You're not wrong. Almost all observational studies could be subject to confounders which if true would invalidate their claimed results. But you seem to be suggesting that just because science is hard we should accept the results of all these flawed studies at face value without further questioning.

The better conclusion is that we should be skeptical of almost all such studies, and always look closely at the data and methodology to see if a confounding effect might contradict the conclusion. We should only give weight to studies that stand up to such scrutiny, and even then we should realize their limitations.

yes...I do not trust the claims of any single observational study. Using observational studies is necessary to help understand a problem in preparation for well-conducted large randomized clinical trials (RCTs). Sometime RCTs are not ethical or feasible (like etiological studies of cancer), and then we need many observational studies in different settings and using different designs showing a large effect that cannot be explained by known or unknown confounders and supported by many experimental studies (e.g., in cells). And even then we accept the results on the principle that is better to be safe than sorry. Many observational studies claimed that HRT protect women from heart disease until the definitive trial was done and it was found HRT actually increases the risk of heart disease!! Hundreds of other examples exist of well-researched and widely accepted hypotheses that turned out to be false in RCTs. For every observational study that claims X causes Y, I could find one that claims that X protects against Y, that X has nothing to do with Y, that Y causes X, that there is no X and no Y.. etc. It is the Wild West of science.