cb‐Barlow 2001.

Study characteristics
Methods	Cohort study ‐ the design of the Vaccine Safety Datalink ‐ from 1 March 1991 to 30 September 1993
Participants	Data are collected from 4 HMOs: the Group Health Cooperative in Seattle; Northwest Kaiser Permanente in Portland, Oregon; Kaiser Permanente of Northern California in Oakland; and Southern California Kaiser Permanente in Torrance. Children (N = 137,457). Children entered the cohort at birth, on the date of their enrolment in the HMO, or at the beginning of a study site’s observation period, whichever came last, and remained in the cohort until the age of 7 years, disenrolment from the HMO, or the end of the observation period, whichever occurred first. Using the automated data, the authors identified 2281 possible first seizures. Using the random‐sampling plan previously described, they selected a total of 1094 children for chart review. 716 of these children were confirmed to have had a first seizure during the study period. The reference group at the time of the seizure was composed of children matched for age, calendar time, and HMO but who had not had a vaccination in the preceding 30 days.
Interventions	Immunisation with MMR vaccine: data on immunisation were derived from automated immunisation tracking systems initially developed to collect information on all routinely administered immunisations.
Outcomes	Risk of febrile seizure within 0 to 7, 8 to 14, 15 to 30 days after immunisation. Potential seizures were identified through the automated data systems of each HMO, on the basis of visits classified according to the ICD‐9‐CM as code 333.2 (myoclonus), code 345 (epilepsy), code 779.0 (convulsions in a newborn), or code 780.3 (convulsions).
Funding Source	Government
Notes	Conclusions: there are significantly elevated risks of febrile seizures after receipt of DTP vaccine or MMR vaccine, but these risks do not appear to be associated with any long‐term, adverse consequences.
Risk of bias
Bias	Authors' judgement	Support for judgement
PCS/RCS ‐ exposed cohort selection	Unclear risk	Based on large HMO ‐ probable selection bias ‐ data on immunisation were derived from automated immunisation tracking system
PCS/RCS ‐ non‐exposed cohort selection	Unclear risk	Drawn from the same population ‐ probable selection bias
PCS/RCS ‐ comparability	Unclear risk	Adjusted by multivariate model
PCS/RCS ‐ assessment of outcome	Unclear risk	Based on hospitalisation record
Summary Risk of Bias assessment	Unclear risk	We had concerns regarding at least 1 domain such that some doubt is raised about the results.