18 November 2011. Here is a typical scenario. A big trial ends, the sponsor announces success or failure, the researchers describe the findings at conferences, and if the press takes an interest in the story, the results may end up on the evening news. But throughout the process, we seldom hear from the patients who participated in the study and their caregivers, or the hospital staff who administered the tests. What did they think about the trial? Two presentations at the 4th International Conference on Clinical Trials on Alzheimer's Disease (CTAD), held 3-5 November in San Diego, California, asked precisely that question. The answer, put simply, is that Alzheimer’s disease (AD) trials are extremely burdensome, sometimes unnecessarily so, and hard to comply with to the bitter end for most elderly people. Another CTAD presentation suggested one solution: home-based assessment methods that relieve some of the burden.
“In the earlier days of AD studies it was very straightforward: three standard memory scales, one to two activities of daily living scales, blood samples and electrocardiograms for safety, and one CT scan at study entry. Subjects visited every three months for half a day or less and everyone was happy,” wrote Roseanne Hogarth, a research coordinator at Hornsby Ku-ring-gai Hospital in Australia, in an e-mail to ARF. At CTAD, Hogarth presented a poster bluntly titled “Protocol from Hell.” It described the frustration of volunteers, caregivers, and research coordinators with the growing complexity of AD trials. Today, AD patients are asked to fill out several lengthy questionnaires testing their cognition; those paper-and-pencil sessions are increasingly accompanied by magnetic resonance imaging and/or lumbar spine punctures to collect cerebrospinal fluid. In addition, researchers usually collect blood samples, and may also take electroencephalography or other measurements. “The designers of these protocols are making full use of the advances in technology, but forget there is a person with dementia in the study who may have very limited abilities,” wrote Hogarth. Study visits for such protocols can take up to eight hours every four weeks, Hogarth added.
In a survey of some 17 principal investigators and study coordinators and 25 trial participants and caregivers at different sites in Australia, Hogarth and Susan Kurrle collected opinions about current AD study procedures. The main complaints from people conducting trials were that informed consent forms are far too long and contain complex language and concepts that are too difficult for people with dementia to understand. Trial staff bemoaned having to undergo continual retraining on tests with which they are familiar, mainly because each sponsor made slight changes to the Standard Assessment Scale. They warned that participants have to do so many procedures per visit that they leave exhausted. In addition, trial staff said, some of the psychometric tests used in protocols are unsuitable for people with significant cognitive impairment. When caregivers and patients were given a chance to weigh in, they mentioned that trials involve too many study visits, cutting into holidays or other activities. They found visits too long, requiring reorganization of weekly routines, and complained that trials involve repetitive questionnaires with redundant information. Study participants also complained that instructions about study drugs are often too difficult to understand.
Besides reducing efficiency and bumping up cost, demanding protocols make it harder to recruit and retain participants. “Some studies are so burdensome, people refuse to join,” wrote Hogarth.
To look more closely at what determines whether someone says yes or no to a trial, Sandrine Andrieu, Bruno Vellas, and colleagues at the University of Toulouse in France conducted...well, a trial. The ACCEPT study followed up, with questionnaires and in-person interviews, both people who accepted and other people who refused participation in an intervention study for the prevention of AD, the Multidomain Alzheimer Preventive Trial (MAPT; see Gillette-Guyonnet et al., 2009). The impetus for the ACCEPT trial, explained Andrieu in her presentation at CTAD, was the notion that people who participate in clinical trials for AD are not necessarily representative of the general population. That discrepancy might help explain why some protective factors identified in epidemiological studies fail translation into interventions in clinical trials. For example, Andrieu mentioned that in 1999, about 12.2 percent of people 65 years and older in France had a high school diploma, yet among participants in the Three-City Study recruited around that time, that number is 22 percent. In 2010, the level of high school-educated French elderly people rose to 17.6 percent, but was 45.4 percent among MAPT participants. Such skewing might affect results, as higher education is a protective factor for AD (see ARF related news story).
In Andrieu’s study presented at CTAD, out of 1,902 cognitively normal people 70 years and older who agreed to participate in the MAPT, 71 percent also completed the ACCEPT questionnaire; 54 percent of the 674 people who declined to join MAPT did the ACCEPT questionnaire. Twenty people from each group were followed up in semi-structured interviews. From those and the questionnaires, Andrieu and colleagues identified several factors that distinguished MAPT participants from the "thanks, but no thanks" group. The older people were, the more likely they were to refuse participation. People who had never married were less likely to participate. People with higher education and high incomes were more likely to participate. People who thought they might be at higher risk of AD, for example, those with a family history of the disease, were more likely to participate.
It made no difference to participation whether people heard about a trial from a relative or friend or the media, or whether someone had a big or small social network. However, one key factor in their motivation was the nature of the relationships patients had with their doctors. The more confidence people had in their doctors, the more likely they were to participate in the prevention trial, Andrieu reported.
The main reason people gave as wanting to participate in the trial was that they were interested in research and wanted to help those with the disease. Some people who participated thought the trial would help them find ways of training their memory and thus benefit them. On the other hand, people who refused said the trial was too long or did not fit into their schedules. Some people had problems with transportation or did not want to go to a hospital. Others refused because they could not choose whether they would end up in the placebo or treatment arm, Andrieu reported.
Perhaps it should not be surprising that people are loathe to go to the hospital, especially for prevention trials whose participants are not sick. Then one way to make clinical trials more palatable to a wider group might be to carry out assessments and tests in their homes. That notion is being tested in a large trial sponsored by the Alzheimer's Disease Cooperative Study (ADCS). The Home-Based Assessment (HBA) Study, which completed enrollment in 2009, recruited 640 cognitive normal participants with a mean age of 80 and older who lived at home. They are required to take a multivitamin twice a day for four years while completing cognitive assessments at monthly, quarterly, or yearly intervals. These work by way of one of three home-based methods. One group of participants mails in questionnaires and answers questions in telephone interviews with a real person; they keep track of their vitamin intake in a written log that they also return by mail. The second group conducts assessments via an automated phone response system, and vitamin compliance is monitored with telephone keypad responses. The third group is the most high tech. Here, participants use desktop computers equipped with touch screens and speech recognition software to record answers. They also use the Med-Tracker, an electronic pillbox that automatically records when they remove a vitamin.
Mary Sano of Mount Sinai School of Medicine in New York City, who co-directs HBA together with Jeffrey Kaye of Oregon Health and Science University in Portland and Steven Ferris at New York University in New York City, told ARF the trial was able to recruit 20 percent minority patients. “This study was designed to try to recruit a sample that is more representative of the population at risk for AD,” said Sano. “We hope the results will develop interest in more studies being home-based.”
“Right now, to participate in a study, you have to be able to travel to a clinic. That’s only possible for people who live near a large academic hospital or university, and those who can drive or have a caregiver who can drive,” said Kaye.
That is not to say that home-based assessments are without complications. Some elderly people live in apartments too small to accommodate a computer; others live in areas that have no broadband or Internet access. In some cases, Kaye said study coordinators did not know how to set up the system.
Sano and Kaye are trying to reduce the burden on patients on all fronts. “We assess several domains of cognitive function, and we try to cut down the number of items in each assessment to eight per domain,” Sano said. “This is huge, because right now tests are too cumbersome.” The scientists are also trying to determine just how often to assess participants. “Doing assessments more frequently gives you more intimate data, but may lead to more dropouts. Patients don’t like doing the assessments too often. But if you do them too infrequently, the danger is that participants can lose interest in the study,” Sano explained.
Beyond getting people to join and stay with the study, home-based monitoring may also provide richer and more accurate data, Kaye argued in his presentation at CTAD. When answers are recorded by a phone or computer system, they are automatically analyzed. “There is less opportunity for data artifacts that arise from different people scoring the answers,” he said. Audio files can be further analyzed for pauses between answers or other speech features that may provide insight into cognitive function. Kaye’s group has also developed door sensors that can monitor how often people go out or open the fridge, and infrared sensors installed on ceilings that can measure a person’s walking speed during everyday activities. Some studies have shown that changes in how quickly a person walks can serve as early markers for mild cognitive impairment (MCI) (Hayes et al., 2008). “Traditional measures typically capture only a snapshot of cognition in an artificial setting,” said Kaye. “Home-based systems allow for more continuous real-time assessments.”—Laura Bonetta.