Disproportionality analysis is coming in JMP Clinical 4.0
While randomized clinical trials are the gold standard for evaluating the efficacy of a new intervention, the available sample size is often insufficient to fully understand its safety profile. The risk a new therapy may pose may not be well understood until it has been on the market for many years, taken by individuals who differ from those studied under the inclusion criteria of the clinical development program.
Spontaneously reported adverse events are collected by regulatory agencies, pharmaceutical companies and device manufacturers to monitor the safety of a product once it reaches the market. The FDA, for example, maintains databases for pharmaceuticals (AERS: Adverse Event Reporting System), vaccines (VAERS: Vaccines Adverse Event Reporting System) and medical devices (MAUDE: Manufacturer and User Facility Device Experience). These data are generally obtained from physicians, patients or the medical literature. Spontaneously reported adverse events present a unique challenge in that there is no measure of total exposure. In other words, there is no clear denominator to define an adverse event incidence for a particular drug. In order to identify potential safety-signals, the rate at which a particular event of interest co-occurs with a given drug is compared to the rate this event occurs without the drug in the event database. This is referred to as disproportionality analysis.
A Case, which can refer to an individual patient or report, can have one or more adverse events associated with it. Each event, in turn, can have one or more drugs associated with it. Events that occur on different Cases are considered distinct for the purposes of determining a total count of adverse events for the analysis. Within a Case,JMP Clinical determines event uniqueness based on the variables supplied for analysis such as the event name and classification, which could refer to MedDRA preferred terms and system organ class; the onset date of the event; or any variables used for stratification, such as event severity. Once the total number of events is determined, JMP Clinical calculates four measures of disproportionality – the reporting odds ratio (Meyboom et al., 1997), the proportional reporting ratio (Evans, Waller & Davis, 2001), the multi-item gamma poisson shrinker (DuMouchel, 1999) and the Bayesian confidence propagation neural network (Bate et al., 1998; Gould, 2003).
The Disproportionality Analysis analytical process (AP) has minimal data requirements to perform an analysis – an event field and a drug field. Though data requirements are minimal, additional specified variables provide a more informative analysis. For example, providing an onset date generates analyses of disproportionality statistics over time to illustrate how robust findings are to accumulating data. The user can also specify one or more variables for stratification. How drugs are prescribed, disease severity and how individuals may respond to treatment can all be influenced by demographic and other background characteristics. Therefore, disproportionality statistics should be calculated among homogeneous groups to avoid inappropriate conclusions (Woo et al., 2008).