A discovery and verification approach to pharmacovigilance using electronic healthcare data.

Introduction: Pharmacovigilance is vital for drug safety. The process typically involves two key steps: initial signal generation from spontaneous reporting systems (SRSs) and subsequent expert review to assess the signals' (potential) causality and decide on the appropriate action. Methods: We propose a novel discovery and verification approach to pharmacovigilance based on electronic healthcare data. We enhance the signal detection phase by introducing an ensemble of methods which generated signals are combined using Borda count ranking; a method designed to emphasize consensus. Ensemble methods tend to perform better when data is noisy and leverage the strengths of individual classifiers, while trying to mitigate some of their limitations. Additionally, we offer the committee of medical experts with the option to perform an in-depth investigation of selected signals through tailored pharmacoepidemiological studies to evaluate their plausibility or spuriousness. To illustrate our approach, we utilize data from the German Pharmacoepidemiological Research Database, focusing on drug reactions to the direct oral anticoagulant rivaroxaban. Results: In this example, the ensemble method is built upon the Bayesian confidence propagation neural network, longitudinal Gamma Poisson shrinker, penalized regression and random forests. We also conduct a pharmacoepidemiological verification study in the form of a nested active comparator case-control study, involving patients diagnosed with atrial fibrillation who initiated anticoagulant treatment between 2011 and 2017. Discussion: The case study reveals our ability to detect known adverse drug reactions and discover new signals. Importantly, the ensemble method is computationally efficient. Hasty false conclusions can be avoided by a verification study, which is, however, time-consuming to carry out. We provide an online tool for easy application: https://borda.bips.eu.

Pediatric chronic pain grading: a revised classification of the severity of pediatric chronic pain.

ABSTRACT: In this study, we describe the development and validation of a revised Pediatric Chronic Pain Grading (P-CPG) for children aged 8 to 17 years that adds emotional impairment to previously used measures of pain intensity and functional impairment. Such a measure enables the assessment of chronic pain severity in different epidemiological and clinical populations, the stratification of treatment according to pain severity, and the monitoring of treatment outcome. The P-CPG was developed using a representative sample of school children with chronic pain (n = 454; M age = 12.95, SD = 2.22). Construct validity and sensitivity to change were examined within a sample of N = 2448 children and adolescents (M age = 12.71, SD = 2.47) comprising 3 subsamples (school n = 1562, primary care n = 129, and tertiary care n = 757) affected by chronic pain to varying extents. Results showed that P-CPG grades differed significantly among the 3 subsamples, with school children being least affected by chronic pain and tertiary care patients being most affected. As P-CPG grade increased, so did pain intensity, functional impairment, pain-related school absence, and emotional impairment. Convergent validity was demonstrated by significant positive correlations between the P-CPG and global ratings of pain severity as well as objective claims data; the latter reflects greater health care costs with increasing P-CPG scores. Sensitivity to change was supported by a significant reduction in baseline P-CPG grades 3 and 6 months after intensive interdisciplinary pain treatment in tertiary care sample. In conclusion, the P-CPG is an appropriate measure of pain severity in children and adolescents with chronic pain in clinical as well as epidemiological settings.

In atrial fibrillation epilepsy risk differs between oral anticoagulants: active comparator, nested case-control study.

AIMS: Atrial fibrillation (AF) is a risk factor for brain infarction, which can lead to epilepsy. We aimed to investigate whether treatment of AF with direct oral anticoagulants (DOACs) affects the risk of epilepsy in comparison to treatment with the vitamin K antagonist phenprocoumon (PPC). METHODS AND RESULTS: We performed an active comparator, nested case-control study based on the German Pharmacoepidemiological Research Database that includes claims data from statutory health insurance providers of about 25 million persons since 2004. In 2011-17, 227 707 AF patients initiated treatment with a DOAC or PPC, of which 1828 cases developed epilepsy on current treatment with an oral anticoagulant. They were matched to 19 084 controls without epilepsy. Patients with DOAC treatment for AF had an overall higher risk of epilepsy with an odds ratio of 1.39, 95% CI (1.24; 1.55) compared to current PPC treatment. Cases had higher baseline CHA2DS2-VASc scores and more frequently a history of stroke than controls. After excluding patients with ischaemic stroke prior to the diagnosis of epilepsy, the risk of epilepsy was still higher on DOACs than on PPC. In contrast, within a cohort of patients with venous thromboembolism, the risk of epilepsy on treatment with DOACs was less elevated [adjusted odds ratio 1.15, 95% CI (0.98; 1.34)]. CONCLUSION: In patients with AF initiating oral anticoagulation, treatment with a DOAC was associated with an increased risk of epilepsy compared to the vitamin K antagonist PPC. Covert brain infarction may explain the observed elevated risk of epilepsy.

Evaluation of the EMPAR study population on the basis of metabolic phenotypes of selected pharmacogenes.

The impact of genetic variability of pharmacogenes as a possible risk factor for adverse drug reactions is elucidated in the EMPAR (Einfluss metabolischer Profile auf die Arzneimitteltherapiesicherheit in der Routineversorgung/English: influence of metabolic profiles on the safety of drug therapy in routine care) study. EMPAR evaluates possible associations of pharmacogenetically predicted metabolic profiles relevant for the metabolism of frequently prescribed cardiovascular drugs. Based on a German study population of 10,748 participants providing access to healthcare claims data and DNA samples for pharmacogenetic assessment, first analyses were performed and evaluated. The aim of this first evaluation was the characterization of the study population with regard to general parameters such as age, gender, comorbidity, and polypharmacy at baseline (baseline year) as well as important combinations of cardiovascular drugs with relevant genetic variants and predicted metabolic phenotypes. The study was registered in the German Clinical Trials Register (DRKS) on July 6, 2018 (DRKS00013909).

Population-based study of autoimmune hepatitis and primary biliary cholangitis in Germany: rising prevalences based on ICD codes, yet deficits in medical treatment.

BACKGROUND: Population-based data on the prevalence of and real-life treatment for the autoimmune liver diseases (AILD), primary biliary cholangitis (PBC) and autoimmune hepatitis (AIH), are scarce, and such knowledge may help to improve patient care. METHODS: Data of 8.1 million individuals having health insurance with the "Techniker Krankenkasse," the largest German health insurer, were analyzed with regard to the prevalence of and prescribed medication for AILD in Germany from 2011 until 2014. Patients with viral hepatitis B infection (HBV) and alcoholic liver cirrhosis (ALC) served as control groups. Case definition was based on ICD codes. RESULTS: The prevalences of PBC and AIH were 36.9/100 000 inhabitants (95 % CI: 35.6-38.2) and 23.0/100 000 inhabitants (95 % CI: 22.0-24.0) in 2014, respectively. The prevalences of AILD increased from 2011 to 2014 (for PBC by 31 % and for AIH by 29 %), with the largest increase for male patients with PBC. In contrast, the prevalence of HBV declined while that of ALC remained stable. The analysis of prescribed real-life treatment revealed considerable deviations from standard treatment recommendations. Specifically, in older patients with PBC or AIH, undertreatment was common. CONCLUSION: The prevalence of PBC and AIH based on ICD codes is increasing in Germany. The analysis of real-life treatment in this large and population-based cohort points to the unmet need to improve the implementation of treatment guidelines for autoimmune liver diseases in the broader medical community.

Influence of metabolic profiles on the safety of drug therapy in routine care in Germany: protocol of the cohort study EMPAR.

INTRODUCTION: Pre-emptive testing of pharmacogenetically relevant single-nucleotide polymorphisms can be an effective tool in the prevention of adverse drug reactions and therapy resistance. However, most of the tests are not used as standard in routine care in Germany because of lacking evidence for the clinical and economical benefit and their impact on the usage of healthcare services. We address this issue by investigating the influence of pharmacogenetic profiles on the use of healthcare services over an extended period of several years using routine care data from a statutory health insurance company. The goal is to provide clinical evidence whether pre-emptive pharmacogenetic testing of metabolic profiles in routine care in Germany is beneficial and cost-effective. METHODS AND ANALYSIS: The EMPAR (Einfluss metabolischer Profile auf die Arzneimitteltherapiesicherheit in der Routineversorgung) study is a non-interventional cohort study conducted to analyse pharmacogenetic risk factors that are important for drug therapy by means of endpoints relevant for healthcare. The analysis is based on pharmacogenetic profiles and statutory health insurance data. We perform pharmacogenetic, pharmacoepidemiological and pharmacoeconomic analyses using health care utilisation scores and machine learning techniques. Therefore, we aim to include about 10 000 patients (≥18 years) insured by the health insurance provider Techniker Krankenkasse. The study focuses on patients with prescriptions of anticoagulants and prescriptions of cholesterol-lowering drugs. Also, a screening for special pharmacogenetic characteristics will be performed in patients with at least one Y57.9! diagnosis (Complication of medical and surgical care: drug or medicament, unspecified). Outcomes include the utilisation of health insurance services, the incidence of incapacity for work and costs for drugs and treatment. ETHICS AND DISSEMINATION: The protocol was approved by the Ethics Committee of the Medical Faculty, University of Bonn (Lfd. Nr. 339/17). The results of this research project will be published in scientific open access journals and at conferences. TRIAL REGISTRATION NUMBER: German Clinical Trials Register, DRKS00013909.

[Chronic pain in children and adolescents: an economic perspective]. / Chronischer Schmerz bei Kindern und Jugendlichen: eine ökonomische Betrachtung.

BACKGROUND: Chronic pain is a frequent and disabling health problem in children and adolescents and is associated with high health care utilization and costs. OBJECTIVE: The aim of this study was to analyze the direct and indirect costs of chronic pain in children and adolescents in monetary terms before and after multimodal pain therapy from a societal perspective. MATERIALS AND METHODS: Health care costs 12 months before and after multimodal pain therapy include direct costs from statutory health insurances and parents as well as indirect costs due to working days lost. RESULTS: Direct median costs before multimodal treatment were 5619 € (min-max: 377-35,509 €) per year. In the year after pain therapy, costs decreased to a median of 3262 € (min-max: 142-42,910 €) (p = 0.001). In all, 55% of patients showed a significant cost reduction, while 18% had a cost increase. CONCLUSIONS: An effective multimodal pain therapy may reduce health care costs in children and adolescents. Further economic studies are needed to evaluate long-term effects of pain therapy for children and adolescents with chronic pain in a controlled design.

Metamizole and the risk of drug-induced agranulocytosis and neutropenia in statutory health insurance data.

The non-opioid analgesic metamizole (dipyrone) is used for the treatment of acute and chronic pain and fever. Agranulocytosis is known as a serious adverse drug reaction of metamizole with potentially fatal outcome. However, its frequency is controversially discussed. The aim of our study was to determine the risk of metamizole-associated agranulocytosis and neutropenia using statutory health insurance data. We analyzed data from a large German health insurance fund in the period from 2010 to 2013. Metamizole-exposed subjects were identified and compared to a propensity score-matched control cohort. A total of 630,285 metamizole-treated subjects and 390,830 matched control subjects were included. In the metamizole cohort, ICD codes for agranulocytosis and neutropenia appeared more often than in non-users. The relative risk for drug-induced agranulocytosis and neutropenia (D70.1) was 3.03 (95% confidence interval, 2.49 to 3.69). The risk for developing drug-induced agranulocytosis and neutropenia after metamizole prescription was 1: 1602 (CI 95%, 1:1926 to 1:1371). Our results confirm the risk estimation of previous studies. However, the outcome of our study may be confounded by an association of metamizole treatment and chemotherapy. Therefore, consequences for treatment have to be drawn with care.

[Guideline-Orientated Prescription of Disease-Modifying Drugs in Patients with Rheumatoid Arthritis - an Analysis Based on Social Health Insurance Claims Data]. / Leitliniengerechte Verordnung von krankheitsmodifizierenden Arzneimitteln bei Patienten mit rheumatoider Arthritis ­ eine Analyse auf Basis von GKV-Routinedaten.

BACKGROUND: With the market entry of biologics, the treatment of rheumatoid arthritis (RA) has changed fundamentally in terms of efficacy and costs. THE AIM: of this study is to analyse the treatment according the guideline of the German society of rheumatology for RA patients with disease-modifying anti-rheumatic drugs (DMARDs) using claims data from the statutory health insurance. MATERIALS AND METHODS: The claims data of the Techniker Krankenkasse were analysed retrospectively for the years 2011 - 2014. Subgroup analyses were used to conduct prescription and treatment differences with respect to guideline-recommended conventional DMARDs and biologics. RESULTS: The study population included 55,538 RA patients (29.7 % incidence, 70.3 % prevalence, 22.3 % M05: Seropositive rheumatoid arthritis, 77.7 % M06: Other rheumatoid arthritis). Only 21,616 insured patients (38.9 %) were prescribed a guideline-recommended conventional DMARD or biologic at least once within one year of/after the first diagnosis. Among incident patients, the coverage rate with disease-modifying drugs was below the prevalence patients (31.5 % vs. 42.1 %). 60.9 % of M05 patients and only 29.7 % of M06 patients received a single DMARD after index diagnosis. If a DMARD has been prescribed, then it was prescribed, on average, within the first quarter of the initial diagnosis. The leading role in the prescription of basic therapies for index medication is provided by the rheumatologist. Nevertheless, 68.3 % of patients consulted a specialist in rheumatology at least once within a year of the first diagnosis. CONCLUSION: The results of this large sample show differences in the guideline recommended prescription of disease-modifying drugs for different subgroups of RA as well as an undersupply in patients not treated by the rheumatologist.

[Detection of drug risks after approval : Methods development for the use of routine statutory health insurance data]. / Aufdeckung von Arzneimittelrisiken nach der Zulassung : Methodenentwicklung zur Nutzung von Routinedaten der gesetzlichen Krankenversicherungen.

Adverse drug reactions are among the leading causes of death. Pharmacovigilance aims to monitor drugs after they have been released to the market in order to detect potential risks. Data sources commonly used to this end are spontaneous reports sent in by doctors or pharmaceutical companies. Reports alone are rather limited when it comes to detecting potential health risks. Routine statutory health insurance data, however, are a richer source since they not only provide a detailed picture of the patients' wellbeing over time, but also contain information on concomitant medication and comorbidities.To take advantage of their potential and to increase drug safety, we will further develop statistical methods that have shown their merit in other fields as a source of inspiration. A plethora of methods have been proposed over the years for spontaneous reporting data: a comprehensive comparison of these methods and their potential use for longitudinal data should be explored. In addition, we show how methods from machine learning could aid in identifying rare risks. We discuss these so-called enrichment analyses and how utilizing pharmaceutical similarities between drugs and similarities between comorbidities could help to construct risk profiles of the patients prone to experience an adverse drug event.Summarizing these methods will further push drug safety research based on healthcare claim data from German health insurances which form, due to their size, longitudinal coverage, and timeliness, an excellent basis for investigating adverse effects of drugs.

The Choice of Transcatheter Aortic Valve Implementation (TAVI): Do Patient Co-morbidity and Hospital Ownership Type Matter?

BACKGROUND: Innovative technologies challenge healthcare systems, as evidence on costs and benefits frequently usually are slow to reflect new technology. We investigated these dynamics for Germany, using the emergence of transcatheter aortic valve implementation (TAVI) as an alternative to conventional aortic valve replacements (CAVR). OBJECTIVE: We focused on the role of patient co-morbidity-which would be a medical explanation for adopting TAVI-and hospital ownership status, hypothesizing that for-profit facilities are more likely to capitalize on the favorable reimbursement conditions of TAVI. METHODS: The analysis uses claims data from the Techniker Krankenkasse, the largest health insurance fund in Germany, for the years 2009-2015, covering 2892 patients with TAVI and 9523 with CAVR. The decision on TAVI versus CAVR was estimated for patient-level data, that is, socioeconomic data as well as co-morbidity. At the hospital level, we included the ownership type. We also controlled for effects of the respective owner (rather than the type of ownership), including a random intercept. RESULTS: While the co-morbidity score of TAVI patients was much higher in the early years, over time, the score almost converged with that of CAVR patients. This is in agreement with emerging evidence that suggests the use of TAVI also leads to better patient outcomes. Our results indicate that the type of ownership does not drive the switch to TAVI. We found little, if any, effect from the respective owner, regardless of ownership type. CONCLUSION: Overall, the effects of co-morbidity suggest that providers acted responsibly when adopting TAVI while evidence was still emerging.

[Prescribing valproate to girls and women of childbearing age in Germany : Analysis of trends based on claims data]. / Valproatverordnungen bei Mädchen und Frauen im gebärfähigen Alter in Deutschland : Untersuchung zeitlicher Trends basierend auf Versichertendaten.

BACKGROUND: Measures to raise awareness of the teratogenic potential of valproate and restrict its use in girls/women of childbearing age have been intensified. For Germany, the impact of these measures on valproate prescription rates remains unknown. OBJECTIVES: Trends in prescribing valproate, the underlying treatment indication, and the specialty of the prescribing physician are analyzed. MATERIALS AND METHODS: With claims data from several statutory health insurance providers from 2004 to 2016 (approximately 3.5 million insured persons per year) considering treatment indication and medical specialties of prescribing physicians, we assessed the rate of girls/women (12 to 50 years) with at least one valproate dispensation per year. RESULTS: The age-standardized rate of girls/women with at least one valproate dispensation declined by 28% between 2004 and 2016 (2.91/1000 vs. 2.09/1000). For 2015, the indications were epilepsy (66.9%), bipolar disorder (13.6%), migraine/headache (5.6%), schizoaffective disorder (4.3%), and other mental disorders (8.9%). Among epilepsy patients, the proportion treated with valproate declined from 26.2 to 16.8%, but changed little in patients with bipolar disorder (9.3% vs. 8.0%). A total of 46.3% of valproate dispensations were issued by neurologists or psychiatrists and 29.6% by general practitioners, internal medicine specialists, or family doctors. CONCLUSIONS: Based on German claims data, a decline of valproate dispensations was shown for epilepsy patients of childbearing age, while the proportion in other indications has hardly changed since 2004.

Robust versus consistent variance estimators in marginal structural Cox models.

In survival analyses, inverse-probability-of-treatment (IPT) and inverse-probability-of-censoring (IPC) weighted estimators of parameters in marginal structural Cox models are often used to estimate treatment effects in the presence of time-dependent confounding and censoring. In most applications, a robust variance estimator of the IPT and IPC weighted estimator is calculated leading to conservative confidence intervals. This estimator assumes that the weights are known rather than estimated from the data. Although a consistent estimator of the asymptotic variance of the IPT and IPC weighted estimator is generally available, applications and thus information on the performance of the consistent estimator are lacking. Reasons might be a cumbersome implementation in statistical software, which is further complicated by missing details on the variance formula. In this paper, we therefore provide a detailed derivation of the variance of the asymptotic distribution of the IPT and IPC weighted estimator and explicitly state the necessary terms to calculate a consistent estimator of this variance. We compare the performance of the robust and consistent variance estimators in an application based on routine health care data and in a simulation study. The simulation reveals no substantial differences between the 2 estimators in medium and large data sets with no unmeasured confounding, but the consistent variance estimator performs poorly in small samples or under unmeasured confounding, if the number of confounders is large. We thus conclude that the robust estimator is more appropriate for all practical purposes.

Guidelines versus reality: is coronary stent application in three-vessel disease standard or the exception?

The national guidelines for treatment of chronic coronary heart disease (CHD) recommend surgical coronary aortic bypass grafting (CABG) rather than percutaneous coronary intervention (PCI) for patients with a coronary three-vessel disease. The epidemiology of three-vessel CHD and data about the application of different revascularisation strategies raise suspicion of deviation from the guidelines in the treatment of those patients. Claims data containing records of almost 10 million patients of the largest German statutory health insurance fund (Techniker Krankenkasse) were utilised to measure adherence to the guidelines for treatment of groups of patients with one-, two-, and three-vessel CHD, respectively. The impact of age, sex, and comorbidity on each patient's revascularisation procedure was investigated as well. There was no significant difference in the rate of PCI between the groups. In conclusion, the hypothesis that patients with a coronary three-vessel disease are not always treated according to the recommendations of the national guidelines could not be disproved by this study. Finally, the results of this study suggest that the best revascularisation strategy for each patient with two- and three-vessel disease should be decided upon by an interdisciplinary discussion between both cardiologists and cardiac surgeons.

Health system responsiveness and chronic disease care - What is the role of disease management programs? An analysis based on cross-sectional survey and administrative claims data.

Health system responsiveness is an important aspect of health systems performance. The concept of responsiveness relates to the interpersonal and contextual aspects of health care. While disease management programs (DMPs) aim to improve the quality of health care (e.g. by improving the coordination of care), it has not been analyzed yet whether these programs improve the perceived health system responsiveness. Our study aims to close this gap by analyzing the differences in the perceived health system responsiveness between DMP-participants and non-participants. We used linked survey- and administrative claims data from 7037 patients with coronary heart disease in Germany. Of those, 5082 were enrolled and 1955 were not enrolled in the DMP. Responsiveness was assessed with an adapted version of the WHO responsiveness questionnaire in a postal survey in 2013. The survey covered 9 dimensions of responsiveness and included 17 items for each, GP and specialist care. Each item had five answer categories (very good - very bad). We handled missing values in the covariates by multiple imputation and applied propensity score matching (PSM) to control for differences between the two groups (DMP/non-DMP). We used Wilcoxon-signed-rank and McNemar test to analyze differences regarding the reported responsiveness. The PSM led to a matched and well balanced sample of 1921 pairs. Overall, DMP-participants rated the responsiveness of care more positive. The main difference was found for the coordination of care at the GP, with 62.0% of 1703 non-participants reporting a "good" or "very good" experience, compared to 69.1% of 1703 participants (p < 0.001). The results of our study indicate an overall high responsiveness for CHD-care, as well for DMP-participants as for non-participants. Yet, the results also clearly indicate that there is still a need to improve the coordination of care.

Initiation and duration of dual antiplatelet therapy after inpatient percutaneous coronary intervention with stent implantation in Germany: An electronic healthcare database cohort study.

BACKGROUND: Studies assessing the routine outpatient dual antiplatelet therapy (DAPT) after percutaneous coronary intervention (PCI) in Germany are scarce. The aim of this study was (i) to investigate the initiation and duration of DAPT after inpatient PCI with stent implantation in Germany, and (ii) to identify factors associated with DAPT discontinuation during the recommended treatment period. METHODS: This retrospective cohort study was based on data from a large German electronic healthcare database of the years 2004 to 2009. The study population comprised four groups of patients with acute coronary syndrome (ACS) or stable angina pectoris undergoing inpatient PCI with either bare metal stent (BMS) or drug eluting stent (DES) implantation between 2005 and 2008. Initiation of outpatient DAPT within a period from 100 days before the PCI to 60 days after the PCI was ascertained. Time until end of treatment was analysed using the Kaplan-Meier method. Factors potentially associated with DAPT discontinuation, like sex, age, cardiovascular comorbidity, contraindications, and other antithrombotic drugs were analysed in a Cox proportional hazard model. RESULTS: The cohort comprised 37,001 patients. Depending on the type of stent and the indication for the PCI, DAPT was initiated in 85 % (ACS/BMS) and 95 % (AP/DES) of all patients. Of those, 12 % (AP/DES) and 64 % (ACS/BMS) discontinued DAPT during the recommended treatment duration. An age of over 80 years (OR 1.2-1.5 compared to patients aged 0-49 years) and the use of phenprocoumon (OR 2.7-5.0 compared to no phenprocoumon) were associated with an increased risk of DAPT discontinuation. CONCLUSIONS: A high proportion of patients with coronary artery disease undergoing inpatient PCI with stent implantation received DAPT. However, DAPT discontinuation during the recommended time span was frequent, particularly in patients suffering from ACS. On the other hand, especially patients with AP and DES were often treated longer than recommended.