Recent advances suggest that miniaturised mid-infrared (MIR) devices could replace more time-consuming, laboratory-based techniques for clinical diagnostics. This work uses Fourier transform infrared spectroscopy to show that the MIR complex refractive index of whole blood varies across a range of haematocrit. This indicates that the use of an evanescent measurement is not sufficient to optically exclude the cellular content of blood in the MIR, as previously assumed. Here, spectral refractive index data is presented in two ways. First, it is given as whole blood with varying haematocrit. Second, it is given as the percentage error that haematocrit introduces to plasma. The maximum error in the effective plasma refractive index due to the haematocrit of healthy adults was 0.25% for the real part n and 11% for the imaginary part k. This implies that calibration measurements of haematocrit can be used to account for errors introduced by the cellular content, enabling plasma spectra and analyte concentrations to be indirectly calculated from a whole blood sample. This methodological advance is of clinical importance as plasma concentration of analytes such as drugs can be determined using MIR without the preprocessing of whole blood.