Artificial intelligence in virtual standardized patients: Combining natural language understanding and rule based dialogue management to improve conversational fidelity.

INTRODUCTION: Advances in natural language understanding have facilitated the development of Virtual Standardized Patients (VSPs) that may soon rival human patients in conversational ability. We describe herein the development of an artificial intelligence (AI) system for VSPs enabling students to practice their history taking skills. METHODS: Our system consists of (1) Automated Speech Recognition (ASR), (2) hybrid AI for question identification, (3) classifier to choose between the two systems, and (4) automated speech generation. We analyzed the accuracy of the ASR, the two AI systems, the classifier, and student feedback with 620 first year medical students from 2018 to 2021. RESULTS: System accuracy improved from ∼75% in 2018 to ∼90% in 2021 as refinements in algorithms and additional training data were utilized. Student feedback was positive, and most students felt that practicing with the VSPs was a worthwhile experience. CONCLUSION: We have developed a novel hybrid dialogue system that enables artificially intelligent VSPs to correctly answer student questions at levels comparable with human SPs. This system allows trainees to practice and refine their history-taking skills before interacting with human patients.

TextEssence: A Tool for Interactive Analysis of Semantic Shifts Between Corpora.

Embeddings of words and concepts capture syntactic and semantic regularities of language; however, they have seen limited use as tools to study characteristics of different corpora and how they relate to one another. We introduce TextEssence, an interactive system designed to enable comparative analysis of corpora using embeddings. TextEssence includes visual, neighbor-based, and similarity-based modes of embedding analysis in a lightweight, web-based interface. We further propose a new measure of embedding confidence based on nearest neighborhood overlap, to assist in identifying high-quality embeddings for corpus analysis. A case study on COVID-19 scientific literature illustrates the utility of the system. TextEssence can be found at https://textessence.github.io.

Automated Coding of Under-Studied Medical Concept Domains: Linking Physical Activity Reports to the International Classification of Functioning, Disability, and Health.

Linking clinical narratives to standardized vocabularies and coding systems is a key component of unlocking the information in medical text for analysis. However, many domains of medical concepts, such as functional outcomes and social determinants of health, lack well-developed terminologies that can support effective coding of medical text. We present a framework for developing natural language processing (NLP) technologies for automated coding of medical information in under-studied domains, and demonstrate its applicability through a case study on physical mobility function. Mobility function is a component of many health measures, from post-acute care and surgical outcomes to chronic frailty and disability, and is represented as one domain of human activity in the International Classification of Functioning, Disability, and Health (ICF). However, mobility and other types of functional activity remain under-studied in the medical informatics literature, and neither the ICF nor commonly-used medical terminologies capture functional status terminology in practice. We investigated two data-driven paradigms, classification and candidate selection, to link narrative observations of mobility status to standardized ICF codes, using a dataset of clinical narratives from physical therapy encounters. Recent advances in language modeling and word embedding were used as features for established machine learning models and a novel deep learning approach, achieving a macro-averaged F-1 score of 84% on linking mobility activity reports to ICF codes. Both classification and candidate selection approaches present distinct strengths for automated coding in under-studied domains, and we highlight that the combination of (i) a small annotated data set; (ii) expert definitions of codes of interest; and (iii) a representative text corpus is sufficient to produce high-performing automated coding systems. This research has implications for continued development of language technologies to analyze functional status information, and the ongoing growth of NLP tools for a variety of specialized applications in clinical care and research.

Ambiguity in medical concept normalization: An analysis of types and coverage in electronic health record datasets.

OBJECTIVES: Normalizing mentions of medical concepts to standardized vocabularies is a fundamental component of clinical text analysis. Ambiguity-words or phrases that may refer to different concepts-has been extensively researched as part of information extraction from biomedical literature, but less is known about the types and frequency of ambiguity in clinical text. This study characterizes the distribution and distinct types of ambiguity exhibited by benchmark clinical concept normalization datasets, in order to identify directions for advancing medical concept normalization research. MATERIALS AND METHODS: We identified ambiguous strings in datasets derived from the 2 available clinical corpora for concept normalization and categorized the distinct types of ambiguity they exhibited. We then compared observed string ambiguity in the datasets with potential ambiguity in the Unified Medical Language System (UMLS) to assess how representative available datasets are of ambiguity in clinical language. RESULTS: We found that <15% of strings were ambiguous within the datasets, while over 50% were ambiguous in the UMLS, indicating only partial coverage of clinical ambiguity. The percentage of strings in common between any pair of datasets ranged from 2% to only 36%; of these, 40% were annotated with different sets of concepts, severely limiting generalization. Finally, we observed 12 distinct types of ambiguity, distributed unequally across the available datasets, reflecting diverse linguistic and medical phenomena. DISCUSSION: Existing datasets are not sufficient to cover the diversity of clinical concept ambiguity, limiting both training and evaluation of normalization methods for clinical text. Additionally, the UMLS offers important semantic information for building and evaluating normalization methods. CONCLUSIONS: Our findings identify 3 opportunities for concept normalization research, including a need for ambiguity-specific clinical datasets and leveraging the rich semantics of the UMLS in new methods and evaluation measures for normalization.

Using virtual standardized patients to accurately assess information gathering skills in medical students.

Introduction: Practicing a medical history using standardized patients is an essential component of medical school curricula. Recent advances in technology now allow for newer approaches for practicing and assessing communication skills. We describe herein a virtual standardized patient (VSP) system that allows students to practice their history taking skills and receive immediate feedback. Methods: Our VSPs consist of artificially intelligent, emotionally responsive 3D characters which communicate with students using natural language. The system categorizes the input questions according to specific domains and summarizes the encounter. Automated assessment by the computer was compared to manual assessment by trained raters to assess accuracy of the grading system. Results: Twenty dialogs chosen randomly from 102 total encounters were analyzed by three human and one computer rater. Overall scores calculated by the computer were not different than those provided by the human raters, and overall accuracy of the computer system was 87%, compared with 90% for human raters. Inter-rater reliability was high across 19 of 21 categories. Conclusions: We have developed a virtual standardized patient system that can understand, respond, categorize, and assess student performance in gathering information during a typical medical history, thus enabling students to practice their history-taking skills and receive immediate feedback.

HARE: a Flexible Highlighting Annotator for Ranking and Exploration.

Exploration and analysis of potential data sources is a significant challenge in the application of NLP techniques to novel information domains. We describe HARE, a system for highlighting relevant information in document collections to support ranking and triage, which provides tools for post-processing and qualitative analysis for model development and tuning. We apply HARE to the use case of narrative descriptions of mobility information in clinical data, and demonstrate its utility in comparing candidate embedding features. We provide a web-based interface for annotation visualization and document ranking, with a modular backend to support interoperability with existing annotation tools.

A Quantitative and Qualitative Evaluation of Sentence Boundary Detection for the Clinical Domain.

Sentence boundary detection (SBD) is a critical preprocessing task for many natural language processing (NLP) applications. However, there has been little work on evaluating how well existing methods for SBD perform in the clinical domain. We evaluate five popular off-the-shelf NLP toolkits on the task of SBD in various kinds of text using a diverse set of corpora, including the GENIA corpus of biomedical abstracts, a corpus of clinical notes used in the 2010 i2b2 shared task, and two general-domain corpora (the British National Corpus and Switchboard). We find that, with the exception of the cTAKES system, the toolkits we evaluate perform noticeably worse on clinical text than on general-domain text. We identify and discuss major classes of errors, and suggest directions for future work to improve SBD methods in the clinical domain. We also make the code used for SBD evaluation in this paper available for download at http://github.com/drgriffis/SBD-Evaluation.

Automatic data source identification for clinical trial eligibility criteria resolution.

Clinical trial coordinators refer to both structured and unstructured sources of data when evaluating a subject for eligibility. While some eligibility criteria can be resolved using structured data, some require manual review of clinical notes. An important step in automating the trial screening process is to be able to identify the right data source for resolving each criterion. In this work, we discuss the creation of an eligibility criteria dataset for clinical trials for patients with two disparate diseases, annotated with the preferred data source for each criterion (i.e., structured or unstructured) by annotators with medical training. The dataset includes 50 heart-failure trials with a total of 766 eligibility criteria and 50 trials for chronic lymphocytic leukemia (CLL) with 677 criteria. Further, we developed machine learning models to predict the preferred data source: kernel methods outperform simpler learning models when used with a combination of lexical, syntactic, semantic, and surface features. Evaluation of these models indicates that the performance is consistent across data from both diagnoses, indicating generalizability of our method. Our findings are an important step towards ongoing efforts for automation of clinical trial screening.

Comparison of UMLS terminologies to identify risk of heart disease using clinical notes.

The second track of the 2014 i2b2 challenge asked participants to automatically identify risk factors for heart disease among diabetic patients using natural language processing techniques for clinical notes. This paper describes a rule-based system developed using a combination of regular expressions, concepts from the Unified Medical Language System (UMLS), and freely-available resources from the community. With a performance (F1=90.7) that is significantly higher than the median (F1=87.20) and close to the top performing system (F1=92.8), it was the best rule-based system of all the submissions in the challenge. We also used this system to evaluate the utility of different terminologies in the UMLS towards the challenge task. Of the 155 terminologies in the UMLS, 129 (76.78%) have no representation in the corpus. The Consumer Health Vocabulary had very good coverage of relevant concepts and was the most useful terminology for the challenge task. While segmenting notes into sections and lists has a significant impact on the performance, identifying negations and experiencer of the medical event results in negligible gain.

Textual inference for eligibility criteria resolution in clinical trials.

Clinical trials are essential for determining whether new interventions are effective. In order to determine the eligibility of patients to enroll into these trials, clinical trial coordinators often perform a manual review of clinical notes in the electronic health record of patients. This is a very time-consuming and exhausting task. Efforts in this process can be expedited if these coordinators are directed toward specific parts of the text that are relevant for eligibility determination. In this study, we describe the creation of a dataset that can be used to evaluate automated methods capable of identifying sentences in a note that are relevant for screening a patient's eligibility in clinical trials. Using this dataset, we also present results for four simple methods in natural language processing that can be used to automate this task. We found that this is a challenging task (maximum F-score=26.25), but it is a promising direction for further research.

A review of approaches to identifying patient phenotype cohorts using electronic health records.

OBJECTIVE: To summarize literature describing approaches aimed at automatically identifying patients with a common phenotype. MATERIALS AND METHODS: We performed a review of studies describing systems or reporting techniques developed for identifying cohorts of patients with specific phenotypes. Every full text article published in (1) Journal of American Medical Informatics Association, (2) Journal of Biomedical Informatics, (3) Proceedings of the Annual American Medical Informatics Association Symposium, and (4) Proceedings of Clinical Research Informatics Conference within the past 3 years was assessed for inclusion in the review. Only articles using automated techniques were included. RESULTS: Ninety-seven articles met our inclusion criteria. Forty-six used natural language processing (NLP)-based techniques, 24 described rule-based systems, 41 used statistical analyses, data mining, or machine learning techniques, while 22 described hybrid systems. Nine articles described the architecture of large-scale systems developed for determining cohort eligibility of patients. DISCUSSION: We observe that there is a rise in the number of studies associated with cohort identification using electronic medical records. Statistical analyses or machine learning, followed by NLP techniques, are gaining popularity over the years in comparison with rule-based systems. CONCLUSIONS: There are a variety of approaches for classifying patients into a particular phenotype. Different techniques and data sources are used, and good performance is reported on datasets at respective institutions. However, no system makes comprehensive use of electronic medical records addressing all of their known weaknesses.

How essential are unstructured clinical narratives and information fusion to clinical trial recruitment?

Electronic health records capture patient information using structured controlled vocabularies and unstructured narrative text. While structured data typically encodes lab values, encounters and medication lists, unstructured data captures the physician's interpretation of the patient's condition, prognosis, and response to therapeutic intervention. In this paper, we demonstrate that information extraction from unstructured clinical narratives is essential to most clinical applications. We perform an empirical study to validate the argument and show that structured data alone is insufficient in resolving eligibility criteria for recruiting patients onto clinical trials for chronic lymphocytic leukemia (CLL) and prostate cancer. Unstructured data is essential to solving 59% of the CLL trial criteria and 77% of the prostate cancer trial criteria. More specifically, for resolving eligibility criteria with temporal constraints, we show the need for temporal reasoning and information integration with medical events within and across unstructured clinical narratives and structured data.

Inter-annotator reliability of medical events, coreferences and temporal relations in clinical narratives by annotators with varying levels of clinical expertise.

The manual annotation of clinical narratives is an important step for training and validating the performance of automated systems that utilize these clinical narratives. We build an annotation specification to capture medical events, and coreferences and temporal relations between medical events in clinical text. Unfortunately, the process of clinical data annotation is both time consuming and costly. Many annotation efforts have used physicians to annotate the data. We investigate using annotators that are current students or graduates from diverse clinical backgrounds with varying levels of clinical experience. In spite of this diversity, the annotation agreement across our team of annotators is high; the average inter-annotator kappa statistic for medical events, coreferences, temporal relations, and medical event concept unique identifiers was 0.843, 0.859, 0.833, and 0.806, respectively. We describe methods towards leveraging the annotations to support temporal reasoning with medical events.

Segmenting words from natural speech: subsegmental variation in segmental cues.

Most computational models of word segmentation are trained and tested on transcripts of speech, rather than the speech itself, and assume that speech is converted into a sequence of symbols prior to word segmentation. We present a way of representing speech corpora that avoids this assumption, and preserves acoustic variation present in speech. We use this new representation to re-evaluate a key computational model of word segmentation. One finding is that high levels of phonetic variability degrade the model's performance. While robustness to phonetic variability may be intrinsically valuable, this finding needs to be complemented by parallel studies of the actual abilities of children to segment phonetically variable speech.

Effects of disfluencies, predictability, and utterance position on word form variation in English conversation.

Function words, especially frequently occurring ones such as (the, that, and, and of), vary widely in pronunciation. Understanding this variation is essential both for cognitive modeling of lexical production and for computer speech recognition and synthesis. This study investigates which factors affect the forms of function words, especially whether they have a fuller pronunciation (e.g., thi, thaet, aend, inverted-v v) or a more reduced or lenited pronunciation (e.g., thax, thixt, n, ax). It is based on over 8000 occurrences of the ten most frequent English function words in a 4-h sample from conversations from the Switchboard corpus. Ordinary linear and logistic regression models were used to examine variation in the length of the words, in the form of their vowel (basic, full, or reduced), and whether final obstruents were present or not. For all these measures, after controlling for segmental context, rate of speech, and other important factors, there are strong independent effects that made high-frequency monosyllabic function words more likely to be longer or have a fuller form (1) when neighboring disfluencies (such as filled pauses uh and um) indicate that the speaker was encountering problems in planning the utterance; (2) when the word is unexpected, i.e., less predictable in context; (3) when the word is either utterance initial or utterance final. Looking at the phenomenon in a different way, frequent function words are more likely to be shorter and to have less-full forms in fluent speech, in predictable positions or multiword collocations, and utterance internally. Also considered are other factors such as sex (women are more likely to use fuller forms, even after controlling for rate of speech, for example), and some of the differences among the ten function words in their response to the factors.