Hacker News Reader

On the one hand, it makes total sense to me that childhood antibiotics could mess up your digestive tract immune system. On the other hand:

"Subgroup analysis showed that antibiotic use <18 years old was a risk factor for CD development in the Chinese (adjusted OR 4.80, 95% CI 1.62–12.24; P = 0.005) but not in Australian populations (OR 1.80, 95% CI 0.33–9.95; P = 0.498)."

If it were a robust finding, it would not fall apart on subgroup analysis.

View on HN

Actually the reverse, we should be skeptical of subgroup analyses, because they are often fishing for results, have issues with multiple comparisons, and are typically underpowered.

Different antibiotics/doses/frequency of prescription in different countries? Haven't read the whole thing yet though.

Just look at the results table. Even for china it looks like only certain age ranges show the effect. Most likely just random.

Reply on news.ycombinator.com