I think also small initial trials often have better results than larger follow up trials (don't know why could be something to do with patient selection?)
I believe each study has to be analyzed on it's own. For example, the early research on CoQ10 did not have the benefit of the knowledge about the quality issues or the absorption issues. Studies that found supplementing with CoQ10 had little to no impact might actually have been a quality or an absorption issue. Therefore, the farther back the studies, the more questions that arise. The same with the genetic studies. What used to be considered clear cut yes/no issues - such as the diseases where homozygous were told they had a disease and heterozygous were told they were carriers. Now there are many studies that depending upon the disease show the heterozygous carriers might also be impacted or have a lower enzyme that coupled with a secondary issue can cause symptoms. Therefore studies that show a disease is inherited in a recessive pattern (needing homozygous mutations - 2x mutations), are now often taken with a grain of salt because depending upon the mutation there might be a spectrum of symptoms. It's reminds me of 40 years ago when kids were diagnosed with autism. However, during that time kids that we now know are autistic spectrum were back then just seen as trouble makers or different. Now we know about the spectrum. Spotting the patients with the severe issues is much easier (low hanging fruit). The researchers tackled those early on. As time has passed, the researchers are now delving into all the diseases and syndromes where people are in the spectrum. Therefore, I believe across the board as time passes, the farther back a study, the more likely there may be issues looking at it with today's knowledge.
I agree, it depends on the results, the subject being studied and many other factors. There was a phase 1 trial of Nilotinib (a chemotherapy drug) on Parkinson's patients, it did amazing, and the mechanism on why it would work was theorized in published studies from several years before it was tried. The long term safety, exorbitant cost and replication are still underway in a larger study but i expect good results. However the Rituximab study got good results the first time out and bad results the second time, though i recall Dr Edwards saying it was confusing why it would work before we found out the disappointing results. Though i hope they try to figure it out because even 30% suggests either a subset or a different treatable disease. And thats another point if early trials show an effect and later trials show a different one its probably worth figuring out why because it can lead to further scientific understanding
