Why public health professionals need biostatistics
May 14, 2018 10:51 AM
More data or big data will not alone provide us with the answers to important and complex medical and public health questions, says Lisa Sullivan, Associate Dean for Education and Professor of Biostatistics at the Boston University School of Public Health.
If you have biostatistics skills, your job prospects look good, says Lisa Sullivan. She is the Associate Dean for Education at the Boston University School of Public Health and the author of Essentials of Biostatistics in Public Health, and she has been teaching biostatistics for nearly 25 years.
Last year, she debuted a one-week intensive course at BU on the essentials of biostatistics for incoming Master of Science students who needed a foundation in biostatistics. “I chose JMP because it really worked with the course – JMP is easy to learn; it is set up in a way that is very intuitive, yet very powerful,” she says.
I asked her a few questions about biostatistics and the role that data visualization and analytics can play in an evidence-based approach to public health.
What are some trends you are seeing in the field of biostatistics for public health professionals?
The job outlook for biostatistics is excellent, and I suspect will only get stronger. That said, successful biostatisticians (and people who work with biostatisticians) need solid skills in the foundational concepts and applications of biostatistics – more data or big data will not alone provide us with the answers to important and complex medical and public health questions.
What are the primary challenges to implementing an evidence-based approach to public health?
In my opinion, the biggest challenge is understanding how to address bias and threats to validity in statistical and epidemiological analyses.
What kinds of public health change are possible through the use of data and statistical results?
Data can be turned into information and knowledge with careful statistical design and analysis. I think most people favor data/evidence-based decision making, but that hinges on good data and solid statistical analysis.
How do you see analytics and data visualization facilitating conversations with public health professionals?
Data visualization is absolutely critical! Presenting data and statistical results to other scientists at professional conferences and in peer-reviewed publications is important but totally insufficient for our research to have impact. We have to make data and statistical results available and accessible to much wider audiences, and effective data visualization is a very important way to do this.
We developed this online course in response to a call from our alumni and others working in public health who want a grounding in biostatistics or who want to hone their skills.
What made you decide to teach with JMP? What do you like about it?
We chose JMP because it is an extremely flexible and powerful tool that will give students a new skill that they can apply to advance their careers. I love the way JMP labels variable types (as we do in the course and in practice) and also directs students toward the appropriate analyses for specific settings. I always feared point-and-click statistical software (fearing that students could easily run the wrong analysis), but JMP is different – it directs the user/investigator but still requires the user/investigator to think through the question and approach to solving statistical problems.
How does a tool like JMP add value for today’s public health professional in epidemiology, clinical research, health care, etc.?
Statistical computing is a skill in tremendous demand by employers in most sectors. Statistical computing with JMP is easy to learn – extremely well-supported with informative trainings, videos and examples – and very powerful. JMP offers solutions for today’s statistical issues and has extremely powerful options for data visualization. JMP guides the user/investigator toward appropriate solutions and readily creates presentation-ready graphics that convey complex messages in data.