Artificial Intelligence in Pharmacoepidemiology: A Systematic Review. Part 1-Overview of Knowledge Discovery Techniques in Artificial Intelligence

Traditional Methods Hold Their Ground Against Machine Learning in Predicting Potentially Inappropriate Medication Use in Older Adults

OBJECTIVES

Machine learning methods have gained much attention in health sciences for predicting various health outcomes but are scarcely used in pharmacoepidemiology. The ability to identify predictors of suboptimal medication use is essential for conducting interventions aimed at improving medication outcomes. It remains uncertain whether machine learning methods could enhance the identification of potentially inappropriate medication use among older adults compared with traditional methods. This study aimed to (1) to compare the performances of machine learning models in predicting use of potentially inappropriate medications and (2) to quantify and compare the relative importance of predictors in a population of community-dwelling older adults (>65 years) in the province of Québec, Canada.

METHODS

We used the Québec Integrated Chronic Disease Surveillance System and selected a cohort of 1 105 295 older adults of whom 533 719 were potentially inappropriate medication users. Potentially inappropriate medications were defined according to the Beers list. We compared performances between 5 popular machine learning models (gradient boosting machines, logistic regression, naive Bayes, neural networks, and random forests) based on receiver operating characteristic curves and other performance criteria, using a set of sociodemographic and medical predictors.

RESULTS

No model clearly outperformed the others. All models except neural networks were in agreement regarding the top predictors (sex and anxiety-depressive disorders and schizophrenia) and the bottom predictors (rurality and social and material deprivation indices).

CONCLUSIONS

Including other types of predictors (eg, unstructured data) may be more useful for increasing performance in prediction of potentially inappropriate medication use.

Drug Repurposing in Crohn's Disease Using Danish Real-World Data

Aim

Drug repurposing, utilizing electronic healthcare records (EHRs), offers a promising alternative by repurposing existing drugs for new therapeutic indications, especially for patients lacking effective therapies. Intestinal fibrosis, a severe complication of Crohn's disease (CD), poses significant challenges, increasing morbidity and mortality without available pharmacological treatments. This article focuses on identifying medications associated with an elevated or reduced risk of fibrosis in CD patients through a population-wide real-world data and artificial intelligence (AI) approach.

Methods

Patients aged 65 or older with a diagnosis of CD from 1996 to 2019 in the Danish EHRs were followed for up to 24 years. The primary outcome was the need of specific surgical procedures, namely proctocolectomy with ileostomy and ileocecal resection as proxies of intestinal fibrosis. The study explored drugs linked to an increased or reduced risk of the study outcome through machine-learning driven survival analysis.

Results

Among the 9179 CD patients, 1029 (11.2%) underwent surgery, primarily men (58.5%), with a mean age of 76 years, 10 drugs were linked to an elevated risk of surgery for proctocolectomy with ileostomy and ileocecal resection. In contrast, 10 drugs were associated with a reduced risk of undergoing surgery for these conditions.

Conclusion

This study focuses on repurposing existing drugs to prevent surgery related to intestinal fibrosis in CD patients, using Danish EHRs and advanced statistical methods. The findings offer valuable insights into potential treatments for this condition, addressing a critical unmet medical need. Further research and clinical trials are warranted to validate the effectiveness of these repurposed drugs in preventing surgery related to intestinal fibrosis in CD patients.

Epidemiological surveillance of drug safety using cumulative sequential analysis in electronic healthcare data

BACKGROUND

Methods for safety signal detection in electronic healthcare data analysing data sequentially are being developed to meet the limitations of spontaneous reporting systems.

OBJECTIVES

This study aims to provide an overview of the literature on sequential analysis of electronic healthcare data and describe the development and testing of a novel epidemiological surveillance system.

METHODS

We searched Medline, Embase, PubMed, Scopus, Web of Science, and the Cochrane Library applying similar in- and exclusion criteria as those of a previous systematic review. The proposed system consisted of repeated cohort studies and was tested in an emulated prospective setting. Two signal evaluations were performed with several sensitivity analyses and a target trial emulation.

FINDINGS

In the literature, 11 studies analysed the data sequentially of which two applied traditional epidemiological methods. Epidemiological surveillance of several exposures and outcomes can be successfully conducted with the newly proposed sequential analysis of electronic healthcare data. Signal evaluation studies confirmed the results of the system.

CONCLUSIONS

Very few studies in the literature analysed data at multiple time points, although this seems to be a prerequisite for testing the methods in a realistic setting. We demonstrated the feasibility of a sequential surveillance system using electronic healthcare data.

Identification and verification of novel immune-related ferroptosis signature with excellent prognostic predictive and clinical guidance value in hepatocellular carcinoma

Background: Immunity and ferroptosis often play a synergistic role in the progression and treatment of hepatocellular carcinoma (HCC). However, few studies have focused on identifying immune-related ferroptosis gene biomarkers. Methods: We performed weighted gene co-expression network analysis (WGCNA) and random forest to identify prognostic differentially expressed immune-related genes (PR-DE-IRGs) highly related to HCC and characteristic prognostic differentially expressed ferroptosis-related genes (PR-DE-FRGs) respectively to run co-expression analysis for prognostic differentially expressed immune-related ferroptosis characteristic genes (PR-DE-IRFeCGs). Lasso regression finally identified 3 PR-DE-IRFeCGs for us to construct a prognostic predictive model. Differential expression and prognostic analysis based on shared data from multiple sources and experimental means were performed to further verify the 3 modeled genes' biological value in HCC. We ran various performance testing methods to test the model's performance and compare it with other similar signatures. Finally, we integrated composite factors to construct a comprehensive quantitative nomogram for accurate prognostic prediction and evaluated its performance. Results: 17 PR-DE-IRFeCGs were identified based on co-expression analysis between the screened 17 PR-DE-FRGs and 34 PR-DE-IRGs. Multi-source sequencing data, QRT-PCR, immunohistochemical staining and testing methods fully confirmed the upregulation and significant prognostic influence of the three PR-DE-IRFeCGs in HCC. The model performed well in the performance tests of multiple methods based on the 5 cohorts. Furthermore, our model outperformed other related models in various performance tests. The immunotherapy and chemotherapy guiding value of our signature and the comprehensive nomogram's excellent performance have also stood the test. Conclusion: We identified a novel PR-DE-IRFeCGs signature with excellent prognostic prediction and clinical guidance value in HCC.

Artificial intelligence-driven prediction of COVID-19-related hospitalization and death: a systematic review

Aim

To perform a systematic review on the use of Artificial Intelligence (AI) techniques for predicting COVID-19 hospitalization and mortality using primary and secondary data sources.

Study eligibility criteria

Cohort, clinical trials, meta-analyses, and observational studies investigating COVID-19 hospitalization or mortality using artificial intelligence techniques were eligible. Articles without a full text available in the English language were excluded.

Data sources

Articles recorded in Ovid MEDLINE from 01/01/2019 to 22/08/2022 were screened.

Data extraction

We extracted information on data sources, AI models, and epidemiological aspects of retrieved studies.

Bias assessment

A bias assessment of AI models was done using PROBAST.

Participants

Patients tested positive for COVID-19.

Results

We included 39 studies related to AI-based prediction of hospitalization and death related to COVID-19. The articles were published in the period 2019-2022, and mostly used Random Forest as the model with the best performance. AI models were trained using cohorts of individuals sampled from populations of European and non-European countries, mostly with cohort sample size <5,000. Data collection generally included information on demographics, clinical records, laboratory results, and pharmacological treatments (i.e., high-dimensional datasets). In most studies, the models were internally validated with cross-validation, but the majority of studies lacked external validation and calibration. Covariates were not prioritized using ensemble approaches in most of the studies, however, models still showed moderately good performances with Area under the Receiver operating characteristic Curve (AUC) values >0.7. According to the assessment with PROBAST, all models had a high risk of bias and/or concern regarding applicability.

Conclusions

A broad range of AI techniques have been used to predict COVID-19 hospitalization and mortality. The studies reported good prediction performance of AI models, however, high risk of bias and/or concern regarding applicability were detected.

A novel approach for pharmacological substantiation of safety signals using plasma concentrations of medication and administrative/healthcare databases: a case study using Danish registries for an FDA warning on lamotrigine

PHARMACOM-EPI is a novel framework to predict plasma concentrations of drugs at the time of occurrence of clinical outcomes. In early 2021, the U.S. Food and Drug Administration (FDA) issued a warning on the antiseizure drug lamotrigine claiming that it has the potential to increase the risk of arrhythmias and related sudden cardiac death due to a pharmacological sodium channel-blocking effect. We hypothesized that the risk of arrhythmias and related death is due to toxicity. We used the PHARMACOM-EPI framework to investigate the relationship between lamotrigine's plasma concentrations and the risk of death in older patients using real-world data. Danish nationwide administrative and healthcare registers were used as data sources and individuals aged 65 years or older during the period 1996 - 2018 were included in the study. According to the PHARMACOM-EPI framework, plasma concentrations of lamotrigine were predicted at the time of death and patients were categorized into non-toxic and toxic groups based on the therapeutic range of lamotrigine (3-15mg/L). Over 1 year of treatment, the incidence rate ratio (IRR) of all-cause mortality was calculated between the propensities score matched toxic and non-toxic groups. In total, 7286 individuals were diagnosed with epilepsy and were exposed to lamotrigine, 432 of which had at least one plasma concentration measurement The pharmacometric model by Chavez et al. was used to predict lamotrigine's plasma concentrations considering the lowest absolute percentage error among identified models (14.25%, 95% CI: 11.68-16.23). The majority of lamotrigine associated deaths were cardiovascular-related and occurred among individuals with plasma concentrations in the toxic range. The IRR of mortality between the toxic group and non-toxic group was 3.37 [95% CI: 1.44-8.32] and the cumulative incidence of all-cause mortality exponentially increased in the toxic range. Application of our novel framework PHARMACOM-EPI provided strong evidence to support our hypothesis that the increased risk of all-cause and cardiovascular death was associated with a toxic plasma concentration level of lamotrigine among older lamotrigine users.

A machine-learning guided method for predicting add-on and switch in secondary data sources: A case study on anti-seizure medications in Danish registries

Purpose: There is a lack of available evidence regarding the treatment pattern of switches and add-ons for individuals aged 65 years or older with epilepsy during the first years from the time they received their first anti-seizure medication because of the lack of valid methods. Therefore, this study aimed to develop an algorithm for identifying switches and add-ons using secondary data sources for anti-seizure medication users. Methods: Danish nationwide databases were used as data sources. Residents in Denmark between 1996 and 2018 who were diagnosed with epilepsy and redeemed their first prescription for anti-seizure medication after epilepsy diagnosis were followed up for 730 days until the end of the follow-up period, death, or emigration to assess switches and add-ons occurred during the follow-up period. The study outcomes were the overall accuracy of the classification of switch or add-on of the newly developed algorithm. Results: In total, 15870 individuals were included in the study population with a median age of 72.9 years, of whom 52.0% were male and 48.0% were female. A total of 988 of the 15879 patients from the study population were present during the 730-day follow-up period, and 988 individuals (6.2%) underwent a total of 1485 medication events with co-exposure to two or more anti-seizure medications. The newly developed algorithmic method correctly identified 9 out of 10 add-ons (overall accuracy 92%) and 9 out of 10 switches (overall accuracy 88%). Conclusion: The majority of switches and add-ons occurred early during the first 2 years of disease and according to clinical recommendations. The newly developed algorithm correctly identified 9 out of 10 switches/add-ons.

Classification and biomarker gene selection of pyroptosis-related gene expression in psoriasis using a random forest algorithm

Background: Psoriasis is a chronic and immune-mediated skin disorder that currently has no cure. Pyroptosis has been proved to be involved in the pathogenesis and progression of psoriasis. However, the role pyroptosis plays in psoriasis remains elusive.

Methods: RNA-sequencing data of psoriasis patients were obtained from the Gene Expression Omnibus (GEO) database, and differentially expressed pyroptosis-related genes (PRGs) between psoriasis patients and normal individuals were obtained. A principal component analysis (PCA) was conducted to determine whether PRGs could be used to distinguish the samples. PRG and immune cell correlation was also investigated. Subsequently, a novel diagnostic model comprising PRGs for psoriasis was constructed using a random forest algorithm (ntree = 400). A receiver operating characteristic (ROC) analysis was used to evaluate the classification performance through both internal and external validation. Consensus clustering analysis was used to investigate whether there was a difference in biological functions within PRG-based subtypes. Finally, the expression of the kernel PRGs were validated in vivo by qRT-PCR.

Results: We identified a total of 39 PRGs, which could distinguish psoriasis samples from normal samples. The process of T cell CD4 memory activated and mast cells resting were correlated with PRGs. Ten PRGs, IL-1β, AIM2, CASP5, DHX9, CASP4, CYCS, CASP1, GZMB, CHMP2B, and CASP8, were subsequently screened using a random forest diagnostic model. ROC analysis revealed that our model has good diagnostic performance in both internal validation (area under the curve [AUC] = 0.930 [95% CI 0.877–0.984]) and external validation (mean AUC = 0.852). PRG subtypes indicated differences in metabolic processes and the MAPK signaling pathway. Finally, the qRT-PCR results demonstrated the apparent dysregulation of PRGs in psoriasis, especially AIM2 and GZMB.

Conclusion: Pyroptosis may play a crucial role in psoriasis and could provide new insights into the diagnosis and underlying mechanisms of psoriasis.

Black Swan Events and Intelligent Automation for Routine Safety Surveillance

Effective identification of previously implausible safety signals is a core component of successful pharmacovigilance. Timely, reliable, and efficient data ingestion and related processing are critical to this. The term 'black swan events' was coined by Taleb to describe events with three attributes: unpredictability, severe and widespread consequences, and retrospective bias. These rare events are not well understood at their emergence but are often rationalized in retrospect as predictable. Pharmacovigilance strives to rapidly respond to potential black swan events associated with medicine or vaccine use. Machine learning (ML) is increasingly being explored in data ingestion tasks. In contrast to rule-based automation approaches, ML can use historical data (i.e., 'training data') to effectively predict emerging data patterns and support effective data intake, processing, and organisation. At first sight, this reliance on previous data might be considered a limitation when building ML models for effective data ingestion in systems that look to focus on the identification of potential black swan events. We argue that, first, some apparent black swan events-although unexpected medically-will exhibit data attributes similar to those of other safety data and not prove algorithmically unpredictable, and, second, standard and emerging ML approaches can still be robust to such data outliers with proper awareness and consideration in ML system design and with the incorporation of specific mitigatory and support strategies. We argue that effective approaches to managing data on potential black swan events are essential for trust and outline several strategies to address data on potential black swan events during data ingestion.

The PHARMACOM-EPI Framework for Integrating Pharmacometric Modelling Into Pharmacoepidemiological Research Using Real-World Data: Application to Assess Death Associated With Valproate

In pharmacoepidemiology, it is usually expected that the observed association should be directly or indirectly related to the pharmacological effects of the drug/s under investigation. Pharmacological effects are, in turn, strongly connected to the pharmacokinetic and pharmacodynamic properties of a drug, which can be characterized and investigated using pharmacometric models. Recently, the use of pharmacometrics has been proposed to provide pharmacological substantiation of pharmacoepidemiological findings derived from real-world data. However, validated frameworks suggesting how to combine these two disciplines for the aforementioned purpose are missing. Therefore, we propose PHARMACOM-EPI, a framework that provides a structured approach on how to identify, characterize, and apply pharmacometric models with practical details on how to choose software, format dataset, handle missing covariates/dosing data, how to perform the external evaluation of pharmacometric models in real-world data, and how to provide pharmacological substantiation of pharmacoepidemiological findings. PHARMACOM-EPI was tested in a proof-of-concept study to pharmacologically substantiate death associated with valproate use in the Danish population aged ≥ 65 years. Pharmacological substantiation of death during a follow-up period of 1 year showed that in all individuals who died (n = 169) individual predictions were within the subtherapeutic range compared with 52.8% of those who did not die (n = 1,084). Of individuals who died, 66.3% (n = 112) had a cause of death possibly related to valproate and 33.7% (n = 57) with well-defined cause of death unlikely related to valproate. This proof-of-concept study showed that PHARMACOM-EPI was able to provide pharmacological substantiation for death associated with valproate use in the study population.

The Relationship Between Valproate and Lamotrigine/Levetiracetam Use and Prognosis in Patients With Epilepsy and Heart Failure: A Danish Register-Based Study

OBJECTIVE

To compare the hazard for all-cause mortality and mortality due to heart failure (HF) between valproate (VPA) and levetiracetam (LEV)/lamotrigine (LTG) users in patients aged ≥ 65 with comorbidities of epilepsy and HF.

METHODS

This was a cohort study using Danish registers during the period from January 1996 to July 2018. The study population included new users of LTG, LEV or VPA. A Cox regression model was used to compute crude and adjusted hazard ratios for the outcome, using an intention-to-treat approach. Average treatment effects (eg, 1-year absolute risks), risk differences and the ratio of risks were computed using the G-formula based on a Cox regression model for the outcomes at the end of the follow-up period.

RESULTS

We included 1345 subjects in the study population. VPA users (n = 696), when compared to LTG/LEV users (n = 649), had an increased hazard of mortality due to HF (hazard ratio [HR] 2.39; 95% CI 1.02-5.60) and to all-cause mortality (HR 1.37; 95% CI 1.01-1.85) in both crude and adjusted analyses. The 1-year absolute risks for all-cause mortality were 29% (95% CI 25%-33%) and 22% (95% CI 18%-26%) for VPA and LTG/LEV users. For mortality due to HF, 1-year absolute risks were 5% (95% CI 3%-7%) and 2% (95% CI 1%-4%) for VPA and LTG/LEV users. The average risk ratio, with LTG/LEV as the reference group, was 1.31 (95% CI 1.02-1.71) for all-cause mortality and 2.35 (95% CI 1.11-5.76) for HF mortality.

CONCLUSION

In older people with HF and epilepsy, treatment with VPA was associated with a higher risk of all-cause and HF mortality compared to treatment with LTG and LEV.

Performance assessment of different machine learning approaches in predicting diabetic ketoacidosis in adults with type 1 diabetes using electronic health records data

Purpose

To assess the performance of different machine learning (ML) approaches in identifying risk factors for diabetic ketoacidosis (DKA) and predicting DKA.

Methods

This study applied flexible ML (XGBoost, distributed random forest [DRF] and feedforward network) and conventional ML approaches (logistic regression and least absolute shrinkage and selection operator [LASSO]) to 3400 DKA cases and 11 780 controls nested in adults with type 1 diabetes identified from Optum® de‐identified Electronic Health Record dataset (2007–2018). Area under the curve (AUC), accuracy, sensitivity and specificity were computed using fivefold cross validation, and their 95% confidence intervals (CI) were established using 1000 bootstrap samples. The importance of predictors was compared across these models.

Results

In the training set, XGBoost and feedforward network yielded higher AUC values (0.89 and 0.86, respectively) than logistic regression (0.83), LASSO (0.83) and DRF (0.81). However, the AUC values were similar (0.82) among these approaches in the test set (95% CI range, 0.80–0.84). While the accuracy values >0.8 and the specificity values >0.9 for all models, the sensitivity values were only 0.4. The differences in these metrics across these models were minimal in the test set. All approaches selected some known risk factors for DKA as the top 10 features. XGBoost and DRF included more laboratory measurements or vital signs compared with conventional ML approaches, while feedforward network included more social demographics.

Conclusions

In our empirical study, all ML approaches demonstrated similar performance, and identified overlapping, but different, top 10 predictors. The difference in selected top predictors needs further research.

Artificial Intelligence in Pharmacoepidemiology: A Systematic Review. Part 2-Comparison of the Performance of Artificial Intelligence and Traditional Pharmacoepidemiological Techniques

Aim: To summarize the evidence on the performance of artificial intelligence vs. traditional pharmacoepidemiological techniques.

Methods: Ovid MEDLINE (01/1950 to 05/2019) was searched to identify observational studies, meta-analyses, and clinical trials using artificial intelligence techniques having a drug as the exposure or the outcome of the study. Only studies with an available full text in the English language were evaluated.

Results: In all, 72 original articles and five reviews were identified via Ovid MEDLINE of which 19 (26.4%) compared the performance of artificial intelligence techniques with traditional pharmacoepidemiological methods. In total, 44 comparisons have been performed in articles that aimed at 1) predicting the needed dosage given the patient’s characteristics (31.8%), 2) predicting the clinical response following a pharmacological treatment (29.5%), 3) predicting the occurrence/severity of adverse drug reactions (20.5%), 4) predicting the propensity score (9.1%), 5) identifying subpopulation more at risk of drug inefficacy (4.5%), 6) predicting drug consumption (2.3%), and 7) predicting drug-induced lengths of stay in hospital (2.3%). In 22 out of 44 (50.0%) comparisons, artificial intelligence performed better than traditional pharmacoepidemiological techniques. Random forest (seven out of 11 comparisons; 63.6%) and artificial neural network (six out of 10 comparisons; 60.0%) were the techniques that in most of the comparisons outperformed traditional pharmacoepidemiological methods.

Conclusion: Only a small fraction of articles compared the performance of artificial intelligence techniques with traditional pharmacoepidemiological methods and not all artificial intelligence techniques have been compared in a Pharmacoepidemiological setting. However, in 50% of comparisons, artificial intelligence performed better than pharmacoepidemiological techniques.

Role of Artificial Intelligence (AI) in Surgery: Introduction, General Principles, and Potential Applications

AI (Artificial intelligence) is an interdisciplinary field aimed at the development of algorithms to endow machines with the capability of executing cognitive tasks. The number of publications regarding AI and surgery has increased dramatically over the last two decades. This phenomenon can partly be explained by the exponential growth in computing power available to the largest AI training runs. AI can be classified into different sub-domains with extensive potential clinical applications in the surgical setting. AI will increasingly become a major component of clinical practice in surgery. The aim of the present Narrative Review is to give a general introduction and summarized overview of AI, as well as to present additional remarks on potential surgical applications and future perspectives in surgery.