INSPECT-SR Aims to Keep Problematic Randomised Trials Out of Evidence Synthesis
Systematic reviews are built on a simple promise: find the eligible trials, judge their bias, and combine the data into a reliable estimate of benefit and harm. INSPECT-SR was developed to address the part of that process that too often slips through the net — trials that are not merely flawed, but fundamentally untrustworthy.
The new tool, Investigating Problematic Clinical Trials in Systematic Reviews, is designed to help researchers assess the trustworthiness of randomised controlled trials before those studies are absorbed into meta-analyses, guidelines, and practice-changing recommendations. The problem is not academic. When a fabricated, falsified, or otherwise compromised trial sits inside a pooled analysis, it can distort the apparent size of benefit, conceal harm, or shift a conclusion entirely.
One sentence captures the urgency plainly: untrustworthy trials can make reliable reviews unreliable.
Why Risk of Bias Tools Are Not Enough
Traditional evidence appraisal focuses on risk of bias, but problematic trials can escape that framework. Serious concerns about trustworthiness may arise from research misconduct, including plagiarism, falsification, or fabrication of data or methods. They may also stem from critical errors in study conduct that are severe enough to undermine the findings without fitting neatly into standard bias domains.
That gap matters because evidence syntheses do not fail only when studies are poorly blinded or incompletely reported. They fail when the underlying data themselves cannot be trusted. If those studies remain in the pool, downstream recommendations may be based on results that never should have entered the analysis in the first place. In practical terms, the damage can last for years, especially when a problematic trial is not retracted promptly — or is never retracted at all.
The article points to a sobering backdrop. Carlisle estimated that 26% of randomised controlled trials submitted to the journal Anaesthesia were problematic. A separate review identified 847 systematic reviews that included one or more retracted randomised controlled trials in meta-analysis. Removing those studies changed the magnitude of effect in 16% of meta-analyses and changed the direction of effect in just over 8%.
A Tool Built for Screening Trustworthiness, Not Just Bias
INSPECT-SR was created to fill a specific methodological need: spotting trials that should be treated as potentially problematic before they affect the evidence base. The tool is intended for use in systematic reviews of randomised controlled trials, where the standard workflow begins with study identification and then moves to appraisal and synthesis. INSPECT-SR adds a trustworthiness screen to that sequence.
The logic is straightforward. A review can only be as sound as the studies it includes. If a trial is so compromised that its data or methods cannot be relied upon, then no amount of statistical precision can rescue the pooled estimate. By giving reviewers a structured way to assess trustworthiness, the tool is meant to reduce the chance that misleading studies are carried forward into guidelines and clinical decision-making.
The accompanying guidance document is part of the same effort. Together, the tool and guidance aim to standardise how reviewers handle trials that raise red flags for integrity, whether those concerns arise from misconduct or from serious critical errors that make the findings unsafe to use.
The Stakes Extend Beyond Publication Ethics
This is not just a matter of correcting the literature after the fact. Evidence synthesis influences treatment recommendations, and treatment recommendations shape bedside care. When a problematic trial is treated as legitimate evidence, the distortion can propagate through meta-analysis, guideline panels, and eventually into decisions made for real patients.
In fields where treatment effects are modest, a single unreliable study can have disproportionate influence. It may tip a pooled estimate over a threshold of apparent efficacy, conceal an important safety concern, or create false confidence in an intervention that does not deserve it. The downstream consequence is not simply an erratum in a journal. It can become a clinical recommendation built on unstable ground.
The authors’ central concern is therefore operational as much as ethical: reviewers need a practical way to identify trials that should be excluded from synthesis because their trustworthiness is in serious doubt. INSPECT-SR is meant to provide exactly that kind of framework.
Evidence Integrity Now Sits Inside the Review Process
The development of INSPECT-SR reflects a larger shift in evidence-based medicine. Reviewers can no longer assume that every randomised trial marked as eligible is fit for inclusion. The literature contains studies that are retracted only after they have already influenced syntheses, and many others may never be formally retracted despite deep concerns about their validity.
That reality changes the job of the systematic reviewer. Assessing outcomes and risk of bias is no longer sufficient on its own when the trial itself may be structurally unreliable. A review process that ignores trustworthiness leaves a blind spot exactly where the highest-stakes errors can occur.
By formalising a trustworthiness assessment tool, INSPECT-SR gives reviewers a way to confront that blind spot directly. The purpose is not to replace existing appraisal methods, but to add a separate layer for the problem of problematic trials — a category that can include both misconduct and serious conduct errors.
The broader implication is concrete. Better screening of trial integrity should reduce the chance that distorted data are embedded in meta-analyses, guideline recommendations, and clinical practice. For systematic reviewers, that means fewer false positives, fewer false assurances, and a more defensible evidence base.
What the Current Evidence Still Leaves Unsettled
The article is clear about the need for a trustworthiness tool, but the field still faces unresolved methodological questions. Problematic studies are difficult to detect, which means prevalence is hard to measure with precision. The available data likely understate the scale of the issue, especially because some trials take years to be recognised as unreliable and others may never be formally removed from the literature.
That uncertainty is part of the challenge INSPECT-SR is meant to address. Reviewers need a reproducible process for deciding when a trial is too compromised to include, and they need guidance that can be applied before unreliable data change the direction of a meta-analysis. The article presents the tool and its guidance as a response to that exact need, in a research environment where evidence integrity can no longer be assumed from publication status alone.