Parenteral artificial nutrition (PAN) is a lifesaving treatment for a large population of patients affected by different diseases, and it consists of intravenous injection of nutritive fluids by means of infusion pumps. Wrong PAN solutions are, unfortunately, often administered, thus threatening the patients’ well-being. Here, we report an optofluidic label-free sensor that can distinguish PAN solutions on the basis of their volumetric refractive index (RI). In our system, a monochromatic light beam, generated by a laser diode, travels obliquely through a transparent, square-section polystyrene channel, is then back-reflected by a mirror, and finally exits the channel in a position that depends on the filling fluid RI. The displacement of the output light spot ΔXexperim is easily detected with a linear, 1-D position sensitive detector (PSD). We initially calibrated the sensor with water-glucose solutions demonstrating a sensitivity S = ΔXexperim/Δn = 13,960 µm/RIU. We then clearly distinguished six commercial PAN solutions, commonly administered to patients. To the best of our knowledge, this is the first reported healthcare sensing platform for remote contactless recognition of PAN fluids, which could be inserted into infusion pumps to improve treatment safety, by checking the compliance to the prescription of the fluid actually delivered to the patient.