INSPECT: Identifying Non-obvious Signals in Protocols for Early Credibility Threats
The CRS is pleased to announce the launch of INSPECT, a new project funded by the Gates Foundation. INSPECT will examine whether it is possible to identify potential red flags for questionable research practices (QRPs) and potential scientific misconduct from clinical trial protocols or statistical analysis plans. Identifying QRPs, misconduct or fraud solely from trial protocol and analysis plan is inherently difficult, and in some cases, impossible. Nevertheless, there is value in exploring whether early indicators can be identified berfore a trial begins. While the SPIRIT guideline provides a structure to ensure complete reporting of trial protocols, even complete and fully compliant protocols may contain signs of methodological weakness, sloppiness or manipulation. Additional characteristics such as suspcious citiation patterns, poorly justified methods, or unplausible authorship or collaboration networks, may also serve as early warining signs.
The first challenge of INSPECT will be to define which QRPs and forms of miscounduct are detectable at the protocol stage, and to develop reliable methods to spot them. This will need both conceptual development, including a mapping of protocol-specific QRPs, and empirical validation using real-world protocols. Techniques from forensic metascience will pay a central role in the project.
Currently no established framework exists to help health research funders identify potential QRPs or red flags before a trial is conducted. Developing such approaches could assist funders prevent resource waste, promote better science and enhance public trust.
INSPECT, which officially launcehd in January 2026, will draw on evidence from published literature as well as expert consultations. We will then develop a structured detection system which links specific QRPs and forms of misconduct detectable in clinical trial protocols to tools and techniques to uncover them. All methods developed or adapted through this project will be thoroughly tested using an in-house database of study protocols.
The project will be led by Dr. Rachel Heyard with support from the entire CRS, the department of Biostatistics and the Design, Analyze and Communicate program at the Gates Foundation.