The Inner Life of a Complex Clinical Trial

Team Meeting at the BHP Photo By Dominic Chavez

Sent: October 30, 2013
Subject: YA TSIE-In the Field

Hi,
This is to confirm that the YA TSIE Study—The Botswana Combination Prevention Project (BCPP) is underway and the team deployed today 30th October in the field at both Ranaka and Digawana. While there have been challenges and some initial delay in study initiation, it is with great confidence in the team and their ability to rise up to challenges that I am confident this shall become one of BHP’s blue chip studies. I wish the team success and God’s Speed as they roll out this immensely challenging study.
Joe
Dr. Joseph Makhema
C.E.O. Botswana Harvard AIDS Institute

Evolving Guidelines

It was the first week of June in 2015. The BCPP leadership team was stressed. There were even more conference calls than usual. Two upcoming events could change the course of the trial. The first was a June 11th meeting of the Data Safety Monitoring Board (DSMB) at the National Institutes of Health (NIH) in Bethesda. The second was the July meeting of the International AIDS Society (IAS) in Vancouver.

In May, preliminary results from the Strategic Timing of Antiretroviral Treatment (START) Trial had been announced. START was designed to determine the best time for HIV-infected individuals to begin antiretroviral therapy (ART). The trial was stopped early because the data showed a strong benefit to starting ART in all HIV-infected individuals as soon as they were diagnosed, regardless of their CD4 count. CD4 cells are white blood cells that play a major role in protecting the body from infection.

Clinic in Dar es Salaam, Tanzania
Clinic in Dar es Salaam, Tanzania

Before the START results, there had always been a question about the risks versus benefits of long-term ART treatment for people who felt healthy and had a high CD4 count. With decisive evidence from START, there was widespread speculation that the World Health Organization (WHO) would now recommend a policy known as Universal Test and Treat (UTT), in which all people with HIV start ART as soon as possible. UTT would be simpler to implement, require less monitoring, be good for individuals, and also be good for a community—people put on treatment earlier would be less likely to infect their partners.

Because the strong START results could be used as a justification to alter existing HIV trials, there was a possibility that the DSMB would stop the Botswana project, but there were also good reasons why they wouldn’t. The WHO hadn’t yet changed guidelines, and even when they did, many countries wouldn’t have the money or resources to implement a UTT strategy.

In 2013, the WHO had changed guidelines to recommend initiation of ART in all HIV-infected adults with CD4 below 500 (cells in a cubic millimeter of blood), up from their prior recommendation of 350. Botswana national guidelines were still at 350. In April 2016, Botswana’s Ministry of Health was expected to move to 500. Now there was a question of whether they would skip 500 and move straight to UTT. Before making that decision, the Ministry would have to consider issues of cost, infrastructure, and competing healthcare needs—including trauma, cancer, and heart disease.

The BCPP team had anticipated the change to 500 and already implemented the change in their study protocol. “If we’re testing a strategy that isn’t keeping up with where the standard of care is headed, it will no longer be an ethical, relevant or useful study,” said Dr. Shahin Lockman, a Harvard investigator. But the START Trial being stopped so early was a surprise to the research community. “In HIV the new data and standard of care change more rapidly than we can often keep up with,” said Lockman. “Some of it should be anticipatable, other elements are not. But you’re also paralyzed if you don’t start a study because you’re waiting for more relevant information to come in.”

Dr. Shahin Lockman
Dr. Shahin Lockman. Photo by Kent Dayton

Every time a change is made to the BCPP, the study protocol must be amended and then reapproved by both the Botswana and the CDC Institutional Review Boards (IRBs). Every change involves a time-consuming, bureaucratic process.

At the June 11th meeting, representatives of all the BCPP partners crowded into an NIH conference room. Lockman and Dr. Max Essex, Principal Investigator of the BCPP and Chair of the Harvard AIDS Initiative (HAI) were there. The DSMB asked the team to look into moving to UTT, but there was no discussion of stopping the trial. The BCPP would help determine whether widespread rapid implementation of treatment was a possibility. Its findings would provide much-needed guidance for expanding large HIV treatment programs. Within a few years, the BCPP researchers hope to provide what Lockman calls “the strongest possible evidence to people who decide how to spend billions of dollars.”

At the Vancouver meeting in July, the WHO announced they would now recommend a policy of UTT.

Everything Changes

“You can’t just design a study and make an assumption that a year or two later you’ll be doing the same thing,” said Essex. “You have to allow for the possibility that the study has to be completely redesigned to get information that will be valuable.”

Dr. Max Essex
Dr. Max Essex. Photo by Kent Dayton

Besides adapting to new research findings, the team must also contend with the different institutional cultures of each of the BCPP partners. Harvard, the CDC, and Botswana’s Ministry of Health must all coordinate schedules, data, and communications to serve a common purpose. “This is a large and complicated study with multiple partners working across different continents and time zones on different components of the overall effort, so communication and rapid decision-making are challenging, especially with the levels of oversight and regulatory review that are required,” said Lockman. “Being quick and effective in making changes is a challenge.”

Prevalence & Power

A large trial like the BCPP must be flexible, while at the same time maintaining statistical rigor. If reality in the field proves to be markedly different from assumptions made in the original study design, the team must adjust in real time. The BCPP biostatisticians constantly monitor data from the field and make adjustments as necessary.

In some villages, for example, the actual HIV prevalence was found to be much higher than the 28% anticipated when the protocol was written. Rates in a few villages were as high as 40%. This meant that in a high-prevalence village, the field team had to test more people than originally anticipated in order to meet their targets for uninfected adults. The process of interviewing, testing, and counseling one person can take several hours. The additional interviews meant that the team remained in a village longer than expected and began work in the next village later than originally scheduled.

In addition to field conditions differing from models, biostatisticians must also contend with changes in the protocol. Though the BCPP study population is over 100,000, the biostatisticians must make sure that, after changes, their study size is still large enough to show statistical significance.

“The biggest problem with a study like this is that you need a lot of people involved to get the statistical power to see a clear difference for the effect that you’re seeking,” said Essex. “If the study has to be changed so often because of new developments and better ways of treating or different ethical standards, then the challenge becomes even greater.” Results from BCPP should be available in 2018.

Title photo: Team meeting at the Botswana Harvard Partnership headquarters. Photo by Dominic Chavez