scholarly journals Text Mining of Adverse Events in Clinical Trials: Deep Learning Approach

10.2196/28632 ◽  
2021 ◽  
Vol 9 (12) ◽  
pp. e28632
Author(s):  
Daphne Chopard ◽  
Matthias S Treder ◽  
Padraig Corcoran ◽  
Nagheen Ahmed ◽  
Claire Johnson ◽  
...  
Keyword(s):  
Clinical Trials ◽  
Adverse Event ◽  
Adverse Events ◽  
Critical Role ◽  
Language Model ◽  
Class Imbalance ◽  
International Classification Of Diseases ◽  
Adverse Event Report ◽  
Serious Adverse Event ◽  
Medical Dictionary

Background Pharmacovigilance and safety reporting, which involve processes for monitoring the use of medicines in clinical trials, play a critical role in the identification of previously unrecognized adverse events or changes in the patterns of adverse events. Objective This study aims to demonstrate the feasibility of automating the coding of adverse events described in the narrative section of the serious adverse event report forms to enable statistical analysis of the aforementioned patterns. Methods We used the Unified Medical Language System (UMLS) as the coding scheme, which integrates 217 source vocabularies, thus enabling coding against other relevant terminologies such as the International Classification of Diseases–10th Revision, Medical Dictionary for Regulatory Activities, and Systematized Nomenclature of Medicine). We used MetaMap, a highly configurable dictionary lookup software, to identify the mentions of the UMLS concepts. We trained a binary classifier using Bidirectional Encoder Representations from Transformers (BERT), a transformer-based language model that captures contextual relationships, to differentiate between mentions of the UMLS concepts that represented adverse events and those that did not. Results The model achieved a high F1 score of 0.8080, despite the class imbalance. This is 10.15 percent points lower than human-like performance but also 17.45 percent points higher than that of the baseline approach. Conclusions These results confirmed that automated coding of adverse events described in the narrative section of serious adverse event reports is feasible. Once coded, adverse events can be statistically analyzed so that any correlations with the trialed medicines can be estimated in a timely fashion.

scholarly journals Text Mining of Adverse Events in Clinical Trials: Deep Learning Approach (Preprint)

2021 ◽  
Author(s):  
Daphne Chopard ◽  
Matthias S Treder ◽  
Padraig Corcoran ◽  
Nagheen Ahmed ◽  
Claire Johnson ◽  
...  
Keyword(s):  
Clinical Trials ◽  
Adverse Event ◽  
Adverse Events ◽  
Critical Role ◽  
Language Model ◽  
Class Imbalance ◽  
International Classification Of Diseases ◽  
Adverse Event Report ◽  
Serious Adverse Event ◽  
Medical Dictionary

BACKGROUND Pharmacovigilance and safety reporting, which involve processes for monitoring the use of medicines in clinical trials, play a critical role in the identification of previously unrecognized adverse events or changes in the patterns of adverse events. OBJECTIVE This study aims to demonstrate the feasibility of automating the coding of adverse events described in the narrative section of the serious adverse event report forms to enable statistical analysis of the aforementioned patterns. METHODS We used the Unified Medical Language System (UMLS) as the coding scheme, which integrates 217 source vocabularies, thus enabling coding against other relevant terminologies such as the International Classification of Diseases–10th Revision, Medical Dictionary for Regulatory Activities, and Systematized Nomenclature of Medicine). We used MetaMap, a highly configurable dictionary lookup software, to identify the mentions of the UMLS concepts. We trained a binary classifier using Bidirectional Encoder Representations from Transformers (BERT), a transformer-based language model that captures contextual relationships, to differentiate between mentions of the UMLS concepts that represented adverse events and those that did not. RESULTS The model achieved a high F1 score of 0.8080, despite the class imbalance. This is 10.15 percent points lower than human-like performance but also 17.45 percent points higher than that of the baseline approach. CONCLUSIONS These results confirmed that automated coding of adverse events described in the narrative section of serious adverse event reports is feasible. Once coded, adverse events can be statistically analyzed so that any correlations with the trialed medicines can be estimated in a timely fashion.

scholarly journals A Bayesian meta-analytic approach for safety signal detection in randomized clinical trials

2017 ◽  
Vol 14 (2) ◽  
pp. 192-200 ◽  
Author(s):  
Motoi Odani ◽  
Satoru Fukimbara ◽  
Tosiya Sato
Keyword(s):  
Clinical Trials ◽  
Adverse Event ◽  
Adverse Events ◽  
Hierarchical Structure ◽  
Meta Analysis ◽  
Event Data ◽  
Analysis Model ◽  
Exact Test ◽  
Adverse Event Data ◽  
Fisher’S Exact Test

Background/Aim: Meta-analyses are frequently performed on adverse event data and are primarily used for improving statistical power to detect safety signals. However, in the evaluation of drug safety for New Drug Applications, simple pooling of adverse event data from multiple clinical trials is still commonly used. We sought to propose a new Bayesian hierarchical meta-analytic approach based on consideration of a hierarchical structure of reported individual adverse event data from multiple randomized clinical trials. Methods: To develop our meta-analysis model, we extended an existing three-stage Bayesian hierarchical model by including an additional stage of the clinical trial level in the hierarchical model; this generated a four-stage Bayesian hierarchical model. We applied the proposed Bayesian meta-analysis models to published adverse event data from three premarketing randomized clinical trials of tadalafil and to a simulation study motivated by the case example to evaluate the characteristics of three alternative models. Results: Comparison of the results from the Bayesian meta-analysis model with those from Fisher’s exact test after simple pooling showed that 6 out of 10 adverse events were the same within a top 10 ranking of individual adverse events with regard to association with treatment. However, more individual adverse events were detected in the Bayesian meta-analysis model than in Fisher’s exact test under the body system “Musculoskeletal and connective tissue disorders.” Moreover, comparison of the overall trend of estimates between the Bayesian model and the standard approach (odds ratios after simple pooling methods) revealed that the posterior median odds ratios for the Bayesian model for most adverse events shrank toward values for no association. Based on the simulation results, the Bayesian meta-analysis model could balance the false detection rate and power to a better extent than Fisher’s exact test. For example, when the threshold value of the posterior probability for signal detection was set to 0.8, the false detection rate was 41% and power was 88% in the Bayesian meta-analysis model, whereas the false detection rate was 56% and power was 86% in Fisher’s exact test. Limitations: Adverse events under the same body system were not necessarily positively related when we used “system organ class” and “preferred term” in the Medical Dictionary for Regulatory Activities as a hierarchical structure of adverse events. For the Bayesian meta-analysis models to be effective, the validity of the hierarchical structure of adverse events and the grouping of adverse events are critical. Conclusion: Our proposed meta-analysis models considered trial effects to avoid confounding by trial and borrowed strength from both within and across body systems to obtain reasonable and stable estimates of an effect measure by considering a hierarchical structure of adverse events.

scholarly journals Safety of SGLT2 Inhibitors: A Pharmacovigilance Study From 2015 to 2020 Based on FDA Adverse Event Report System Database

2021 ◽  
Author(s):  
Xiang Zhou ◽  
Xiaofei Ye ◽  
Yinghong Zhai ◽  
Fangyuan Hu ◽  
Yongqing Gao ◽  
...  
Keyword(s):  
Adverse Event ◽  
Adverse Events ◽  
Diabetic Ketoacidosis ◽  
Adverse Event Reporting System ◽  
Kidney Injury ◽  
Adverse Event Reporting ◽  
Adverse Event Report ◽  
System Database ◽  
Report System ◽  
Euglycemic Diabetic Ketoacidosis

Aim: With the widespread use of SGLT2i, various adverse events (AEs) have been reported. This study aimed to describe the distribution of SGLT2i-related AEs in different systems, quantify the association of important medical events (IMEs) and SGLT2i regimens, and build a signal profile of SGLT2i- induced IMEs. Methods: Data from 2015 Q1 to 2020 Q4 in the FDA Adverse Event Reporting System database (FAERS) were selected to conduct disproportionality analysis. Two signal indicators, the reported odds ratio (ROR) and information component (IC), were used to evaluate the correlation between SGLT2i and IMEs. The lower end of the 95% confidence interval of IC (IC025) exceeding zero was deemed a signal. For ROR, it was defined a signal if ROR025 over one, with at least 3 cases. Results: A total of 45,771,436 records were involved, including 111,564 records related to SGLT2i, with 38,366 records of SGLT2i-induced IMEs. Overall, SGLT2i was significantly associated with IMEs (IC=0.36, 95% CI: 0.35-0.38; ROR=1.44, 95% CI: 1.42-1.46). Most SGLT2i-related adverse events occurred in monotherapy (92.93%). Diabetic ketoacidosis was the most IMEs. Specifically, acute osteomyelitis has the strongest signal of all SGLT2i (IC025=7.83), and it was unique to canagliflozin. Diabetic ketoacidosis, acute kidney injury, ketoacidosis, Fournier’s gangrene, and euglycemic diabetic ketoacidosis were common to the four FDA-approved SGLT2i. Conclusion: Our study demonstrated that different SGLT2i regimens lead to different important adverse events, but there are overlapping events. Early identification and management of SGLT2i-associated IMEs are essential for clinical practice.

Reconsenting patients with cancer on clinical trials: Does added risk influence continued participation?

2009 ◽  
Vol 27 (15_suppl) ◽  
pp. e15610-e15610
Author(s):  
A. Elegbede ◽  
A. Andrei ◽  
A. Andrei ◽  
K. D. Holen
Keyword(s):  
Clinical Trial ◽  
Clinical Trials ◽  
Adverse Event ◽  
Side Effects ◽  
Adverse Events ◽  
Study Participant ◽  
P Value ◽  
Study Participation ◽  
Grade 5 ◽  
Study Results

e15610 Background: The general policy endorsed by multiple professional societies and cooperative groups regarding patients on cancer clinical trials states that subjects should be informed of new adverse events or significant developments during study participation and re-consented to continue on study. However, no information is known as to the effect of re-consenting on a patients’ decision to continue study participation. Our research question addresses how the severity of reported risk to other study participants will impact the subjects’ decision to continue participation in a clinical trial. Methods: We surveyed 34 patients with gastrointestinal (GI) tumors all of whom were currently enrolled in a clinical trial. The survey portrayed hypothetical adverse reactions affecting another study participant ranging from Grade 1 to Grade 5 according to the National Cancer Institutes Common Terminology Criteria for Adverse Effects v. 3.0. The survey asked about subjects’ opinions of the theoretical adverse event categorized as “would not be concerned,” “would be concerned, but would continue the study,” and “would discontinue the study.” Results: Patients willingness to continue the study was highest at Grade 1 with 97% of all participants. However, willingness to continue participation progressively declined as the severity of adverse events increased such that only 44% of participants would continue participation with a reported Grade 5 adverse event. Conclusions: Among surveyed GI cancer patients, willingness to continue participation in a clinical trial declined significantly as the severity of adverse events increased from Grade 1 to Grade 3 - 5 (p-value < 0.001. This could be due to multiple factors, including the terminal nature of the patients’ cancer, the side effects of study therapy and the patients’ response to study treatment. This data could produce a reasonable adverse event grade cut-off for re-consenting patients regarding new side effects. No significant financial relationships to disclose.

scholarly journals Procedural Pain and Serious Adverse Events with Bone Marrow Aspiration and Biopsy: Real World Experience from India

Blood ◽  
2020 ◽  
Vol 136 (Supplement 1) ◽  
pp. 38-38
Author(s):  
Chepsy C Philip ◽  
Inderjit Singh ◽  
Rachel Thaper ◽  
Alice David ◽  
Suvir Singh ◽  
...  
Keyword(s):  
Bone Marrow ◽  
Adverse Event ◽  
Adverse Events ◽  
Pain Perception ◽  
Bone Marrow Aspiration ◽  
Procedural Pain ◽  
Routine Practice ◽  
Serious Adverse Events ◽  
Serious Adverse Event ◽  
Factors Associated

Background: Bone Marrow aspiration and Biopsy (BMAB) is perceived by patients as a painful procedure with fearsome complications. Though informed as safe and well tolerated; there is limited data about the complications and degree of pain experienced by patients undergoing BMAB.[1] Further scarce is data from the developing world where procedural fear discourages patients from pursuing treatment and diagnosis.[2] Methods Aims: To estimate the level of pain and frequency of serious adverse eventsexperienced by patients undergoing BMAB at our center. We also attempted to identify factors associated with increased pain perception. Study setting: This study was conducted at a tertiary level teaching hospital, the Christian Medical College & Hospital, Ludhiana. Ethics approval was obtained from the Institutional research committee (CMC/1495). Study period: 01 April 2015 through 30 Nov 2019 Study Design: This is a comparative cross sectional study where comparison of those with relatively more pain to those with less was done to elicit the factors associated with pain perception. Study Population: All consecutive patients who underwent a BMAB and provided informed consent which was taken pre-procedure, were included. We excluded patients who underwent the procedure under general anesthesia. Logistics of the Study: The BMAB was performed variably by Consultant Physicians, Trainee Physicians and Physician Assistants. All patients were pre-medicated with tramadol intravenous pre-procedure, and the preferred approach was from the posterior superior iliac spine (PSIS) in a left lateral decubitus under local anesthesia with lignocain. Patients were sent home or returned to their ward after upto 60 minutes of observation. A serious adverse event was considered as one requiring a prolonged observation beyond routine practice or extending to an admission to manage adverse events following and related to the BMAB. Data sources and variables Information regarding age at diagnosis, address and sex, indication to perform the BMAB, coded as malignant and non-malignant was collected from each patient. Number of prior procedures and details regarding food intake were collected as recalled by the patient. Level of pain was noted soon after the BMAB using a combined Wong-Baker grimace with numeric pain scale by the patient themselves. Statistical Analysis: Descriptive statistics were used to characterize variables. Univariate and Multivariate Logistic Regression were used to identify factors associated with higher pain severity (Score &gt;2). Results: A total of 942 BMAB procedures were performed in this period. Baseline characteristics as tabulated below (Table1). Although the Mean + SD pain score was only 2.7 + 1.39, fourteen patients (1.48%) reported severe pain (&gt;8). The following risk-factors were associated with increased pain on multivariate analysis: those experiencing their first BMAB procedure had very low odds of pain (OR (95 % CI): 0.23 (0.15-0.37)). However, when more than one attempt of biopsy was made, the odds of pain was much higher (OR (95 % CI): 1.62 (1.29-2.05)). Food and drink intake prior to procedure was associated only at the univariate level. Those who did not take any food prior to procedure had very high odds of pain (odds ratio (OR) 1.81 (95 % CI 1.01-3.22)). However, those who took juice had very low odds (OR (95 % CI): 0.619 (0.43-0.90)). Nine (0.95%) serious adverse events were reported. There were no deaths. The major serious adverse event was hemorrhage resulting from pseudo-aneurysm of the posterior iliac artery, which comprised 2 of the 9 serious adverse events. Other serious adverse events included persistent vomiting and severe aching pain in the ipsilateral leg. Conclusions: In our analysis BMAB is associated with a low level of procedural pain and is safe. The pain perception was not influenced by the operator. Factors associated with decreased pain perception were first procedural BMAB experience and successful completion of the procedure in the first attempt. Having at least a snack or a juice pre-procedure could reduce pain perception. Serious adverse events are rare in our experience. Disclosures No relevant conflicts of interest to declare.

Real-World Safety of Emicizumab: The First Interim Analysis of the European Haemophilia Safety Surveillance (EUHASS) Database

Blood ◽  
2020 ◽  
Vol 136 (Supplement 1) ◽  
pp. 29-30
Author(s):  
Aijing Shang ◽  
Nives Selak Bienz ◽  
Ravi Gadiraju ◽  
Tiffany Chang ◽  
Peter Kuebler
Keyword(s):  
Clinical Trials ◽  
Adverse Event ◽  
Adverse Events ◽  
Real World ◽  
Event Data ◽  
Publicly Traded ◽  
Real World Data ◽  
World Data ◽  
Safety Surveillance ◽  
Adverse Event Data

Introduction: Long-term data from the HAVEN 1-4 clinical trials reaffirmed the safety and efficacy of emicizumab (HEMLIBRA®) prophylaxis in persons with hemophilia A (PwHA; Callaghan M et al. ISTH 2019 presentation). However, early data from the first Phase III trial, HAVEN 1, identified a risk for thrombotic microangiopathy (TMA) or thrombotic events (TEs) when emicizumab was used alongside activated prothrombin complex concentrate (aPCC [FEIBA]; dosed on average a cumulative amount of &gt;100 U/kg/24 hours for ≥24 hours) leading to a warning in the label and ongoing safety monitoring. European Haemophilia Safety Surveillance (EUHASS) is a large pharmacovigilance program that monitors the safety of treatments for inherited bleeding disorders. The EUHASS registry includes real-world data on the use of emicizumab in a broad, representative population of PwHA. The objective of this analysis was to summarize thrombotic, TMA, and anaphylaxis events reported to EUHASS in association with emicizumab prophylaxis. Methods: EUHASS is an investigator-led program with 86 participating centers in 27 countries, with centers reporting information on all the individuals they treat, thus minimizing selection bias. Adverse event data were collected from all PwHA in EUHASS who received emicizumab prophylaxis during 2018. EUHASS adverse event data are not collected according to Medical Dictionary for Regulatory Affairs (MedDRA) classification; however, for this exploratory analysis, events were coded at MedDRA preferred term level as far as possible; endpoints are provided as descriptive statistics. Results: Data from 148 PwHA treated with emicizumab in 2018 were included in this analysis. Concurrent treatments included recombinant activated factor VII (rFVIIa; NovoSeven®; n=23 PwHA), factor VIII, (FVIII products other than Obizur®; n=9 PwHA) and aPCC (n=1 PwHA). Two adverse events were reported in 2018 (Table 1). One event was an acute reaction (rash), reported 48 hours after dosing of a PwHA treated with emicizumab only. He recovered from the rash; the frequency was 0.7% (1/148; 95% confidence interval [CI] 0.02-3.71%). The second event was a TE-a myocardial infarction that occurred 10 hours after emicizumab dosing in a PwHA age &gt;65 years receiving emicizumab and aPCC. The frequency of TE events was calculated as 0.7% (1/148; 95% CI 0.02-3.71%). No TMA or anaphylaxis events were reported. Conclusions: Among PwHA treated with emicizumab at centers participating in EUHASS during 2018, only two adverse events were reported and there were no cases of TMA or anaphylaxis. While a full assessment is reserved for the final analysis, these interim real-world data are not inconsistent with the adverse event profile of emicizumab observed in clinical trials. No new or emerging safety signals for emicizumab were identified. However, this analysis was limited by the low number of emicizumab treated PwHA-especially in those without FVIII inhibitors, and relatively short exposure time to emicizumab. Disclosures Shang: F. Hoffmann-La Roche Ltd: Current Employment, Current equity holder in publicly-traded company, Other: All authors received support for third party writing assistance, furnished by Scott Battle, PhD, provided by F. Hoffmann-La Roche, Basel, Switzerland.. Selak Bienz:F. Hoffmann-La Roche Ltd: Current Employment. Gadiraju:I am 50% shareholder in my own private limited company (Ravi Gadiraju Pharma Ltd): Current equity holder in private company; F. Hoffmann-La Roche Ltd, Safety Scientist (Mar 19 to current): Current Employment; Britannia Pharmaceuticals, Senior PV officer (Feb 17 to Mar 19): Ended employment in the past 24 months. Chang:Genentech, Inc.: Current Employment, Current equity holder in publicly-traded company. Kuebler:Genentech, Inc.: Current Employment, Current equity holder in publicly-traded company.

Serious adverse event reporting in investigator‐initiated clinical trials

2016 ◽  
Vol 204 (6) ◽  
pp. 231-233 ◽  
Author(s):  
Sophie Wallace ◽  
Paul S Myles ◽  
Nikolajs Zeps ◽  
John R Zalcberg
Keyword(s):  
Clinical Trials ◽  
Adverse Event ◽  
Adverse Event Reporting ◽  
Event Reporting ◽  
Serious Adverse Event

scholarly journals Discovering Outliers of Potential Drug Toxicities Using a Large-scale Data-driven Approach

2016 ◽  
Vol 15 ◽  
pp. CIN.S39549
Author(s):  
Jake Luo ◽  
Ron A. Cisler
Keyword(s):  
Clinical Trial ◽  
Clinical Trials ◽  
Adverse Event ◽  
Adverse Effects ◽  
Adverse Events ◽  
Large Scale ◽  
Data Driven ◽  
Cancer Drugs ◽  
Data Driven Approach ◽  
Grubbs Test

We systematically compared the adverse effects of cancer drugs to detect event outliers across different clinical trials using a data-driven approach. Because many cancer drugs are toxic to patients, better understanding of adverse events of cancer drugs is critical for developing therapies that could minimize the toxic effects. However, due to the large variabilities of adverse events across different cancer drugs, methods to efficiently compare adverse effects across different cancer drugs are lacking. To address this challenge, we present an exploration study that integrates multiple adverse event reports from clinical trials in order to systematically compare adverse events across different cancer drugs. To demonstrate our methods, we first collected data on 186,339 clinical trials from ClinicalTrials.gov and selected 30 common cancer drugs. We identified 1602 cancer trials that studied the selected cancer drugs. Our methods effectively extracted 12,922 distinct adverse events from the clinical trial reports. Using the extracted data, we ranked all 12,922 adverse events based on their prevalence in the clinical trials, such as nausea 82%, fatigue 77%, and vomiting 75.97%. To detect the significant drug outliers that could have a statistically high possibility of causing an event, we used the boxplot method to visualize adverse event outliers across different drugs and applied Grubbs’ test to evaluate the significance. Analyses showed that by systematically integrating cross-trial data from multiple clinical trial reports, adverse event outliers associated with cancer drugs can be detected. The method was demonstrated by detecting the following four statistically significant adverse event cases: the association of the drug axitinib with hypertension (Grubbs’ test, P < 0.001), the association of the drug imatinib with muscle spasm ( P < 0.001), the association of the drug vorinostat with deep vein thrombosis ( P < 0.001), and the association of the drug afatinib with paronychia ( P < 0.01).

Sign in / Sign up

Export Citation Format

Share Document