Using statistics to fill in the gaps

I have been really into watching medical reality tv shows over the past few years. While shows like 'Boston Med' do have a good dose of the reality tv sensationalism, they're also full of interesting medical statistics, that have resonance because you see the patient on the screen in front of you, who is up against those particular numbers. I find the statistics at times disquieting or disturbing - but mostly just extremely interesting! For example, did you know that if a tooth is knocked out and placed back in its hole within an hour, there is a 79% chance that the tooth will simply 'replant' itself? I definitely didn't!

...and, unfortunately, my personal experience with the medical system has born out that doctors themselves are often unaware of these statistics. They sometimes fail to act appropriate because they believe that a symptom is never associated with a particular condition, or that a disease always appears in a certain way, when in reality this may only be true 90% of the time (leaving out the experiences of a significant minority of patients).

Lifelines
Source: http://www.cs.umd.edu/hcil/temporalviz/

This is why I find the increasing use of data visualization software within the medical system so interesting, as it has the potential to use statistics in a way that allows medical professionals to understand what the numbers really are, so they can determine what they might mean for a patient. A simple, self-generating system like Lifelines can allow healthcare workers to explore patterns and trends across multiple health records and hospitals. This is the opposite of a random probability sample, as discussed by Luker, because it brings together all of the information that is available (since records have to be kept on each patient). I like to think of the 'sample' as being one single doctor's own experience with a condition (which may only be an experience of patients with textbook presentations), and the data visualization software as presenting all of the experiences of all the doctors in a given area.

Data visualization software therefore become a wonderful way to use statistics to identify problems, and find solutions, that otherwise wouldn't have come to mind.


Sources:

Lifelines for Visualizing Patient Records: http://www.cs.umd.edu/hcil/lifelines/

Luker, Kristin. Salsa Dancing into the Social Sciences. London; Cambridge: Harvard University Press, 2008.

Czochrowska, E.M. et al. "Outcome of Tooth Transplantation: Survival and Success Rates." American Journal of Orthodontics and Dentofacial Orthopedics  121, no. 2 (2002), 110-119. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11840123