SEARCH channels the power of group science and community engagement to guide a global effort to end AIDS
Now approaching its 35th year, the HIV/AIDS epidemic has already claimed 39 million lives. An estimated 30 million people are currently living with HIV worldwide, more than two-thirds of them in Africa. But 2015 looks to be a year of change, opportunity, and challenge for the global AIDS response effort. On the treatment front, recent results from the Temprano clinical trial, a seven-year study in Côte d’Ivoire, showed that immediate treatment dramatically decreases the risk of severe illness in persons with HIV infection. HIV treatment has also been shown in randomized trials to have a secondary benefit of reducing transmission from an HIV-positive person to an uninfected sexual partner by as much as 96 percent.
As scientific knowledge about best treatment and prevention practices continue to expand, an end to the epidemic may be within reach. One of the major challenges is a question of scale and resource allocation. What do the prevention and treatment strategies look like in actual communities, where best practices quickly run up against the challenges of stigma, practicalities of resource-poor environments, and the general messiness of human living?
A broad consortium with the appropriate acronym SEARCH is already seeking answers to these questions. Sustainable East Africa Research in Community Health is a collaboration between U.S. and African universities, government organizations, and community stakeholders that proposes to stop the spread of HIV in Africa and then inform the global effort to end AIDS.
“SEARCH will be able to inform how to do it, how much it costs, and how much of an impact it can have. That’s the sort of information really needed to get an international response to stop AIDS.”
The SEARCH study includes 32 communities of roughly 10,000 persons each, in Uganda and Kenya, the largest cohort of current HIV studies in Africa. The consortium itself is also numerous and broad. Led by Dr. Diane Havlir, professor of medicine at UCSF, and Dr. Moses Kamya, professor and chair of the Department of Medicine at Makerere University (MU) in Kampala, Uganda, the project includes investigators from UCSF, MU, UC Berkeley, the Kenya Medical Research Institute, and the Infectious Disease Research Collaboration. The NIH, PEPFAR, World Bank, WHO, UNAIDS, and Gilead are all sponsors of the study. And external advisory boards in Kenya and Uganda include the ministries of health, community members and leaders, HIV care organizations, the private sector, and the ministries of finance.
“This is a great example of group science—no one scientist or one individual could tackle the SEARCH project alone,” says Havlir. “We needed to establish a cooperative group and include the communities themselves from the very beginning, and every step of the way.”
Advances in science and policy point to a new model for care
At its core, the SEARCH Consortium is evaluating a “test and treat” strategy, which posits that HIV/AIDS can be eliminated if all adults are tested regularly and all infected persons are put on antiretroviral therapy (ART)—regardless of the level of the virus in their bloodstream. There have been some challenges and concerns surrounding earlier treatment of HIV, including drug resistance, long-term toxicity of medications, and cost of treatment.
“In earlier days, there were more side effects and the drugs were harder to take and required more frequent and exact dosing to avoid developing resistance,” says Maya Petersen PhD ’07, assistant professor of biostatistics and epidemiology at the UC Berkeley School of Public Health and study statistician in the SEARCH Consortium. “There was a worry that, if you started people too early when they feel fine and seem to be healthy, that you could end up doing more harm to them through toxicity or wiping out their drug options.”
But, with treatment options expanding and the science of HIV replication advancing, policy is shifting in favor of earlier treatment. In October 2014, UNAIDS launched a 90-90-90 campaign, with the ultimate goal of ending transmission of AIDS by 2030. The 90-90-90 goals mean that 90 percent of all people living with HIV know their diagnosis; 90 percent of those people are receiving antiretroviral treatment; and 90 percent of those on HIV treatment will have an undetectable viral load.
“When the study started, test and treat wasn’t global policy,” says Petersen. “But increasingly, international consensus is building in favor of this approach, which is great. SEARCH will be able to inform how to do it, how much it costs, and how much of an impact it can have. That’s the sort of information really needed to get an international response to stop AIDS.”
Africa’s double burden
Between HIV/AIDS, malaria, tuberculosis, cholera, rotavirus, and others—infectious diseases are still the number-one killer in Africa, especially of children. But the burden of chronic disease in Africa is on the rise. In fact, the World Health Organization (WHO) projects that, between 2010 and 2020, the continent will experience the largest increase in death rates from cardiovascular disease, cancer, respiratory disease, and diabetes. The age-specific mortality rates from chronic diseases as a whole are actually higher in sub-Saharan Africa than in virtually all other regions of the world, in both men and women.
What to do when faced with this “double burden” of disease? Big, complex problems require bold, scalable solutions.
The SEARCH Consortium delivers both. Its vision does not end with HIV/AIDS. The study intervention is designed to improve the entire continuum of care, to address not just HIV, but other infectious and chronic diseases as well, including malaria, tuberculosis, hypertension, and diabetes.
“We’re saying, okay, if we’re going to tackle HIV, we really need to be doing this within a broader framework of public health and economic development,” says Petersen. “We are integrating HIV testing and care with other chronic disease service, diagnosis, and ultimately other chronic disease care as well. There is a great need for that in these communities.”
Big science with a human face
Beginning in 2011, the SEARCH consortium ran a pilot program involving a community in rural Uganda, and showed a promising drop in population-wide levels of HIV detectable in the blood of infected individuals, an indirect measure of the potential for new HIV infections in a community. The scaled-up study—involving more than 320,000 individuals in Kenya and Uganda—got underway in 2013. The goal is to regularly test each community in its entirety for HIV, and quickly put all HIV-positive persons on treatment to suppress viral replication, both improving their own health and dramatically reducing the chance that they transmit HIV to others. A key component of this strategy is the community health campaigns, which include HIV screening, but also other health screenings. Children get vitamins and deworming medications. In some communities, there’s even a raffle for the adults and small prizes for each of the children.
“We asked the communities, ‘What would you like to see?'” Havlir says about designing the health campaigns. “A community said to us, we’d like to know about hypertension and we’d like to know about diabetes. So we included services to check blood pressure and blood glucose, because that made good public health sense and because it was important to the community.”
Including a broad array of health screenings has multiple benefits to the project.
“It’s really designed as a health fair, not an HIV fair,” says Petersen. “The idea is both to destigmatize HIV in these communities and to recognize that it is only one of many health problems faced in these communities, and that these problems are best addressed synergistically.”
SEARCH convenes a health campaign in all 32 communities at baseline and after three and five years of follow-up. In the intervention communities, campaigns will be held annually, together with more frequent testing of key high-risk populations. Community members who test positive for HIV will get counseling and an appointment linked to care and to begin receiving ART immediately using a streamlined and patient-centered care model. Control communities will receive the current standard of care, which is continually evolving.
While the idea behind the intervention—comprehensive testing and streamlined treatment as early as possible—is relatively straightforward, the organization, implementation, and analysis of a study on such a broad scale is almost overwhelmingly complex.
“When people talk about big science, this is big science,” says Petersen. “There’s no way that this project could be done without such a broad consortium. I’ve certainly never had an experience like this ever; it’s a phenomenal experience.”
Design and analysis is what UC Berkeley has been bringing to the UCSF–MU-led project, and from the very beginning. In addition to Petersen, the Berkeley biostatistics team includes Professor Mark van der Laan and PhD student Laura Balzer.
“We needed people who could tackle a really challenging statistical and methodological project, and Maya, Mark, and Laura are those people,” says Havlir. “We wanted to reach out to people who are developing new, cuttingedge methods.”
The UC Berkeley biostatistics team has been instrumental in helping to design the study in order to ask the best questions, develop the best tools, and achieve the best analysis and outcomes.
“A more traditional model of working with statisticians might be that you get a bit of consulting here about power and a bit of consulting there about analysis,” says Petersen. “I strongly believe that to do really transformative research, you need committed high-level thinking about these design and statistical issues early on. It has to be in full integration with the rest of the study team, it can’t be done in isolation; you don’t have the same results.”
‘Pop-up’ statistical problems present a challenge
In addition to tracking new HIV infections, SEARCH is also looking at the health, economic, and education outcomes, including mortality, mother-to-child HIV transmission, AIDS, tuberculosis, HIV drug resistance, adult and child employment levels, asset holdings, school attendance levels, programmatic costs, and cost effectiveness of the intervention.
SEARCH is a cluster randomized community trial, meaning the community, not the individual, is the unit of randomization. At baseline, the 32 communities were matched into 16 pairs based on key health, geographic, and ethnographic variables. One member of each pair has been randomly selected to receive the intervention, while the other the control. Balzer, who is completing her thesis on cluster randomized trials, considers this randomization at the community level, as opposed to the individual level, to be a stimulating challenge.
“It could be that HIV incidence and transmission is totally different between an agricultural community in Eastern Uganda and a fishing village in Kenya.”
“It makes the analysis a lot more fun and complicated,” she says. “As we know, people are messy and not independent from each other. We all interact and impact each other’s health in different ways. So we are doing our best to take the potential dependencies into account with the analysis.”
There are a lot of other factors the biostatisticians must take into account, including population mixing and migration over time, and how the changing WHO guidelines on treatment might affect their control groups. “We would never say, ‘Well, the World Health Organization says you should treat earlier, but you’re in our trial, so you shouldn’t,’ ” says Petersen.
“HIV policy is rapidly evolving, which is fantastic for public health,” agrees Balzer. “From a statistics standpoint, treatment guidelines are continually changing, and we have to anticipate these changes in the study design and analysis plan.”
They’ve also considered how to account for the heterogeneity across study sites. “It could be that HIV incidence and transmission is totally different between an agricultural community in Eastern Uganda and a fishing village in Kenya,” says Balzer. “How do we best capture the intervention’s impact and its implementation in all the communities? It’s been an exciting challenge.”
Balzer has been particularly interested in the pair matching of the communities, and what it can bring to the study in terms of strength of analysis and acceptability to the communities themselves and the broader health community. In SEARCH, the communities have been matched on baseline predictors of HIV incidence (the primary outcome).
“In most settings, pair-matching can really help you gain more power to detect an intervention effect,” she says. “It can also help protect the validity of your study and increase the understandability of your study.”
In September 2014, Balzer traveled to Uganda for two weeks to see the SEARCH study in action. She is including trial design and analysis strategies in her thesis, and also in a paper published in Statistics in Medicine.
“It was great to meet and work with our in-country partners,” she says, “and also a learning experience to see the variables in our data sets in person—what this all really means and whom it really impacts.”
Collaboration keeps it all together
“People will ask me, ‘Do you really think you can end the AIDS epidemic?'” says Havlir. “And my answer is ‘Yes, we can start shutting down the epidemic if we take a community-based approach. We are trying to define the key elements of that approach.'”
SEARCH is designed to return results that are directly relevant to policy makers, in terms of limited resources and return on investment. It’s also designed to be responsive to the input and needs of each community. It has the potential to be a landmark effort that brings a wealth of information on strategies around delivering health care, not just for HIV, but for many other diseases.
“This study asks, ‘Is it time to move to an aggressive population-based treatment/prevention strategy in this part of the world?’ ” says Petersen. “And if it is, how do we do it, and how do we do it most efficiently and effectively? The goal is that we come out of this five-year project and say, ‘Wow, we can shut down HIV, and this is how we do it. Let’s do it!’ ”
Can housing reduce HIV-related risk behaviors among San Francisco homeless youth?
In 2013, San Francisco’s first city-funded Permanent Supportive Housing (PSH) building for youth opened at 5th and Harrison Streets. Run by Community Housing Partnership (CHP), a nonprofit organization that provides supportive housing to formerly homeless individuals and families, the 5th Street Apartments offer 44 youths a place to live. They range in age from 18 to 25 years and used to be homeless or at high risk of being homeless. Many are young adults who have aged out of the foster care system.
In a city where youth homelessness is a persistent challenge, the 5th Street Apartments, also known as 5H, offer a chance to conduct a unique natural experiment. When the San Francisco Mayor’s Office of Housing Opportunity, Partnerships & Engagement (HOPE) approached Colette (Coco) Auerswald MD, MS ’89, about leading a collaborative longitudinal evaluation of youth in PSH in San Francisco, she knew it would be a great opportunity to inform policy and solutions to youth homelessness locally and statewide.
“PSH represents a structural change that provides housing as a given first, rather than a traditional, graduated approach,” says Auerswald, an associate professor in the UC Berkeley-UCSF Joint Medical Program, who has been conducting collaborative research among homeless youth for more than 15 years. “While it has been shown to be remarkably successful among chronically homeless adults, the effects and appropriateness for young adults remain unknown.”
The 5H Project—a collaboration of San Francisco city government agencies, CHP and other housing and homelessness providers, and UC Berkeley School of Public Health researchers—proposes to collect data to fill in the these gaps in the existing research. The project is evaluating whether housing provision reduces HIV and substance use risk behaviors and, more broadly, whether PSH for youth has the potential to have enduring impacts on shaping a positive health trajectory through life.
Currently operating on pilot funds from the UC Berkeley- UCSF Robert Wood Johnson Health and Society Scholars program, the 5H Project has already enrolled 37 of the 44 5H residents.
“We’re testing participants for HIV, Hepatitis C, and other STIs, as well as assessing markers of stress and cellular aging,” says Jess Lin MPH ’10, who is coordinating the project. “We’re also collecting survey and in-depth ethnographic data about topics including their HIV risk behavior, overall health, social support, and economic and educational challenges and opportunities.”
In addition to the data collection, the project has a collaborative relationship with the staff and youth residents, and hosts monthly community-building events open to all residents.
“The 5H project is an effort of great collective will, longstanding partnerships, and complementary skills,” says Auerswald, “but it’s currently strung together on a shoestring.”
Auerswald and Lin are in the process of seeking more funding, which would allow the pilot project to expand into a five-year study encompassing residents of more recently opened PSH buildings and a set of comparison youth not residing in PSH. This rigorous evaluation could inform and fundamentally shift policy and programs for homeless youth.