Mayo Clinic Records-Linkage: Contributions to Neuroepidemiology

Abstract

A records-linkage system has been established for residents of Rochester, Minnesota and surrounding Olmsted County in the United States. This includes medical data sources both within and outside the Mayo Clinic (located in Rochester) and involves the cooperation of all medical facilities in the area. Rochester residents have access to high quality medical care and most serious health problems come to medical attention. Although the population is relatively small, there is a high level of case ascertainment and follow-up. Such a records-linkage system provides a rich data base for studies of neurologic disease in a well-defined population. In many instances, Rochester data provide the only estimates of morbidity for a large number of neurological and neuromuscular disorders in the United States.

The Scope of Neuroepidemiology: from Stone Age to Stockholm

“…Studies by Kurlandst al. [3] have shown that the prognosis of multiple sclerosis is not as severe as suggested by several clinical case series. In counseling those affected by multiple sclerosis as to what might be expected in the future, such information is extremely valuable and reassur ing to patients.…”

“…The fact that all of these records are usually stored in different places makes it extremely difficult to trace this information in order to obtain a complete medical profile of any given individual. The highly successful application of a records-linkage system to neurologic diseases is described in a future issue of this journal by Kurland et al [3], The prepaid medical plans represent another important potential source of morbidity information. Examples of such medical care systems are the Hospital Insurance Plan of New York and the Kaiser-Permanente system.…”

Incidence of Acute Transverse Myelitis in Rochester, Minnesota, 1970–1980, and Implications with Respect to Influenza Vaccine

“…The diagnoses made in all these health-care institutions, including information provided by death certificates and the high autopsy rate (60-70% among local residents who die), have been coded into a central computerized diagnostic index covering several decades. The potential of this data resource has been described elsewhere [34,35]. The capabilities of case ascertainment and patient follow-up afforded by this records-linkage system offer unique opportunities for studying diseases in a circumscribed population over a long period of time.…”

“…The usual criteria for Rochester residence were applied to each acceptable case; the ascertainment of residency status was carried out as previously described [34]. To be eli gible, the patient was required to have been a bona fide resident of Rochester prior to the onset of symptoms.…”

Hypothesis Testing in Neuroepidemiology: Experiments of Nature and Experiments of Man

“…In situations in which all cases of a particular disease in a well-defined population have been identified, these indi viduals can then participate in a case-control study. Controls can be selected from the population not affected by disease [5], This procedure minimizes problems of selection bias in choosing cases and controls. With case-control studies, we cannot measure disease incidence and can only estimate the relative risk.…”