The Case for Endpoint Adjudication Through Automation
The clinical trial industry is tackling workflow challenges through automation. For years, IT companies have developed important tools to help sponsors and CROs manage the data they collect, ensuring that every system interaction is properly logged so analysis can find and correct shortcomings in the data-gathering process.
Unfortunately, digital data gathering remains a complex problem, and being able to just “replay” the actions taken by trial managers, coordinators and investigators doesn’t simplify the problem enough.
Take, for instance, adverse event or endpoint adjudication processes. The determination whether an event is the result of the drug being tested is an elaborate process that involves collecting documents from investigator sites, determining whether they are complete, and putting together subject dossiers.
These dossiers are then used to enable adjudication workflows. In some cases, adjudicating an event is a kind of “voting” process involving individual “adjudicators” and clinical event committees (CEC) to break ties in the voting process. At times, the dossier is distributed to an odd number of adjudicators voting independently, where majority wins. Other methods may allow an even number of adjudicators to communicate with each other under certain constrained situations, enabling one or more of them to change their vote to achieve concordance. The types of workflows available to adjudicate events are almost infinitely variable, and reflect the preference of trial managers, individual adjudicators or the chairs of particular CECs.
The complexity of these adjudication workflows makes it difficult to validate the technical implementations that manage these processes. Moreover, different groups within sponsors and CROs adjudicate the same type of event differently. Further exacerbating the situation is the fact that adjudicating events of the same type across different clinical trials that are managed by the same operations group vary as well.
In short, there are no standard practices.
A death event in oncology protocol 1 managed by CRO X for sponsor Y is likely adjudicated differently by the same CRO for the same sponsor for oncology protocol 2 in the same drug program. Using different adjudication methods in cases like this prevents everyone, from sponsors all the way to regulators, to develop any kind of predictability or repeatability, creating obfuscation that makes it harder to determine outcomes. This, in turn, greatly increases risk in clinical trials. Stories abound about clinical trials whose compound seems to have failed, only to be rescued by re-adjudication processes that could better understand endpoints.
The answer to this problem is not obvious. The tension between flexibility and predictability is high, with sponsors and CROs arguing that every trial is different, and regulatory agencies needing an objective way to analyze results and reach valid conclusions in increasingly complex studies.
Driving clarity must be the primary goal of adjudication. Anything standing in the way of that clarity must be broken down. Like all science, pursuing and achieving objectivity is the only path toward successful outcomes.