I, Data Factory

On this matter, I have a much more thorough entry forthcoming. But, to apply pressure to ensure that I actually finish that post, and more importantly, to get the ball rolling on this exciting new project, I offer this unedited mission statement.

Flummoxed by the inadequacies of conventional medicine at offering sound lifestyle recommendations for individuals with autoimmune diseases and inspired by Raymond Kurzweil and Sergey Brin, both of whom have balked at the rigid use of the scientific method in medicine, suggesting instead that medicine can be approached as a data mining problem, I have decided to turn myself into a data factory.

For the last year I have kept a somewhat well maintained journal of my medicines, relapses, and periods of relief. Yesterday I came to the startling conclusion that this data in this form is pathetically useless. Reading through it, I could hope to identify a bona fide cure, if one existed. But I couldn’t possibly hope to discern, for example, whether a vegetarian diet conferred a 20% reduction in the duration or severity of flares. My journals diligently documented daily events but provided little basis for analysis across long periods of time.

In the new incarnation of my record keeping, I plan to maintain measures of progress (some objective i.e. number of episodes of limb numbness, no of fasciculation attacks, as well as a wide number of environmental variables, ranging from granular dietary data, to exercise, location, medicines, etc. Surely, some trivial correlations will emerge - hours slept would correlate highly with symptom severity. Not because sleep causes symptoms but because fatigue is itself a symptom during exacerbations.

Ultimately this exercise in pattern recognition can serve both as a hypothesis generator and enable me to explore my health statistically, slicing up data across periods of time, making month over month comparisons etc, far beyond the capabilities of my old analog record-keeping system. If only I had begun this path when I first fell ill, I could have racked up over a 1000 data points, and potentially derived statistically significant insights.