This paper presents a novel, non-invasive approach for distinguishing between patients with various medical conditions based on volatile organic compound (VOC) signatures. Cotton swabs were used to swab the nose, navel, and ear of 24 patients, after which the odor of the cotton swabs was analyzed using Laser-based photoacoustic spectroscopy (LPAS). The LPAS system was equipped with three quantum cascade laser (QCL) modules covering wavelengths from 5.6 μm to 12.9 μm. The resulting spectra were preprocessed using a Savitzky-Golay filter and subjected to principal component analysis (PCA). Projection of the data onto the first two principal components revealed group-level structure corresponding to healthy individuals (H), patients with Parkinson’s disease (PD), COVID-19 (C), and mild cognitive impairment (MC). Investigations into possible correlations with demographic factors and lifestyle habits showed no consistent association with the principal components, with the exception of age, which is expected, as diseases like PD and MC primarily affect older individuals. While this excludes strong confounding effects, confirming true independence requires a larger cohort in future work. Moreover, it could be shown that as little as five wavelengths are sufficient to separate the samples. These findings suggest that odor analysis of cotton swabs combined with mid-infrared LPAS spectroscopy and multivariate analysis may represent a promising approach for future non-invasive screening technologies.