I agree. This is clearly not a double blind experiment [1]. And self reporting studies are more prone to produce false results, because the people are optimistic and want to recover and want to feel better. Also people want to be kind and make the doctors happy, so they unconsciously ignore some of the bad side effects.

And also people confuse normal variations with treatment consequences, so perhaps the persons in the treatment group classify that as an improvement and the persons in the control group classify that as normal. [2]

In this case, perhaps the comparison is worse because fasting has some side effects and you can get answers like: "I fell a little dizzy, but not as dizzy as the previous week while I was fasting." Should you classify that as an improvement?

[1] Is it possible to do a double blind experiment for this?

[2] Do they have control group?