Doctor who blew whistle on defunded study speaks

It’s rare for the U.S. government to revoke grants – but it happened recently, according to a report this week by the Kentucky Center for Investigative Reporting (KyCIR). As the report notes, in March the government revoked $914,000 in funding awarded to Susan Harkema at the University of Louisville in Kentucky, after discovering problems with a study that examined whether the muscle relaxant baclofen helps paralyzed patients move on treadmills. (The university has denied it lost any government funding; a representative of Louisville Public Media, which houses the KyCIR, is standing by the story.) All of this has not been news to Steve Williams, a physician now based at the University of Washington, who has been raising questions about the study for years.

Retraction Watch: What was your role in the study in question, that’s now been defunded by the National Institute on Disability, Independent Living, and Rehabilitation Research (NIDILIRR)?

Steve Williams: I was the study physician who evaluated patients for enrollment.

RW: According to the KyCIR, the study began recruiting patients in 2012. When did you start getting concerned about how it was being conducted?

SW: I became concerned in the summer of 2015 when I learned a research participant from Vermont left the study before completion.  Participants were supposed to complete 80 sessions of locomotor training after they had been randomized to a treatment arm of the study (with or without Baclofen).  The participant left the study, then emailed me thanking me for the care he had received while at Frazier Rehab.  I asked him why he had left and he replied saying he had been asked to pay for the remaining locomotor training sessions ($15,000) once his insurance had refused to pay.  He had completed 40/80 sessions and his insurance refused to pay for the remaining 40 sessions. Dr. Harkema met with him and asked him to pay to participate in the remaining 40 sessions of the study.  But the study was a sub-contracted portion of a larger data collection effort called a “Model System” grant.  The Model System grant sub-contracted money to Dr. Harkema to pay for locomotor training sessions if a participant’s insurance refused payment.

Once I knew this information, I spoke with the principal investigator (PI) of the Model System grant and gave him this information.  He tracked the sub-contracted money through accounting and found the money that had been sent to Dr. Harkema over the preceding three years was in an account over which she had sole signatory authority, and none of the money–almost $300,000–had been spent.  This was very interesting as the study was way behind recruitment goals.  The goal for recruitment was 60 patients over four years and in year three Dr. Harkema had only enrolled 20 participants and less than half of those participants had completed the study.  The question arose whether other participants had been asked to pay for participating in the study, and if they had been asked, had they dropped out for the same reason as the participant who had emailed me.

The informed consent for the study specifically states that study participants will bear no expenses other than expenses that may arise from treatment for complications they experience as a result of participating in the study.  There is no mention in the informed consent of the participants needing to pay for treadmill training should their insurance refuse payment once they are enrolled.

RW: Did you have other concerns about the study?

SW: About this same time a study participant with paraplegia who had completed the study visited an internist at Frazier Rehab and complained of pain in her leg at night.  She was wearing a long skirt, and when he pulled the skirt back to examine her knee, he was horrified to find that the lower half of her leg was displaced laterally.  He asked if her leg had always looked this way and she replied no. The leg had only looked this way after participating in the study.  He sent her for an x-ray and found that her leg was broken.  He then brought this concern to the research physical therapist who was coordinating the study, and when he looked on the “ITW” system (the record on which therapy notes were kept), he corroborated the following story.

This patient had recounted how there had been one day when she had been participating in the locomotor training portion of the study, and the therapist treating her had heard a loud pop in her knee area.  She had immediately developed swelling of the knee and an x-ray had been ordered.  She never had the study performed.  Despite never having been “cleared” for return to study participation, she had continued to participate in the treadmill training and had ultimately completed the protocol.  All on a broken leg.  As stated above, ITW notes confirm the story of this patient having heard a loud “pop” and developed knee swelling that day on the treadmill.  However, as the patient is insensate, she’d never considered the resulting pain a “big deal.”  This serious adverse event (SAE) had never been adequately evaluated or reported, and due to the time lapse between the event (which remains unreported) and the patient’s report of pain to the internist, the division chief of Rehabilitation Medicine—who is currently the study physician but who had never before been involved in this study—retroactively reviewed the chart and attested to the fact that the one (broken leg) had nothing to do with the other (study participation).

RW: When did you report your concerns, who did you report them to, and what was their response?

SW: Once I became aware of the Vermont participant being asked to pay for the remainder of his treadmill training sessions, and the participant having the broken leg, in September 2015 the PI of the study and I reported our findings to the funding agency, NIDRR (National Institute on Disability and Rehabilitation Research), a Division of the Health and Human Services.  The study’s project director at NIDRR, Dawn Carlson, PhD, immediately decided the study should be halted as it was biased by Dr. Harkema asking the participant to pay for his treadmill training sessions once his insurance refused payment for the remaining treadmill training sessions of the study.  As well, the enrollment numbers for the study were very low.

Simultaneously I had reported my concerns to the Office for Protection of Human Subjects at the University of Louisville (U of L) and to the University’s institutional review board (IRB).  The findings from the U of L IRB’s investigation were devastating.  The Baclofen study had 151 potential adverse events that had not been reported to the IRB.  There were no records of changes in dosing of Baclofen in the study.  Participants had signed the wrong consent forms, etc. Yet, the U of L IRB stated in their investigation report they were happy to continue the Baclofen study and the other studies they had examined as all of their findings had been corrected by Dr. Harkema.  They stated in their investigatory report that the University would, however, defer the decision of allowing the study to continue to NIDRR.

NIDRR sent a letter to William Pierce, the VP of Research and Innovation at UofL saying the study was defunded and should be closed.

RW: Harkema is adamant that no patients were harmed by the problems with the project. Do you agree with that?

SW: No.  As above.

There was also an additional patient who had a broken leg early in the Baclofen study before I was employed at U of L.  This was reported as a Unanticipated Problem Involving Risks to Subjects or Others (UPIRTSO).

RW: We hear many stories about the troubles whistleblowers face when they report allegations about misconduct. Did you suffer at all personally or professionally when you came forward? If so, in what way?

SW: I chose to leave the University of Louisville to protect my professional reputation on October 6, 2015.  In the preceding months I had applied to an open position for the Chief of Rehabilitation at the University of Washington in Seattle (the top rehabilitation medicine program in the country).  I was fortunate that I was offered the job prior to my reporting my findings to the U of L IRB and NIDRR.  I did not suffer any professional troubles as a result of my reporting my concerns.  However, I did take a significant pay cut to move to Seattle ($150,000 per year and lost over $500,000 in the sale of my home in Louisville).

Note: In the original Kentucky Center for Investigative Reporting story, Harkema denies knowing patients were asked to pay to be in the study, but says they knew it was a possibility. She also said the concerns stemmed from a “few disgruntled ex-colleagues [who] blitzed regulators with exaggerated accusations,” and says any allegation about the safety of the enrolled patients was “unfounded.”

 Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our new daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.