FEATURE

Trials and Errors

As bias infiltrates today’s clinical research landscape, patients pay the price

Advertisement

 

By Madeleine Burry
October 3, 2025 | VOLUME 3, ISSUE 3

Without ongoing research to identify new medications and best practices for interventions, health care stands still. But it’s not only important that research continues. Ceaseless efforts to stamp out bias within studies and clinical trials are also key.

That means careful consideration of what gets researched, who leads and participates in studies, and how studies are implemented.

“It’s fundamental that we try to minimize [bias in medical research] because there could be implications and consequences related to efficacy or safety. So the more transparent and less biased the research, the better the care that will be given to patients,” said Dominique Demolle, PhD, Chief Executive Officer at Cognivia, a manufacturer of analytical tools to improve clinical trials.

Dominique Demolle, PhD

Dominique Demolle, PhD
Chief Executive Officer
Cognivia

Types of bias in clinical research

Even with safeguards in place, bias creeps in: Study participants may give immoderate responses to questions (extreme response bias), participation can lead to behavior changes (the Hawthorne effect), and study makeup can favor certain groups (sampling bias). And there are several other common sources of bias.

A lack of diversity in participants

There’s a long history of clinical trials made up solely or primarily of White men. To counteract this, in 1993 the National Institutes of Health (NIH) began requiring women and minorities to be included in clinical studies.1,2

Research that limits diversity does a disservice—people’s reactions to medications and interventions can vary based on age, sex, or ethnicity.

“We have made some advances [in improving diversity in clinical research], but there's still more work to be done,” said Anita Phung, MRGCP, MRCP, MBBS, a general practitioner who helps oversee clinical trials at Lindus Health, a clinical research organization. A review of 20 years of registered clinical trials in the United States revealed that only 43% of studies reported data around ethnicity, and, when provided, 79% of enrollees were White individuals.3

Anita Phung, MRGCP, MRCP, MBBS

Anita Phung, MRGCP, MRCP, MBBS
General Practitioner
Lindus Health

Designing unbiased studies can be challenging. Researchers balance fair representation of various demographics with keeping the sample size affordable, Dr. Demolle said. And some groups can be hard to recruit for a variety of reasons, including lower health literacy levels and socioeconomic factors (such as lower income or curtailed time due to caretaking responsibilities).

Publishing bias

In clinical research, positive outcomes are published more frequently than negative ones. “If we really unpick the human psyche, part of it is maybe pride, dignity, and the [desire for] credibility and prestige behind being able to publish positive results,” Dr. Phung said.

When negative outcomes aren’t publicized via journals, findings in a given area can appear rosier than reality. This becomes a problem when meta-analyses are conducted, Dr. Demolle said.

“It’s very, very important to publish all the negative trials. Ultimately, the more data there [are], the more we can appraise [them] and be critical about [them]... [and they] can contribute to the wider evidence base,” Dr. Phung said. In some ways, research is wasted if it’s not published and shared, she noted.


“The more transparent and less biased the research, the better the care that will be given to patients.”


But there are positive steps afoot here, Dr. Demolle said. As of 2016, clinical trials receiving funding from the NIH must share a summary of their results.4 That’s better, although not perfect: In terms of crafting a representative meta-analysis, a synopsis “is not the same as publication in a peer-reviewed journal,” Dr. Demolle noted.

Funding bias

A finite amount of funding is available for clinical research.

That leads to trade-offs: Government entities may opt to funnel money in directions that serve the greatest number of people (resulting in less funds available for rare diseases), while the private sector may prioritize funding research with the potential for profit rather than toward health issues that primarily affect marginalized communities. Another aspect of funding bias to note: When researchers receive funding from private sector sources, it can skew the details of a study from design through publication.5

Then there’s the question of whose research gets funded.

After a researcher applies for a grant through the NIH, there’s an array of steps taken, including prospective research being assigned an impact score.6 “You don’t get to see who received what score,” said Dr. N., a pulmonologist and assistant professor at a teaching hospital in Massachusetts, who requested anonymity. “You only see the ultimate outcomes. You only see the disparities in who gets funded.”

Just as people of color are underrepresented as study participants, so too are they underrepresented as leaders in research. In 2020, only 2% of NIH research project grants were awarded to Black scientists and only 5% to Latinx scientists.7 For people in these groups, “when their grants go before the NIH, they have a higher odds of [not being approved] even after you control for factors like their prior productivity and the quality of their letters and the number of years they've been doing research,” said Dr. B., a pulmonologist and critical care practitioner and assistant professor in California, who also requested anonymity.


“We have made some advances [in improving diversity in clinical research], but there's still more work to be done.”


In 1989, the NIH introduced diversity supplements, which help fund projects for people of marginalized and underrepresented groups, in an attempt to change this pattern. Dr. B. applied for and received this funding in 2021. This program is designed to address the question of “How do we create a system that's more fair?” and also to prioritize the study of systemic equity issues, Dr. B. said.

Explicit bias isn’t the most prevalent form of bias in medicine and research, Dr. N. said. Rather, it’s the commitment to continue doing things as they’ve been done instead of focusing on equity, he said.

A singular moment for health research in the United States

As we’ve seen this year, research funding can become highly politicized. Under the current administration, the NIH has cut $1.8 billion in research grants—specifically targeting areas of research that focus on diversity, equity, and inclusion; HIV; COVID-19; and transgender health.8 These cuts have affected thousands of research grants.9 The impact of these sudden changes are real for individuals and institutions engaged in research—and for the patients and health care system as a whole.

A grant to study race-neutral PFT guidelines

Take race-neutral lung function interpretation, for example. There’s a known difference in levels of FVC between Black and White individuals in the United States.10 Despite this difference, there are advantages to using a race-neutral pulmonary function test (PFT), Dr. B. said. “They've looked at transplant and found it would have changed the prioritization of who gets [a] lung transplant. They've looked [at] COPD and found that it more accurately predicts exacerbation and hospitalization risk as well as symptom burden,” he said. This isn’t a finding that solely benefits [Black] people but rather something that’s relevant to all, Dr. B. pointed out. “This race-neutral equation was more effective.”


“[When] there's representation in research and there's representation in the participants, it's going to lead to better outcomes overall.”


As a next step, Dr. B. wanted to study another important respiratory disease to see how it would be affected by shifting to a race-neutral metric.

However, his research is currently on hold. “Technically, my diversity supplement expired and just wasn't extended. But it was a kind of situation where, at least as I understand, they routinely had been offering these extensions and just didn't because of the changes in policy,” Dr. B. explained. As a result, his institution is now trying to marshal internal funding resources, and he’s seeking out alternate funding.

“It really slowed down my productivity,” Dr. B. said.

A grant to improve lung cancer screening guidelines

Also on hold: Dr. N.’s research to improve lung cancer screening rates.

Less than 20% of eligible individuals get a low-dose CT scan to screen for lung cancer.11,12 This “is [a] really stark comparison to other cancer screening exams,” Dr. N. said. Providers may not be recommending the insurance-covered screening because expanded recommendation guidelines are new, and time is short during appointments.13 “And lung cancer is unique. There's a lot of stigma around smoking,” Dr. N. said. Fear of the results can also be a blocker in getting screened.

This combination of a screening test that can reduce cancer mortality and a low adoption rate leaves room for opportunity. Getting recommendations from trusted messengers—people who come from within one’s community—could lead to increased uptake of low-dose CT scans and tobacco treatment, Dr. N. said. Plus, it would remove one item from providers’ to-do lists during brief appointments with patients. Dr. N. and his team have already developed a messaging prototype and training process for health workers.

When Dr. N. submitted his grant application to the National Cancer Institute, he received the highest-possible impact score. His funding was ultimately provided via a diversity mechanism (Dr. N. is from an underrepresented group in medicine). But now these efforts to increase uptake in lung cancer screening are on hold because his funding was cut.

“We're ready to start training patient navigators and roll out the trial, except now we don't have funding to pay the patient navigators or the research participants,” Dr. N. said. An extra layer of frustration is his belief that “if I wasn't a diversity applicant, my grant would have been funded, and it wouldn't be under threat of being cut,” he said.

Funding cut ramifications

Clinical research funding cuts have a ripple effect, from people in health care to patients—which is, of course, all of us at some point in our lives.

One day, a project is funded; the next, funding is cut. It’s an emotional experience.


“If I wasn't a diversity applicant, my grant would have been funded.”


One of the first things Dr. N. felt was doubt. “You start questioning yourself as an individual. Is it because my work is not meaningful?” he said. He’s frustrated for the people who worked with him to design the intervention, as well as for his patient navigator partners and potential study participants.

Ongoing legal battles, an overall lack of clarity, and digging to find alternate funding can make the experience even more dizzying.

A disservice to health

There’s value to studying marginalized and vulnerable populations for everyone who needs health care.

Consider the proverbial canary in a coal mine. These songbirds were not sent underground for their melodic noises. Rather, it’s “because they had a more frail respiratory system,” Dr. B. said. When poisonous gasses amassed, canaries would be affected first, pausing their singing and indicating that an evacuation was merited for humans. Dr. B. continued with his analogy: “When there's a problem in our health care system, eventually it will impact everybody. But you can often see those impacts first or most easily among the populations that are already in that marginalized position.”


“When there's a problem in our health care system, eventually it will impact everybody.”


A burden on the health ecosystem

Clinical research is a long, cooperative endeavor.

“There is no scientific question that you can ask today and answer tomorrow to change the state of health care,” said Bert Hartog, PhD, the founder of Hartog Digital Health Consulting, who has decades of experience in clinical trial development. Time is required to know what questions to ask, to design and perform research, and to understand the results.

Dominique Demolle, PhD

Bert Hartog, PhD
Founder
Hartog Digital Health Consulting

Interrupting that process has an effect like reducing oxygen levels, he said. “That doesn't mean it will die, because you can survive on reduced oxygen, but you will thrive much less, and you'll be less healthy,” he said. Research output may be reduced and less strong. The effects will potentially be uneven—just as reducing oxygen levels is more challenging for some individuals than others, reducing funding will be harder for some programs than others, Dr. Hartog said.

The impact of politically motivated fiscal cuts in a targeted area of medical science is further hindering the pursuit of sound, diverse, unbiased, science-based medicine. Clinicians and researchers do not yet know how to navigate, avoid, or overcome this new reality.

But the ripple effect will come down to patients, as it always does. “Some people will suffer because their treatment will not be developed,” Dr. Hartog said. “And what is being developed will take longer.”

What Else Would Improve Research?

Talk to health researchers about the various biases found in their line of work, and they’ll want to share what can be done to improve study designs. Here are a few ideas.

Use technology

Conversations via video. Wearable technology reporting vitals. Digital surveys.

Technology makes it easier to gather data without requiring patients to visit a clinic, which can be time-consuming.

“I specifically design decentralized clinical trials in order to make trials more accessible,” said Anita Phung, MRGCP, MRCP, MBBS, a general practitioner who helps oversee clinical trials at Lindus Health, a clinical research organization. These types of trials can benefit women, who are often primary caretakers of children and older parents; people who are short on time for any nonessential activities; and people who may be balancing multiple hourly jobs to make a living, she said.

But use of technology shouldn’t take the place of conversation and connection. Spending time addressing concerns is also important, Dr. Phung said.

Collect more data

Studies are better now—although still not perfect—about collecting demographic data of participants. But Dominique Demolle, PhD, Chief Executive Officer at Cognivia, a manufacturer of analytical tools to improve clinical trials, would love to see even more information collected about patients that are not just objective measurements. Details about factors that will influence a person’s adherence to treatment and even details on their personality can be meaningful. After all, the placebo effect is powerful.

Consider this: An individual who gets medication and swallows it daily as prescribed may have a different reaction than an individual who reads every page in the packaging insert before taking their first pill, Dr. Demolle said.

Consider the patient journey

Good trial design requires empathy, Dr. Phung said, and deep consideration of the participant journey. When she examines the schedule for a trial, she tries to understand the clinical rationale and what’s being measured at every moment. And then she asks, “Are there ways that we can do this better?” Perhaps that means having someone get blood tested in their home or a local health clinic rather than traveling to the study site.

When surveys are involved, Dr. Phung tries to imagine how it feels to complete them herself, while maintaining scientific integrity. Some of the questions she asks are, “Is it too intrusive? Is it too much? What's the frequency of it?”

“Participating in a trial can be a real burden,” said Bert Hartog, PhD, founder of Hartog Digital Health Consulting. It can impose both physical and emotional imposition, he said. Decades ago, informed consent in trials used to consist of sharing the risks and benefits; now, education is woven in, with explanations of what is going to happen in the trial and why it’s important for the research.

“We want maximum engagement, really. We don't want to scare people away,” Dr. Phung said.


References

  1. Mastroianni AC, Faden R, Federman D. Women and Health Research: Ethical and Legal Issues of Including Women in Clinical Studies: Volume I. Washington, DC: National Academies Press (US); 1994. https://www.ncbi.nlm.nih.gov/books/NBK236531/
  2. Bibbins-Domingo K, Helman A. Improving Representation in Clinical Trials and Research: Building Research Equity for Women and Underrepresented Groups. Washington, DC: National Academies Press (US); 2022. https://www.ncbi.nlm.nih.gov/books/NBK584392/#
  3. Turner BE, Steinberg JR, Weeks BT, Rodriguez F, Cullen MR. Race/ethnicity reporting and representation in US clinical trials: a cohort study. Lancet. 2022;11:100252. doi:1016/j.lana.2022.100252
  4. NIH Policy on the Dissemination of NIH-Funded Clinical Trial Information. National Institutes of Health. 2016. https://grants.nih.gov/grants/guide/notice-files/NOT-OD-16-149.html
  5. Fabbri A, Lai A, Grundy Q, Bero LA, The influence of industry sponsorship on the research agenda: a scoping review. Am J Public Health. 2018;108(11):e9-e16. doi:2105/AJPH.2018.304677
  6. Scoring & Summary Statements. National Institute of Allergy and Infectious Diseases. https://www.niaid.nih.gov/grants-contracts/scoring-summary-statements
  7. Hill KA, Desai MM, Chaudhry SI, Nguyen M, Boatright D. NIH diversity supplement awards by year and administering institute. JAMA. 2021;326(23):2427-2429. doi:1001/jama.2021.19360
  8. Bendix A. Trump administration cut more than $1.8 billion in NIH grants. NBC News. 2025. https://www.nbcnews.com/health/health-news/trump-administration-cut-18-billion-nih-grants-rcna205568
  9. NIH Grant Terminations in 2025. Grant Witness. 2025. https://airtable.com/appjhyo9NTvJLocRy/shrNto1NNp9eJlgpA?Ffj6Q=sfsOr1bXR7f3fLQwK&Ffj6Q%3Aview=plasNjSudszRmPvEo
  10. Baugh A, Adegunsoye A, Connolly M, et al. Towards a race-neutral system of pulmonary function test results interpretation. Chest. 2023;164(3):P727-P733. doi:1016/j.chest.2023.06.005
  11. Bassett M. Lung Cancer Screening Rates Nearly 4 Times Lower Than Breast, Colon Cancers. MedPage Today. 2025. https://www.medpagetoday.com/pulmonology/lungcancer/114945
  12. Potter AL, Kothagundla S, Haridas C, Chang AEB, Sequist LV, Yang CFJ. Preventive health care use among adults eligible for lung cancer screening in the US. JAMA. 2025;333(18):1635-1638. doi:10.1001/jama.2025.2157
  13. Lung Cancer: Screening. US Preventive Services Task Force. 2021. https://www.uspreventiveservicestaskforce.org/uspstf/recommendation/lung-cancer-screening

 


Read more from this issue