In the SCCS design, the analysis only includes individuals who were both vaccinated and had an event of interest during the observation period. The rate of endpoints per day is compared between an ‘at risk’ period and a control period, which is far enough removed from the time of vaccination find more that it is unlikely for a vaccine to have caused the
endpoint [16]. For each individual, the index date for the exposure is the date of vaccination. Follow-up time for each individual is divided into three distinct intervals: an exposed period (or ‘at risk’ period), an unexposed period (or control period), and a washout period in between the exposed and unexposed periods. Our selection of the ‘at-risk’ and control periods was based on our previous study of ER visits and/or hospitalizations following 2-, 4-, 6-, and 12-month immunizations [9] and [10]. For the 2-, 4- and 6-month immunizations, the ‘at-risk’ period was 0 to 2 days following vaccination and the control period was 9 to 18 days post-vaccination. For the 12-month vaccination, the ‘at-risk’ period was 4 to 12 days post-vaccination and the control period was 20 to 28 days post-vaccination. We calculated the relative incidence of the composite endpoint (ER visits and/or hospital
admissions) in the exposed period versus the unexposed period using a fixed effects conditional Poisson regression model. The regression model controlled for exposure period and individual http://www.selleckchem.com/products/Everolimus(RAD001).html patients, thereby allowing each individual to serve as his/her own control. To control for the dependence of multiple events occurring close together in time (e.g. an ER visit leading to an
admission, or serial ER visits), each individual was classified as having ‘one or more events’ or ‘no events’ in each of the ‘at-risk’ and control Terminal deoxynucleotidyl transferase periods. In order to determine whether the relative incidence of the composite endpoint varied between males and females, we included a risk by sex interaction term in the SCCS conditional Poisson model. A likelihood ratio test is used to compare the full model including the interaction term to the reduced model without the interaction term in order to test whether the interaction term is statistically significant [16]. The parameter estimate of this interaction term can be exponentiated to yield a “relative incidence ratio” (RIR) which is equivalent to the ratio of relative incidence in females to the relative incidence in males: an intuitive measure of the magnitude of the difference in relative incidences for females versus males. This RIR has the added benefit of allowing us to overcome the impact of the healthy vaccinee effect, the decision by parents and health care providers to forgo vaccination when a child is acutely ill resulting in the administration of vaccines to children who are in a comparatively healthy state [7] and [8].