This curious 2006 article has an interview with one researcher in one column and an interview with another researcher in another column. Their results differed.
Pro probiotic. “Children with [infectious acute diarrhea] who took Lactobacillus [various strains and species, in nutritional supplement form, not in yogurt form] had a shorter duration of diarrhea (on average 0.7 days shorter) than those who took placebo. Also, they had fewer episodes of diarrhea, i.e. fewer stools, on the second day of treatment than those in the placebo group. Interestingly, the children who took higher doses of Lactobacillus had shorter duration of diarrhea, and it seems that a daily dose of at least 10 billion viable bacteria is necessary to have a beneficial effect.”
Anti probiotic. “I published a big study looking at Lactobacillus GG in kids with Crohn’s disease who were already doing fairly well on medication. We put them on the probiotic or a placebo for two years. We followed them for two years and looked for whether the probiotic group had a lower rate of relapse and whether there were any differences between the two groups. We didn’t find any differences.”
Seth, you may have already seen it, but Art Ayers’ most recent post also talks about probiotics:
coolinginflammation.blogspot.com/2012/06/dr-oz-on-gut-flora-repair.html
Okay, I don’t know how much of a difference is the “supplement factor” [vs natural food], but to me, the idea of beneficial effects of certain diverse bacterium is beyond the need for a trial like this. It simply makes sense according to all the physiological evidence we have, and these large-scale trials are too easily corrupted to tease out much useful information.
Seth: My guess is that there are lots of things that can go wrong — lots of of ways taking probiotics that don’t work. You need the right dose, you need enough diversity, maybe the bacteria can die. Or don’t work optimally. For example, is taking two probiotics better than taking one? Until you can measure the benefits you can’t answer questions like that.