A healthy person has difficulty determining whether an intervention such as a nutritional supplement has any benefit. The difference is likely smaller than the normal variation in his subjective and objective health. This lengthens the duration and expense of determinative trials. However, a strongly negative effect will be easy to notice.
Illness is therefore a valuable experimental opportunity. Illness reduces activity level and thus causal noise. Illness increases self-monitoring due to incapacitation. Illness makes a moderately positive effect easier to notice, because the relative improvement in health is dramatic. A negative effect is still detectable enough, since one can always get worse until dead.
Many interventions take the form of ingestibles. One great source of difficulty I've found in mitigating my IBS/IBD is that my gut transit time varies from several days to several hours. My diet causes minimal gut load, which results in a normal transit time of several days. IBS-D flares obviously accelerate this.
Long gut transit times make causal attribution of ingestible effects difficult. The reverse is true of short gut transit times. Thus the best time to test an ingestible is when transit time is already short and one's gut is purged.
Obviously this doesn't work if the cause of rapid transit is temporary or already resolved. Then natural recovery is conflated with ingestible benefit. So one needs a stable regimen that causes the rapid transit time. In my case, I merely need to stop taking my usual treatments to achieve this state. Others may need to eat some food that normally disagrees with them.
So if spicy food makes one ill, one can temporarily go on a spicy diet to rapidly test various ingestible medicines or supplements for significant benefits. (Beware of gut acclimation to spiciness skewing results!)