WASHINGTON (AP) -- A fog is slowly creeping over Dr. William Deutsch's brain. He would try a risky experiment, even one requiring brain surgery, in hopes of at least stalling his incurable dementia. But scientists have little to offer.Families battling Alzheimer's disease and similar dementias increasingly are calling for a shot at riskier therapies that might bring bigger benefits than today's pretty safe but largely disappointing drugs.It's a conundrum: Dementia robs its victims of the ability to fully consent to medical experiments. When loved ones can do so for them varies in part according to how much risk is involved.Critics say that's one reason scientists and regulators have generally treated Alzheimer's more like a chronic disease than the killer it is, tolerating fewer side effects than for, say, cancer therapies. Further chilling was the 2002 halt of a study of a vaccine designed to attack Alzheimer's brain-clogging gunk that caused serious brain inflammation in a few patients.The tide seems to be turning. Scientists are tentatively exploring a handful of bold, sometimes invasive, approaches -- from another vaccine attempt to a gene therapy that requires brain injections -- prompting an ethical debate about how to study a vulnerable population.Patients and caregivers want a voice in those decisions, and often accept more risk than doctors predict, was the overarching message when the Alzheimer's Association arranged an unusual meeting between families, drug developers and the Food and Drug Administration."No one has ever survived this disease," Frank Broyles, the University of Arkansas' athletics director whose wife recently died of Alzheimer's, reminded the meeting."We would take any risk," even a drug with a 50 percent chance of death if there were an equal shot at benefit, he said of his own family.Added Deutsch, a former New York podiatrist in the early stages of a similar dementia: "You're going to have to have some more invasive procedure done to get dramatic results. ... You should at least be given the choice."The first formal study of family risk tolerance, published in this month's journal Neurology, supports those calls. University of Michigan researchers wondered: Is it OK for grandma to undergo a spinal tap when she doesn't understand why? What if the research won't help her but might lead to future Alzheimer's treatments?Using 10 research scenarios, they surveyed 229 elderly people at increased risk of developing Alzheimer's because a close relative had it. Most accepted even the greatest potential for side effects, such as gene therapy or a painful spinal tap, for themselves, and were only slightly less risk-tolerant when deciding for a loved one.Yet until now, "the patient community's been conspicuous by its absence," Dr. Russell Katz, FDA's neurologic drugs chief, told the recent Alzheimer's Association meeting, to which a reporter was invited.In contrast, a handful of cancer survivors helps the FDA privately deliberate what cancer drug research to allow, and the agency is starting a similar pilot project in Parkinson's disease -- two patient groups that requested a role.It's easy to say you'll accept more risk in hope of more gain. How to quantify that is the vexing problem of regulators like Katz. Does risk mean pain? Premature death? And what benefit justifies those risks, improved functioning or merely hope?An experiment is just that: Nobody knows the full risks until participants have been treated, something difficult for desperate families to grasp, cautioned Dr. Steven DeKosky, the University of Pittsburgh's neurology chief.Still, Katz said FDA ultimately allows most Alzheimer's proposals.Among the most invasive: Scientists at Rush University in Chicago, Illinois, recently injected six patients' brains with a virus carrying a nerve growth factor that might reverse deterioration. A slightly different gene therapy approach, in California, produced promising preliminary results although two patients experienced dangerous bleeding in their brains. The Chicago researchers are tracking patient reactions, and a follow-up study could begin in a year.Broyles and others complain that such research moves too slowly. Consider Elan Corp.'s 2002 vaccine, ultimately found to slow memory decline somewhat and clear out some plaque, even though the experiment was halted before participants received a full dose. It took FDA more than three years to allow Elan to study a potentially safer version.Caregivers differ on how much risk they'll tolerate; one called it gut-wrenching to watch a loved one suffer even a minor side effect.But Deutsch, still able to measure how his abilities are melting away, boils the question down to family choice: "I don't think it's the scientists' job to pick who's going to get the treatment."Copyright 2005 The Associated Press. All rights reserved.This material may not be published, broadcast, rewritten, or redistributed.