publications - Verifying that cited results are still current before publishing

Assume you are knee-deep in a project, going through a series of papers that are seemingly relevant to your project. You found this ~10 year old paper, which looks very promising indeed. A lot of interesting results, good discussion, etc.

At this stage, I think it's obvious that you should check whether or not the results and conclusions in the paper are still valid by searching for any correspondence related to the article (e.g., pointing out that some claims do not hold, etc). However, I have yet to find anyone that digs into such information. You normally find a paper, read it, find more interesting papers through references and keep on digging in this manner until you have accumulated "enough" articles to form an educated opinion.

My questions are as follows:

• 
  

  Is this (or any other) type of quality assurance process, when it comes to cited literature, common practice?

  
  


• 
  

  Is there a way to streamline this process? PubMed does not seem to include all correspondence related to each article.

  
  
  

---

_{I should perhaps note that, if you are in biomedical sciences (especially related to complex diseases), about half of the publications turn out to be wrong or misleading after some years, and perhaps a quarter more are shown to be incomplete in its findings. I believe it's in the nature of biology, in contrast to more human defined sciences like mathematics or computer science. There are no formulas or calculations to check the integrity of the work published.}