in a significant portion of Veterans earlier than structured data.

Using natural language processing on the free text of clinical documents to screen for evidence of homelessness among US veterans

Information retrieval algorithms based on natural language processing (NLP) of the free text of medical records have been used to find documents of interest from databases. Homelessness is a high priority non-medical diagnosis that is noted in electronic medical records of Veterans in Veterans Affairs (VA) facilities. Using a human-reviewed reference standard corpus of clinical documents of Veterans with evidence of homelessness and those without, an open-source NLP tool (Automated Retrieval Console v2.0, ARC) was trained to classify documents. The best performing model based on document level work-flow performed well on a test set (Precision 94%, Recall 97%, F-Measure 96). Processing of a naïve set of 10,000 randomly selected documents from the VA using this best performing model yielded 463 documents flagged as positive, indicating a 4.7% prevalence of homelessness. Human review noted a precision of 70% for these flags resulting in an adjusted prevalence of homelessness of 3.3% which matches current VA estimates. Further refinements are underway to improve the performance. We demonstrate an effective and rapid lifecycle of using an off-the-shelf NLP tool for screening targets of interest from medical records.

Understanding adoption of a personal health record in rural health care clinics: revealing barriers and facilitators of adoption including attributions about potential patient portal users and self-reported characteristics of early adopting users

Personal health records (PHRs) are important for improving patient care. An important prerequisite to realize benefits of PHR use is patient recruitment. To understand clinic barriers to adoption, we used Rogers' Diffusion of Innovations theory to frame an examination of clinic staff perceptions of a new PHR and perceptions of likely patient portal users. Clinic staff reported many relative advantages and observable benefits of the PHR but also some distinct problems. Attributions about potential patient users included demographic, computer use, and personality characteristics staff expected in likely users. Analysis of patient survey data of early adopters compared to non-users revealed discrepancies between clinic staff expectations and early adopters' self-reports. Implications for improving adoption of PHRs include ensuring compatibility with existing systems and avoiding recruitment biases.

Clinical informatics sub-specialty board certification

Increased funding for health information technology and the advance of electronic health records in hospitals and practices have created the need for a new specialist: the clinical informatician. Clinical informatics was recognized in 2011 as the latest subspecialty in medicine by the American Board of Medical Specialties. This article reviews the need for this new specialty as well as the steps necessary for its creation. The content and training requirements for clinical informatics are discussed as well as eligibility criteria for taking the board examination. Training programs as well as board preparation are addressed along with the expected impact that this new field will have on the practice of medicine.

Validating a strategy for psychosocial phenotyping using a large corpus of clinical text

OBJECTIVE

To develop algorithms to improve efficiency of patient phenotyping using natural language processing (NLP) on text data. Of a large number of note titles available in our database, we sought to determine those with highest yield and precision for psychosocial concepts.

MATERIALS AND METHODS

From a database of over 1 billion documents from US Department of Veterans Affairs medical facilities, a random sample of 1500 documents from each of 218 enterprise note titles were chosen. Psychosocial concepts were extracted using a UIMA-AS-based NLP pipeline (v3NLP), using a lexicon of relevant concepts with negation and template format annotators. Human reviewers evaluated a subset of documents for false positives and sensitivity. High-yield documents were identified by hit rate and precision. Reasons for false positivity were characterized.

RESULTS

A total of 58 707 psychosocial concepts were identified from 316 355 documents for an overall hit rate of 0.2 concepts per document (median 0.1, range 1.6-0). Of 6031 concepts reviewed from a high-yield set of note titles, the overall precision for all concept categories was 80%, with variability among note titles and concept categories. Reasons for false positivity included templating, negation, context, and alternate meaning of words. The sensitivity of the NLP system was noted to be 49% (95% CI 43% to 55%).

CONCLUSIONS

Phenotyping using NLP need not involve the entire document corpus. Our methods offer a generalizable strategy for scaling NLP pipelines to large free text corpora with complex linguistic annotations in attempts to identify patients of a certain phenotype.

Admission decisions and outcomes of community-acquired pneumonia in the homeless population: a review of 172 patients in an urban setting

OBJECTIVES

We compared admission rates, outcomes, and performance of the CURB-65 mortality prediction score of homeless patients and nonhomeless patients with community-acquired pneumonia (CAP).

METHODS

We compared homeless (n = 172) and nonhomeless (n = 1897) patients presenting to a Salt Lake City, Utah, emergency department with CAP from 1996 to 2006. In the homeless cohort, we measured referral from and follow-up with the local homeless health care clinic and arrangement of medical housing.

RESULTS

Homeless patients were younger (44 vs 59 years; P < .001) and had lower CURB-65 scores and higher hospitalization risk (severity-adjusted odds ratio = 1.89; 95% confidence interval = 1.33, 2.69) than did nonhomeless patients, with a similar length of stay, median inpatient cost, and median outpatient cost, even after severity adjustment. Of homeless patients, 22% were referred from the homeless health care clinic to the emergency department; 54% of outpatients and 51% of hospital patients were referred back to the clinic, and medical housing was arranged for 23%.

CONCLUSIONS

A large cohort of homeless patients with CAP demonstrated higher hospitalization risk than but similar length of stay and costs as nonhomeless patients. The strong relationship between the hospital and homeless health care clinic may have contributed to this finding.

Quantitative models of the dose-response and time course of inhalational anthrax in humans

Anthrax poses a community health risk due to accidental or intentional aerosol release. Reliable quantitative dose-response analyses are required to estimate the magnitude and timeline of potential consequences and the effect of public health intervention strategies under specific scenarios. Analyses of available data from exposures and infections of humans and non-human primates are often contradictory. We review existing quantitative inhalational anthrax dose-response models in light of criteria we propose for a model to be useful and defensible. To satisfy these criteria, we extend an existing mechanistic competing-risks model to create a novel Exposure–Infection–Symptomatic illness–Death (EISD) model and use experimental non-human primate data and human epidemiological data to optimize parameter values. The best fit to these data leads to estimates of a dose leading to infection in 50% of susceptible humans (ID₅₀) of 11,000 spores (95% confidence interval 7,200–17,000), ID₁₀ of 1,700 (1,100–2,600), and ID₁ of 160 (100–250). These estimates suggest that use of a threshold to human infection of 600 spores (as suggested in the literature) underestimates the infectivity of low doses, while an existing estimate of a 1% infection rate for a single spore overestimates low dose infectivity. We estimate the median time from exposure to onset of symptoms (incubation period) among untreated cases to be 9.9 days (7.7–13.1) for exposure to ID₅₀, 11.8 days (9.5–15.0) for ID₁₀, and 12.1 days (9.9–15.3) for ID₁. Our model is the first to provide incubation period estimates that are independently consistent with data from the largest known human outbreak. This model refines previous estimates of the distribution of early onset cases after a release and provides support for the recommended 60-day course of prophylactic antibiotic treatment for individuals exposed to low doses.

Anthrax poses a potential community health risk due to accidental or intentional aerosol release. We address the need for a transparent and defensible quantitative dose-response model for inhalational anthrax that is useful for risk assessors in estimating the magnitude and timeline of potential public health consequences should a release occur. Our synthesis of relevant data and previous modeling efforts identifies areas of improvement among many commonly cited dose-response models and estimates. To address those deficiencies, we provide a new model that is based on clear, transparent assumptions and published data from human and non-human primate exposures. Our resulting estimates provide important insight into the infectivity to humans of low inhaled doses of anthrax spores and the timeline of infections after an exposure event. These insights are critical to assessment of the impacts of delays in responding to a large scale aerosol release, as well as the recommended course of antibiotic administration to those potentially exposed.

"Sitting on pins and needles": characterization of symptom descriptions in clinical notes"

Patients report their symptoms and subjective experiences in their own words. These expressions may be clinically meaningful yet are difficult to capture using automated methods. We annotated subjective symptom expressions in 750 clinical notes from the Veterans Affairs EHR. Within each document, subjective symptom expressions were compared to mentions of symptoms in clinical terms and to the assigned ICD-9-CM codes for the encounter. A total of 543 subjective symptom expressions were identified, of which 66.5% were categorized as mental/behavioral experiences and 33.5% somatic experiences. Only two subjective expressions were coded using ICD-9-CM. Subjective expressions were restated in semantically related clinical terms in 246 (45.3%) instances. Nearly one third (31%) of subjective expressions were not coded or restated in standard terminology. The results highlight the diversity of symptom descriptions and the opportunities to further develop natural language processing to extract symptom expressions that are unobtainable by other automated methods.

Perspectives and concerns regarding antimicrobial agent shortages among infectious disease specialists

Antimicrobial shortages have made treating certain infections more difficult. A web-based survey asking about experience with antimicrobial drug shortages was distributed in 2011 to 1328 infectious diseases physician members of the Emerging Infectious Diseases Network of the Infectious Diseases Society of America. A majority (78%) of 627 respondents reported needing to modify antimicrobial choices because of drug shortages within the past 2 years. Antimicrobials most often reported as not available or available but in short supply were trimethoprim-sulfamethoxazole injection (by 65% of respondents), amikacin (by 58%), aztreonam (by 31%), and foscarnet (by 22%). Most respondents (55%) reporting a shortage indicated that the shortage adversely affected patient outcomes and that they were forced to use alternative and second line agents which were either less effective, more toxic, or more costly. Most (70%) indicated that they learned about the shortage from contact with the pharmacy after trying to prescribe a drug in short supply. More effective means of informing physicians about drug shortages is critical to lessen the impact on patient care.

Evaluation of 2009 pandemic influenza A (H1N1) exposures and illness among physicians in training

BACKGROUND

A cluster of influenza-like illness (ILI) among physicians in training during the 2009 influenza A (H1N1) pandemic (pH1N1) led to a health hazard evaluation.

METHODS

We conducted a cross-sectional study to examine exposures, infection control practices, ILI prevalence, and transmission among physicians in training at 4 affiliated hospitals during the pandemic. We administered an electronic survey and met with physicians in training and hospital personnel.

RESULTS

Of the 88 responding physicians, 85% reported exposure to pH1N1. Exposures occurred at work from patients or coworkers and outside of work from coworkers, household members, or the community. Thirteen cases of ILI were reported in May-June 2009; 10 respondents reported working while ill (duration, 1-4 days). Between 13% and 88% of respondents knew which personal protective equipment (PPE) was recommended when caring for influenza patients at the 4 hospitals. The most common reasons for not using PPE were not knowing that a patient had pH1N1 or ILI and not having PPE readily available.

CONCLUSIONS

Physicians in training have gaps in their knowledge of and adherence to recommended PPE and compliance with work restrictions. Our findings underscore the importance of installing isolation precaution signage, making PPE readily available near patients with influenza, and facilitating work restrictions for ill health care personnel.

The EpiCanvas infectious disease weather map: an interactive visual exploration of temporal and spatial correlations

Advances in surveillance science have supported public health agencies in tracking and responding to disease outbreaks. Increasingly, epidemiologists have been tasked with interpreting multiple streams of heterogeneous data arising from varied surveillance systems. As a result public health personnel have experienced an overload of plots and charts as information visualization techniques have not kept pace with the rapid expansion in data availability. This study sought to advance the science of public health surveillance data visualization by conceptualizing a visual paradigm that provides an 'epidemiological canvas' for detection, monitoring, exploration and discovery of regional infectious disease activity and developing a software prototype of an 'infectious disease weather map'. Design objectives were elucidated and the conceptual model was developed using cognitive task analysis with public health epidemiologists. The software prototype was pilot tested using retrospective data from a large, regional pediatric hospital, and gastrointestinal and respiratory disease outbreaks were re-created as a proof of concept.

Analysis of False Positive Errors of an Acute Respiratory Infection Text Classifier due to Contextual Features

Text classifiers have been used for biosurveillance tasks to identify patients with diseases or conditions of interest. When compared to a clinical reference standard of 280 cases of Acute Respiratory Infection (ARI), a text classifier consisting of simple rules and NegEx plus string matching for specific concepts of interest produced 569 (4%) false positive (FP) cases. Using instance level manual annotation we estimate the prevalence of contextual attributes and error types leading to FP cases. Errors were due to (1) Deletion errors from abbreviations, spelling mistakes and missing synonyms (57%); (2) Insertion errors from templated document structures such as check boxes, and lists of signs and symptoms (36%) and; (3) Substitution errors from irrelevant concepts and alternate meanings for the same word (6%). We demonstrate that specific concept attributes contribute to false positive cases. These results will inform modifications and adaptations to improve text classifier performance.

Pilot study of linking Web-based supplemental interpretive information to laboratory test reports

Electronic medical records have the ability to link to reference material, providing clinicians with immediate access to information relevant to patient care. Adding relevant links to laboratory test results could add value while minimizing the volume of ancillary text presented. We provided Web-based universal resource locator (URL) links with all results of 7 laboratory tests ordered at ARUP Laboratories (Salt Lake City, UT). URL links provided were modified 7 months later, and use between initial and subsequent URLs was tracked to establish frequency and duration of access to supplemental Web information. Monthly Web-site hit rates for individual tests varied from 0.00% to 3.00% (median, 0.12%). Rare and specialty tests averaged higher hit rates. There was no decay in hit rate 9 months after URLs were removed from test reports. We conclude that links to reference material are accessed by clinicians. The use of Web links months after links were no longer published raises an important issue of long-term maintenance and the resources required to support these features.

Developing a manually annotated clinical document corpus to identify phenotypic information for inflammatory bowel disease

BACKGROUND

Natural Language Processing (NLP) systems can be used for specific Information Extraction (IE) tasks such as extracting phenotypic data from the electronic medical record (EMR). These data are useful for translational research and are often found only in free text clinical notes. A key required step for IE is the manual annotation of clinical corpora and the creation of a reference standard for (1) training and validation tasks and (2) to focus and clarify NLP system requirements. These tasks are time consuming, expensive, and require considerable effort on the part of human reviewers.

METHODS

Using a set of clinical documents from the VA EMR for a particular use case of interest we identify specific challenges and present several opportunities for annotation tasks. We demonstrate specific methods using an open source annotation tool, a customized annotation schema, and a corpus of clinical documents for patients known to have a diagnosis of Inflammatory Bowel Disease (IBD). We report clinician annotator agreement at the document, concept, and concept attribute level. We estimate concept yield in terms of annotated concepts within specific note sections and document types.

RESULTS

Annotator agreement at the document level for documents that contained concepts of interest for IBD using estimated Kappa statistic (95% CI) was very high at 0.87 (0.82, 0.93). At the concept level, F-measure ranged from 0.61 to 0.83. However, agreement varied greatly at the specific concept attribute level. For this particular use case (IBD), clinical documents producing the highest concept yield per document included GI clinic notes and primary care notes. Within the various types of notes, the highest concept yield was in sections representing patient assessment and history of presenting illness. Ancillary service documents and family history and plan note sections produced the lowest concept yield.

CONCLUSION

Challenges include defining and building appropriate annotation schemas, adequately training clinician annotators, and determining the appropriate level of information to be annotated. Opportunities include narrowing the focus of information extraction to use case specific note types and sections, especially in cases where NLP systems will be used to extract information from large repositories of electronic clinical note documents.

Role of Otolaryngologists in Health Care for the Homeless

Objectives:

This study describes the prevalence of head and neck disease in patients seen at an urban, federally qualified clinic for those experiencing homelessness and determines the need for otolaryngologists in providing health care for the homeless.

Methods:

All adult patient visits from 2000 to 2004 were reviewed for ICD-9 diagnostic codes representing otolaryngological diseases. Chart review was performed for patients referred to a volunteer otolaryngologist to confirm diagnoses and referrals.

Results:

During the study period, the clinic served 11,690 unique patients in 59,060 patient visits. Otolaryngological diagnoses accounted for 8,959 of these total visits (15.2%), 94% of which were managed by primary care providers. The most common diagnoses were “upper respiratory infections,” “symptoms, signs, and ill-defined conditions of the head and neck, cough, dysphagia,” “other diseases of upper respiratory tract,” and “disorders of the ear/mastoid process.” An on-site volunteer otolaryngologist provided consultation on 6% of all patients with otolaryngological diagnoses, averaging 108 consultations per year; 39 patients were subsequently referred to community specialists for surgery and audiology services.

Conclusions:

There is a small but significant need for otolaryngologists in providing health care for homeless individuals in an urban setting. The authors encourage otolaryngologists and homeless health-care clinics to establish a relationship for volunteer services.

Developing a manually annotated clinical document corpus to identify phenotypic information for inflammatory bowel disease

BACKGROUND

Natural Language Processing (NLP) systems can be used for specific Information Extraction (IE) tasks such as extracting phenotypic data from the electronic medical record (EMR). These data are useful for translational research and are often found only in free text clinical notes. A key required step for IE is the manual annotation of clinical corpora and the creation of a reference standard for (1) training and validation tasks and (2) to focus and clarify NLP system requirements. These tasks are time consuming, expensive, and require considerable effort on the part of human reviewers.

METHODS

Using a set of clinical documents from the VA EMR for a particular use case of interest we identify specific challenges and present several opportunities for annotation tasks. We demonstrate specific methods using an open source annotation tool, a customized annotation schema, and a corpus of clinical documents for patients known to have a diagnosis of Inflammatory Bowel Disease (IBD). We report clinician annotator agreement at the document, concept, and concept attribute level. We estimate concept yield in terms of annotated concepts within specific note sections and document types.

RESULTS

Annotator agreement at the document level for documents that contained concepts of interest for IBD using estimated Kappa statistic (95% CI) was very high at 0.87 (0.82, 0.93). At the concept level, F-measure ranged from 0.61 to 0.83. However, agreement varied greatly at the specific concept attribute level. For this particular use case (IBD), clinical documents producing the highest concept yield per document included GI clinic notes and primary care notes. Within the various types of notes, the highest concept yield was in sections representing patient assessment and history of presenting illness. Ancillary service documents and family history and plan note sections produced the lowest concept yield.

CONCLUSIONS

Challenges include defining and building appropriate annotation schemas, adequately training clinician annotators, and determining the appropriate level of information to be annotated. Opportunities include narrowing the focus of information extraction to use case specific note types and sections, especially in cases where NLP systems will be used to extract information from large repositories of electronic clinical note documents.

Composite patient reports: a laboratory informatics perspective and pilot project for personalized medicine and translational research

Clinical laboratories are a strong and integral partner in personalized health care. Laboratory information systems hold a vast amount of data representing human phenotypes, genotypes, biomarkers, progression of disease and response to therapy. These structured and unstructured free text data are critical for patient care and a resource for personalized medicine and translational research. Laboratory data are integrated into many electronic medical records that provide "summary reports" and "trending" to visualize longitudinal patient data. However, these generic reports are not sufficient to manage complex sub-specialty patients. There is an urgent need for end-user driven composite reports for the care of such patients. Using multiple myeloma as a model, this pilot was performed to assess the needs of stakeholders and create a customized report. This laboratory informatics solution is delivered at the point of care through the hospital EMR. Future work will involve further integration with hospital systems to promote clinical decision support and translational research.

Infectious disease exposures and contact tracing in homeless shelters

An outbreak among homeless shelter users of a communicable disease with a short generation time would pose serious public health challenges. Data from Toronto were used to examine the number of shelter residents potentially exposed in the event of such an outbreak. A shelter user had contact with a mean of 97 other residents (range, 1-292) in one day and a mean of 120 (range, 2-624) in eight days. After a single week, contact tracing becomes difficult due to the challenge of locating homeless people who have left the shelter system. Over an 8-day period, individuals who used more than one shelter had contact with an average of 98 more other shelter residents than those who stayed in a single shelter had. At the onset of a serious outbreak, it may be desirable to institute policies that strongly encourage individuals to remain at their current shelter for the duration of the outbreak.

Optimizing A syndromic surveillance text classifier for influenza-like illness: Does document source matter?

Syndromic surveillance systems that incorporate electronic free-text data have primarily focused on extracting concepts of interest from chief complaint text, emergency department visit notes, and nurse triage notes. Due to availability and access, there has been limited work in the area of surveilling the full text of all electronic note documents compared with more specific document sources. This study provides an evaluation of the performance of a text classifier for detection of influenza-like illness (ILI) by document sources that are commonly used for biosurveillance by comparing them to routine visit notes, and a full electronic note corpus approach. Evaluating the performance of an automated text classifier for syndromic surveillance by source document will inform decisions regarding electronic textual data sources for potential use by automated biosurveillance systems. Even when a full electronic medical record is available, commonly available surveillance source documents provide acceptable statistical performance for automated ILI surveillance.

Application of Natural Language Processing to VA Electronic Health Records to Identify Phenotypic Characteristics for Clinical and Research Purposes

Informatics tools to extract and analyze clinical information on patients have lagged behind data-mining developments in bioinformatics. While the analyses of an individual's partial or complete genotype is nearly a reality, the phenotypic characteristics that accompany the genotype are not well known and largely inaccessible in free-text patient health records. As the adoption of electronic medical records increases, there exists an urgent need to extract pertinent phenotypic information and make that available to clinicians and researchers. This usually requires the data to be in a structured format that is both searchable and amenable to computation. Using inflammatory bowel disease as an example, this study demonstrates the utility of a natural language processing system (MedLEE) in mining clinical notes in the paperless VA Health Care System. This adaptation of MedLEE is useful for identifying patients with specific clinical conditions, those at risk for or those with symptoms suggestive of those conditions.

Validity of electronic medical record-based rules for the early detection of meningitis and encephalitis

Diseases of the central nervous system (CNS) such as meningitis or encephalitis may represent events of public health interest due to emerging infections and/or NIH/CDC Category B priority pathogens. Apart from influencing treatment and management of the index case, some diagnoses such as meningococcal meningitis warrant an immediate public health response. Others such as West Nile Virus may require public education and vector control. Thus early detection of CNS syndromes is of benefit to patients, providers and public health. While computer-based surveillance methods have been used with success in the early detection of respiratory syndromes, there is little data on their use in CNS syndromes. This study analyzed the validity of a hospital emergency department computer-based surveillance system in the early detection of meningitis and encephalitis and determined the test characteristics of selected computer-based rules.

Hospital electronic medical record-based public health surveillance system deployed during the 2002 Winter Olympic Games

BACKGROUND

Several computer biosurveillance systems are in place to detect events of public health (PH) significance; however, most lack access to timely and detailed patient-level data and investigation of alerts places a strain on PH resources.

METHODS

Hospital-based infection control professionals led a multi-disciplinary team to develop a computer rule-based system that relies on the patient's electronic medical record. The rules operated on HL7 messages transmitted by clinical computing systems and encompassed a variety of types of patient-level data, including laboratory test ordering and results, radiology ordering and reports, emergency room and outpatient clinic visits, and hospital admissions. Laboratory data were mapped to standard vocabularies, and radiology data were processed using natural language-processing algorithms before the rules were applied to filter for events of PH interest. For each rule, statistical process controls were applied to generate alerts when levels exceeded two standard deviations above the mean. The system was deployed at a large hospital in Salt Lake City during the 2002 Winter Olympic Games, and it was accessed 3 times a day to perform surveillance. Daily reports were provided to local PH agencies after preliminary investigation of the alerts.

RESULTS

Of the 24 rules monitored, 9 generated alerts on 11 different occasions. The only significant event of PH interest that was noted during the surveillance period was an increase in influenza during the Games. The positive predictive value of the rules varied with a high value (89%) noted for identification of pneumonia from chest radiograph reports by natural language-processing algorithms.

CONCLUSIONS

With the assistance of a novel computer-based surveillance system linked to the electronic medical record that uses objective, quantifiable events and access to patient data, infection control practitioners could play a front-line role in biosurveillance and facilitate bidirectional communication with PH agencies.