Validity of the Grossarth-Maticek and Eysenck personality-stress model of disease: An empirical prospective cohort study

•2197 women and 919 men completed the combined short interpersonal reactions inventory with nine subscores calculated.•After 23.4 years, protocols were matched against dates and causes of death from the Australian National Death Index (1108 out of 3027 individuals with useable SIRI scores had died).•Type 2 (CHD-prone) and Type 4 (healthy) SIRI scores were significantly associated (p < 0.05) with all-cause and cardiovascular mortality but not with any-cancer mortality.•Despite criticisms of the Grossarth-Maticek and Eysenck personality-stress model, our study found empirical support for some SIRI subtypes.•Consistent with two previous studies, inverse associations of Type 4 (healthy) scores with all-cause mortality were found, and Type 2 scores predicted CHD mortality.•No significant relationship was found between Type 1 scores and cancer mortality.