PURPOSE: To develop novel analytical approaches for identifying both miscibility and phase separation in hot-melt extruded formulations. METHODS: Felodipine-Eudragit E PO solid dispersions were prepared using hot-melt extrusion. The fresh and aged formulations were characterised using scanning electron microscopy, differential scanning calorimetry, heat capacity (C(p)) measurements using modulated temperature DSC and nuclear magnetic resonance relaxometry. RESULTS: The solubility of the drug in polymer was predicted as being < or =10% w/w using a novel model proposed in this study. Freshly prepared HME formulations were found to show no evidence for phase separation despite drug loadings greatly in excess of this figure. Conventional DSC showed limitations in directly detecting phase separation. However, a novel use of C(p) measurements indicated that extensive phase separation into crystalline domains was present in all aged samples, a conclusion supported by SEM studies. The NMR relaxometry study confirmed the existence of phase separation in all aged formulations and also allowed the estimation of separated domains sizes in different formulations. CONCLUSIONS: This study has presented a series of novel approaches for the identification, quantification and prediction of phase separation in HME formulations. Supersaturation of drug in the polymer caused the phase separation of the aged felodipine-Eudragit E PO formulations.