The aim of the study was to investigate how settings and registry characteristics affect the prevalence and nature of multimorbidity in elderly individuals. We used data from three population-based studies, two general practitioner registries, one hospital discharge register, and one nursing home registry to estimate the prevalence of multimorbidity. Individuals aged 55 years and over were included. Results: Multimorbidity was most prevalent in nursing homes (82%), followed by the general population and general practitioner registries (56%–72%) and the hospital setting (22%). There were large differences in the nature of multimorbidity between settings. Combinations of hypertension, heart disease, and osteoarthritis were dominant in the population-based setting, whereas hypertension in combination with osteoarthritis, obesity, disorders of lipid metabolism, and diabetes dominated in the general practitioner setting. In the hospital setting, combinations of heart diseases had the highest prevalence. Combinations of dementia, hypertension, and stroke were dominant within the nursing home setting. Conclusion: This study shows that setting and registry characteristics have an important influence on the outcome of multimorbidity studies. We recommend provision of at least information about the setting, the (list of) conditions included, the data collection method, and the time frame used, when reporting about the size and nature of multimorbidity. (aut.ref.)