about my results

[Cheng YuanLung]

I don't know if it is proper to ask the question here. Hope you don't mind.

Recently, I did an analysis on the factors related to cardiovascular(CV) risk and found confusing results.

H pylori with ulcer is associated with great CV risk than H pylori infection without ulcer.
However, ulcer patients without and with H pylori have non-significant CV risk difference.

My colleague told me the presence of H pylori infection is not related to a higher CV risk.
Ulcer disease is the main factor for the increasing CV risk as CV risk is the same with or without the presence of H pylori.
Once peptic ulcers occur in patients, their CV risk increases regardless of what caused the ulcers.

I find it odd, but i don't know how to better explain my results. Could you help me, please?

Thank you for your help in advance.

[Carlo Lazzaro]

Cheng:
as per your post, my guess is that you performed a regression with CV risk as regressand and the interaction between ulcer and H pylori infection among the predictors
Without taking a look at what you type and Stata gave you back, helpful advice are difficult to provide (for me at least).
That said, I would consider that the contribution of each predictor in explaining variation of the dependent variable is conditional on the role of (adjusted for) the other predictors.
Or a third variable can mediate the results of your interaction.