Here’s an uncomfortable thought experiment. Imagine a medication that’s FDA-approved, prescribed to millions of people with dementia, and has been the standard of care for decades. Now imagine new research suggesting that, on average, it doesn’t reduce dementia symptoms at all and may be shortening lives instead.

That’s exactly what economist Robert J. Brent found when he ran a rigorous cost-benefit analysis on the five FDA-approved Alzheimer’s medications: donepezil, galantamine, memantine, rivastigmine, and tacrine. Known collectively as ADM (Alzheimer’s Disease Medications), these drugs work by boosting a brain chemical called acetylcholine, which is supposed to help with memory and thinking.

The theory is sound. The results, using a massive national dataset of real dementia patients, were not what anyone expected.

Running the real numbers

Brent’s approach isn’t just “did symptoms improve, yes or no.” It’s a full cost-benefit analysis, the same rigorous framework economists use to evaluate whether a public investment is worth funding. It comes down to three steps: does the intervention change dementia symptoms, does that change something people actually care about (in this case, how long they live), and what is that change worth in dollar terms?

For the ADM, step one already went sideways. Instead of reducing dementia symptoms, the medications were associated with a rise in them. And when researchers traced that through to mortality, both the direct effect of the drugs and the indirect effect (via worsened dementia) pointed the same way: more deaths, not fewer.

Converting that into dollars using the Value of a Statistical Life, a standard $3.27 million figure used across public health and regulatory economics, the picture gets stark. The direct mortality effect alone represented a loss of roughly $61,000 per person. Add the indirect effect through worsened dementia, and total losses reached about $76,000 per person. Add in the cost of the medications themselves, roughly $32,000 per person over a lifetime, and the full picture lands at a loss of about $108,000 per person, on average.

In plain terms: for the average patient in this data, taking these medications was worse than doing nothing.

Is this a fluke?

It’s a fair question, and Brent asked it too. The result was so unexpected that a second, entirely different statistical method, one designed to mimic the gold-standard randomized clinical trial that medicine relies on, was run as a check. If anything, it made the finding worse: the adverse mortality effect nearly doubled.

So why would drugs designed to help end up hurting? The book points to a few likely culprits. These medications are known to carry real side effects: nausea, cardiac issues, and in rare cases a dangerous “cholinergic crisis” from overstimulation of the nervous system. One of the five, tacrine, was pulled from the US market back in 2012 over liver damage. The side effects aren’t secret; what’s new here is evidence that they may be far more common, and far more consequential, than the rare-event framing usually attached to them.

What this doesn’t mean

This isn’t a claim that every single person on these medications is being harmed, or that anyone should stop a prescribed treatment without talking to their doctor. It’s a population-level economic finding: on average, across a large, real-world sample of older adults with dementia, the costs outweighed the benefits, and the benefits were negative before costs even entered the picture.

What it does mean is something bigger for how we think about dementia care. If the flagship, FDA-approved pharmaceutical option isn’t paying off, then the conversation needs to shift toward what actually does work, much of it far less flashy than a pill. Things like corrective vision, hearing aids, Medicare access, and avoiding unnecessary nursing home placement all showed measurable, positive effects on dementia symptoms in Brent’s research, at a fraction of the cost.

The takeaway isn’t that nothing can be done for dementia. It’s that the thing everyone assumed was “the treatment” may not be the answer, and the real answers might be hiding in plain sight, in decisions far more mundane than a prescription pad.