The biostatistics section at the Department of Public Health helps to analyse and interpret the data gathered by the researchers at AU Health. We handle both clinical data and data extracted from Denmark’s many public-health registers, such as the National Prescription Database and the National Patient Register. These electronic resources supply information that is essential for comprehensive population-wide studies.


Statistical analysis is used to get a grip on the large amounts of data that researchers assemble to shed light on a particular topic, and it indicates:

  • What information is significant?
  • What is not significant?
  • Can many numbers be described in few numbers, without losing important information?

Biostatistics primarily deals with the statistical methods used in the health sciences and biology. Our field develops new methods, assesses their characteristics, and applies them to help resolve new, complex problems.


A team player for the whole faculty

Employees at the biostatistics section are involved in research projects at all of departments of the university’s Health faculty, contributing to more than 40 projects each year. These include two comprehensive population studies concerning the prevalence of autism and diabetes, respectively. We have also assisted with risk assessments that will be a significant element in an upcoming study of patients with prostate cancer.

Data sources are never static. New and improved measurement methods are constantly being introduced, and new and more detailed electronic health databases being established, and this means there is a consistent need for ongoing development of new statistical methods. With the rapid pace at which computing capabilities continue to grow, we are constantly pushing the limits of the demands our calculation methods put on processing power. We therefore continually optimize our skills in order to be at the cutting edge of the field – and we pass these new methods on: All PhD students at Health take a mandatory course in methodology and statistics as part of their PhD programme.



  1. Pharmacoepidemiology: We are developing better methods to determine the treatment duration, based solely on when a prescription was filled and the amount of medication supplied. Based on the Danish National Prescription Database.
  2. Autism: Studies of heritability, factors during pregnancy and delivery, and early symptoms relating to autism spectrum disorders (ASD). Based on data from public registers. These studies aim to help find the causes of ASD and contribute to early detection of the condition.
  3. Prostate cancer: Calculation of individual life-time risk for prostate cancer patients. Based on genetic profiling.
  4. Assistance to the Danish Medicines Agency: Our assistance is occasionally requested regarding the approval of new medication. We assess applicant compliance with the recommended statistical principles for clinical trials.



Demonstration that the increased prevalence of autism in Denmark is related to earlier diagnosing (Parner ET , et al. Arch Pediatr Adolesc Med. 2008;162:1150–6)

Detailed description of the occurrence of caries on dental surfaces among Danish children and adolescents (Parner ET , et al. Eur J Oral Sci. 2007;115:491–6)

Development and application of methods to assess selection bias in cohort studies (Nohr EA, et al. Epidemiology. 2006;17:413–8)

The diabetes epidemic of the 1990s shown to be caused not by obesity, but by fewer diabetics dying than the number of new diabetics being diagnosed (Støvring H, et al. Lancet. 2003;362:537–8)

Development on the mathematical model to describe the occurrence of exposurerelated cancer (Pierce D, Væth M. Biostatistics. 2003;4:231–48)



By using statistical models that reflect the health-science problem being examined, biostatistics makes it possible to distinguish biological variation, measurement errors, and other random variations from the systematic trends – the trends being what researchers are usually interested in identifying. To avoid misinterpretation, biostatistics focusses on diagnostic methods to reveal whether the parameters applied to a model are suitable in the specific context. These diagnostic methods are often graphical representations of noteworthy characteristics in the data. In constructing

appropriate models, it is a crucial biostatistical skill to be able to adapt one’s mindset to the relevant field and communicate the results in a form that is understandable and useful to other health researchers.

  • Mathematics: statistics is a mathematical discipline
  • Probability theory: uncertainties are described as probabilities
  • Calculations: statistical calculations are done using computers with special software
  • Graphical representations: a good chart is often the best way to describe results
  • Common sense: some statistical methods are simply systematized common sense
  • Communication: sharing knowledge amongst project participants ensures that analyses are relevant.