Framework for Implementation of Poisson MaxSPRT Technique with Variations for Vaccine Safety

For direct, continuous, and sequential drug and vaccine safety surveillance, the Maximized Sequential Probability Ratio Test (MaxSPRT) was developed by the Centers for Disease Control and Prevention (CDC) (Kulldorff et al, 2011). Its predictive value and power to detect signals and the ability to monitor adverse events continuously have made it an emerging technique for vaccine adverse event surveillance. Moreover, being able to use a statistical method e.g. MaxSPRT in the absence of dose distributed denominator is a practical advantage for spontaneous reporting systems to function as stand-alone signal detection systems. In this paper, we present a comprehensive framework for implementing MaxSPRT for vaccine safety surveillance and Poisson data. We analysed the literature regarding MaxSPRT and sequential analysis. Our analysis revealed numerous variations of MaxSPRT, adapted to the specific requirements and objectives of the users. Variations are due to differing types of data and purpose of use, including whether used for epidemiological surveillance or for regulatory monitoring. This paper provides a comprehensive guide for organisations contemplating the implementation of MaxSPRT. It synthesises existing literature on MaxSPRT, identifies variations based on specific requirements, and describes an implementation framework. We offer a detailed explanation of the steps and challenges associated with the implementation of MaxSPRT on the adverse event following immunisation (AEFI) reporting database of Surveillance of Adverse Events Following Vaccination in the Community, Victoria, Australia (SAEFVIC), the largest jurisdictional reporting service by volume in Australia. It also proposes some techniques and measures to deal with the challenges associated with the implementation process.