Evaluating automated approaches to anaphylaxis case classification using unstructured data from the FDA Sentinel System

A general framework for developing computable clinical phenotype algorithms

OBJECTIVE

To present a general framework providing high-level guidance to developers of computable algorithms for identifying patients with specific clinical conditions (phenotypes) through a variety of approaches, including but not limited to machine learning and natural language processing methods to incorporate rich electronic health record data.

MATERIALS AND METHODS

Drawing on extensive prior phenotyping experiences and insights derived from 3 algorithm development projects conducted specifically for this purpose, our team with expertise in clinical medicine, statistics, informatics, pharmacoepidemiology, and healthcare data science methods conceptualized stages of development and corresponding sets of principles, strategies, and practical guidelines for improving the algorithm development process.

RESULTS

We propose 5 stages of algorithm development and corresponding principles, strategies, and guidelines: (1) assessing fitness-for-purpose, (2) creating gold standard data, (3) feature engineering, (4) model development, and (5) model evaluation.

DISCUSSION AND CONCLUSION

This framework is intended to provide practical guidance and serve as a basis for future elaboration and extension.

Using machine learning to improve anaphylaxis case identification in medical claims data

Objectives

Anaphylaxis is a severe life-threatening allergic reaction, and its accurate identification in healthcare databases can harness the potential of "Big Data" for healthcare or public health purposes.

Materials and methods

This study used claims data obtained between October 1, 2015 and February 28, 2019 from the CMS database to examine the utility of machine learning in identifying incident anaphylaxis cases. We created a feature selection pipeline to identify critical features between different datasets. Then a variety of unsupervised and supervised methods were used (eg, Sammon mapping and eXtreme Gradient Boosting) to train models on datasets of differing data quality, which reflects the varying availability and potential rarity of ground truth data in medical databases.

Results

Resulting machine learning model accuracies ranged from 47.7% to 94.4% when tested on ground truth data. Finally, we found new features to help experts enhance existing case-finding algorithms.

Discussion

Developing precise algorithms to detect medical outcomes in claims can be a laborious and expensive process, particularly for conditions presented and coded diversely. We found it beneficial to filter out highly potent codes used for data curation to identify underlying patterns and features. To improve rule-based algorithms where necessary, researchers could use model explainers to determine noteworthy features, which could then be shared with experts and included in the algorithm.

Conclusion

Our work suggests machine learning models can perform at similar levels as a previously published expert case-finding algorithm, while also having the potential to improve performance or streamline algorithm construction processes by identifying new relevant features for algorithm construction.

The Impact of Artificial Intelligence on Allergy Diagnosis and Treatment

PURPOSE OF REVIEW

Artificial intelligence (AI), be it neuronal networks, machine learning or deep learning, has numerous beneficial effects on healthcare systems; however, its potential applications and diagnostic capabilities for immunologic diseases have yet to be explored. Understanding AI systems can help healthcare workers better assimilate artificial intelligence into their practice and unravel its potential in diagnostics, clinical research, and disease management.

RECENT FINDINGS

We reviewed recent advancements in AI systems and their integration in healthcare systems, along with their potential benefits in the diagnosis and management of diseases. We explored machine learning as employed in allergy diagnosis and its learning patterns from patient datasets, as well as the possible advantages of using AI in the field of research related to allergic reactions and even remote monitoring. Considering the ethical challenges and privacy concerns raised by clinicians and patients with regard to integrating AI in healthcare, we explored the new guidelines adapted by regulatory bodies. Despite these challenges, AI appears to have been successfully incorporated into various healthcare systems and is providing patient-centered solutions while simultaneously assisting healthcare workers. Artificial intelligence offers new hope in the field of immunologic disease diagnosis, monitoring, and management and thus has the potential to revolutionize healthcare systems.

BERT-based language model for accurate drug adverse event extraction from social media: implementation, evaluation, and contributions to pharmacovigilance practices

Introduction

Social media platforms serve as a valuable resource for users to share health-related information, aiding in the monitoring of adverse events linked to medications and treatments in drug safety surveillance. However, extracting drug-related adverse events accurately and efficiently from social media poses challenges in both natural language processing research and the pharmacovigilance domain.

Method

Recognizing the lack of detailed implementation and evaluation of Bidirectional Encoder Representations from Transformers (BERT)-based models for drug adverse event extraction on social media, we developed a BERT-based language model tailored to identifying drug adverse events in this context. Our model utilized publicly available labeled adverse event data from the ADE-Corpus-V2. Constructing the BERT-based model involved optimizing key hyperparameters, such as the number of training epochs, batch size, and learning rate. Through ten hold-out evaluations on ADE-Corpus-V2 data and external social media datasets, our model consistently demonstrated high accuracy in drug adverse event detection.

Result

The hold-out evaluations resulted in average F1 scores of 0.8575, 0.9049, and 0.9813 for detecting words of adverse events, words in adverse events, and words not in adverse events, respectively. External validation using human-labeled adverse event tweets data from SMM4H further substantiated the effectiveness of our model, yielding F1 scores 0.8127, 0.8068, and 0.9790 for detecting words of adverse events, words in adverse events, and words not in adverse events, respectively.

Discussion

This study not only showcases the effectiveness of BERT-based language models in accurately identifying drug-related adverse events in the dynamic landscape of social media data, but also addresses the need for the implementation of a comprehensive study design and evaluation. By doing so, we contribute to the advancement of pharmacovigilance practices and methodologies in the context of emerging information sources like social media.

Six Years of the US Food and Drug Administration's Postmarket Active Risk Identification and Analysis System in the Sentinel Initiative: Implications for Real World Evidence Generation

Congress mandated the creation of a postmarket Active Risk Identification and Analysis (ARIA) system containing data on 100 million individuals for monitoring risks associated with drug and biologic products using data from disparate sources to complement the US Food and Drug Administration's (FDA's) existing postmarket capabilities. We report on the first 6 years of ARIA utilization in the Sentinel System (2016-2021). The FDA has used the ARIA system to evaluate 133 safety concerns; 54 of these evaluations have closed with regulatory determinations, whereas the rest remain in progress. If the ARIA system and the FDA's Adverse Event Reporting System are deemed insufficient to address a safety concern, then the FDA may issue a postmarket requirement to a product's manufacturer. One hundred ninety-seven ARIA insufficiency determinations have been made. The most common situation for which ARIA was found to be insufficient is the evaluation of adverse pregnancy and fetal outcomes following in utero drug exposure, followed by neoplasms and death. ARIA was most likely to be sufficient for thromboembolic events, which have high positive predictive value in claims data alone and do not require supplemental clinical data. The lessons learned from this experience illustrate the continued challenges using administrative claims data, especially to define novel clinical outcomes. This analysis can help to identify where more granular clinical data are needed to fill gaps to improve the use of real-world data for drug safety analyses and provide insights into what is needed to efficiently generate high-quality real-world evidence for efficacy.

Trends and opportunities in computable clinical phenotyping: A scoping review

Identifying patient cohorts meeting the criteria of specific phenotypes is essential in biomedicine and particularly timely in precision medicine. Many research groups deliver pipelines that automatically retrieve and analyze data elements from one or more sources to automate this task and deliver high-performing computable phenotypes. We applied a systematic approach based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines to conduct a thorough scoping review on computable clinical phenotyping. Five databases were searched using a query that combined the concepts of automation, clinical context, and phenotyping. Subsequently, four reviewers screened 7960 records (after removing over 4000 duplicates) and selected 139 that satisfied the inclusion criteria. This dataset was analyzed to extract information on target use cases, data-related topics, phenotyping methodologies, evaluation strategies, and portability of developed solutions. Most studies supported patient cohort selection without discussing the application to specific use cases, such as precision medicine. Electronic Health Records were the primary source in 87.1 % (N = 121) of all studies, and International Classification of Diseases codes were heavily used in 55.4 % (N = 77) of all studies, however, only 25.9 % (N = 36) of the records described compliance with a common data model. In terms of the presented methods, traditional Machine Learning (ML) was the dominant method, often combined with natural language processing and other approaches, while external validation and portability of computable phenotypes were pursued in many cases. These findings revealed that defining target use cases precisely, moving away from sole ML strategies, and evaluating the proposed solutions in the real setting are essential opportunities for future work. There is also momentum and an emerging need for computable phenotyping to support clinical and epidemiological research and precision medicine.

Improving Methods of Identifying Anaphylaxis for Medical Product Safety Surveillance Using Natural Language Processing and Machine Learning

We sought to determine whether machine learning and natural language processing (NLP) applied to electronic medical records could improve performance of automated health-care claims-based algorithms to identify anaphylaxis events using data on 516 patients with outpatient, emergency department, or inpatient anaphylaxis diagnosis codes during 2015-2019 in 2 integrated health-care institutions in the Northwest United States. We used one site's manually reviewed gold-standard outcomes data for model development and the other's for external validation based on cross-validated area under the receiver operating characteristic curve (AUC), positive predictive value (PPV), and sensitivity. In the development site 154 (64%) of 239 potential events met adjudication criteria for anaphylaxis compared with 180 (65%) of 277 in the validation site. Logistic regression models using only structured claims data achieved a cross-validated AUC of 0.58 (95% CI: 0.54, 0.63). Machine learning improved cross-validated AUC to 0.62 (0.58, 0.66); incorporating NLP-derived covariates further increased cross-validated AUCs to 0.70 (0.66, 0.75) in development and 0.67 (0.63, 0.71) in external validation data. A classification threshold with cross-validated PPV of 79% and cross-validated sensitivity of 66% in development data had cross-validated PPV of 78% and cross-validated sensitivity of 56% in external data. Machine learning and NLP-derived data improved identification of validated anaphylaxis events.

Challenges and opportunities for mining adverse drug reactions: perspectives from pharma, regulatory agencies, healthcare providers and consumers

Monitoring drug safety is a central concern throughout the drug life cycle. Information about toxicity and adverse events is generated at every stage of this life cycle, and stakeholders have a strong interest in applying text mining and artificial intelligence (AI) methods to manage the ever-increasing volume of this information. Recognizing the importance of these applications and the role of challenge evaluations to drive progress in text mining, the organizers of BioCreative VII (Critical Assessment of Information Extraction in Biology) convened a panel of experts to explore 'Challenges in Mining Drug Adverse Reactions'. This article is an outgrowth of the panel; each panelist has highlighted specific text mining application(s), based on their research and their experiences in organizing text mining challenge evaluations. While these highlighted applications only sample the complexity of this problem space, they reveal both opportunities and challenges for text mining to aid in the complex process of drug discovery, testing, marketing and post-market surveillance. Stakeholders are eager to embrace natural language processing and AI tools to help in this process, provided that these tools can be demonstrated to add value to stakeholder workflows. This creates an opportunity for the BioCreative community to work in partnership with regulatory agencies, pharma and the text mining community to identify next steps for future challenge evaluations.

W. Emerging technologies and their impact on regulatory science

There is an evolution and increasing need for the utilization of emerging cellular, molecular and in silico technologies and novel approaches for safety assessment of food, drugs, and personal care products. Convergence of these emerging technologies is also enabling rapid advances and approaches that may impact regulatory decisions and approvals. Although the development of emerging technologies may allow rapid advances in regulatory decision making, there is concern that these new technologies have not been thoroughly evaluated to determine if they are ready for regulatory application, singularly or in combinations. The magnitude of these combined technical advances may outpace the ability to assess fit for purpose and to allow routine application of these new methods for regulatory purposes. There is a need to develop strategies to evaluate the new technologies to determine which ones are ready for regulatory use. The opportunity to apply these potentially faster, more accurate, and cost-effective approaches remains an important goal to facilitate their incorporation into regulatory use. However, without a clear strategy to evaluate emerging technologies rapidly and appropriately, the value of these efforts may go unrecognized or may take longer. It is important for the regulatory science field to keep up with the research in these technically advanced areas and to understand the science behind these new approaches. The regulatory field must understand the critical quality attributes of these novel approaches and learn from each other's experience so that workforces can be trained to prepare for emerging global regulatory challenges. Moreover, it is essential that the regulatory community must work with the technology developers to harness collective capabilities towards developing a strategy for evaluation of these new and novel assessment tools.

"Artificial Intelligence" for Pharmacovigilance: Ready for Prime Time?

There is great interest in the application of 'artificial intelligence' (AI) to pharmacovigilance (PV). Although US FDA is broadly exploring the use of AI for PV, we focus on the application of AI to the processing and evaluation of Individual Case Safety Reports (ICSRs) submitted to the FDA Adverse Event Reporting System (FAERS). We describe a general framework for considering the readiness of AI for PV, followed by some examples of the application of AI to ICSR processing and evaluation in industry and FDA. We conclude that AI can usefully be applied to some aspects of ICSR processing and evaluation, but the performance of current AI algorithms requires a 'human-in-the-loop' to ensure good quality. We identify outstanding scientific and policy issues to be addressed before the full potential of AI can be exploited for ICSR processing and evaluation, including approaches to quality assurance of 'human-in-the-loop' AI systems, large-scale, publicly available training datasets, a well-defined and computable 'cognitive framework', a formal sociotechnical framework for applying AI to PV, and development of best practices for applying AI to PV. Practical experience with stepwise implementation of AI for ICSR processing and evaluation will likely provide important lessons that will inform the necessary policy and regulatory framework to facilitate widespread adoption and provide a foundation for further development of AI approaches to other aspects of PV.

Broadening the reach of the FDA Sentinel system: A roadmap for integrating electronic health record data in a causal analysis framework

The Sentinel System is a major component of the United States Food and Drug Administration's (FDA) approach to active medical product safety surveillance. While Sentinel has historically relied on large quantities of health insurance claims data, leveraging longitudinal electronic health records (EHRs) that contain more detailed clinical information, as structured and unstructured features, may address some of the current gaps in capabilities. We identify key challenges when using EHR data to investigate medical product safety in a scalable and accelerated way, outline potential solutions, and describe the Sentinel Innovation Center's initiatives to put solutions into practice by expanding and strengthening the existing system with a query-ready, large-scale data infrastructure of linked EHR and claims data. We describe our initiatives in four strategic priority areas: (1) data infrastructure, (2) feature engineering, (3) causal inference, and (4) detection analytics, with the goal of incorporating emerging data science innovations to maximize the utility of EHR data for medical product safety surveillance.

Validity of hospital ICD-10-GM codes to identify anaphylaxis

PURPOSE

Anaphylaxis (ANA) is an important adverse drug reaction. We examined positive predictive values (PPV) and other test characteristics of ICD-10-GM code algorithms for detecting ANA as used in a multinational safety study (PASS).

METHODS

We performed a cross-sectional study on routine data from a German academic hospital (2004-2019, age ≥ 18). Chart review was used for case verification. Potential cases were identified from the hospital administration system. The main outcome required at least one of the following: any type of specific in-hospital code (T78.2, T88.6, and T80.5) OR specific outpatient code in combination with a symptom code OR in-hospital non-specific code (T78.4, T88.7, and Y57.9) in combination with two symptom codes. PPV were calculated with 95% confidence interval. Sensitivity analyses modified type of codes, unit of analysis, verification criteria and time period. The most specific algorithm used only primary codes for ANA (numbers added in brackets).

RESULTS

Four hundred and sixteen eligible cases were evaluated, and 78 (37) potential ANA cases were identified. PPV were 62.8% (95% CI 51.1-73.5) (main) and 77.4% (58.9-90.4) (most specific). PPV from all modifications ranged from 12.9% to 80.6%. The sensitivity of the main algorithm was 66.2%, specificity 91.5%, and negative predictive value 92.6%. Corresponding figures for the most specific algorithm were 32.4%, 98.0%, and 87.0%.

CONCLUSIONS

The PPV of the main algorithm seems of acceptable validity for use in comparative safety research but will underestimate absolute risks by about a third. Restriction to primary discharge codes markedly improves PPV to the expense of reducing sensitivity.

Development and validation of a predictive model algorithm to identify anaphylaxis in adults with type 2 diabetes in U.S. administrative claims data

PURPOSE

To use medical record adjudication and predictive modeling methods to develop and validate an algorithm to identify anaphylaxis among adults with type 2 diabetes (T2D) in administrative claims.

METHODS

A conventional screening algorithm that prioritized sensitivity to identify potential anaphylaxis cases was developed and consisted of diagnosis codes for anaphylaxis or relevant signs and symptoms. This algorithm was applied to adults with T2D in the HealthCore Integrated Research Database (HIRD) from 2016 to 2018. Clinical experts adjudicated anaphylaxis case status from redacted medical records. We used confirmed case status as an outcome for predictive models developed using lasso regression with 10-fold cross-validation to identify predictors and estimate the probability of confirmed anaphylaxis.

RESULTS

Clinical adjudicators reviewed medical records with sufficient information from 272 adults identified by the anaphylaxis screening algorithm, which had an estimated Positive Predictive Value (PPV) of 65% (95% confidence interval [CI]: 60%-71%). The predictive model algorithm had a c-statistic of 0.95. The model's probability threshold of 0.60 excluded 89% (84/94) of false positives identified by the screening algorithm, with a PPV of 94% (95% CI: 91%-98%). The model excluded very few true positives (15 of 178), and identified 92% (95% CI: 87%-96%) of the cases selected by the screening algorithm.

CONCLUSIONS

Predictive modeling techniques yielded an accurate algorithm with high PPV and sensitivity for identifying anaphylaxis in administrative claims. This algorithm could be considered in future safety studies using similar claims data to reduce potential outcome misclassification.

Using and improving distributed data networks to generate actionable evidence: the case of real-world outcomes in the Food and Drug Administration's Sentinel system

The US Food and Drug Administration (FDA) Sentinel System uses a distributed data network, a common data model, curated real-world data, and distributed analytic tools to generate evidence for FDA decision-making. Sentinel system needs include analytic flexibility, transparency, and reproducibility while protecting patient privacy. Based on over a decade of experience, a critical system limitation is the inability to identify enough medical conditions of interest in observational data to a satisfactory level of accuracy. Improving the system's ability to use computable phenotypes will require an "all of the above" approach that improves use of electronic health data while incorporating the growing array of complementary electronic health record data sources. FDA recently funded a Sentinel System Innovation Center and a Community Building and Outreach Center that will provide a platform for collaboration across disciplines to promote better use of real-world data for decision-making.

The use of natural language processing to identify vaccine-related anaphylaxis at five health care systems in the Vaccine Safety Datalink

PURPOSE

The objective was to develop a natural language processing (NLP) algorithm to identify vaccine-related anaphylaxis from plain-text clinical notes, and to implement the algorithm at five health care systems in the Vaccine Safety Datalink.

METHODS

The NLP algorithm was developed using an internal NLP tool and training dataset of 311 potential anaphylaxis cases from Kaiser Permanente Southern California (KPSC). We applied the algorithm to the notes of another 731 potential cases (423 from KPSC; 308 from other sites) with relevant codes (ICD-9-CM diagnosis codes for anaphylaxis, vaccine adverse reactions, and allergic reactions; Healthcare Common Procedure Coding System codes for epinephrine administration). NLP results were compared against a reference standard of chart reviewed and adjudicated cases. The algorithm was then separately applied to the notes of 6 427 359 KPSC vaccination visits (9 402 194 vaccine doses) without relevant codes.

RESULTS

At KPSC, NLP identified 12 of 16 true vaccine-related cases and achieved a sensitivity of 75.0%, specificity of 98.5%, positive predictive value (PPV) of 66.7%, and negative predictive value of 99.0% when applied to notes of patients with relevant diagnosis codes. NLP did not identify the five true cases at other sites. When NLP was applied to the notes of KPSC patients without relevant codes, it captured eight additional true cases confirmed by chart review and adjudication.

CONCLUSIONS

The current study demonstrated the potential to apply rule-based NLP algorithms to clinical notes to identify anaphylaxis cases. Increasing the size of training data, including clinical notes from all participating study sites in the training data, and preprocessing the clinical notes to handle special characters could improve the performance of the NLP algorithms. We recommend adding an NLP process followed by manual chart review in future vaccine safety studies to improve sensitivity and efficiency.

Advances in drug allergy, urticaria, angioedema, and anaphylaxis in 2018

Many notable advances in drug allergy, urticaria, angioedema, and anaphylaxis were reported in 2018. Broad-spectrum antibiotic use and, consequently, antibiotic resistance are widespread, and algorithms to clarify β-lactam allergy and optimize antibiotic use were described. Meaningful data emerged on the pathogenesis of delayed drug hypersensitivity reactions. Progress not only in defining biomarkers but also in understanding the effect on quality of life and developing better treatments has been made for patients with chronic idiopathic urticaria. Patients with hereditary angioedema (HAE) have gained additional access to highly efficacious therapies, with associated improvements in quality of life, and some progress was made in our understanding of recurrent angioedema in patients with normal laboratory results. Guidelines have defined clear goals to help providers optimize therapies in patients with HAE. The epidemiology and triggers of anaphylaxis and the mechanisms underlying anaphylaxis were elucidated further. In summary, these disorders (and labels) cause substantial burdens for individual persons and even society. Fortunately, publications in 2018 have informed on advancements in diagnosis and management and have provided better understanding of mechanisms that potentially could yield new therapies. This progress should lead to better health outcomes and paths forward in patients with drug allergy, urticaria, HAE, and anaphylaxis.

Challenges and Opportunities for Using Big Health Care Data to Advance Medical Science and Public Health

Methodological advancements in epidemiology, biostatistics, and data science have strengthened the research world's ability to use data captured from electronic health records (EHRs) to address pressing medical questions, but gaps remain. We describe methods investments that are needed to curate EHR data toward research quality and to integrate complementary data sources when EHR data alone are insufficient for research goals. We highlight new methods and directions for improving the integrity of medical evidence generated from pragmatic trials, observational studies, and predictive modeling. We also discuss needed methods contributions to further ease data sharing across multisite EHR data networks. Throughout, we identify opportunities for training and for bolstering collaboration among subject matter experts, methodologists, practicing clinicians, and health system leaders to help ensure that methods problems are identified and resulting advances are translated into mainstream research practice more quickly.