Epidemiological designs for vaccine safety assessment: Methods and pitfalls

Statistics Modelling and Economics Department, Health Protection Services, Health Protection Agency, 61 Colindale Avenue, London, NW9 5EQ, UK.
Biologicals (Impact Factor: 1.21). 10/2011; 40(5):389-92. DOI: 10.1016/j.biologicals.2011.08.010
Source: PubMed


Three commonly used designs for vaccine safety assessment post licensure are cohort, case-control and self-controlled case series. These methods are often used with routine health databases and immunisation registries. This paper considers the issues that may arise when designing an epidemiological study, such as understanding the vaccine safety question, case definition and finding, limitations of data sources, uncontrolled confounding, and pitfalls that apply to the individual designs. The example of MMR and autism, where all three designs have been used, is presented to help consider these issues.

5 Reads
  • [Show abstract] [Hide abstract]
    ABSTRACT: Background Carotid angioplasty and stent (CAS) placement has emerged as an attractive revascularization strategy for patients with internal carotid artery stenosis. However, the effectiveness and safety of CAS were not fully evaluated, mainly because of methodological difficulties in finding an appropriate comparison group. Methods Patients who underwent CAS were identified from Taiwan's National Health Insurance claims database between 2005 and 2008. The incidence rate of ischemic stroke after CAS was compared with that of the year prior to the procedure using a self-controlled case series analysis and a conditional Poisson regression model. Logistic regression was conducted to identify factors associated with poor outcome. Results A total of 1258 patients who had undergone CAS were included, and 73 cases (5.8%) of death or ischemic stroke occurred during the index hospitalization. Within 1 year after CAS, 74 patients died and 80 experienced an ischemic stroke. Of the 1184 patients who were followed for 360 days, the rate ratio for ischemic stroke decreased to 0.21 (95% CI: 0.08–0.51) between 31 and 180 days, and 0.10 (95% CI: 0.03–0.32) between 181 and 360 days. Statin therapy was associated with a reduced risk of death or ischemic stroke in the 1st month (odds ratio of 0.53; 95% CI: 0.32–0.90). Conversely, the use of nonsteroidal anti-inflammatory agents, possibly histamine-2 receptor blockers, and CAS performed by low-volume operators were associated with a twofold increased risk. Conclusion CAS reduced the long-term risk for ischemic stroke. Self-controlled case series analysis might be an appropriate design for evaluating device safety and effectiveness.
    Journal of the Formosan Medical Association 01/2014; 114(3). DOI:10.1016/j.jfma.2014.05.001 · 1.97 Impact Factor