A conditional maximized sequential probability ratio test for Pharmacovigilance

Department of Population Medicine, Harvard Medical School and Harvard Pilgrim Health Care Institute, 133 Brookline Ave., 6th floor, Boston, MA 02215, USA.
Statistics in Medicine (Impact Factor: 1.83). 01/2010; 29(2):284-95. DOI: 10.1002/sim.3780
Source: PubMed


The importance of post-marketing surveillance for drug and vaccine safety is well recognized as rare but serious adverse events may not be detected in pre-approval clinical trials. In such surveillance, a sequential test is preferable, in order to detect potential problems as soon as possible. Various sequential probability ratio tests (SPRT) have been applied in near real-time vaccine and drug safety surveillance, including Wald's classical SPRT with a single alternative and the Poisson-based maximized SPRT (MaxSPRT) with a composite alternative. These methods require that the expected number of events under the null hypothesis is known as a function of time t. In practice, the expected counts are usually estimated from historical data. When a large sample size from the historical data is lacking, the SPRTs are biased due to the variance in the estimate of the expected number of events. We present a conditional maximized sequential probability ratio test (CMaxSPRT), which adjusts for the uncertainty in the expected counts. Our test incorporates the randomness and variability from both the historical data and the surveillance population. Evaluations of the statistical power for CMaxSPRT are presented under different scenarios.

20 Reads
  • [Show abstract] [Hide abstract]
    ABSTRACT: A synchronized (SYNC) network has been designed as a spread spectrum overlay of an existing communications system, EXCOM. These networks will share a direct multiple-access satellite transponder. The basic network, EXCOM, utilizes frequency division multiple access (FDMA) carriers with quadrature phase shift keyed (QPSK) modulation to which the SYNC network must appear as a stable noise floor across the entire transponder bandwidth. The SYNC network contains a centralized control system interfacing directly with the EXCOM control system. The EXCOM control system will specify the transponder power level available to the SYNC network. The SYNC control system must maintain this usage level to within ±1.0 dB, or the EXCOM control system will be triggered to readjust the EXCOM FDMA carrier power levels. To investigate the feasibility of meeting this requirement within the context of the SYNC network, a simulation has been performed. This paper out-lines the SYNC control problem and the predicted performance of the SYNC control system. The results suggest that the system is capable of satisfying the ±1.0 dB constraint placed on it by the EXCOM control system under the conditions assumed.
    No preview · Conference Paper · Nov 1985
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The emergence of pandemic H1N1 influenza in 2009 has prompted public health responses, including production and licensure of new influenza A (H1N1) 2009 monovalent vaccines. Safety monitoring is a critical component of vaccination programs. As proof-of-concept, the authors mimicked near real-time prospective surveillance for prespecified neurologic and allergic adverse events among enrollees in 8 medical care organizations (the Vaccine Safety Datalink Project) who received seasonal trivalent inactivated influenza vaccine during the 2005/06-2007/08 influenza seasons. In self-controlled case series analysis, the risk of adverse events in a prespecified exposure period following vaccination was compared with the risk in 1 control period for the same individual either before or after vaccination. In difference-in-difference analysis, the relative risk in exposed versus control periods each season was compared with the relative risk in previous seasons since 2000/01. The authors used Poisson-based analysis to compare the risk of Guillain-Barré syndrome following vaccination in each season with that in previous seasons. Maximized sequential probability ratio tests were used to adjust for repeated analyses on weekly data. With administration of 1,195,552 doses to children under age 18 years and 4,773,956 doses to adults, no elevated risk of adverse events was identified. Near real-time surveillance for selected adverse events can be implemented prospectively to rapidly assess seasonal and pandemic influenza vaccine safety.
    Preview · Article · Dec 2009 · American journal of epidemiology
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: To describe the Vaccine Safety Datalink (VSD) project's experience with population-based, active surveillance for vaccine safety and draw lessons that may be useful for similar efforts. The VSD comprises a population of 9.2 million people annually in 8 geographically diverse US health care organizations. Data on vaccinations and diagnoses are updated and extracted weekly. The safety of 5 vaccines was monitored, each with 5 to 7 prespecified outcomes. With sequential analytic methods, the number of cases of each outcome was compared with the number of cases observed in a comparison group or the number expected on the basis of background rates. If the test statistic exceeded a threshold, it was a signal of a possible vaccine-safety problem. Signals were investigated by using temporal scan statistics and analyses such as logistic regression. Ten signals appeared over 3 years of surveillance: 1 signal was reported to external stakeholders and ultimately led to a change in national vaccination policy, and 9 signals were found to be spurious after rigorous internal investigation. Causes of spurious signals included imprecision in estimated background rates, changes in true incidence or coding over time, other confounding, inappropriate comparison groups, miscoding of outcomes in electronic medical records, and chance. In the absence of signals, estimates of adverse-event rates, relative risks, and attributable risks from up-to-date VSD data have provided rapid assessment of vaccine safety to policy-makers when concerns about a specific vaccine have arisen elsewhere. Care with data quality, outcome definitions, comparison groups, and length of surveillance are required to enable detection of true safety problems while minimizing false signals. Some causes of false signals in the VSD system were preventable and have been corrected, whereas others will be unavoidable in any active surveillance system. Temporal scan statistics, analyses to control for confounding, and chart review are indispensable tools in signal investigation. The VSD's experience may inform new systems for active safety surveillance.
    Full-text · Article · May 2011 · PEDIATRICS
Show more