Should doctors prescribe more weekend?

How seasons and other cyclical events affect chronic illness (PART 1)

What makes us sick and what can we do to prevent or alleviate sickness? Answering this question usually depends on the individual, as there is not a single cure for every disease. In addition, it seems that the same disease manifests itself differently at the individual level. Here at Flaredown, we try to offer insights into how a person’s lifestyle affects their overall well-being and the contribution of user-defined factors (e.g. food, activity, medicine) that positively (make you less sick) or negatively (make you more sick) affect their status. But what if there are lifestyle factors that transcend individuals and diseases? Wouldn’t these factors ultimately be the panacea (cure-all), even if only modestly improving health? Exactly this was our question.

We quickly realized that many disease-specific treatments (e.g. medicines and personal routines) could be ruled out: Medicines (unlike food and sleep) are taken only be few individuals related to their medical condition and the data for most the tracked medications and personal routines would be too sparse (at least at this point) for analyses between individuals across the whole database. We will revisit this point in the future, stay tuned!

But there are many natural rhythms that influence our daily life. Examples include but are not limited to: season, day of the week, and lunar cycle. Obviously, these ideas are not original, as seasonality (e.g. 1, 2), and lunar phase (e.g. 3, 4) have been previously studied for their correlation with certain diseases or conditions. Our advantage at Flaredown is the availability of long term data for several thousand individuals (as of March 2018) and we can explore these natural rhythms at a large scale. We surmised that the individual contribution of natural rhythms (e.g. seaon, lunar cycle, weekday) impacts the overall well-being of our users.

Overview of natural rhythms contributing to overall well-being. Schematic showing that potentially many concurrently happening natural rhythms individually influence the overall well-being.

We investigated changes in the “trackable_value” (range: 0–4; best = 0, worst =4), a surrogate for disease/condition severity, for all users combined. For this analysis we explored the effect on “trackable_value” of different rhythmical factors that were easily identifiable: season, weekday, lunar cycle.

Mean “trackable_value” for different rhythmical factors. Mean and standard error of the mean plotted, n >500

The results suggest that there are no robust changes across all diseases/conditions associated with any of the rhythmical factors. Perhaps this is not surprising as different diseases/conditions may have unique patterns and when combined, any disease/condition-specific patterns get lost in the noise.

The “migraines” example

If we only use data from migraines and run the same analysis, we can see that there are some modest benefits associated with the rhythmical factors (note: I didn’t run any statistical analysis, since this study was largely observational). Severity seems to be lower on average during fall/summer and during full moon.

Mean “trackable_value” for different rhythmical factors only considering “migraines”. Mean and standard error of the mean plotted and colorized by mean value

However, this analysis only covers migraines and other diseases/conditions might exhibit different patterns. For example, one condition might have the highest disease severity during full moon (think werewolf) and others might have the highest disease severity during new moon.

Every Disease is unique but some are unique together

We explored (from a bird’s eyes view) all diseases/conditions in our database that met a frequency cutoff (e.g. migraines had over 5000 entries). It appears that many diseases/conditions cluster together (i.e. produce similar effects during defined rhythms) and show defined patterns.

Scaled and centered mean “trackable_value” for different rhythmical factors for all conditions above 500 entries.

Each season seems to uniquely worsen some diseases/conditions (below; top left). Acid reflux seems to be worst on Saturdays (could this be related to people going out on the weekend?) and anxiety, pain and several other conditions cluster together and seem to improve on the weekend (both top right panel). The lunar cycle (bottom panels) also seems to influence severity of conditions/symptoms, most notably the migraine example (as discussed above). Bottomline: season, weekday, lunar cycle potentially affect our health. So individuals suffering from pain and anxiety might benefit from “doctors prescribing more weekend”.

So what?

…we can’t stop winter or Monday from occurring! While this is true, maybe the medicine dose can be rhythmically-adjusted. Many medicines come with unpleasant side-effects and the medicine dose could potentially be adjusted to the season, phase of the moon, or day of the week. Changing geographical locations could allow for adjusting seasonal effects (e.g. more summer and less winter) and pollen exposure (5). Others have also noticed/studied effects of cyclical rhythms on several diseases/conditions (see References 1, 2, 3, 4, above). With mounting evidence it might be interesting to explore what ultimately causes these effects and how we might benefit from understanding their etiology.

Further, we can use these patterns to help predict severity outcomes in the future and allow people to adjust their lifestyle to alleviate disease symptoms. Ultimately, we want to help people understand their bodies and diseases/conditions better and provide guidance on how to improve their health by adjusting their lifestyle.