Within research projects, the BfArM also awards bachelor's, master's and doctoral theses - in cooperation with the University of Bonn, but also with other universities. The research projects focus on important and current research topics thematically related to the tasks of the BfArM.

Julia Wicherski is a doctoral candidate at BfArM, where she is researching the serious side effects associated with the use of fluoroquinolones (drugs from the group of synthetic antibiotics). Her goal is to determine more precisely the factors that lead to the occurrence of these side effects. In the following article, she describes why real-world data play a crucial role here.

Within the framework of my doctoral studies in drug sciences, the BfArM offers me the opportunity to be part of the interdisciplinary team of the department of pharmacoepidemiology in the Research Division. There I work as a research assistant in the implementation of the project "FQrisk". With FQrisk, we aim to contribute to the evidence of serious side effects, so-called adverse drug reactions (ADRs), resulting from the use of fluoroquinolones (FQs) in routine care. The results of the project can contribute to the drug therapy safety of FQ and comparable antibiotics.

Fluoroquinolones are synthetic antibiotic agents with a broad spectrum of activity and are used in particular for serious bacterial infections. Besides this antibiotic effect, side effects which may affect multiple organs or organ system classes have been observed. For example, these side effects include tendon ruptures or peripheral neuropathies.
Since many of these side effects are long-lasting and possibly even irreversible, a comprehensive European risk assessment procedure was initiated by the BfArM in 2017, which ended with far-reaching restrictions on the approved areas of application, specification of the remaining indications, and new notes and warnings on the possible side effects in 2019.

Fluoroquinolones: highly effective antibiotics with risks

This issue was widely communicated to the public. Media reports described the dangers and risks posed by fluoroquinolones and critically questioned why these antibiotics may continue to be used despite the occurrence of severe side effects.

However, it is important to know that fluoroquinolones, as highly effective antibiotics with a broad spectrum of activity (including against Gram-negative and Gram-positive bacteria), still represent an important treatment option. Especially in cases where other antibiotics are not sufficiently effective or in cases where bacteria developed resistance to other antibiotics, fluoroquinolones are valuable drugs.

For optimal and safe treatment of patients with fluoroquinolones, it is important to look at physician prescribing practices in routine care. In this way, it is possible to investigate which factors play a role in the occurrence of serious and partly irreversible side effects. These findings can be used to improve drug therapy safety of fluoroquinolones. Data from medical routine care belong to the so-called Real World Data (RWD). They are generated in the real world under everyday conditions. In the form of anonymized health insurance billing data, these data from medical routine care can be used to analyze the use and specific risks of drugs.

"FQrisk" research project: Minimizing the risk of serious side effects

This is exactly what is happening in the "FQrisk" research project. We are investigating the association between the use of fluoroquinolones (FQ) and the occurrence of serious ADRs in routine care. The use of German Real World Data (RWD) will clarify and extend the results from the European risk assessment process and previous RWD studies. For example, due to the size of the study sample , we will be able to detect rare ADRs with good statistical power, describe differences between multiple comparator antibiotics, and analyze ADRs which occur over a longer follow-up period. Moreover, based on the European risk assessment procedure, modifying factors on the occurrence and severity of ADRs and on the use of health care services will be analyzed. This approach makes it possible to contribute to patient safety in order to be able to minimize the risk of serious ADRs resulting from an FQ prescription.

BfArM: Regulatory research and optimal conditions

The BfArM offers the optimal conditions for this research project, as it is one of the few regulatory drug authorities with its own research department in Europe. At the BfArM, different interdisciplinary research groups cooperate with leading national, European and international universities and non-university research institutions. The regulatory research of the BfArM includes the investigation of public health-relevant, current research questions regarding the approval and safety of drugs. A particular focus is on the pharmacoepidemiological evaluation of large, population-based data sets (RWD). "Real world" evidence-generating studies are increasingly central to regulatory decision-making processes.

Based on my studies in health sciences/public health, I was already familiar with the research activities of the BfArM. I became interested in clinical epidemiology very early in my bachelor studies, so that I dedicated my bachelor thesis to a topic of descriptive clinical epidemiology related to pharmaceuticals. For my master's thesis, it was important for me to expand my methodological knowledge, so I decided to conduct a systematic review and meta-analysis. In doing so, I became aware of the scientific discourse on the validity of clinical trials and RWD studies. At this point, it was clear to me that I would like to work on a topic of clinical epidemiology within the framework of my doctoral studies and thereby, I would like to learn more about the significance of clinical trials and RWD studies, and actively contribute to this discussion in a results-oriented manner. Therefore, I applied for a position as a PhD student in the research department at the BfArM. I already kept an eye on the job advertisements from the BfArM while writing my master’s thesis. Finally, I became aware of the suitable position through the job portal of the German Society for Epidemiology (DGEpi). I am very happy to be a research associate and doctoral student at the BfArM since November 2019.

Since then, I have been working on the research project FQrisk in the research group Pharmacoepidemiology of Prof. Britta Hänisch. She is the head of the Research Division at the BfArM as well as of the two Pharmacoepidemiology research groups involved in the project, at the BfArM and at the German Center for Neurodegenerative Diseases (DZNE) in Bonn.

Scientific Institute of the AOK provided access to suitable data

The idea and motivation for the project stemmed from the previous regulatory events surrounding fluoroquinolones. The already existing long-standing cooperation of the research group Pharmacoepidemiology with the Scientific Institute of the AOK (WIdO) provided access to the appropriate RWD, so that I could directly start with the project.

At the beginning of such a research project, there is always the development of an adequate study design. The network of epidemiological, pharmacological and clinical expertise plays a special role in the successful implementation of the project. In the interdisciplinary project team, colleagues from WIdO provide routine claims data from the AOK (a large German statutory health insurance) and contribute their professional expertise to project planning and implementation.

The perspective of clinical practice is represented by two medical experts from the fields of infectious diseases and general medicine. Methodological, data-analytical and pharmacological expertise as well as experience in the implementation of pharmacoepidemiological projects are contributed to the project by the team of the research groups Pharmacoepidemiology at the BfArM and at the DZNE.

The members of the research groups regularly participate in various (inter-) national training courses and conferences. Thus, we constantly learn new skills, expand our existing knowledge, are in exchange with other scientists, can develop new networks and integrate the latest knowledge into our work. In addition, Prof. Britta Hänisch involves us in her activities in various (inter-) national working groups, for example on the topic of RWD. Thus, she enables me as a doctoral student to gain challenging insights and further knowledge for my further scientific career, which last but not least has a direct positive impact on the elaboration of the PhD project.

Goal: Increase therapy safety

Back to the project - The core of my work is the application of epidemiological and statistical methods for the analysis of AOK health insurance data. These data include, among others, information on prescribed drugs and medical devices, diagnosed diseases and symptoms, and performed treatments and surgeries. Therefore, health insurance data are a valuable RWD source for deriving insights for regulatory decision-making processes. Regarding FQrisk, evidence will be generated to increase drug therapy safety of this antibiotic class in routine care. It is important to note that data privacy and the principle of data parsimony take a central role. Accordingly, the analysis of these data is limited to variables relevant to the analysis with the highest possible degree of aggregation. It is not possible to draw conclusions about the insured individuals.

Prior to the actual data analysis, we carry out a comprehensive data quality control . An example of this is the plausibility check. This means that we describe the absolute and relative distributions of the variables in detail and check them for plausibility and biological explainability. An example of this is the age distribution of the persons insured. If we consider only adult insurants in our data set, it would be implausible if someone in the data set were only 15 years old. Likewise, the age of 150 years would be implausible.

Results are also prepared for technical laypersons

After data quality control, preparations for multiple data analysis take place. These preparations include checking all assumptions for the selected analysis methods. If the defined assumptions for the correct implementation of the special method are fulfilled, the models are programmed in our statistical programs. I myself use the statistical program R for this purpose. As a quality assurance measure of our work, we aim to program our analyses in at least two different applications in the research group. For example, the statistical program SAS is also suitable for pharmacoepidemiological analyses of large data sets. To check the robustness of our results, we perform additional subgroup and sensitivity analyses of the data. All results are discussed within the project team. Finally, the project and the results found, including critical discussion, are summarized in a manuscript and submitted for publication in a peer-reviewed journal. It is common practice to publish in peer-reviewed journals. This means that the article submitted is presented to different uninvolved experts as reviewers. They have the opportunity to send us comments and requests for additions, which we take into account before the journal publishes our article.

In addition to distributing our results to other scientists and decision-makers in the health care system, we also aim to prepare and communicate our results in a clear way to non-specialist experts. Therefore, it remains interesting for everyone to follow the project and the research activities of the BfArM in general.

Julia Wicherski

2012-2014: Vocational baccalaureate for health and social affairs, Berufskolleg des Kreises Kleve, Germany
2014-2017: B.Sc. Health Sciences at the HAW Hamburg, Germany; Bachelorthesis "Prevalence and incidence of gastritis and heartburn and their drug therapy in the Heinz Nixdorf Recall Study" in cooperation with the Center for Clinical Epidemiology (ZKE) in Essen, Germany
2017-2019: M.Sc. Public Health at the University of Bielefeld, Germany, Master thesis "Association between Breakfast Skipping & Body Weight - A Systematic Review & Meta-Analysis" in cooperation with the German Diabetes Center (DDZ) in Düsseldorf, Germany
Since 11/2019: Doctoral student at the BfArM and doctoral student in pharmaceutical sciences at the University of Bonn, Germany
Since 11/2020: Member of the speaker team of the junior research group epidemiology in the DGEpi