301 first-admitted hospitalized patients with paranoid psychoses have been studied by Retterstøl over a period of 5-18 years. Common Scandinavian diagnostic procedures were used. About 200 are still alive, and these subjects are at present being interviewed semistructurally by Opjordsmoen using a modification of SADS-L, and making a total follow-up period of 22-37 years. The diagnoses are confirmed according to ICD-9, RDC, DSM-III and some specific groups of delusional disorders (DD) operationalized by Winokur and Kendler. All interviews have been carried out non-blind to the diagnoses which will make a bias possible. However, in paranoid cases, it is an advantage for establishing contact and a conductive atmosphere to know something about the patient beforehand. Based upon our own experiences and reviewing the literature, we point to some important methodological aspects regarding follow-up studies in delusional persons. The suspiciousness, misinterpretation, dissimulation, rationalization and convincing argumentation seen in many paranoid cases, call for a skilled investigator and a clinical approach. However, operational criteria, new diagnostic concepts and standard procedures for follow-up interviewing and outcome assessments will make comparison for international readers easier.