Infrared (IR) spectroscopy of serum/plasma represents an alluring molecular diagnostic tool, especially for cancer, as it can provide a molecular fingerprint of clinical samples based on vibrational modes of chemical bonds. However, despite the superior performance, the routine adoption of this technique for clinical settings has remained elusive. This is due to the potential confounding factors that are often overlooked and pose a significant barrier to clinical translation. In this Perspective, we summarize the concerns associated with various confounding factors, such as fluid sampling, optical effects, hemolysis, abnormal cardiovascular and/or hepatic functions, infections, alcoholism, diet style, age, and gender of a patient or normal control cohort, and improper selection of numerical methods that ultimately would lead to improper spectral diagnosis. We also propose some precautionary measures to overcome the challenges associated with these confounding factors.