Complex statistical calculations are a prerequisite
Meet Morten Overgaard
“I learn a little more about something almost every day, and that encourages me to believe I’m doing the right thing.”
How and why do statistical methods work? And can we develop new, suitable methods for specific situations? These are some of the questions that interest Associate Professor Morten Overgaard from the department’s Biostatistical Research Unit.
Complex statistical calculations are a prerequisite for valid analyses of health data
“It is important for research in health sciences to have statistical methods that provide easily interpretable results that can be expected to hit the target and make the best possible use of the available data. In this way, my work is about securing the tools and a correct foundation for other researchers to be able to deliver results with a more direct societal impact. In slightly popular terms, you could say I play my colleagues a good hand by ensuring the quality of their methods and data,” explains Morten Overgaard.
He is mainly concerned with methods of handling survival data, which consist, amongst other things, of investigating whether a patient’s prospects for survival are better with one treatment than another.
“To give an example, we could take as our point of departure the concept of censoring. When you wish to say something about survival, it is often a central problem that as a researcher, you can only follow the study participants during a particular period, not until they are all dead,” explains Morten Overgaard, and elaborates:
“The concept of censoring covers, amongst other things, the situation in which the researcher knows that a participant has survived for a certain period of time, without knowing when the participant actually ends up dying afterwards. And this is where my colleagues and I enter the picture. The correct handling of this and similar problems is crucial for the ability of my research colleagues to get something useful out of their analyses. So we help them to assess what can be said about survival in such a situation – what are, for example, good measures? And how can they be estimated with this type of data?”
Particularly fascinated by survival analysis
Morten Overgaard wishes to achieve an even deeper understanding of a number of statistical topics, e.g. within the field of survival analysis, which he can subsequently share with others.
“It is my impression that there is a degree of confusion about the necessary prerequisites for certain methods of survival analysis, and I would like to contribute to combating this confusion,” he says, and explains:
“It is a question of evaluating the situations in which the methods can be expected to produce reasonable results and work as intended. If, for example, the prerequisites for using the method have not been met, then there is a risk of bias and of drawing the wrong conclusions from the analyses.”
This is precisely what Morten Overgaard has been working on both during and after his PhD studies, during which he succeeded in contributing to the understanding of the prerequisites and generality of the so-called pseudo-observation method.
Morten Overgaard was born in 1987, and he has a background in mathematics and (theoretical) statistics. He is particularly pleased that his position at the Department of Public Health allows him to cultivate the more theoretical understanding of methods together with their specific applications, as it is very much the specific applications that motivate the theoretical work of Morten Overgaard.