Let’s start with the good first when thinking about peer review.
The process of peer review can actually be a great way to identify fraudulent papers and research. The referees who read through the paper should have enough familiarity with the field to pinpoint any discrepancies. These foundational flaws can help in finding problems with the research or the overall premise. However, in many fields (especially scientific ones), peer review does not end with publication. The process continues by trying to replicate the findings and investigate the research procedures in depth. With this added level of peer review and research replication, many fraudulent studies have been proven false (think Andrew Wakefield and his vaccines equal autism hoopla).
Yet, peer review does obviously have its flaws like everything else. The referees are really just people in the field themselves. They have their own biases and assumptions which leak into their own work and can bleed into the works of others. Seemingly original research can sometimes get the short end of the stick when it comes to peer review. This problem is exacerbated with the added multitude of supporting sources—meaning truly original research has to grab from a limited number of foundational sources but still does not have the precedence that helps when being published in a peer review journal.
Overall, it can be a mixed bag with peer review. The type of paper created should certainly be taken into account since some forms of research are better for peer review than others.
Lancet retraction of Andrew Wakefield paper