Wang, Liwei

Abstract Error analysis is a critical step in evaluating and improving clinical concept extraction models, the most common clinical natural language processing (NLP) task. Unlike corpus annotation, which follows standardized protocols for creating gold-standard datasets, error analysis requires nuanced judgment grounded in both clinical expertise and NLP knowledge. This is especially important given the heterogeneity of clinical text, where variations in documentation style, note structure, and terminology can substantially influence model behavior. Despite its importance, there is currently no standardized, user-level framework to support systematic error analysis in clinical concept extraction task. In this study, we developed and validated MedError, a machine-assisted, human-in-the-loop framework designed to standardize and enhance error analysis for clinical concept extraction tasks. We collected and manually curated a corpus of 1,187 unique errors from a total of 4,237 notes across three different distinct hospitals. The error categories were defined using our previously validated error taxonomy and included 480 false negatives and 707 false positives across 25 error types and 48 clinical concept categories. We evaluated the performance of three proprietary and three open-source large language models (LLMs) in automatically classifying these errors into 26 and 15 predefined categories. We further developed a machine-assisted framework, MedError, which integrates best practices in error analysis, LLM-assisted classification and reasoning, and a user-friendly interface to enable more efficient, reproducible, and context-aware error analysis. The framework supports both single-site and federated multisite error analysis, facilitating the effective translation of clinical NLP systems into real-world settings.

Abstract Current discussion surrounding the clinical capabilities of generative language models (GLMs) predominantly center around multiple-choice question-answer (MCQA) benchmarks derived from clinical licensing examinations. While accepted for human examinees, characteristics unique to GLMs bring into question the validity of such benchmarks. Here, we validate four benchmarks using eight GLMs, ablating for parameter size and reasoning capabilities, validating via prompt permutation three key assumptions that underpin the generalizability of MCQA-based assessments: that knowledge is applied, not memorized, that semantic consistency will lead to consistent answers, and that situations with no answers can be recognized. While large models are more resilient to our perturbations compared to small models, we globally invalidate these assumptions, with implications for reasoning models. Additionally, despite retaining the knowledge, small models are prone to answer memorization. All models exhibit significant failure in null-answer scenarios. We then suggest several adaptations for more robust benchmark designs more reflective of real-world conditions.

Abstract Background Error analysis plays a crucial role in clinical concept extraction, a fundamental subtask within clinical natural language processing (NLP). The process typically involves a manual review of error types, such as contextual and linguistic factors contributing to their occurrence, and the identification of underlying causes to refine the NLP model and improve its performance. Conducting error analysis can be complex, requiring a combination of NLP expertise and domain-specific knowledge. Due to the high heterogeneity of electronic health record (EHR) settings across different institutions, challenges may arise when attempting to standardize and reproduce the error analysis process. Objectives This study aims to facilitate a collaborative effort to establish common definitions and taxonomies for capturing diverse error types, fostering community consensus on error analysis for clinical concept extraction tasks. Materials and Methods We iteratively developed and evaluated an error taxonomy based on existing literature, standards, real-world data, multisite case evaluations, and community feedback. The finalized taxonomy was released in both .dtd and .owl formats at the Open Health Natural Language Processing Consortium. The taxonomy is compatible with several different open-source annotation tools, including MAE, Brat, and MedTator. Results The resulting error taxonomy comprises 43 distinct error classes, organized into 6 error dimensions and 4 properties, including model type (symbolic and statistical machine learning), evaluation subject (model and human), evaluation level (patient, document, sentence, and concept), and annotation examples. Internal and external evaluations revealed strong variations in error types across methodological approaches, tasks, and EHR settings. Key points emerged from community feedback, including the need to enhancing clarity, generalizability, and usability of the taxonomy, along with dissemination strategies. Conclusion The proposed taxonomy can facilitate the acceleration and standardization of the error analysis process in multi-site settings, thus improving the provenance, interpretability, and portability of NLP models. Future researchers could explore the potential direction of developing automated or semi-automated methods to assist in the classification and standardization of error analysis.

Abstract Objective Transformer-based language models are prevailing in the clinical domain due to their excellent performance on clinical NLP tasks. The generalizability of those models is usually ignored during the model development process. This study evaluated the generalizability of CancerBERT, a Transformer-based clinical NLP model, along with classic machine learning models, i.e., conditional random field (CRF), bi-directional long short-term memory CRF (BiLSTM-CRF), across different clinical institutes through a breast cancer phenotype extraction task. Materials and methods Two clinical corpora of breast cancer patients were collected from the electronic health records from the University of Minnesota (UMN) and Mayo Clinic (MC), and annotated following the same guideline. We developed three types of NLP models (i.e., CRF, BiLSTM-CRF and CancerBERT) to extract cancer phenotypes from clinical texts. We evaluated the generalizability of models on different test sets with different learning strategies (model transfer vs locally trained). The entity coverage score was assessed with their association with the model performances. Results We manually annotated 200 and 161 clinical documents at UMN and MC. The corpora of the two institutes were found to have higher similarity between the target entities than the overall corpora. The CancerBERT models obtained the best performances among the independent test sets from two clinical institutes and the permutation test set. The CancerBERT model developed in one institute and further fine-tuned in another institute achieved reasonable performance compared to the model developed on local data (micro-F1: 0.925 vs 0.932). Conclusions The results indicate the CancerBERT model has superior learning ability and generalizability among the three types of clinical NLP models for our named entity recognition task. It has the advantage to recognize complex entities, e.g., entities with different labels.

Abstract Despite recent methodology advancements in clinical natural language processing (NLP), the adoption of clinical NLP models within the translational research community remains hindered by process heterogeneity and human factor variations. Concurrently, these factors also dramatically increase the difficulty in developing NLP models in multi-site settings, which is necessary for algorithm robustness and generalizability. Here, we reported on our experience developing an NLP solution for Coronavirus Disease 2019 (COVID-19) signs and symptom extraction in an open NLP framework from a subset of sites participating in the National COVID Cohort (N3C). We then empirically highlight the benefits of multi-site data for both symbolic and statistical methods, as well as highlight the need for federated annotation and evaluation to resolve several pitfalls encountered in the course of these efforts.

AbstractAn increasing number of studies have reported using natural language processing (NLP) to assist observational research by extracting clinical information from electronic health records (EHRs). Currently, no standardized reporting guidelines for NLP‐assisted observational studies exist. The absence of detailed reporting guidelines may create ambiguity in the use of NLP‐derived content, knowledge gaps in the current research reporting practices, and reproducibility challenges. To address these issues, we conducted a scoping review of NLP‐assisted observational clinical studies and examined their reporting practices, focusing on NLP methodology and evaluation. Through our investigation, we discovered a high variation regarding the reporting practices, such as inconsistent use of references for measurement studies, variation in the reporting location (reference, appendix, and manuscript), and different granularity of NLP methodology and evaluation details. To promote the wide adoption and utilization of NLP solutions in clinical research, we outline several perspectives that align with the six principles released by the World Health Organization (WHO) that guide the ethical use of artificial intelligence for health.

PURPOSEThe advancement of natural language processing (NLP) has promoted the use of detailed textual data in electronic health records (EHRs) to support cancer research and to facilitate patient care. In this review, we aim to assess EHR for cancer research and patient care by using the Minimal Common Oncology Data Elements (mCODE), which is a community-driven effort to define a minimal set of data elements for cancer research and practice. Specifically, we aim to assess the alignment of NLP-extracted data elements with mCODE and review existing NLP methodologies for extracting said data elements.METHODSPublished literature studies were searched to retrieve cancer-related NLP articles that were written in English and published between January 2010 and September 2020 from main literature databases. After the retrieval, articles with EHRs as the data source were manually identified. A charting form was developed for relevant study analysis and used to categorize data including four main topics: metadata, EHR data and targeted cancer types, NLP methodology, and oncology data elements and standards.RESULTSA total of 123 publications were selected finally and included in our analysis. We found that cancer research and patient care require some data elements beyond mCODE as expected. Transparency and reproductivity are not sufficient in NLP methods, and inconsistency in NLP evaluation exists.CONCLUSIONWe conducted a comprehensive review of cancer NLP for research and patient care using EHRs data. Issues and barriers for wide adoption of cancer NLP were identified and discussed.

Abstract Objective Accurate extraction of breast cancer patients’ phenotypes is important for clinical decision support and clinical research. This study developed and evaluated cancer domain pretrained CancerBERT models for extracting breast cancer phenotypes from clinical texts. We also investigated the effect of customized cancer-related vocabulary on the performance of CancerBERT models. Materials and Methods A cancer-related corpus of breast cancer patients was extracted from the electronic health records of a local hospital. We annotated named entities in 200 pathology reports and 50 clinical notes for 8 cancer phenotypes for fine-tuning and evaluation. We kept pretraining the BlueBERT model on the cancer corpus with expanded vocabularies (using both term frequency-based and manually reviewed methods) to obtain CancerBERT models. The CancerBERT models were evaluated and compared with other baseline models on the cancer phenotype extraction task. Results All CancerBERT models outperformed all other models on the cancer phenotyping NER task. Both CancerBERT models with customized vocabularies outperformed the CancerBERT with the original BERT vocabulary. The CancerBERT model with manually reviewed customized vocabulary achieved the best performance with macro F1 scores equal to 0.876 (95% CI, 0.873–0.879) and 0.904 (95% CI, 0.902–0.906) for exact match and lenient match, respectively. Conclusions The CancerBERT models were developed to extract the cancer phenotypes in clinical notes and pathology reports. The results validated that using customized vocabulary may further improve the performances of domain specific BERT models in clinical NLP tasks. The CancerBERT models developed in the study would further help clinical decision support.