Novel data-mining methodologies for adverse drug event discovery and analysis

Signal detection of drugs associated with obstructive and central sleep apnoea

We aim to discover new safety signals of drug-induced sleep apnoea (SA), a global health problem affecting approximately 1 billion people worldwide. We first conducted a series of sequence symmetry analyses (SSA) in a cohort composed from all patients who received a first SA diagnosis or treatment between 2006 and 2018 in the Echantillon Généraliste des Bénéficaires (EGB), a random sample of the French healthcare database. We used two primary outcomes to estimate the sequence ratio (SR) for all drug classes available in France: a sensitive one (diagnosis or treatment of SA) and a specific one (Positive Airway Pressure (PAP) therapy). We then performed disproportionality analyses using the "Bayesian neural network method" on all cases of sleep apnoea (MedDRA high level term) reported up to November 2023 in the World Health Organisation (WHO) pharmacovigilance database. Among the 728,167 individuals, 46,193 had an incident diagnosis or treatment for SA and 17,080 had started an incident treatment by PAP therapy. Fifty-eight drug classes had a significant SR, with 7 considered highly plausible: opium alkaloids and derivatives, benzodiazepine derivatives, other centrally acting agents, other anxiolytics, carbamic acid esters, quinine and derivatives and antivertigo preparations; with consistent signals found for the first 3 drug classes in the disproportionality analysis. In this signal detection study, we found that opioids, benzodiazepines (but not Z-drugs) and myorelaxing agents are associated with the onset or aggravation of SA. Moreover, a new safety signal for antivertigo preparations such as betahistine emerged and needs to be further explored.

Safeguarding Patients in the AI Era: Ethics at the Forefront of Pharmacovigilance

Artificial intelligence is increasingly being used in pharmacovigilance. However, the use of artificial intelligence in pharmacovigilance raises ethical concerns related to fairness, non-discrimination, compliance, and responsibility as the central ethical principles in risk assessment and regulatory requirements. This paper explores these concerns and provides a roadmap to how to address these challenges by considering data collection, privacy protection, transparency and accountability, model training, and explainability in artificial intelligence decision making for drug safety surveillance. A number of responsible approaches have been identified including an ethics framework and best practices to enhance artificial intelligence use in healthcare. The document also recognizes some initiatives that have demonstrated the importance of ethics in artificial intelligence pharmacovigilance. Nevertheless, the major needs mentioned in this paper are transparency, accountability, data protection, and fairness, which stress the necessity of collaboration to construct a cognitive framework aimed at integrating ethical artificial intelligence into pharmacovigilance. In conclusion, innovation should be balanced with ethical responsibility to enhance public health outcomes as well as patient safety.

Long-term hepatobiliary disorder associated with trastuzumab emtansine pharmacovigilance study using the FDA Adverse Event Reporting System database

Trastuzumab emtansine (T-DM1) is widely utilized as a second-line and subsequent treatment for metastatic HER2+ breast cancer and has shown promise in early breast cancer treatment, particularly in adjuvant settings for residual disease after neoadjuvant chemotherapy. However, concerns have arisen regarding long-term hepatic adverse drug reactions (ADRs) not identified in clinical trials. We investigated potential safety signals of T-DM1 in hepatobiliary disorders and the time-to-onset of ADRs using the FDA Adverse Event Reporting System (FAERS) database. Suspected ADRs were extracted and divided into two groups: T-DM1 (N = 3387) and other drugs (N = 11,833,701). Potential signal for T-DM1 in hepatobiliary disorder were identified (reporting odds ratio [ROR] = 5.66, 95% confidence interval [CI] = 5.11-6.27; information component [IC] = 2.35, 95% Credibility Interval [Crl] = 2.18-2.51). A breast cancer indicated subgroup analysis (2519 T-DM1; 172,329 other drugs) also identified a potential safety signal (ROR = 3.28, 95% CI = 2.92-3.68; IC = 1.53, 95%CrI = 1.35-1.71). The median time-to-onset for T-DM1-associated hepatobiliary disorders was 41 days. For prolonged and chronic hepatobiliary disorders, median times were 322.5 and 301.5 days, respectively. These findings highlight the need for further research to inform clinical decisions on optimal T-DM1 treatment duration, balancing benefits with potential adverse reactions.

In-silico approaches to assessing multiple high-level drug-drug and drug-disease adverse drug effects

INTRODUCTION

Pharmacovigilance plays a pivotal role in monitoring adverse events (AEs) related to chemical substances in human/animal populations. With increasing spontaneous-reporting systems, researchers turned to in-silico approaches to efficiently analyze drug safety profiles. Here, we review in-silico methods employed for assessing multiple drug-drug/drug-disease AEs covered by comparative analyses and visualization strategies.

AREAS COVERED

Disproportionality, involving multi-stage statistical methodologies and data processing, identifies safety signals among drug-AE pairs. By stratifying data based on disease indications/demographics, researchers address confounders and assess drug safety. Comparative analyses, including clustering techniques and visualization techniques, assess drug similarities, patterns, and trends, calculate correlations, and identify distinct toxicities. Furthermore, we conducted a thorough Scopus search on 'pharmacovigilance,' yielding 5,836 publications spanning 2003 to 2023.

EXPERT OPINION

Pharmacovigilance relies on diverse data sources, presenting challenges in the integration of in-silico approaches and requiring compliance with regulations and AI adoption. Systematic use of statistical analyses enables identifications of potential risks with drugs. Frequentist and Bayesian methods are used in disproportionalities, each with its strengths and weaknesses. Integration of pharmacogenomics with pharmacovigilance enables personalized medicine, with AI further enhancing patient engagement. This multidisciplinary approach holds promise, improving drug efficacy and safety, and should be a core mission of One-Health studies.

New exploration of signal detection of Regional Risks from the perspective of data mining: a pharmacovigilance analysis based on spontaneous reporting data in Zhenjiang, China

BACKGROUND

This study aimed to adopt the conventional signal detection methods to explore a new way of risk identification and to mine important drug risks from the perspective of big data based on Zhenjiang Adverse Event Reporting System (ZAERS).

RESEARCH DESIGN AND METHODS

Data were extracted from ZAERS database between 2012 and 2022. The risks of all the reported drug event combinations were identified at the preferred term level and the standardized MedDRA query level using disproportionality analysis. Then, we conducted signal assessment according to the descriptions of drug labels.

RESULTS

In total 41,473 ADE were reported and there were 12 risky signals. Signal assessment indicates the suspected causal associations in clindamycin-taste and smell disorders, valsartan-hepatic enzyme increased and valsartan-edema peripheral; the specific manifestations of allergic reactions triggered by clindamycin, cefotaxime, cefazodime, ShexiangZhuanggu plaster, ShexiangZhuifeng plaster, and Yanhuning need to be refined in drug labels. In addition, the drug labels of NiuHuangShangQing tablet/capsule, Fuyanxiao capsule, and BiYanLing tablet should be improved.

CONCLUSIONS

In this study, we attempted a new way to find potential drug risks using small spontaneous reporting data. Our findings also suggested the need for more precise identification of allergic risks and the improvement of traditional Chinese medicine labels.

Use of Real-World Data and Machine Learning to Screen for Maternal and Paternal Characteristics Associated with Cardiac Malformations

Effective prevention of cardiac malformations, a leading cause of infant morbidity, is constrained by limited understanding of etiology. The study objective was to screen for associations between maternal and paternal characteristics and cardiac malformations. We selected 720,381 pregnancies linked to live-born infants (n=9,076 cardiac malformations) in 2011-2021 MarketScan US insurance claims data. Odds ratios were estimated with clinical diagnostic and medication codes using logistic regression. Screening of 2,000 associations selected 81 associated codes at the 5% false discovery rate. Grouping of selected codes, using latent semantic analysis and the Apriori-SD algorithm, identified elevated risk with known risk factors, including maternal diabetes and chronic hypertension. Less recognized potential signals included maternal fingolimod or azathioprine use. Signals identified might be explained by confounding, measurement error, and selection bias and warrant further investigation. The screening methods employed identified known risk factors, suggesting potential utility for identifying novel risk factors for other pregnancy outcomes.

The Reporting of a Disproportionality Analysis for Drug Safety Signal Detection Using Individual Case Safety Reports in PharmacoVigilance (READUS-PV): Development and Statement

BACKGROUND AND AIM

Disproportionality analyses using reports of suspected adverse drug reactions are the most commonly used quantitative methods for detecting safety signals in pharmacovigilance. However, their methods and results are generally poorly reported in published articles and existing guidelines do not capture the specific features of disproportionality analyses. We here describe the development of a guideline (REporting of A Disproportionality analysis for drUg Safety signal detection using individual case safety reports in PharmacoVigilance [READUS-PV]) for reporting the results of disproportionality analyses in articles and abstracts.

METHODS

We established a group of 34 international experts from universities, the pharmaceutical industry, and regulatory agencies, with expertise in pharmacovigilance, disproportionality analyses, and assessment of safety signals. We followed a three-step process to develop the checklist: (1) an open-text survey to generate a first list of items; (2) an online Delphi method to select and rephrase the most important items; (3) a final online consensus meeting.

RESULTS

Among the panel members, 33 experts responded to round 1 and 30 to round 2 of the Delphi and 25 participated to the consensus meeting. Overall, 60 recommendations for the main body of the manuscript and 13 recommendations for the abstracts were retained by participants after the Delphi method. After merging of some items together and the online consensus meeting, the READUS-PV guidelines comprise a checklist of 32 recommendations, in 14 items, for the reporting of disproportionality analyses in the main body text and four items, comprising 12 recommendations, for abstracts.

CONCLUSIONS

The READUS-PV guidelines will support authors, editors, peer-reviewers, and users of disproportionality analyses using individual case safety report databases. Adopting these guidelines will lead to more transparent, comprehensive, and accurate reporting and interpretation of disproportionality analyses, facilitating the integration with other sources of evidence.

Integrating clinical pharmacology and artificial intelligence: potential benefits, challenges, and role of clinical pharmacologists

INTRODUCTION

The integration of artificial intelligence (AI) into clinical pharmacology could be a potential approach for accelerating drug discovery and development, improving patient care, and streamlining medical research processes.

AREAS COVERED

We reviewed the current state of AI applications in clinical pharmacology, focusing on drug discovery and development, precision medicine, pharmacovigilance, and other ventures. Key AI applications in clinical pharmacology are examined, including machine learning, natural language processing, deep learning, and reinforcement learning etc. Additionally, the evolving role of clinical pharmacologists, ethical considerations, and challenges in implementing AI in clinical pharmacology are discussed.

EXPERT OPINION

The AI could be instrumental in accelerating drug discovery, predicting drug safety and efficacy, and optimizing clinical trial designs. It can play a vital role in precision medicine by helping in personalized drug dosing, treatment selection, and predicting drug response based on genetic, clinical, and environmental factors. The role of AI in pharmacovigilance, such as signal detection and adverse event prediction, is also promising. The collaboration between clinical pharmacologists and AI experts also poses certain ethical and practical challenges. Clinical pharmacologists can be instrumental in shaping the future of AI-driven clinical pharmacology and contribute to the improvement of healthcare systems.

Exploring the safety profile of tremelimumab: an analysis of the FDA adverse event reporting system

BACKGROUND

Despite the approval of tremelimumab in 2022, there is a lack of pharmacovigilance studies investigating its safety profile in real-world settings using the FDA Adverse Event Reporting System (FAERS) database.

AIM

This pharmacovigilance study aimed to comprehensively explore the adverse events (AEs) associated with tremelimumab using data mining techniques on the FAERS database.

METHOD

The study utilized data from the FAERS database, covering the period from the first quarter of 2004 to the third quarter of 2022. Disproportionality analysis, the Benjamini Hochberg adjustment method and volcano plots were used to identify and evaluate AE signals associated with tremelimumab.

RESULTS

The study uncovered 233 AE cases associated with tremelimumab. Among these cases, pyrexia (n = 39), biliary tract infection (n = 23), and sepsis (n = 21) were the three main AEs associated with tremelimumab use. The study also investigated the system organ classes associated with tremelimumab-related AEs. The top three classes were gastrointestinal disorders (17.9%), infections and infestations (16.6%), and general disorders and administration site infections (11.2%). Several AEs were identified that were not listed on the drug label of tremelimumab. These AEs included pyrexia, biliary tract infection, sepsis, dyspnea, infusion site infection, hiccup, appendicitis, hypotension, dehydration, localised oedema, presyncope, superficial thrombophlebitis and thrombotic microangiopathy.

CONCLUSION

This pharmacovigilance study identified several potential adverse events signals related to tremelimumab including some adverse events not listed on the drug label. However, further basic and clinical research studies are needed to validate these results.

Multimorbidity in Atherosclerotic Cardiovascular Disease and Its Associations With Adverse Cardiovascular Events and Healthcare Costs: A Real-World Evidence Study

Background: Atherosclerotic cardiovascular disease (ASCVD) remains the leading cause of mortality and disability in the United States and worldwide. Objective: To assess the multimorbidity burden and its associations with adverse cardiovascular events (ACE) and healthcare costs among patients with ASCVD. Methods: This is a retrospective observational cohort study using Aetna claims database. Patients with ASCVD were identified during the study period (1/1/2018-10/31/2021). The earliest ASCVD diagnosis date was identified as the index date. Qualified patients were ≥18 years of age and had ≥12 months of health plan enrollment before and after the index date. Comorbid conditions were assessed using all data available within 12 months prior to and including the index date. Association rule mining was applied to identify comorbid condition combinations. ACEs and healthcare costs were assessed using all data within 12 months after the index date. Multivariable generalized linear models were performed to examine the associations between multimorbidity and ACEs and healthcare costs. Results: Of 223 923 patients with ASCVD (mean [SD] age, 73.6 [10.7] years; 42.2% female), 98.5% had ≥2, and 80.2% had ≥5 comorbid conditions. The most common comorbid condition dyad was hypertension-hyperlipidemia (78.7%). The most common triad was hypertension-hyperlipidemia-pain disorders (61.1%). The most common quartet was hypertension-hyperlipidemia-pain disorders-diabetes (30.2%). The most common quintet was hypertension-hyperlipidemia-pain disorders-diabetes-obesity (16%). The most common sextet was hypertension-hyperlipidemia-pain disorders-diabetes-obesity-osteoarthritis (7.6%). The mean [SD] number of comorbid conditions was 7.1 [3.2]. The multimorbidity burden tended to increase in older age groups and was comparatively higher in females and in those with higher social vulnerability. The increased number of comorbid conditions was significantly associated with increased ACEs and increased healthcare costs. Discussion: Extremely prevalent multimorbidity should be considered in the context of clinical decision-making to optimize secondary prevention of ASCVD. Conclusions: Multimorbidity was extremely prevalent among patients with ASCVD. Multimorbidity patterns varied considerably across ASCVD patients and by age, gender, and social vulnerability status. Multimorbidity was strongly associated with ACEs and healthcare costs.

Clinician challenges to evidence-based prescribing for heart failure and reduced ejection fraction: A qualitative evaluation

BACKGROUND

Reasons for suboptimal prescribing for heart failure with reduced ejection fraction (HFrEF) have been identified, but it is unclear if they remain relevant with recent advances in healthcare delivery and technologies. This study aimed to identify and understand current clinician-perceived challenges to prescribing guideline-directed HFrEF medications.

METHODS

We conducted content analysis methodology, including interviews and member-checking focus groups with primary care and cardiology clinicians. Interview guides were informed by the Cabana Framework.

RESULTS

We conducted interviews with 33 clinicians (13 cardiology specialists, 22 physicians) and member checking with 10 of these. We identified four levels of challenges from the clinician perspective. Clinician level challenges included misconceptions about guideline recommendations, clinician assumptions (e.g., drug cost or affordability), and clinical inertia. Patient-clinician level challenges included misalignment of priorities and insufficient communication. Clinician-clinician level challenges were primarily between generalists and specialists, including lack of role clarity, competing priorities of providing focused versus holistic care, and contrasting confidence regarding safety of newer drugs. Policy and system/organisation level challenges included insufficient access to timely/reliable patient data, and unintended care gaps for medications without financially incentivized metrics.

CONCLUSION

This study presents current challenges faced by cardiology and primary care which can be used to strategically design interventions to improve guideline-directed care for HFrEF. The findings support the persistence of many challenges and also sheds light on new challenges. New challenges identified include conflicting perspectives between generalists and specialists, hesitancy to prescribe newer medications due to safety concerns, and unintended consequences related to value-based reimbursement metrics for select medications.

Development and Application of a Data-Driven Signal Detection Method for Surveillance of Adverse Event Variability Across Manufacturing Lots of Biologics

INTRODUCTION

Postmarketing drug safety surveillance research has focused on the product-patient interaction as the primary source of variability in clinical outcomes. However, the inherent complexity of pharmaceutical manufacturing and distribution, especially of biologic drugs, also underscores the importance of risks related to variability in manufacturing and supply chain conditions that could potentially impact clinical outcomes. We propose a data-driven signal detection method called HMMScan to monitor for manufacturing lot-dependent changes in adverse event (AE) rates, and herein apply it to a biologic drug.

METHODS

The HMMScan method chooses the best-fitting candidate from a family of probabilistic Hidden Markov Models to detect temporal correlations in per lot AE rates that could signal clinically relevant variability in manufacturing and supply chain conditions. Additionally, HMMScan indicates the particular lots most likely to be related to risky states of the manufacturing or supply chain condition. The HMMScan method was validated on extensive simulated data and applied to three actual lot sequences of a major biologic drug by combining lot metadata from the manufacturer with AE reports from the US FDA Adverse Event Reporting System (FAERS).

RESULTS

Extensive method validation on simulated data indicated that HMMScan is able to correctly detect the presence or absence of variable manufacturing and supply chain conditions for contiguous sequences of 100 lots or more when changes in these conditions have a meaningful impact on AE rates. Applying the HMMScan method to FAERS data, two of the three actual lot sequences examined exhibited evidence of potential manufacturing or supply chain-related variability.

CONCLUSIONS

HMMScan could be utilized by both manufacturers and regulators to automate lot variability monitoring and inform targeted root-cause analysis. Broad application of HMMScan would rely on a well-developed data input pipeline. The proposed method is implemented in an open-source GitHub repository.

Pulmonary hypertension associated with diazoxide: the SUR1 paradox

The ATP-sensitive potassium channels and their regulatory subunits, sulfonylurea receptor 1 (SUR1/Kir6.2) and SUR2/Kir6.1, contribute to the pathophysiology of pulmonary hypertension (PH). Loss-of-function pathogenic variants in the ABCC8 gene, which encodes for SUR1, have been associated with heritable pulmonary arterial hypertension. Conversely, activation of SUR1 and SUR2 leads to the relaxation of pulmonary arteries and reduces cell proliferation and migration. Diazoxide, a SUR1 activator, has been shown to alleviate experimental PH, suggesting its potential as a therapeutic option. However, there are paradoxical reports of diazoxide-induced PH in infants. This review explores the role of SUR1/2 in the pathophysiology of PH and the contradictory effects of diazoxide on the pulmonary vascular bed. Additionally, we conducted a comprehensive literature review of cases of diazoxide-associated PH and analysed data from the World Health Organization pharmacovigilance database (VigiBase). Significant disproportionality signals link diazoxide to PH, while no other SUR activators have been connected with pulmonary vascular disease. Diazoxide-associated PH seems to be dose-dependent and potentially related to acute effects on the pulmonary vascular bed. Further research is required to decipher the differing pulmonary vascular consequences of diazoxide in different age populations and experimental models.

Myocarditis and Pericarditis Post-mRNA COVID-19 Vaccination: Insights from a Pharmacovigilance Perspective

Concerns remain regarding the rare cardiovascular adverse events, myocarditis and pericarditis (myo/pericarditis), particularly in younger individuals following mRNA COVID-19 vaccination. Our study aimed to comprehensively assess potential safety signals related to these cardiac events following the primary and booster doses, with a specific focus on younger populations, including children as young as 6 months of age. Using the Vaccine Adverse Events Reporting System (VAERS), the United States national passive surveillance system, we conducted a retrospective pharmacovigilance study analyzing spontaneous reports of myo/pericarditis. We employed both frequentist and Bayesian methods and conducted subgroup analyses by age, sex, and vaccine dose. We observed a higher reporting rate of myo/pericarditis following the primary vaccine series, particularly in males and mainly after the second dose. However, booster doses demonstrated a lower number of reported cases, with no significant signals detected after the fourth or fifth doses. In children and young adults, we observed notable age and sex differences in the reporting of myo/pericarditis cases. Males in the 12-17 and 18-24-year-old age groups had the highest number of cases, with significant signals for both males and females after the second dose. We also identified an increased reporting for a spectrum of cardiovascular symptoms such as chest pain and dyspnea, which increased with age, and were reported more frequently than myo/pericarditis. The present study identified signals of myo/pericarditis and related cardiovascular symptoms after mRNA COVID-19 vaccination, especially among children and adolescents. These findings underline the importance for continued vaccine surveillance and the need for further studies to confirm these results and to determine their clinical implications in public health decision-making, especially for younger populations.

The association between prescription drugs and colorectal cancer prognosis: a nationwide cohort study using a medication-wide association study

BACKGROUND

With the availability of health insurance claim data, pharmacovigilance for various drugs has been suggested; however, it is necessary to establish an appropriate analysis method. To detect unintended drug effects and to generate new hypotheses, we conducted a hypothesis-free study to systematically examine the relationship between all prescription nonanticancer drugs and the mortality of colorectal cancer patients.

METHODS

We used the Korean National Health Insurance Service-National Sample Cohort database. A total of 2,618 colorectal cancer patients diagnosed between 2004 and 2015 were divided into drug discovery and drug validation sets (1:1) through random sampling. Drugs were classified using the Anatomical Therapeutic Chemical (ATC) classification system: 76 drugs classified as ATC level 2 and 332 drugs classified as ATC level 4 were included in the analysis. We used a Cox proportional hazard model adjusted for sex, age, colorectal cancer treatment, and comorbidities. The relationship between all prescription nonanticancer drugs and the mortality of colorectal cancer patients was analyzed, controlling for multiple comparisons with the false discovery rate.

RESULTS

We found that one ATC level-2 drug (drugs that act on the nervous system, including parasympathomimetics, addictive disorder drugs, and antivertigo drugs) showed a protective effect related to colorectal cancer prognosis. At the ATC level 4 classification, 4 drugs were significant: two had a protective effect (anticholinesterases and opioid anesthetics), and the other two had a detrimental effect (magnesium compounds and Pregnen [4] derivatives).

CONCLUSIONS

In this hypothesis-free study, we identified four drugs linked to colorectal cancer prognosis. The MWAS method can be useful in real-world data analysis.

An Early Adverse Drug Event Detection Approach with False Discovery Rate Control

Adverse drug event (ADE) is a significant challenge in clinical practice. Many ADEs have not been identified timely after the approval of the corresponding drugs. Despite the use of drug similarity network demonstrates early success on improving ADE detection, false discovery rate (FDR) control remains unclear in its application. Additionally, performance of early ADE detection has not been explicitly investigated under the time-to-event framework. In this manuscript, we propose to use the drug similarity based posterior probability of null hypothesis for early ADE detection. The proposed approach is also able to control FDR for monitoring a large number of ADEs of multiple drugs. The proposed approach outperforms existing approaches on mining labeled ADEs in the US FDA's Adverse Event Reporting System (FAERS) data, especially in the first few years after the drug initial reporting time. Additionally, the proposed approach is able to identify more labeled ADEs and has significantly lower time to ADE detection. In simulation study, the proposed approach demonstrates proper FDR control, as well as has better true positive rate and an excellent true negative rate. In our exemplified FAERS analysis, the proposed approach detects new ADE signals and identifies ADE signals in a timelier fashion than existing approach. In conclusion, the proposed approach is able to both reduce the time and improve the FDR control for ADE detection.

Sample Bias in Web-Based Patient-Generated Health Data of Dutch Patients With Gastrointestinal Stromal Tumor: Survey Study

BACKGROUND

Increasingly, social media is being recognized as a potential resource for patient-generated health data, for example, for pharmacovigilance. Although the representativeness of the web-based patient population is often noted as a concern, studies in this field are limited.

OBJECTIVE

This study aimed to investigate the sample bias of patient-centered social media in Dutch patients with gastrointestinal stromal tumor (GIST).

METHODS

A population-based survey was conducted in the Netherlands among 328 patients with GIST diagnosed 2-13 years ago to investigate their digital communication use with fellow patients. A logistic regression analysis was used to analyze clinical and demographic differences between forum users and nonusers.

RESULTS

Overall, 17.9% (59/328) of survey respondents reported having contact with fellow patients via social media. Moreover, 78% (46/59) of forum users made use of GIST patient forums. We found no statistically significant differences for age, sex, socioeconomic status, and time since diagnosis between forum users (n=46) and nonusers (n=273). Patient forum users did differ significantly in (self-reported) treatment phase from nonusers (P=.001). Of the 46 forum users, only 2 (4%) were cured and not being monitored; 3 (7%) were on adjuvant, curative treatment; 19 (41%) were being monitored after adjuvant treatment; and 22 (48%) were on palliative treatment. In contrast, of the 273 patients who did not use disease-specific forums to communicate with fellow patients, 56 (20.5%) were cured and not being monitored, 31 (11.3%) were on curative treatment, 139 (50.9%) were being monitored after treatment, and 42 (15.3%) were on palliative treatment. The odds of being on a patient forum were 2.8 times as high for a patient who is being monitored compared with a patient that is considered cured. The odds of being on a patient forum were 1.9 times as high for patients who were on curative (adjuvant) treatment and 10 times as high for patients who were in the palliative phase compared with patients who were considered cured. Forum users also reported a lower level of social functioning (84.8 out of 100) than nonusers (93.8 out of 100; P=.008).

CONCLUSIONS

Forum users showed no particular bias on the most important demographic variables of age, sex, socioeconomic status, and time since diagnosis. This may reflect the narrowing digital divide. Overrepresentation and underrepresentation of patients with GIST in different treatment phases on social media should be taken into account when sourcing patient forums for patient-generated health data. A further investigation of the sample bias in other web-based patient populations is warranted.

Identifying new drugs associated with pulmonary arterial hypertension: A WHO pharmacovigilance database disproportionality analysis

Since the 1960s, several drugs have been linked to the onset or aggravation of pulmonary arterial hypertension (PAH): dasatinib, some amphetamine-like appetite suppressants (aminorex, fenfluramine, dexfenfluramine, benfluorex) and recreational drugs (methamphetamine). Moreover, in numerous cases, the implication of other drugs with PAH have been suggested, but the precise identification of iatrogenic aetiologies of PAH is challenging given the scarcity of this disease and the potential long latency period between drug intake and PAH onset. In this context, we used the World Health Organization's pharmacovigilance database, VigiBase, to generate new hypotheses about drug associated PAH.

METHODS

We used VigiBase, the largest pharmacovigilance database worldwide to generate disproportionality signals through the Bayesian neural network method. All disproportionality signals were further independently reviewed by experts in pulmonary arterial hypertension, pharmacovigilance and vascular pharmacology and their plausibility ranked according to World Health Organization causality categories.

RESULTS

We included 2184 idiopathic PAH cases, yielding a total of 93 disproportionality signals. Among them, 25 signals were considered very likely, 15 probable, 28 possible and 25 unlikely. Notably, we identified 4 new protein kinases inhibitors (lapatinib, lorlatinib, ponatinib and ruxolitinib), 1 angiogenesis inhibitor (bevacizumab), and several chemotherapeutics (etoposide, trastuzumab), antimetabolites (cytarabine, fludarabine, fluorouracil, gemcitabine) and immunosuppressants (leflunomide, thalidomide, ciclosporin).

CONCLUSION

Such signals represent plausible adverse drug reactions considering the knowledge of iatrogenic PAH, the drugs' biological and pharmacological activity and the characteristics of the reported case. Although confirmatory studies need to be performed, the signals identified may help clinicians envisage an iatrogenic aetiology when faced with a patient who develops PAH.

Prevalence of multimorbidity combinations and their association with medical costs and poor health: A population-based study of U.S. adults

Background

Multimorbidity is common, but the prevalence and burden of the specific combinations of coexisting disease has not been systematically examined in the general U.S. adult population.

Objective

To identify and estimate the burden of highly prevalent combinations of chronic conditions that are treated among one million or more adults in the United States.

Methods

Cross-sectional analysis of U.S. households in the Medical Expenditure Panel Survey (MEPS), 2016-2019, a large nationally-representative sample of the community-dwelling population. Association rule mining was used to identify the most common combinations of 20 chronic conditions that have high relevance, impact, and prevalence in primary care. The main measures and outcomes were annual treated prevalence, total medical expenditures, and perceived poor health. Logistic regression models with poor health as the outcome and each multimorbidity combination as the exposure were used to calculate adjusted odds ratios and 95% confidence intervals.

Results

Frequent pattern mining yielded 223 unique combinations of chronic disease, including 74 two-way (dyad), 115 three-way (triad), and 34 four-way combinations that are treated in one million or more U.S. adults. Hypertension-hyperlipidemia was the most common two-way combination occurring in 30.8 million adults. The combination of diabetes-arthritis-cardiovascular disease was associated with the highest median annual medical expenditures ($23,850, interquartile range: $11,593-$44,616), and the combination of diabetes-arthritis-asthma/COPD had the highest age-race-sex adjusted odds ratio of poor self-rated health (adjusted odd ratio: 6.9, 95%CI: 5.4-8.8).

Conclusion

This study demonstrates that many multimorbidity combinations are highly prevalent among U.S. adults, yet most research and practice-guidelines remain single disease focused. Highly prevalent and burdensome multimorbidity combinations could be prioritized for evidence-based research on optimal prevention and treatment strategies.

Changing paradigms in detecting rare adverse drug reactions: from disproportionality analysis, old and new, to machine learning

PLAIN LANGUAGE SUMMARYYour physician, pharmacist, nurse, or even you can voluntarily report suspected adverse events associated with drugs. The FDA Adverse Reporting System (FAERS) and the WHO Vigibase are large databases that store individual reports of adverse drug reactions (ADRs). While some ADRs are very common, others are seen rarely. Detecting rare and very rare ADRs is extremely difficult but very important for the safe use of drugs. Databases such as FAERS and WHO Vigibase contain a large amount of data and are commonly used for analysis applying a statistical method called disproportionately analysis. This type of analysis determines whether there is a higher-than-expected number of adverse reactions for a particular drug. In the future, machine learning will complement this process by applying algorithms to the data, constructing and refining rules of inference, and building predictive models of ADRs. This paradigm change in testing for ADRs is expected to provide a better understanding of the factors impacting drug safety.

Artificial Intelligent Context-Aware Machine-Learning Tool to Detect Adverse Drug Events from Social Media Platforms

INTRODUCTION

Pharmacovigilance (PV) has proven to detect post-marketing adverse drug events (ADE). Previous research used the natural language processing (NLP) tool to extract unstructured texts relevant to ADEs. However, texts without context reduce the efficiency of such algorithms. Our objective was to develop and validate an innovative NLP tool, aTarantula, using a context-aware machine-learning algorithm to detect existing ADEs from social media using an aggregated lexicon.

METHOD

aTarantula utilized FastText embeddings and an aggregated lexicon to extract contextual data from three patient forums (i.e., MedHelp, MedsChat, and PatientInfo) taking warfarin. The lexicon used warfarin package inserts and synonyms of warfarin ADEs from UMLS and FAERS databases. Data was stored on SQLite and then refined and manually checked by three clinical pharmacists for validation.

RESULTS

Multiple organ systems where the most frequent ADE were reported at 1.50%, followed by CNS side effects at 1.19%. Lymphatic system ADEs were the least common side effect reported at 0.09%. The overall Spearman rank correlation coefficient between patient-reported data from the forums and FAERS was 0.19. As determined by pharmacist validation, aTarantula had a sensitivity of 84.2% and a specificity of 98%. Three clinical pharmacists manually validated our results. Finally, we created an aggregated lexicon for mining ADEs from social media.

CONCLUSION

We successfully developed aTarantula, a machine-learning algorithmn based on artificial intelligence to extract warfarin-related ADEs from online social discussion forums automatically. Our study shows that it is feasible to use aTarantula to detect ADEs. Future researchers can validate aTarantula on the diverse dataset.

Potential safety signals for antibacterial agents from the Brazilian national pharmacovigilance database (Vigimed/VigiFlow)

Antibacterial drugs are a widely used drug class due to the frequency of infectious diseases globally. Risks knowledge should ground these medicines' selection. Data mining in large databases is essential to identify early safety signals and to support pharmacovigilance systems. We conducted a cross-sectional study to assess adverse drug events related to antibiotics reporting between December 2018 and December 2021 in the Brazilian database (Vigimed/VigiFlow). We used the Reporting Odds Ratio (ROR) disproportionality analysis method to identify disproportionate reporting signals (SDR), referring to statistical combinations between drugs and adverse events. Vancomycin was the most reported antibiotic (n = 1,733), followed by ceftriaxone (n = 1,277) and piperacillin and tazobactam (n = 1,024). We detected 294 safety signals related to antibacterials. We identified azithromycin leading in the number of safety signals (n = 49), followed by polymyxin B (n = 25). Of these, 95 were not provided for in the drug label and had little or no reports in the medical literature. Three serious events are associated with ceftazidime and avibactam, a new drug in the Brazilian market. We also found suicide attempts as a sign associated with amoxicillin/clavulanate. Gait disturbance, a worrying event, especially in the elderly, was associated with azithromycin. Our findings may help guide further pharmacoepidemiologic studies and monitoring safety signals in pharmacovigilance.

Develop an ADR prediction system of Chinese herbal injections containing Panax notoginseng saponin: a nested case-control study using machine learning

OBJECTIVE

This study aimed to develop an adverse drug reactions (ADR) antecedent prediction system using machine learning algorithms to provide the reference for security usage of Chinese herbal injections containing Panax notoginseng saponin in clinical practice.

DESIGN

A nested case-control study.

SETTING

National Center for ADR Monitoring and the Electronic Medical Record (EMR) system.

PARTICIPANTS

All patients were from five medical institutions in Sichuan Province from January 2010 to December 2018.

MAIN OUTCOMES/MEASURES

Data of patients with ADR who used Chinese herbal injections containing Panax notoginseng saponin were collected from the National Center for ADR Monitoring. A nested case-control study was used to randomly match patients without ADR from the EMR system by the ratio of 1:4. Eighteen machine learning algorithms were applied for the development of ADR prediction models. Area under curve (AUC), accuracy, precision, recall rate and F1 value were used to evaluate the predictive performance of the model. An ADR prediction system was established by the best model selected from the 1080 models.

RESULTS

A total of 530 patients from five medical institutions were included, and 1080 ADR prediction models were developed. Among these models, the AUC of the best capable one was 0.9141 and the accuracy was 0.8947. According to the best model, a prediction system, which can provide early identification of patients at risk for the ADR of Panax notoginseng saponin, has been established.

CONCLUSION

The prediction system developed based on the machine learning model in this study had good predictive performance and potential clinical application.

Detection of statin-induced rhabdomyolysis and muscular related adverse events through data mining technique

BACKGROUND AND OBJECTIVE

Rhabdomyolysis (RM) is a life-threatening adverse drug reaction in which statins are the one commonly related to RM. The study aimed to explore the association between statin used and RM or other muscular related adverse events. In addition, drug interaction with statins were also assessed.

METHODS

All extracted prescriptions were grouped as lipophilic and hydrophilic statins. RM outcome was identified by electronically screening and later ascertaining by chart review. The study proposed 4 models, i.e., logistic regression (LR), Bayesian network (BN), random forests (RF), and extreme gradient boosting (XGBoost). Features were selected using multiple processes, i.e., bootstrapping, expert opinions, and univariate analysis.

RESULTS

A total of 939 patients who used statins were identified consisting 15, 9, and 19 per 10,000 persons for overall outcome prevalence, using statin alone, and co-administrations, respectively. Common statins were simvastatin, atorvastatin, and rosuvastatin. The proposed models had high sensitivity, i.e., 0.85, 0.90, 0.95 and 0.95 for LR, BN, RF, and XGBoost, respectively. The area under the receiver operating characteristic was significantly higher in LR than BN, i.e., 0.80 (0.79, 0.81) and 0.73 (0.72, 0.74), but a little lower than the RF [0.817 (95% CI 0.811, 0.824)] and XGBoost [0.819 (95% CI 0.812, 0.825)]. The LR model indicated that a combination of high-dose lipophilic statin, clarithromycin, and antifungals was 16.22 (1.78, 148.23) times higher odds of RM than taking high-dose lipophilic statin alone.

CONCLUSIONS

The study suggested that statin uses may have drug interactions with others including clarithromycin and antifungal drugs in inducing RM. A prospective evaluation of the model should be further assessed with well planned data monitoring. Applying LR in hospital system might be useful in warning drug interaction during prescribing.

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.

Hypnotics and infections: disproportionality analysis of the U.S. Food & Drug Administration adverse event reporting system database

STUDY OBJECTIVES

There is no consensus information on infections associated with nonbenzodiazepines. Knowledge about infections related to newly marketed hypnotics (orexin receptor antagonists and melatonin receptor agonists) is scarce. The study aimed to detect infection signals for nonbenzodiazepines, orexin receptor antagonists, and melatonin receptor agonists by analyzing data from the U.S. Food & Drug Administration adverse event reporting system.

METHODS

A disproportionality analysis was performed to quantitatively detect infection signals for hypnotics by calculating the reporting odds ratio and the 95% confidence interval. Data registered in the U.S. Food & Drug Administration adverse event reporting system from 2010-2020 were retrieved.

RESULTS

A total of 3,092 patients with infection were extracted for the 3 classes of hypnotic drugs. Nonbenzodiazepines were associated with a higher disproportionality of infections (reporting odds ratio: 1.10; 95% confidence interval, 1.06-1.14). The association of infections was not present for melatonin receptor agonists (reporting odds ratio: 0.86; 95% confidence interval, 0.74-1.00) and orexin receptor antagonists (reporting odds ratio: 0.19; 95% confidence interval, 0.15-0.25). Significant reporting associations were identified for nonbenzodiazepines concerning the categories of bone and joint infections, dental and oral soft tissue infections, upper respiratory tract infections, and urinary tract infections.

CONCLUSIONS

Nonbenzodiazepines had a positive signal for infections, while orexin receptor antagonists and melatonin receptor agonists had a negative signal. More research needs to be conducted to confirm this relationship.

CITATION

Meng L, Huang J, He Q, et al. Hypnotics and infections: disproportionality analysis of the U.S. Food & Drug Administration adverse event reporting system database. J Clin Sleep Med. 2022;18(9):2229-2235.

Generalised weibull model-based approaches to detect non-constant hazard to signal adverse drug reactions in longitudinal data

Pharmacovigilance is the process of monitoring the emergence of harm from a medicine once it has been licensed and is in use. The aim is to identify new adverse drug reactions (ADRs) or changes in frequency of known ADRs. The last decade has seen increased interest for the use of electronic health records (EHRs) in pharmacovigilance. The causal mechanism of an ADR will often result in the occurrence being time dependent. We propose identifying signals for ADRs based on detecting a variation in hazard of an event using a time-to-event approach. Cornelius et al. proposed a method based on the Weibull Shape Parameter (WSP) and demonstrated this to have optimal performance for ADRs occurring shortly after taking treatment or delayed ADRs, and introduced censoring at varying time points to increase performance for intermediate ADRs. We now propose two new approaches which combined perform equally well across all time periods. The performance of this new approach is illustrated through an EHR Bisphosphonates dataset and a simulation study. One new approach is based on the power generalised Weibull distribution (pWSP) introduced by Bagdonavicius and Nikulin alongside an extended version of the WSP test, which includes one censored dataset resulting in improved detection across time period (dWSP). In the Bisphosphonates example, the pWSP and dWSP tests correctly signalled two known ADRs, and signal one adverse event for which no evidence of association with the drug exist. A combined test involving both pWSP and dWSP is reliable independently of the time of occurrence of ADRs.

Comparison of Pretraining Models and Strategies for Health-Related Social Media Text Classification

Pretrained contextual language models proposed in the recent past have been reported to achieve state-of-the-art performances in many natural language processing (NLP) tasks, including those involving health-related social media data. We sought to evaluate the effectiveness of different pretrained transformer-based models for social media-based health-related text classification tasks. An additional objective was to explore and propose effective pretraining strategies to improve machine learning performance on such datasets and tasks. We benchmarked six transformer-based models that were pretrained with texts from different domains and sources—BERT, RoBERTa, BERTweet, TwitterBERT, BioClinical_BERT, and BioBERT—on 22 social media-based health-related text classification tasks. For the top-performing models, we explored the possibility of further boosting performance by comparing several pretraining strategies: domain-adaptive pretraining (DAPT), source-adaptive pretraining (SAPT), and a novel approach called topic specific pretraining (TSPT). We also attempted to interpret the impacts of distinct pretraining strategies by visualizing document-level embeddings at different stages of the training process. RoBERTa outperformed BERTweet on most tasks, and better than others. BERT, TwitterBERT, BioClinical_BERT and BioBERT consistently underperformed. For pretraining strategies, SAPT performed better or comparable to the off-the-shelf models, and significantly outperformed DAPT. SAPT + TSPT showed consistently high performance, with statistically significant improvement in three tasks. Our findings demonstrate that RoBERTa and BERTweet are excellent off-the-shelf models for health-related social media text classification, and extended pretraining using SAPT and TSPT can further improve performance.

Detecting drug-drug interactions between therapies for COVID-19 and concomitant medications through the FDA adverse event reporting system

Background: COVID-19 patients with underlying medical conditions are vulnerable to drug-drug interactions (DDI) due to the use of multiple medications. We conducted a discovery-driven data analysis to identify potential DDIs and associated adverse events (AEs) in COVID-19 patients from the FDA Adverse Event Reporting System (FAERS), a source of post-market drug safety.

Materials and Methods: We investigated 18,589 COVID-19 AEs reported in the FAERS database between 2020 and 2021. We applied multivariate logistic regression to account for potential confounding factors, including age, gender, and the number of unique drug exposures. The significance of the DDIs was determined using both additive and multiplicative measures of interaction. We compared our findings with the Liverpool database and conducted a Monte Carlo simulation to validate the identified DDIs.

Results: Out of 11,337 COVID-19 drug-Co-medication-AE combinations investigated, our methods identified 424 signals statistically significant, covering 176 drug-drug pairs, composed of 13 COVID-19 drugs and 60 co-medications. Out of the 176 drug-drug pairs, 20 were found to exist in the Liverpool database. The empirical p-value obtained based on 1,000 Monte Carlo simulations was less than 0.001. Remdesivir was discovered to interact with the largest number of concomitant drugs (41). Hydroxychloroquine was detected to be associated with most AEs (39). Furthermore, we identified 323 gender- and 254 age-specific DDI signals.

Conclusion: The results, particularly those not found in the Liverpool database, suggest a subsequent need for further pharmacoepidemiology and/or pharmacology studies.

Signaling COVID-19 Vaccine Adverse Events

INTRODUCTION

Statistical signal detection is a crucial tool for rapidly identifying potential risks associated with pharmaceutical products. The unprecedented environment created by the coronavirus disease 2019 (COVID-19) pandemic for vaccine surveillance predisposes commonly applied signal detection methodologies to a statistical issue called the masking effect, in which signals for a vaccine of interest are hidden by the presence of other reported vaccines. This masking effect may in turn limit or delay our understanding of the risks associated with new and established vaccines.

OBJECTIVE

The aim is to investigate the problem of masking in the context of COVID-19 vaccine signal detection, assessing its impact, extent, and root causes.

METHODS

Based on data underlying the Vaccine Adverse Event Reporting System, three commonly applied statistical signal detection methodologies, and a more advanced regression-based methodology, we investigate the temporal evolution of signals corresponding to five largely recognized adverse events and two potentially new adverse events.

RESULTS

The results demonstrate that signals of adverse events related to COVID-19 vaccines may be undetected or delayed due to masking when generated by methodologies currently utilized by pharmacovigilance organizations, and that a class of advanced methodologies can partially alleviate the problem. The results indicate that while masking is rare relative to all possible statistical associations, it is much more likely to occur in COVID-19 vaccine signaling, and that its extent, direction, impact, and roots are not static, but rather changing in accordance with the changing nature of data.

CONCLUSIONS

Masking is an addressable problem that merits careful consideration, especially in situations such as COVID-19 vaccine safety surveillance and other emergency use authorization products.

A Comprehensive Review of Computational Methods For Drug-Drug Interaction Detection

The detection of drug-drug interactions (DDIs) is a crucial task for drug safety surveillance, which provides effective and safe co-prescriptions of multiple drugs. Since laboratory researches are often complicated, costly and time-consuming, it's urgent to develop computational approaches to detect drug-drug interactions. In this paper, we conduct a comprehensive review of state-of-the-art computational methods falling into three categories: literature-based extraction methods, machine learning-based prediction methods and pharmacovigilance-based data mining methods. Literature-based extraction methods detect DDIs from published literature using natural language processing techniques; machine learning-based prediction methods build prediction models based on the known DDIs in databases and predict novel ones; pharmacovigilance-based data mining methods usually apply statistical techniques on various electronic data to detect drug-drug interaction signals. We first present the taxonomy of drug-drug interaction detection methods and provide the outlines of three categories of methods. Afterwards, we respectively introduce research backgrounds and data sources of three categories, and illustrate their representative approaches as well as evaluation metrics. Finally, we discuss the current challenges of existing methods and highlight potential opportunities for future directions.

Automated gathering of real-world data from online patient forums can complement pharmacovigilance for rare cancers

Current methods of pharmacovigilance result in severe under-reporting of adverse drug events (ADEs). Patient forums have the potential to complement current pharmacovigilance practices by providing real-time uncensored and unsolicited information. We are the first to explore the value of patient forums for rare cancers. To this end, we conduct a case study on a patient forum for Gastrointestinal Stromal Tumor patients. We have developed machine learning algorithms to automatically extract and aggregate side effects from messages on open online discussion forums. We show that patient forum data can provide suggestions for which ADEs impact quality of life the most: For many side effects the relative reporting rate differs decidedly from that of the registration trials, including for example cognitive impairment and alopecia as side effects of avapritinib. We also show that our methods can provide real-world data for long-term ADEs, such as osteoporosis and tremors for imatinib, and novel ADEs not found in registration trials, such as dry eyes and muscle cramping for imatinib. We thus posit that automated pharmacovigilance from patient forums can provide real-world data for ADEs and should be employed as input for medical hypotheses for rare cancers.

A data-driven pipeline to extract potential adverse drug reactions through prescription, procedures and medical diagnoses analysis: application to a cohort study of 2,010 patients taking hydroxychloroquine with an 11-year follow-up

Context

Real-life data consist of exhaustive data which are not subject to selection bias. These data enable to study drug-safety profiles but are underused because of their temporality, necessitating complex models (i.e., safety depends on the dose, timing, and duration of treatment). We aimed to create a data-driven pipeline strategy that manages the complex temporality of real-life data to highlight the safety profile of a given drug.

Methods

We proposed to apply the weighted cumulative exposure (WCE) statistical model to all health events occurring after a drug introduction (in this paper HCQ) and performed bootstrap to select relevant diagnoses, drugs and interventions which could reflect an adverse drug reactions (ADRs). We applied this data-driven pipeline on a French national medico-administrative database to extract the safety profile of hydroxychloroquine (HCQ) from a cohort of 2,010 patients.

Results

The proposed method selected eight drugs (metopimazine, anethole trithione, tropicamide, alendronic acid & colecalciferol, hydrocortisone, chlormadinone, valsartan and tixocortol), twelve procedures (six ophthalmic procedures, two dental procedures, two skin lesions procedures and osteodensitometry procedure) and two medical diagnoses (systemic lupus erythematous, unspecified and discoid lupus erythematous) to be significantly associated with HCQ exposure.

Conclusion

We provide a method extracting the broad spectrum of diagnoses, drugs and interventions associated to any given drug, potentially highlighting ADRs. Applied to hydroxychloroquine, this method extracted among others already known ADRs.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12874-022-01628-3.

• The challenge of drug-safety signal detection methods is to handle four types of difficulties:

○ The data source, the study of long-term adverse drug reactions or effects not suspected by healthcare professionals, requires the use of a real-life data source.

○ The consideration of a broad spectrum of potential adverse drug reactions (ADRs), and not only candidate ADRs.

○ The temporal impact (meaning that safety depends on the dose, date and duration of treatment).

○ The difference between true ADRs and disease natural course.

• We aimed to create a data-driven pipeline strategy, without any assumption of any ADRs, which take into account the complex temporality of real-life data to provide the safety profile of a given drug.

• Our pipeline used three sources of real-life data to establish a safety profile of a given drug: drug prescriptions, procedures and medical diagnoses.

• We successfully applied our data-driven pipeline strategy to hydroxychloroquine (HCQ). Our pipeline enabled us to find diagnoses, drugs and interventions related to HCQ and which could reflect an ADR due to HCQ or the disease itself.

• This data-driven pipeline strategy may be of interest to other experts involved in the pharmacovigilance discipline.

Machine Learning in Causal Inference: Application in Pharmacovigilance

Monitoring adverse drug events or pharmacovigilance has been promoted by the World Health Organization to assure the safety of medicines through a timely and reliable information exchange regarding drug safety issues. We aim to discuss the application of machine learning methods as well as causal inference paradigms in pharmacovigilance. We first reviewed data sources for pharmacovigilance. Then, we examined traditional causal inference paradigms, their applications in pharmacovigilance, and how machine learning methods and causal inference paradigms were integrated to enhance the performance of traditional causal inference paradigms. Finally, we summarized issues with currently mainstream correlation-based machine learning models and how the machine learning community has tried to address these issues by incorporating causal inference paradigms. Our literature search revealed that most existing data sources and tasks for pharmacovigilance were not designed for causal inference. Additionally, pharmacovigilance was lagging in adopting machine learning-causal inference integrated models. We highlight several currently trending directions or gaps to integrate causal inference with machine learning in pharmacovigilance research. Finally, our literature search revealed that the adoption of causal paradigms can mitigate known issues with machine learning models. We foresee that the pharmacovigilance domain can benefit from the progress in the machine learning field.

A smart hospital-driven approach to precision pharmacovigilance

Researchers, regulatory agencies, and the pharmaceutical industry are moving towards precision pharmacovigilance as a comprehensive framework for drug safety assessment, at the service of the individual patient, by clustering specific risk groups in different databases. This article explores its implementation by focusing on: (i) designing a new data collection infrastructure, (ii) exploring new computational methods suitable for drug safety data, and (iii) providing a computer-aided framework for distributed clinical decisions with the aim of compiling a personalized information leaflet with specific reference to a drug's risks and adverse drug reactions. These goals can be achieved by using 'smart hospitals' as the principal data sources and by employing methods of precision medicine and medical statistics to supplement current public health decisions.

Safety Profile of Antipsychotic Drugs: Analysis Based on a Provincial Spontaneous Reporting Systems Database

Introduction: Antipsychotic drugs are the main therapy for schizophrenia and have been widely used in mental disorder fields. However, the research on the safety of antipsychotic drugs in the real-world is rare. The purpose of this research is to evaluate the safety of antipsychotic drugs based on real-world data. Methods: ADR reports collected by the Henan Adverse Drug Reaction Monitoring Center from 2016 to 2020 were analyzed. We described the safety of antipsychotic drugs by descriptive analysis and four signal mining methods. Meanwhile, the risk factors for serious adverse reactions of antipsychotics were identified. Results: A total of 3363 ADR reports related to antipsychotics were included. We found that the number of adverse drug reaction reports and the proportion of serious adverse reactions have increased year by year from 2016 to 2020. Most adverse drug reactions occurred within 3 months after taking the medicine. The symptoms caused by typical antipsychotics and atypical antipsychotics were different and dyskinesia was more common in typical antipsychotics. Most patients improved or recovered after treatment or intervention while only one patient had sequelae. Low-level hospitals, psychiatric hospitals, youth, and old age could increase the risk of serious adverse reactions. Four off-label signals were found through signal mining, including amisulpride-pollakiuria, ziprasidone-dyspnoea, quetiapine-urinary incontinence, olanzapine-hepatic function abnormal. Conclusion: We found that most ADRs occurred within 3 months after taking the medicine, so close observation was required for patients during the first 3 months of treatment. The ADRs of antipsychotics involved multiple organ-system damages but were not serious. It might be recommended to take alternative drugs after a serious ADR occurred. The symptoms caused by typical APDs and atypical APDs were different. For patients with typical APDs, dyskinesia was more common and should be given special attention. Statistics showed that low-level hospitals, psychiatric hospitals, youth, and old age were risk factors for serious ADRs. The four off-label signals obtained by signal mining should be paid special attention, including amisulpride-pollakiuria, ziprasidone-dyspnoea, quetiapine-urinary incontinence, and olanzapine-hepatic function abnormal.

Assessing the effect of a regional integrated care model over ten years using quality indicators based on claims data - the basic statistical methodology of the INTEGRAL project

BACKGROUND

The regional integrated health care model "Healthy Kinzigtal" started in 2006 with the goal of optimizing health care and economic efficiency. The INTEGRAL project aimed at evaluating the effect of this model on the quality of care over the first 10 years.

METHODS

This methodological protocol supplements the study protocol and the main publication of the project. Comparing quality indicators based on claims data between the intervention region and 13 structurally similar control regions constitutes the basic scientific approach. Methodological key issues in performing such a comparison are identified and solutions are presented.

RESULTS

A key step in the analysis is the assessment of a potential trend in prevalence for a single quality indicator over time in the intervention region compared to the corresponding trends in the control regions. This step has to take into account that there may be a common - not necessarily linear - trend in the indicator over time and that trends can also appear by chance. Conceptual and statistical approaches were developed to handle this key step and to assess in addition the overall evidence for an intervention effect across all indicators. The methodology can be extended in several directions of interest.

CONCLUSIONS

We believe that our approach can handle the major statistical challenges: population differences are addressed by standardization; we offer transparency with respect to the derivation of the key figures; global time trends and structural changes do not invalidate the analyses; the regional variation in time trends is taken into account. Overall, the project demanded substantial efforts to ensure adequateness, validity and transparency.

High-Dimensional Propensity Score-Adjusted Case-Crossover for Discovering Adverse Drug Reactions from Computerized Administrative Healthcare Databases

INTRODUCTION

Increasing availability of medico-administrative databases has prompted the development of automated pharmacovigilance signal detection methodologies. Self-controlled approaches have recently been proposed. They account for time-independent confounding factors that may not be recorded. So far, large numbers of drugs have been screened either univariately or with LASSO penalized regressions.

OBJECTIVE

We propose and assess a new method that combines the case-crossover self-controlled design with propensity scores (propensity score-adjusted case-crossover) built from high-dimensional data-driven variable selection, to account for co-medications or possibly other measured confounders.

METHODS

Comparison with the univariate and LASSO case-crossover was performed from simulations and a real-data study. Multiple regressions (LASSO, propensity score-adjusted case-crossover) accounted for co-medications and no other covariates. For the univariate and propensity score-adjusted case-crossover methods, the detection threshold was based on a false discovery rate procedure, while for LASSO, it relied on the Akaike Information Criterion. For the real-data study, two drug safety experts evaluated the signals generated from the analysis of 4099 patients with acute myocardial infarction from the French national health database.

RESULTS

On simulations, our approach ranked the signals similarly to the LASSO and better than the univariate method while controlling the false discovery rate at the prespecified level, contrary to the univariate method. The LASSO provided the best sensitivity at the cost of larger false discovery rate estimates. On the application, our approach showed similar performances to the LASSO and better performances than the univariate method. It highlighted 43 signals out of 609 drug candidates: 22 (51%) were considered as potentially pharmacologically relevant, including seven (16%) regarded as highly relevant.

CONCLUSIONS

Our findings show the interest of a propensity score combined with a case-crossover for pharmacovigilance. They also confirm that indication bias remains a challenge when mining medico-administrative databases.

Combining a Pharmacological Network Model with a Bayesian Signal Detection Algorithm to Improve the Detection of Adverse Drug Events

Introduction: Improving adverse drug event (ADE) detection is important for post-marketing drug safety surveillance. Existing statistical approaches can be further optimized owing to their high efficiency and low cost.

Objective: The objective of this study was to evaluate the proposed approach for use in pharmacovigilance, the early detection of potential ADEs, and the improvement of drug safety.

Methods: We developed a novel integrated approach, the Bayesian signal detection algorithm, based on the pharmacological network model (IC_PNM) using the FDA Adverse Event Reporting System (FAERS) data published from 2004 to 2009 and from 2014 to 2019Q2, PubChem, and DrugBank database. First, we used a pharmacological network model to generate the probabilities for drug-ADE associations, which comprised the proper prior information component (IC). We then defined the probability of the propensity score adjustment based on a logistic regression model to control for the confounding bias. Finally, we chose the Side Effect Resource (SIDER) and the Observational Medical Outcomes Partnership (OMOP) data to evaluate the detection performance and robustness of the IC_PNM compared with the statistical approaches [disproportionality analysis (DPA)] by using the area under the receiver operator characteristics curve (AUC) and Youden’s index.

Results: Of the statistical approaches implemented, the IC_PNM showed the best performance (AUC, 0.8291; Youden’s index, 0.5836). Meanwhile, the AUCs of the IC, EBGM, ROR, and PRR were 0.7343, 0.7231, 0.6828, and 0.6721, respectively.

Conclusion: The proposed IC_PNM combined the strengths of the pharmacological network model and the Bayesian signal detection algorithm and performed better in detecting true drug-ADE associations. It also detected newer ADE signals than a DPA and may be complementary to the existing statistical approaches.

Analysis of Neurological Adverse Events Reported in VigiBase From COVID-19 Vaccines

BACKGROUND

Fifteen COVID-19 vaccines have been granted emergency approval before the completion of conventional phases of clinical trials. The present study aimed to analyze the neurological adverse events (AEs) post-COVID-19 vaccination and focuses on determining the association of AEs with the vaccine.

METHODOLOGY

The neurological AEs reported for COVID-19 vaccines in the WHO pharmacovigilance database (VigiBase) were extracted from the System Organ Classes - neurological disorders and investigations. Descriptive statistics are reported as percentage and frequency and the disproportionality analysis was also conducted.

RESULTS

 For the neurological system, 19,529 AEs were reported. Of these, 15,638 events were reported from BNT162b2 vaccine, 2,751 from AZD1222 vaccine, 1,075 from mRNA-1273 vaccine, eight from Vero vaccine, two from Covaxin, and for 55 AEs, vaccine name was not mentioned. The reason for more AEs reported with BNT162b2 can be maximum vaccination with BNT162b2 vaccine in the study period. According to the disproportionality analysis based on IC025 value, ageusia, anosmia, burning sensation, dizziness, facial paralysis, headache, hypoaesthesia, lethargy, migraine, neuralgia, paresis, parosmia, poor sleep quality, seizure, transient ischemic attack, and tremor are some of the AEs that can be associated with the administration of the vaccine.

CONCLUSION

The vaccines should be monitored for these AEs till the causality of these AEs with COVID-19 vaccines is established through further long-term follow-up studies. These neurological AEs reported in VigiBase should not be taken as conclusive and mass vaccination should be carried out to control the pandemic until a definite link of these adverse effects is established.

Does Social Media Accelerate Product Recalls? Evidence from the Pharmaceutical Industry

Social media has become a vital platform for voicing product-related experiences that may not only reveal product defects, but also impose pressure on firms to act more promptly than before. This study scrutinizes the rarely studied relationship between these voices and the speed of product recalls in the context of the pharmaceutical industry in which social media pharmacovigilance is becoming increasingly important for the detection of drug safety signals. Using Federal Drug Administration drug enforcement reports and social media data crawled from online forums and Twitter, we investigate whether social media can accelerate the product recall process in the context of drug recalls. Results based on discrete-time survival analyses suggest that more adverse drug reaction discussions on social media lead to a higher hazard rate of the drug being recalled and, thus, a shorter time to recall. To better understand the underlying mechanism, we propose the information effect, which captures how extracting information from social media helps detect more signals and mine signals faster to accelerate product recalls, and the publicity effect, which captures how firms and government agencies are pressured by public concerns to initiate speedy recalls. Estimation results from two mechanism tests support the existence of these conceptualized channels underlying the acceleration hypothesis of social media. This study offers new insights for firms and policymakers concerning the power of social media and its influence on product recalls.

Development of a Data-Driven Reference Standard for Adverse Drug Reaction (RS-ADR) Signal Assessment (Preprint)

Background

Pharmacovigilance using real-world data (RWD), such as multicenter electronic health records (EHRs), yields massively parallel adverse drug reaction (ADR) signals. However, proper validation of computationally detected ADR signals is not possible due to the lack of a reference standard for positive and negative associations.

Objective

This study aimed to develop a reference standard for ADR (RS-ADR) to streamline the systematic detection, assessment, and understanding of almost all drug-ADR associations suggested by RWD analyses.

Methods

We integrated well-known reference sets for drug-ADR pairs, including Side Effect Resource, Observational Medical Outcomes Partnership, and EU-ADR. We created a pharmacovigilance dictionary using controlled vocabularies and systematically annotated EHR data. Drug-ADR associations computed from MetaLAB and MetaNurse analyses of multicenter EHRs and extracted from the Food and Drug Administration Adverse Event Reporting System were integrated as “empirically determined” positive and negative reference sets by means of cross-validation between institutions.

Results

The RS-ADR consisted of 1344 drugs, 4485 ADRs, and 6,027,840 drug-ADR pairs with positive and negative consensus votes as pharmacovigilance reference sets. After the curation of the initial version of RS-ADR, novel ADR signals such as “famotidine–hepatic function abnormal” were detected and reasonably validated by RS-ADR. Although the validation of the entire reference standard is challenging, especially with this initial version, the reference standard will improve as more RWD participate in the consensus voting with advanced pharmacovigilance dictionaries and analytic algorithms. One can check if a drug-ADR pair has been reported by our web-based search interface for RS-ADRs.

Conclusions

RS-ADRs enriched with the pharmacovigilance dictionary, ADR knowledge, and real-world evidence from EHRs may streamline the systematic detection, evaluation, and causality assessment of computationally detected ADR signals.

New adaptive lasso approaches for variable selection in automated pharmacovigilance signal detection

Background

Adverse effects of drugs are often identified after market introduction. Post-marketing pharmacovigilance aims to detect them as early as possible and relies on spontaneous reporting systems collecting suspicious cases. Signal detection tools have been developed to mine these large databases and counts of reports are analysed with disproportionality methods. To address disproportionality method biases, recent methods apply to individual observations taking into account all exposures for the same patient. In particular, the logistic lasso provides an efficient variable selection framework, yet the choice of the regularization parameter is a challenging issue and the lasso variable selection may give inconsistent results.

Methods

We propose a new signal detection methodology based on the adaptive lasso. We derived two new adaptive weights from (i) a lasso regression using the Bayesian Information Criterion (BIC), and (ii) the class-imbalanced subsampling lasso (CISL), an extension of stability selection. The BIC is used in the adaptive lasso stage for variable selection. We performed an extensive simulation study and an application to real data, where we compared our methods to the existing adaptive lasso, and recent detection approaches based on lasso regression or propensity scores in high dimension. For both studies, we evaluate the methods in terms of false discoveries and sensitivity.

Results

In the simulations and the application, both proposed adaptive weights show equivalent or better performances than the other competitors, with an advantage for the CISL-based adaptive weights. CISL and lasso regression using BIC are solid alternatives.

Conclusion

Our proposed adaptive lasso is an appealing methodology for signal detection in pharmacovigilance. Although we cannot rely on test theory, our approaches show a low and stable False Discovery Rate in all simulation settings. All methods evaluated in this work are implemented in the adapt4pv R package.

Supplementary Information

The online version contains supplementary material available at (10.1186/s12874-021-01450-3).

Consensus clustering for case series identification and adverse event profiles in pharmacovigilance

OBJECTIVE

To describe and evaluate vigiGroup - a consensus clustering algorithm which can identify groups of individual case reports referring to similar suspected adverse drug reactions and describe associated adverse event profiles, accounting for co-reported adverse event terms.

MATERIALS AND METHODS

Consensus clustering is achieved by grouping pairs of reports that are repeatedly placed together in the same clusters across a set of mixture model-based cluster analyses. The latter use empirical Bayes statistical shrinkage for improved performance. As baseline comparison, we considered a regular mixture model-based cluster analysis. Three randomly selected drugs in VigiBase, the World Health Organization's global database of Individual Case Safety Reports were analyzed: sumatriptan, ambroxol and tacrolimus. Clustering stability was assessed using the adjusted Rand index, ranging between -1 and +1, and clinical coherence was assessed through an intruder detection analysis.

RESULTS

For the three drugs considered, vigiGroup achieved stable and coherent results with adjusted Rand indices between +0.80 and +0.92, and intruder detection rates between 86% and 94%. Consensus clustering improved both stability and clinical coherence compared to mixture model-based clustering alone. Statistical shrinkage improved the stability of clusters compared to the baseline mixture model, as well as the cross-validated log-likelihood.

CONCLUSIONS

The proposed algorithm can achieve adequate stability and clinical coherence in clustering individual case reports, thereby enabling better identification of case series and associated adverse event profiles in pharmacovigilance. The use of empirical Bayes shrinkage and consensus clustering each led to meaningful improvements in performance.

Inclusion of unexposed subjects improves the precision and power of self-controlled case series method

The self-controlled case series is an important method in the studies of the safety of biopharmaceutical products. It uses the conditional Poisson model to make comparison within persons. In models without adjustment for age (or other time-varying covariates), cases who are never exposed to the product do not contribute any information to the estimation. We provide analytic proof and simulation results that the inclusion of unexposed cases in the conditional Poisson model with age adjustment reduces the asymptotic variance of the estimator of the exposure effect and increases power. We re-analysed a vaccine safety dataset to illustrate.

Surveillance of Antidepressant Safety (SADS): Active Signal Detection of Serious Medical Events Following SSRI and SNRI Initiation Using Big Healthcare Data

Introduction

The current process for generating evidence in pharmacovigilance has several limitations, which often lead to delays in the evaluation of drug-associated risks.

Objectives

In this study, we proposed and tested a near real-time epidemiological surveillance system using sequential, cumulative analyses focusing on the detection and preliminary risk quantification of potential safety signals following initiation of selective serotonin reuptake inhibitors (SSRIs) and serotonin-norepinephrine reuptake inhibitors (SNRIs).

Methods

We emulated an active surveillance system in an historical setting by conducting repeated annual cohort studies using nationwide Danish healthcare data (1996–2016). Outcomes were selected from the European Medicines Agency's Designated Medical Event list, summaries of product characteristics, and the literature. We followed patients for a maximum of 6 months from treatment initiation to the event of interest or censoring. We performed Cox regression analyses adjusted for standard sets of covariates. Potential safety signals were visualized using heat maps and cumulative hazard ratio (HR) plots over time.

Results

In the total study population, 969,667 new users were included and followed for 461,506 person-years. We detected potential safety signals with incidence rates as low as 0.9 per 10,000 person-years. Having eight different exposure drugs and 51 medical events, we identified 31 unique combinations of potential safety signals with a positive association to the event of interest in the exposed group. We proposed that these signals were designated for further evaluation once they appeared in a prospective setting. In total, 21 (67.7%) of these were not present in the current summaries of product characteristics.

Conclusion

The study demonstrated the feasibility of performing epidemiological surveillance using sequential, cumulative analyses. Larger populations are needed to evaluate rare events and infrequently used antidepressants.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40264-021-01110-x.

Adversarial neural network with sentiment-aware attention for detecting adverse drug reactions

Adverse drug reaction (ADR) detection is an important issue in drug safety. ADRs are health threats caused by medication. Identifying ADRs in a timely manner can reduce harm to patients and can also assist doctors in the rational use of drugs. Many studies have investigated potential ADRs based on social media due to the openness and timeliness of this resource; however, they have ignored the fine-grained emotional expression in social media text. In addition, the benchmark datasets from social media are usually small, which can result in the problem of over-fitting. In this paper, we propose the Adversarial Neural Network with Sentiment-aware Attention (ANNSA) model, which enhances the sentimental element in social media and improves the performance of neural networks via data augmentation. Specifically, a sentiment-aware attention mechanism is proposed to extract the word-level sentiment features associated with sentiment words and learn task-related information by optimizing a task-specific loss. For low-resource datasets, we use an adversarial training approach to generate perturbations of the word embeddings via an implicit regularization technique. ANNSA was tested on three social media ADR detection datasets, namely, Twitter, TwiMed (Twitter) and CADEC. The experimental results indicated the ability to achieve F1 values of 48.84%, 64.18% and 83.06%, respectively, comparable to the best results reported for state-of-the-art methods. Our study demonstrates that sentiment words are highly correlated with ADRs and that word-level sentiment features can assist in detecting ADRs from social media datasets.

Drug-Wide Association Study (DWAS): Challenges and Opportunities

Cancer risk associations with commonly prescribed medications have been mainly evaluated in hypothesis-driven studies that focus on one drug at a time. Agnostic drug-wide association studies (DWAS) offer an alternative approach to simultaneously evaluate associations between a large number of drugs with one or more cancers using large-scale electronic health records. Although cancer DWAS approaches are promising, a number of challenges limit their applicability. This includes the high likelihood of false positivity; lack of biological considerations; and methodological shortcomings, such as inability to tightly control for confounders. As such, the value of DWAS is currently restricted to hypothesis generation with detected signals needing further evaluation. In this commentary, we discuss those challenges in more detail and summarize the approaches to overcome them by using published cancer DWAS studies, including the accompanied article by Støer and colleagues. Despite current concerns, DWAS future is filled with opportunities for developing innovative analytic methods and techniques that incorporate pharmacology, epidemiology, cancer biology, and genetics.See related article by Støer et al., p. 682.

A meta-epidemiological study found lack of transparency and poor reporting of disproportionality analyses for signal detection in pharmacovigilance databases

OBJECTIVE

To review and appraise methods and reporting characteristics of pharmacovigilance disproportionality analyses.

STUDY DESIGN AND SETTING

We randomly selected 100 disproportionality analyses indexed in Medline found during a systematic literature search. We then extracted and synthetized methodological and reporting characteristics using 7 key items: 1) title transparency; 2) protocol pre-registration; 3) date of data extraction and analysis; 4) outcome, population, exposure and comparator definitions; 5) adjustment and stratification of results; 6) method and threshold for signal detection; 7) secondary and sensitivity analyses.

RESULTS

We found that methods used to generate disproportionality signals were extremely heterogeneous; there were nearly as many unique analyses as studies. The authors used various populations, methods, signal detection thresholds, adjustment or stratification variables, generally without justification for their choice or pre-specification in protocols. Moreover, 78% of studies failed to report methods for case, adverse drug reactions or comparator selection and 32 studies did not define the threshold for signal generation.

CONCLUSION

Our survey raises major concerns regarding all aspects of disproportionality analyses that could lead to misleading results and generate unjustified alarms. We advocate for a strong and transparent rationale for variable selection, choice of population and comparators pre-specified in a protocol and assessed by sensitivity analyses.