Search

Meta-analysis of related trials can provide an overall measure of safety-signal accounting for variability across studies. In addition to an overall measure, researchers may often be interested in study-specific measures to assess safety of the product. Likelihood ratio tests (LRT) methods serve this purpose by identifying studies that appear to show a safety concern. In this paper, we present a Bayesian approach. Despite having good statistical properties, the LRT methods may not be suitable for the meta-analysis of randomized controlled trials (RCTs) when there are several studies with zero events in at least one arm.In this article, we describe a Bayesian framework using a Zero-inflated binomial model with spike-and-slab parameterization for the treatment effects. In addition to providing an overall meta-analytic estimate, this method provides posterior probability of a safety-signal for each study.We illustrate the approach using two published data sets comprising several randomized controlled trials (RCTs) each and compare the model performance for different choices of priors for treatment effect.The proposed Bayesian methodological framework is useful to identify potential signal for single adverse event and to determine overall meta-analytic estimate of the magnitude of the signal. Practitioners may consider this approach as an alternative to the frequentist's LRT approach discussed in Jung et al. (J Biopharm Stat 31:47-54, 2020) when there are zero events in either the treatment arm or the control arm. In the future, this approach can be further extended to accommodate multiple adverse events.

With more and more medical device databases being developed, there is an increasing interest in learning the geographical patterns of medical-device-related adverse events (AEs). For a specific medical device and an adverse event (AE) of interest, our aim is to detect a spatial-cluster signal that has a significantly higher AE rate than the AE rates for other regions, when the exposure information is available.We develop a likelihood ratio test (LRT) method incorporating exposure information, by geographical region, for spatial-cluster signal detection when the underlying observed health outcome associated with the medical device of interest is a Poisson-modeled count data (e.g., an outcome of AE count).An extensive simulation study shows that this method has good power, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV). The application of the method is demonstrated by two hypothetical case studies regarding a medical device that is used for patients who have reached end-stage heart failure.A methodological framework for exploring geographic patterns in device safety surveillance is discussed, including safety data collection, statistical tools, and display of the analysis results. The proposed statistical method can be used for spatial-cluster signal detection for an AE of interest from medical device registries or other databases that have patient-level geographical information.

With a record number of medical devices approved or cleared, it is important to understand the performance of devices once they are on the market. Using data from multiple medical devices and multiple sites, the problem of interest in this article is to detect if a device is a signal; that is, if a device performs significantly different from other devices of the same class, when the outcome of interest is a continuous variable. We develop a normal likelihood ratio test (LRT) method, henceforth referred to as normal-LRT, by incorporating sample size information into the methodological framework, to detect device signals using multi-site and multi-device data. It is shown via extensive simulation that the proposed method controls the type-I error and false discovery rate (FDR), while having good power and sensitivity. This method is applied to a hypothetical case study, in which 6 medical devices of the same class are compared. The normal-LRT method can be considered as a tool for device signal detection using data from multi-site and multi-device when the outcome of interest is a continuous measurement.