Peter Townshend of The Who concluded his baby-boomer anthem, "My Generation," with these words: "I hope I die before I get old."

And my boomer generation may well still wish for that.

I am 62 -- old enough to cash in my 401(k), too young for Medicare -- and standing with my peers on the edge of a dementia precipice.

Alzheimer's disease and other forms of dementia afflict up to 5 million people in the United States and about 26 million people worldwide. By 2050, there could be 13 million cases of Alzheimer's alone among U.S. baby boomers and the aging Generations X and Y, according to the National Institutes of Health.

Some reports have the global prevalence of Alzheimer's growing to 100 million people by midcentury. The U.S. comptroller general estimates that annual long-term care costs for elderly people -- which includes treatment for dementia -- could quadruple by 2050 to $379 billion.

How should President Barack Obama and his health-care policymakers, who are working to overhaul our system, prepare for my generation's future? Based on my experience, they can begin by finding a way to end the over-dependence on drugs in treating dementia.

As a psychologist who works in nursing homes, I am intimately aware of the large number of residents who take one or both of two Food and Drug Administration-approved drugs for dementia -- known generically as donepezil and memantine, which together account for more than 90 percent of the antidementia drug market. The most popular brand-name versions, Aricept and Namenda, make up 75 percent of the market.

I'm also aware of the huge and growing expenditures for these medications -- close to $3 billion annually worldwide for Aricept and more than $500 million for Namenda. Big Pharma spends as many billions of dollars on promotion as it does on research and development.

Examine the documents supporting the FDA's approval of Aricept, and you will see upon what a slim reed this drug's empire was built. Those taking the drug scored, on average, three points better on a 70-item cognitive assessment scale. That's about a 4 percent difference, mostly reflecting a slower decline rather than positive improvement. And the differences disappear when the drug is discontinued -- indicating that the drugs "do not represent a change in the underlying disease."

At best, these effects may be only marginally more effective against dementia than garlic was against the Black Death in the 14th century.

What we do know today, from studies and observation, is that donepezil, memantine and drugs like them fall short on cure and comfort. Even on Aricept's Web site, the claims are sketchy on the drug's effectiveness when it comes to cognition: "People who took Aricept did better on thinking tests than those who took a sugar pill."

How much better? The company doesn't say.

Many studies of the effects of drugs for dementia also speak about statistical significance, but statistical significance can be highly overrated if the differences aren't meaningful. Take my extremely nearsighted wife, for example. Suppose a drug enabled her to read the giant E at the top of an eye chart without her glasses, but none of the smaller letters. Her eyesight would show statistically significant enhancement, but -- despite her being a much better driver than me -- I'd still refuse to ride in a car she was driving if she wasn't wearing her glasses.

There are similar effects at play with antidementia drugs.

In 2004, Richard Gray of the University of Birmingham in Britain compared hundreds of patients with mild to moderate dementia who were taking Aricept or a placebo. The drug did improve mental functioning, but at disappointingly small levels -- about one point on a 60-point scale. More important, there was no delay in the dementia's progression or the rate of patients' institutionalization. And there were no significant differences in mood, behavior or cost of care.

Based on results such as these, the British National Institute for Clinical Excellence -- the functional equivalent of the U.S.'s FDA -- recommended in 2005 that Britain's National Health Service greatly restrict the use of drugs for dementia. Donepezil can be prescribed only by a psychiatrist or a neurologist, and its use is restricted to cases of mild to moderate -- not severe -- dementia. And memantine is restricted to clinical trials.

Could the thousands of dollars spent annually per patient and the billions overall be better directed?

Yes, says Gray: "Doctors and health-care funders need to question whether it would be better to invest in more doctors and nurses and better social support rather than spending huge sums of money prescribing these expensive drugs."

A survey released in 2002 by the Kaiser Foundation found that the staffs in a typical nursing home spend a total of about two hours and 20 minutes a day with each resident. For the remaining 21 hours and 40 minutes, residents are left to their own -- mostly medicated -- devices.

Where is the comfort in that?

Some proponents of drug therapy argue that despite some disappointing results, the drugs do slow the worsening of symptoms for some people. But in our medicalized institutions for frail and elderly people, drugs are the first recourse for most problems. And often the second and third recourse.

In the United States, those over 65 consume 30 percent of the prescription drugs, according to a 2004 report. Dementia sufferers in nursing homes are not only taking donepezil and memantine but other similarly questionable drugs for depression, anxiety, psychosis or for simply being ornery. Many of those without dementia are also on a variety of mind- and mood-altering drugs.

There is some comfort in believing, as our medieval ancestors did, that a tangible nostrum -- like a pearl-hued donepezil tablet -- will do some good, but it may be more comforting simply to comfort.

Instead of drugs, I'd bet many patients are wishing someone would just say the words of another ancient rock anthem: "I want to hold your hand."

Rosofsky is a psychologist and the author.

(March 22, 2009)