You just need to use the 3-fold rule. If the study wasn’t double blinded and the effect is not at least 3-fold then it is most likely to not hold up over time.