Single-Arm Trials Can Provide Randomized Real-World Evidence: The Random Invitation Single-Arm Trial Design.

Abstract

Single-arm trials can be used to explore the feasibility, implementation, and effects of treatment. They typically use opportunistic convenience sampling to find potential participants. Their main limitations for health care decision making are lack of generalizability and the poor quality of the comparative evidence they produce. The authors propose a single-arm trial design that can provide greater generalizability and higher quality of comparative evidence than traditional single-arm trials, called a random invitation single-arm trial or RISAT. A RISAT has 4 essential components. First, it has a data infrastructure for routinely collected real-world data (RWD) where participants have provided consent to have their data used for research (for example, a registry or electronic medical record database). Second, a subset of those participants are randomly invited to take part in the RISAT. Those not invited are not contacted. Third, invitees are offered the specific intervention (such as a novel treatment), to which they consent or not. Fourth, all invitees are followed prospectively regardless of their acceptance of the intervention. For an optional randomized comparison, RISATs can use the RWD infrastructure to measure outcomes from invitees and noninvitees. The authors describe the advantages and challenges of this approach, including inferential issues and biases and comparison with other designs. They show how RISATs allow fairer access to participation, improve applicability and generalizability of results compared with traditional single-arm trials, and provide a form of randomized real-world evidence. At negligible added cost beyond already existing infrastructure, this approach may catalyze the early generation of evidence of higher value than that produced with traditional single-arm trials, increasing the credibility and validity of accelerated drug approval processes and enabling better health care decisions.