The Centre for Biostatistics provides support for researchers in Health Sciences in the Dunedin Campus of the University of Otago.
We support Health Sciences researchers with the highest quality biostatistical advice and research collaboration. All our biostatisticians are academic staff who have expertise in health sciences disciplines and are experts in research methods. Contact us with your request using our form:
Once you have submitted your request you can expect to be contacted by one of our biostatisticians within 10 working days. If you are working to a deadline please include this information in your request. We would strongly encourage you to discuss your project with us earlier rather than later to ensure the statistical design and analysis is of high quality. We will do whatever possible to meet deadlines, however we may not have capacity to prioritise last-minute requests.
- Visit Our people to learn more about our staff and their expertise
We follow the ethical guidelines published by the American Statistical Association. Please read these guidelines before you meet with your biostatistician as they contain important information on what you can expect when working with us.
The biostatisticians (as a group) have enjoyed long standing collaborations with people in the Health Sciences Division for many decades. Our core business is collaborating with and providing advice to researchers in health related areas. We welcome new collaborations with Health researchers.
We have developed a checklist to help clarify how people should engage with the Centre.
Checklist for researchers collaborating with the Centre for Biostatistics (PDF)
Over the last two years the Centre for Biostatistics has been clarifying and setting in place policies and procedures in order to deliver on its vision to be world leaders in providing biostatistical collaboration and advice for health related research.
Centre for Boistatistics vision and mission statements
We have noticed that many students seeking our help do not have sufficient statistical knowledge to obtain a high-quality solution. Therefore, we are making it mandatory for all students to undertake our Introduction to Biostatistics for Clinical Researchers course before contacting us.
Sign up for the Introductory Biostatistics for Clinical Researchers course
We have also noticed an increased number of undergraduate or honours students approaching us – in these cases the supervisor needs to approach us, not the student. This is important to ensure the right information is obtained from the supervisor who has responsibility for these projects.
If you have any queries or concerns, please contact us:
If you are a research student seeking a biostatistical supervisor for your research project, please make this clear when you submit your request form.
Where we are not supervisors for your research project, please note we have a firm policy that we do not undertake analyses. What we can offer you is up to 2 hours of advice on design and/or analysis per year. We will be offering a range of opportunities for students to support your own statistical development.
We offer a range of short courses on biostatistics and data analyses. This includes courses on sample size estimation, basic biostatistics for health researchers, data management and formatting of messy multiple datasets to suit particular analyses, and use of Stata software. Our course duration varies from a couple of hours to a few days depending on the course. We are currently developing courses on some other areas too, and the details will be updated here from time to time. We are also happy to receive suggestions on other topics you would like to be covered in future courses.
For more information:
Short courses, forums and tools
- Information about upcoming short courses (Health Research South website)
- More about our Monthly Biostatistical Forums
- Biostatistics tools, tips and guides
Due to the strong demand for our assistance we prioritise new projects in the following order:
- Funded grants or grant proposals where there is funding for biostatistical support
- Collaborative research
- Research students
All new requests within these categories are followed up in the order that they are submitted. Existing projects are prioritised as negotiated with our biostatistician.
We follow the International Committee of Medical Journal Editors (ICMJE) authorship guidelines. In your first meeting with our biostatistician the appropriateness of authorship will be discussed and agreed upon, subject to renegotiation if the role changes substantially. In general we would not expect authorship to be given for brief advice or where the contribution is minimal.
Most collaborative projects would generally fit the requirement for authorship as our biostatistician will not only be heavily involved in the design and/or analysis but will also contribute to the drafting of the manuscript with the necessary interpretation of the statistical results.
Next forum: 22 October 2019
The Biostatistics team all set for the February forum: Ari, Andrew, Ella, Robin, Claire and Jill.
We invite all staff and students who have an interest in biostatistics to attend our Biostatistics Forum. The Forum is a regular discussion group which focuses on published articles, particular methods, and issues in biostatistics. We have a 20-minute presentation with plenty of time for discussion.
We meet monthly, usually on the third Tuesday, from 4.00 to 4.50pm in room 033/036, ground floor, Adams Building, Frederick Street, Dunedin.
We are also keen to hear from staff or students who would be interested in presenting a session.
2019 Forum calendar
We meet from 4.00 to 4.50pm in room 033/036, ground floor, Adams Building, Frederick Street, Dunedin.
We are inviting all university staff and students who have an interest in biostatistics to attend. It is hosted by the Centre for Biostatistics in the Division of Health Sciences. We are also keen to hear from anyone (staff or students) who would be interested in leading a session.
Remaining 2019 forum dates:
- 26 November
22 October 2019 – Understanding statistics from overheard conversations
Although online communities don’t always bring out the best in people, they can be very edifying places to hang out. In my work, I have greatly benefited from blogposts and online discussions, even without taking part myself. In this talk, I discuss some of the things I have learned about statistics from listening in to online conversations between statisticians, especially at datamethods.
Matt Parry, Senior Lecturer; Associate Dean (International), Mathematics & Statistics; Sciences Divisional Office) presented this forum.
17 September 2019 - Redefine Statistical Significance
A group of researchers are calling for our level of significance to be reduced from 0.05 to 0.005 for new findings. I was all on board until I dug a bit deeper into the paper. Maybe instead we ought to propose that statisticians receive journalistic training so that both sides of an argument get a fair hearing...
Lisa Avery from the Department of Mathematics and Statistics lead the discussion.
Read the research article, Redefine statistical significance.
20 August 2019 Is it always necessary to have a representative sample?
People often think that having a representative sample is integral to scientific research. In this month’s biostatistics forum, we will discuss what does representativeness mean? Is it always necessary? The following two papers were referred to in this presentation by Jimmy Zeng:
Why representativeness should be avoided.
Commentary: Representativeness is usually not necessary and often should be avoided.
23 July 2019 When should we give up?
Interim analyses are planned for many clinical trials and some observational studies, looking for evidence about effects or associations before the sample size target is achieved. A particular type of interim analysis looks at futility–asking whether the study is likely to be able to achieve its objectives when all data are eventually in. While most often seen in clinical trials, such interim analyses for futility can be used for any study design where data is gathered incrementally. We will discuss what futility interim analyses are, when they might be appropriate (if ever), how they could be done, and what implications they could have for study findings when they are done. A short pro/con debate was published some time ago at https://doi.org/10.1186/cc3013 and we will use this to start the discussion. While the debate uses a scenario involving clinical trials in critical care medicine, the issues extend much more broadly and should interest anyone who has ever asked if it’s worth continuing to collect data for their study.
Senior Research Fellow Andrew Gray lead this discussion.
Expanding on Associate Professor Robin Turner’s discussion last month on sample size calculations, Dr Ella Iosua will lead this month’s forum on the inherent challenges with small samples and how these can potentially be negotiated statistically.
In a number of disciplines such as clinical psychology and neuroscience, sample sizes are very limited ( n<30 but often n<10). This has important implications for the appropriateness of the statistical analysis which is predominantly predicated on large sample theory. Subsequently, do small samples immediately discredit the research completely or is there still some merit in continuing with such prevalent practices?
Use of proper statistical techniques for research studies with small samples is a paper that introduces some of the issues succinctly, and will be used as a starting point to facilitate discussion.
As a biostatistician one of the most commonly asked questions we face is “How many do I need?”. Whilst it is a seemingly simple question to work out the minimum sample size, it is however, based on a series of guesses that may or may not reflect reality. This can raise questions about whether it was useful to calculate in the first place. The following paper Breaking Free of Sample Size Dogma to Perform Innovative Translational Research summarises some of these challenges and we discussed these along with the perspective we have from being on the other side of the desk for sample size questions. Associate Professor Robin Turner led this discussion.
Missing data plagues health research. Should we ignore it or impute it? Or do something else that sounds statistically complex? Dr Jill Haszard will lead a discussion about these issues, with reference to the following paper: Statistics Notes: Missing outcomes in randomised trials.
March 2019 forum - Sharing data from clinical trials
The second Biostatistics Forum was facilitated by Dr Ari Samaranayaka. Attendees shared their views in a freely and mutually helpful way on the topic of sharing data from clinical trials.
A recent editorial in the Bulletin of the World Health Organisation, that was also simultaneously published in dozens of other health research journals, says the International Committee of Medical Journal Editors (ICMJE) believes there is an ethical obligation to responsibly share data generated by interventional clinical trials. As the first step of meeting this obligation, they have specified minimum standards on data sharing for manuscripts submitted to journals and for applications for trial registrations.
Attendees discussed the pros and cons of sharing individual level data with wider research community. The views encompass benefits of information sharing, related ethical and practical issues, and management structures required to eliminate possible inappropriate use of data.
WHO Bulletin editorial (PDF 110 KB)
February 2019 forum - p-values
On 19 February we introduced the Biostatistics Forum and discussed the paper from the American Statistical Association: The ASA's Statement on p-Values: Context, Process, and Purpose.
Participants from a range of disciplines soak up the expertise at the February forum.