Biomedical Data Science
Research & Innovation > Research management > Biomedical Data Science

Biomedical Data Science

We offer a broad spectrum of biostatistical expertise, from clinical research studies to a variety of application-oriented methodological issues. You are invited to contact us if you are:

  • employee or student of Paracelsus Medical Private University (PMU)
  • member of the University Hospital Nuremberg
  • member of the University Hospital Salzburg (SALK)
  • external scientists and clinicians of any university on a national or international level.

We present two teams "Biostatistics and Publication/Planning of Clinical Studies" and "Biostatistics and Big Medical Data", each with a comprehensive content focus, and can accordingly offer the following research collaborations (details on how to contact us are given with the respective topics).

Scientific cooperation for publication and analysis of medical/clincial studies

We make time for you!

During a non-binding initial meeting, we will jointly discuss the next steps in your clinical research study, how to proceed towards publication of your research results.

If research data are already available, we offer professional data analysis of your research questions using state-of-the-art biostatistical and data analytical methods.

This includes the data preparation, data analysis and evaluation, discussion of results, preparation of publication- and print-ready figures, the description of the statistical methods in the paper, as well as the defense towards reviewers in all research journals.

We are also at your disposal for the planning of the study, including a priori or a posteriori sample size computations, ethics committee applications, randomization plans, modern study designs, matchings and all other methodological aspects.

The goal is the publication of your research results in the worldwide best medical journals.

Please find a detailed list of ongoing or past publications of clinical studies in all areas in medicine which comprise more than 220 publications!

A detailed list is given here: https://orcid.org/0000-0002-7696-1479 

For more information on the process and contact, please visit our comprehensive website:

We are at your disposal for all your questions!

Contact:

PD MMag. Dr. Wolfgang Hitzl
Head of Team Biostatistics and Publication of Medical Studies; Evaluation of medical studies/Machine learning

Research Management & Technology Transfer, Paracelsus Medical Private University Salzburg
Strubergasse 16
A-5020 Salzburg, Austria

Phone: +43 699 14420032
E-mail: wolfgang.hitzl@pmu.ac.at
ORCID: https://orcid.org/0000-0002-7696-1479

Modern sample size planning, ethics committee applications, randomization plans, study designs

Do you need a sample size planning?

How many patients should be included to achieve sufficient power to find clinically relevant effects or differences?

During an initial meeting, we will jointly discuss the next steps in your medical study and how we can answer the above question.

For example, the following points:

  • Modern sample size planning (a priori sample size computations also with Monte Carlo simulations).
  • Discussion of possible candidates of primary endpoints for your clinical study and extensive discussion of pros and cons of various number sample size scenarios.
  • Full support of your ethics committee applications, especially submission of proposal, supplement of statistical methods sections.
  • Clinical trial randomization: randomization plans to prevent bias to maximize power and minimize selection and allocation bias.
  • Matching algorithms of study groups, e.g. prospensity score matching algorithms.
  • Consulting and discussion of modern professional study designs, e.g. adaptive designs, group sequential designs, internal study pilot designs, two-stage adaptive designs.

Our goal ist is to plan and design your clinical study right from the start in such a way that the hypotheses can be proven with maximum power!

We are at your disposal for all your questions!

For more information on the process and contact, please visit our extensive websiteBiostatistics and publication of clinical research studies:

Contact:

PD MMag. Dr. Wolfgang Hitzl
Head of Team Biostatistics and Publication/Planning of Medical Studies
Evaluation of medical studies/Machine learning

Research Management & Technology Transfer, Paracelsus Medical Private University Salzburg
Strubergasse 16
A-5020 Salzburg

Phone: +43 699 14420032

E-mail: wolfgang.hitzl@pmu.ac.at

Big Medical Data, Machine Learning und Artifical Intelligence

In the medical field, there is a treasure trove of data (registries, examination devices, health apps, etc.) waiting to be mined - but of course in compliance with methodologically sound principles and "good statistical practice"! 

In the Big Data context (- but not only there! -), methods from the fields of Machine Learning or Artificial Intelligence are often used. In "Machine learning/AI", modern algorithms generate a statistical model based on training data, which is then carefully tested. Ideally, particular structures in the data are identified in this process. These algorithms can be applied to the analysis of demographic and clinical routine data, but also, for example, to automated diagnostic procedures for imaging data (e.g. "image classification/image regression"), or to reveal structures in such data (e.g. "medical image segmentation"). We use different modern methods from the fields of "deep learning" or more generally "machine learning" (e.g. support vector machines, random forest models, Bayes classifiers, classification trees, etc.).

Here are some typical examples from practice:

  • Ophthalmology: detection rates of vessels, exudates and hemorrhages in patients with diabetic retinopathy using neural networks.
  • Internal medicine: prognosis of pulmonary embolism and ventricular hypertrophy, myocardial infarction.
  • Pulmonology/respiratory medicine: prognosis of whether a patient can be successfully weaned from mechanical ventilation.
  • Urology: prostate screening.
  • Dermatology: neural networks to predict the occurrence of melanoma.
  • Epileptic seizure prediction or algorithm-based automated epilepsy diagnostics.
  • And many more...

The technical possibilities are enormous – but, most importantly, these research questions have to be tackled with a methodologically sound approach that takes clinical needs into account!

Contact: Georg Zimmermann (Head of Biostatistics and Big Medical Data)

Method development (e.g. small case numbers) and systematic reviews / meta-analyses

Are you active in a research area where you are frequently confronted with certain methodological challenges and you consequently want to build specific statistical-methodological expertise? Of course, you don't need to reinvent the wheel every time. But sometimes there are seemingly simple problems, but after a thorough methodological examination it turns out that new (statistical) methods have to be developed to ultimately solve the application problem satisfactorily.

You find it difficult to judge whether the concrete problem in your medical research field is "interesting enough" for such a methodological investigation? Then please contact us! We have already conducted several projects where we have been working on particular methodological problems that arose from medical datasets or clinical research questions (e.g. life expectancy comparisons, or methods for small sample sizes), but we are also open to new, exciting problems from your field of application!

Likewise, we are generally happy to support you in projects that require extensive, long-term methodologically sound supervision. This is especially true for systematic reviews and/or meta-analyses, where we are happy to offer methodological support based on our experience (e.g., Georg Zimmermann is a member of the Core Group Guideline Development in the European Reference Network EpiCare).

 

Contact: Georg Zimmermann (Head of Biostatistics and Big Medical Data)

 

 

Contact

For us, the personal conversation has priority, because potential misunderstandings, questions, individual goals and concrete forms of cooperation can best be clarified in an open, joint discussion, as experience has shown! Therefore, the mode of getting in touch with us is kept as simple as possible - the few points you should please note are summarized under the following links:

We look forward to the exchanging ideas with you!