Editors Flag Concerns Over Fuzzy Logic Method for Ranking Orphan Drugs

A widely noticed study that proposed a new mathematical technique for evaluating orphan drugs has been placed under an Expression of Concern by the editors of PLOS One, a development that highlights how the integrity of a paper’s citation base can trigger formal scrutiny even when the paper itself has not been retracted. The notice, published on October 12, 2026, applies to a 2024 article by Hu and Pang that introduced a multi-criteria group decision-making, or MCGDM, method built on correlation coefficients within a probabilistic hesitant fuzzy environment and applied it to the clinical comprehensive evaluation of orphan drugs. The editors’ stated reason is specific and narrow: three of the references cited in the original paper were retracted after the paper appeared.

An Expression of Concern is one of the standard instruments in the toolkit of scholarly publishing for handling doubts about a published work. It sits between doing nothing and issuing a full retraction. Editors typically turn to it when an investigation is incomplete, when the problems identified do not by themselves invalidate the paper’s central findings, or when the parties responsible cannot be reached. In this case, the PLOS One editors did not allege problems with the paper’s own data, methods, or conclusions. Instead, they informed readers that references 29, 86, and 88 of the article had been retracted after the article was published, and they advised readers to interpret the paper with caution. That phrasing matters: the notice preserves the paper’s place in the literature while signaling to anyone relying on it that part of its evidentiary scaffolding has since collapsed.

The original study, published in May 2024 in PLOS One, addressed a genuinely difficult problem in medicine and health policy. Orphan drugs, developed to treat rare diseases, present evaluators with an unusually tangled set of considerations. Patient populations are small, clinical trial data are often limited, prices are frequently high, and the social value of providing any treatment at all can weigh heavily against conventional measures of cost-effectiveness. Deciding which orphan drugs deserve funding, reimbursement, or priority in a clinical program therefore requires weighing multiple, partly conflicting criteria simultaneously, and often involves groups of experts whose judgments may be uncertain, incomplete, or hesitant.

This is precisely the setting in which fuzzy decision theory has grown popular. Classical decision analysis assumes that experts can assign crisp numbers to alternatives, but real clinical judgment rarely works that way. A specialist assessing a new therapy for a rare disease may believe several possible effectiveness ratings at once, or may feel that some outcomes are more likely than others without being able to commit to a single value. Hesitant fuzzy sets were introduced to capture exactly this: rather than forcing one number, they allow a set of possible membership values. Probabilistic hesitant fuzzy sets go a step further by attaching probabilities to each of those possible values, so that an expert’s hesitation is not just recorded but weighted by how plausible each option is considered to be.

Within that framework, the authors of the 2024 paper built their method around correlation coefficients. In fuzzy multi-criteria analysis, correlation measures are used to compare how closely the profile of one alternative, such as a candidate orphan drug, matches the profile of an ideal or reference alternative. By computing correlation coefficients between each drug’s aggregated expert assessments and ideal positive and negative reference points, the technique can rank the drugs according to how strongly they resemble the ideal choice and how weakly they resemble the worst case. The group dimension of the method, reflected in the MCGDM label, means that the assessments of multiple experts are aggregated, with the probabilistic structure helping to preserve information about disagreement and uncertainty rather than averaging it away prematurely.

The authors applied this machinery to the clinical comprehensive evaluation of orphan drugs, positioning the technique as a decision-support tool for settings in which committees must compare treatments for rare conditions under uncertainty. Work of this kind belongs to a broad and active research stream: over the past decade, researchers have proposed a proliferation of fuzzy extensions, including interval-valued, dual, and probabilistic variants, each paired with aggregation operators, distance measures, or correlation measures and demonstrated on case studies ranging from medical decision-making to supplier selection. The appeal is obvious for orphan drugs, where the data are sparse and expert judgment carries real weight, and where a structured method can make committee decisions more transparent and reproducible.

Yet the Expression of Concern illustrates a subtler vulnerability of this literature. Methods papers in applied fuzzy decision theory typically cite dozens of prior works that define the underlying sets, operators, and measures, and that supply earlier applications. If a meaningful fraction of that citation base rests on retracted papers, the conceptual foundations the new method inherits may themselves be compromised. A retracted reference can mean anything from duplicated text and unreliable peer review to fabricated data, and readers cannot tell from the notice alone which applies to references 29, 86, and 88. What the editors made clear is the timing: the retractions occurred after the Hu and Pang paper was published, meaning the authors could not have known about them at submission. That distinction is central to why the paper received a notice of concern rather than an immediate retraction.

For the research community, the practical consequences are worth spelling out. Citations to the 2024 paper, and any downstream work that adopts its correlation-based MCGDM technique, now carry an added layer of caution. Reviewers and editors encountering the method in new submissions are likely to expect authors to acknowledge the Expression of Concern and to justify whether the technique’s foundations remain sound. Practitioners who might have considered the method for real-world orphan drug evaluation, such as health technology assessment bodies weighing treatments for rare diseases, now face a reason to pause, since the notice explicitly advises interpreting the article with caution. In bibliometric terms, the notice also becomes a permanent, machine-readable flag attached to the paper’s DOI, ensuring that databases and reference managers surface it to future readers.

The episode also reflects broader trends in research integrity enforcement. Journals increasingly monitor their published record after publication, tracking retraction notices across the literature and cross-checking them against their own reference lists. When retracted papers are found in a citation network, editors must decide whether the retraction merely touches peripheral background material or strikes at the heart of the paper’s claims. The graduated response, ranging from a correction to an Expression of Concern to a full retraction, is meant to calibrate that judgment. PLOS One, a mega-journal that publishes thousands of papers annually, has been a frequent user of such notices, and the present case follows the standard format: a short editorial statement, the specific references involved, and an explicit advisory to readers.

For now, the 2024 paper by Hu and Pang remains in the literature, its method formally described and its application to orphan drug evaluation still readable, but under a cloud of editorial caution. The Expression of Concern does not retract the technique, dispute its mathematics, or question its clinical case study; it flags that three of the works the paper relied on have since been withdrawn from the scientific record. Whether the notice is later upgraded to a retraction, resolved with corrections, or simply stands as a permanent caution will depend on further editorial assessment. In the meantime, the case serves as a reminder that in modern science, the reliability of a paper is tied not only to its own data and reasoning but to the health of the entire web of scholarship it stands on, and that publishers are increasingly willing to weave visible warning threads into that web when parts of it unravel.

Subject of Research: An Expression of Concern issued on a multi-criteria group decision-making study using probabilistic hesitant fuzzy correlation coefficients for clinical evaluation of orphan drugs

Article Title: Expression of Concern: A novel MCGDM technique based on correlation coefficients under probabilistic hesitant fuzzy environment and its application in clinical comprehensive evaluation of orphan drugs

Article References: The PLOS One Editors (2026). Expression of Concern: A novel MCGDM technique based on correlation coefficients under probabilistic hesitant fuzzy environment and its application in clinical comprehensive evaluation of orphan drugs. PLOS One, 21(10), e0359700. https://doi.org/10.1371/journal.pone.0359700

Image Credits: AI Generated

DOI: 10.1371/journal.pone.0359700

Keywords: PLOS One, Expression of Concern, multi-criteria decision-making, probabilistic hesitant fuzzy sets, correlation coefficients, orphan drugs, rare diseases, clinical evaluation, research integrity, retracted references, scientific publishing, Expression