Dataset Information

Complementing Observational Signals with Literature-Derived Distributed Representations for Post-Marketing Drug Surveillance.

ABSTRACT:

Introduction

As a result of the well documented limitations of data collected by spontaneous reporting systems (SRS), such as bias and under-reporting, a number of authors have evaluated the utility of other data sources for the purpose of pharmacovigilance, including the biomedical literature. Previous work has demonstrated the utility of literature-derived distributed representations (concept embeddings) with machine learning for the purpose of drug side-effect prediction. In terms of data sources, these methods are complementary, observing drug safety from two different perspectives (knowledge extracted from the literature and statistics from SRS data). However, the combined utility of these pharmacovigilance methods has yet to be evaluated.

Objective

This research investigates the utility of directly or indirectly combining an observational signal from SRS with literature-derived distributed representations into a single feature vector or in an ensemble approach for downstream machine learning (logistic regression).

Methods

Leveraging a recently developed representation scheme, concept embeddings were generated from relational connections extracted from the literature and composed to represent drug and associated adverse reactions, as defined by two reference standards of positive (likely causal) and negative (no causal evidence) pairs. Embeddings were presented with and without common measures of observational signal from SRS sources to logistic regressors, and performance was evaluated with the receiver operating characteristic (ROC) area under the curve (AUC) metric.

Results

ROC AUC performance with these composite models improves up to???20% over SRS-based disproportionality metrics alone and exceeds the best prior results reported in the literature when models leverage both sources of information.

Conclusions

Results from this study support the hypothesis that knowledge extracted from the literature can enhance the performance of SRS-based methods (and vice versa). Across reference sets, using literature and SRS information together performed better than using either source alone, providing strong support for the complementary nature of these approaches to post-marketing drug surveillance.

SUBMITTER: Mower J

PROVIDER: S-EPMC7243821 | biostudies-literature | 2020 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Complementing Observational Signals with Literature-Derived Distributed Representations for Post-Marketing Drug Surveillance.

Mower Justin J Cohen Trevor T Subramanian Devika D

Drug safety 20200101 1

<h4>Introduction</h4>As a result of the well documented limitations of data collected by spontaneous reporting systems (SRS), such as bias and under-reporting, a number of authors have evaluated the utility of other data sources for the purpose of pharmacovigilance, including the biomedical literature. Previous work has demonstrated the utility of literature-derived distributed representations (concept embeddings) with machine learning for the purpose of drug side-effect prediction. In terms of ...[more]

PMID: 31646442

Similar Datasets

Project description:Highlights • Active post-marketing safety surveillance on ROTASIIL, an oral rotavirus vaccine, was conducted in 9940 infants in India.• Three doses of ROTASIIL were administered at 6, 10 and 14 weeks of age, along with other EPI vaccines.• Most of the AEs were pyrexia and injection site reactions caused by the concomitant injectable vaccines.• No case of intussusception (IS) was reported.• ROTASIIL was safe and well tolerated. Background ROTASIIL, an oral live attenuated bovine-human reassortant pentavalent rotavirus vaccine, was approved in 2017. This post-marketing surveillance (PMS) was conducted to collect real-world data on the safety of ROTASIIL in India. Methods Observational, active PMS was conducted in approximately 10,000 infants aged ≥ 6 weeks. ROTASIIL was administered as a 3-dose regimen, at least 4 weeks apart, beginning at ≥ 6 weeks of age concomitantly with other Expanded Programme on Immunization (EPI) vaccines. Participants were followed for one month after the last dose. The adverse events (AEs) and serious adverse events (SAEs), including intussusception (IS) reported during the follow up period were collected. Findings A total of 9940 infants were enrolled and were considered for safety analysis. Around 9913 (99.7 %) infants received 2 doses, while 9893 (99.5 %) infants completed all three doses. Total 3693 AEs were reported in 2516 (25.3 %) participants. Most of these AEs were pyrexia (78.01 % of events) and injection-site reactions (19.14 % of events). Nearly all AEs were causally unrelated to orally administered ROTASIIL and could be caused by the concomitant injectable vaccines. Only 4 AEs (2 events of vomiting and 1 event each of discomfort and pyrexia) in 4 (<0.1 %) participants could be related to ROTASIIL. AEs were of mild or moderate severity and all resolved without any sequelae. A total of 2 SAEs (acute otitis media and skull fracture) were reported in 2 (<0.1 %) participants and were not related to ROTASIIL and recovered without sequelae. No case of IS was reported. Interpretation ROTASIIL was safe and well tolerated in this study. No safety concerns were reported. Funding The study was funded by SIIPL which is the manufacturer of the study product.

Project description:BackgroundAbiraterone acetate plus prednisolone is approved to treat patients with castration-resistant prostate cancer. This study evaluated the safety and efficacy of abiraterone acetate plus prednisolone in castration-resistant prostate cancer patients with or without previous chemotherapy in a real-world setting in Japan.MethodsThis study was an observational, prospective, post-marketing surveillance. Castration-resistant prostate cancer patients, who initiated abiraterone acetate after its approval in Japan, were enrolled. Data were collected during an observation period of 12 months and a follow-up period of another 12 months. Adverse events and adverse drug reactions were evaluated for safety. Prostate-specific antigen levels and overall survival were evaluated for efficacy.ResultsFrom 141 participating institutions, 497 patients were registered: 492 patients including 180 chemotherapy-naïve, 311 chemotherapy-experienced and one off-label-use patient received abiraterone and were evaluated for safety. Adverse events were observed in 225/492 patients (45.7%), adverse drug reactions in 131/492 patients (26.6%) and serious adverse drug reactions in 61/492 patients (12.4%). The most commonly observed adverse drug reaction was abnormal hepatic function (6.5%), followed by hypokalemia (3.0%) and decreased appetite (2.0%). At week 12, 110/432 patients (25.5%) achieved ≥50% decrease from baseline in prostate-specific antigen, and the proportion was higher in chemotherapy-naïve patients (56/161 patients; 34.8%) compared with chemotherapy-experienced patients (54/271 patients; 19.9%, P < 0.001). Survival rates at 24 months were 68.3% (295/432 patients), 73.9% (119/161 chemotherapy-naïve patients) and 64.9% (176/271 chemotherapy-experienced patients).ConclusionsThis large-scale, real-world, post-marketing surveillance study confirmed the safety and efficacy of abiraterone acetate plus prednisolone in Japanese castration-resistant prostate cancer patients with or without previous chemotherapy.

Project description:IntroductionIn 2017, a quadrivalent inactivated split-virion influenza vaccine (QIV; Vaxigrip Tetra®, Sanofi) was licensed in South Korea for active immunization against influenza A and influenza B viruses in individuals aged 3 years or older, which was subsequently extended to individuals aged 6 months or older in 2018. Post-marketing surveillance trials are mandatory in South Korea to retain drug licensure. Here, we assessed the safety of QIV in routine clinical practice in South Korea.MethodsThis was an open, multicenter, observational, active safety surveillance study conducted between 20 June 2017 and 19 June 2021 at 10 study sites in South Korea in individuals aged 3 years or older who received a single dose of QIV during a routine healthcare visit. The participants or their legally acceptable representatives were instructed to record any adverse reactions (solicited events) and unsolicited non-serious adverse events (AE) in diary cards, and notify study investigators in case of serious adverse events (SAE).ResultsOverall, 663 participants were included in this study. There were no AEs leading to study termination, and no SAEs reported. Injection site pain (278 [41.9%]) was the most frequent solicited injection site reaction, with myalgia (250 [37.7%]) and malaise (236 [35.6%]) the most frequent solicited systemic reactions. Grade 3 solicited injection site and systemic reactions were reported by 8 (1.2%) and 13 (2.0%) participants, respectively; most participants with solicited reactions recovered without the need for further action. Overall, 39 (5.9%) participants experienced 49 unsolicited non-serious AEs with the most frequently reported being nasopharyngitis (19 [2.9%]). Grade 3 unsolicited non-serious adverse events were reported in 1 (0.2%) participant. None of the unsolicited non-serious AEs were considered to be related to QIV.ConclusionThis post-marketing surveillance study confirms that QIV is well tolerated and has an acceptable safety profile in routine practice in South Korea. No unexpected safety concerns were identified.Trial registrationClinicalTrials.gov identifier NCT05406180.

Dataset Information

Complementing Observational Signals with Literature-Derived Distributed Representations for Post-Marketing Drug Surveillance.

Introduction

Objective

Methods

Results

Conclusions

Publications

Complementing Observational Signals with Literature-Derived Distributed Representations for Post-Marketing Drug Surveillance.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets