Active surveillance to detect all patients with Guillain-Barré syndrome who had had onset of illness from July 1, 1976 through April 30, 1977 was undertaken in Michigan after indications that the syndrome might be associated with the National influenza immunization Program of 1976-1977. Hospital record room librarians, neurologists, and neurosurgeons reported the greatest number of cases; coded hospital discharge records were the best means of ascertaining case occurrence. This differed from national surveillance, which relied essentially on reports that neurologists and other clinicians sent to state epidemiologists and then to the Centers for disease Control; hospital discharge lists were not systemically reviewed nationally. A total of 79 of the Michigan cases were in persons who had not received swine influenza vaccine, while 46 cases were in persons who had received it. For unvaccinated adults, the incidence of Guillain-Barré syndrome during the 10-month surveillance period was 0.36 cases per 10(6) person-weeks; for adults with onset within six weeks of vaccination, it was 2.31 cases per 10(6) person-weeks. After six weeks post-vaccination, the rate decreased to 0.17 cases per 10(6) person-weeks. The attributable risk for acquiring Guillain-Barré syndrome within six weeks after receiving swine influenza vaccine was 11.70 cases per 10(6) persons vaccinated.