The bottles come out of the kitchen cupboard and cover most of the table. Eleven of them, plus a supermarket tub of ibuprofen, plus a sleep aid a neighbour recommended two winters ago. Her daughter finds two that look like the same drug under different names, one that ran out in March, and one carrying a prescriber's name nobody in the family recognises.
Ask who is responsible for that collection and the honest answer is nobody in particular. A cardiologist chose three and knows those three well. The family doctor started four. A urologist added one, and a hospital added two more on the way out the door. Every decision was defensible on the day somebody made it. The list, considered as a list, was never decided by anyone at all.
How a list grows without a decision
Medicines arrive one at a time, each inside its own appointment, each with a reason attached. Nothing in the ordinary structure of care makes any single person accountable for the total. Adding is somebody's job. Subtracting is nobody's.
The clearest version of this is the prescribing cascade, a pattern Paula Rochon and Jerry Gurwitz described in the BMJ in 1997. A side effect gets read as a new complaint, so a second drug treats it. A blood pressure tablet causes ankle swelling, the swelling earns a diuretic, the diuretic disturbs sleep, and the sleep problem earns something sedating that raises the odds of a fall. Each step follows from the one before. Nobody made a mistake you could point at.
Meanwhile the record fragments. The cardiologist's notes hold the cardiology drugs. A discharge summary holds whatever the ward changed, which the family doctor may not have read, the same gap that turns leaving hospital into an unattended handoff. The only place every item appears together is the cupboard, and the cupboard is not in anyone's chart.
The number that doubled in twenty years
This is now the ordinary condition of later life. An analysis of the National Health and Nutrition Examination Survey, published in Global Health Research and Policy in 2023, tracked how many people in the United States were taking five or more prescription drugs at once. Among adults aged 65 and over, that share climbed from 23.5 percent in 1999 and 2000 to 44.1 percent by 2017 and 2018. For people with heart disease it reached 61.7 percent.
Read those figures carefully, because the easy reading is wrong. Five drugs is not a diagnosis of bad care. Part of the rise is medicine working: conditions that once killed people quickly are now managed for decades, and managing them takes tablets. The number is evidence about arithmetic. Somewhere between four drugs and eleven, a list stops being something anyone can hold in their head during a fifteen-minute appointment.
A criteria list, not a ban list
Geriatrics has a tool for the narrower question of which drugs deserve a second look in older bodies. The American Geriatrics Society maintains the Beers Criteria, a catalogue of medicines that carry more risk after a certain age than the available alternatives do. The 2023 update, announced in May of that year, came from a panel that reviewed more than 1,500 studies published between 2017 and 2022. It sorts what it finds into five groups, from drugs most older adults should avoid through to drugs needing different dosing when kidneys work poorly.
The tool gets misread as a blacklist, and its own authors keep saying it is not one. Todd Semla, who co-chaired the panel, put the limit plainly: the criteria should never on their own dictate how medicines are prescribed, and should never justify restricting anybody's coverage. Plenty of people take a listed drug for a sound reason. What the catalogue is good for is starting that conversation, about one specific bottle.
The review nobody schedules
An annual sit-down over the whole list is not a novel idea. Medicare Part D plans are required to run a medication therapy management programme, and for the members it enrols that includes a yearly one-to-one review with a pharmacist, ending in a written summary the person keeps.
The catch is who qualifies. Plans target members with several chronic conditions, several Part D drugs, and spending likely to pass an annual threshold, which leaves out plenty of people carrying eight bottles. And a benefit that exists is not a benefit that happens: the offer arrives as a letter or an unfamiliar phone number and gets filed as marketing. So in practice the review happens because a family member decides it will.
Building one list the household controls
The work is clerical long before it is clinical. What follows prepares a conversation with professionals rather than replacing one.
- Put everything in one bag, including what nobody prescribed. Actual bottles beat a remembered list. Bring the supplements, the shop-shelf painkillers, the eye drops, the herbal sleep aid. Interactions do not care which counter something came from.
- Write four columns, not one. Drug, dose, what it is for, who started it and roughly when. The last two are the columns nobody has, and the ones that make a review possible.
- Ask the pharmacist first. One pharmacy sees every fill from every prescriber, which is more of the picture than any single clinic holds, and asking needs no appointment.
- Question the newest arrival before the oldest. If a symptom appeared within weeks of a new drug, say those two facts out loud next to each other. That sentence is what interrupts a cascade.
- For each item, ask what would show it is still working. A drug started for a reason that has since passed, or one whose benefit nobody can name a measure for, is the useful thing to raise.
- Ask what stopping would involve. Not whether to stop, but the process: which drugs need tapering, what to watch for, when to check back.
- Keep the dated list where the routine already goes. Inside the cupboard door, beside the pill organiser, photographed on two phones. Records parked outside the loops a household already runs go stale inside a month, and an undated list gets distrusted by the next professional who reads it.
Notice how little of that is medical. It is inventory, provenance and a diary entry. Clinical judgement stays where it belongs, but it cannot be exercised on information nobody has assembled.
A question for the next appointment
If you help somebody with their medicines, spend an hour this month building the four-column list from the actual bottles. Then carry one question into the next appointment and ask it about a named drug rather than about the pile: this one, started in 2023, for that reason, is it still earning its place?
Ask a second question of the system while you are there. Among everyone writing prescriptions for this person, who is looking at the whole list? Sometimes the answer names someone and you can stop worrying. Often the question is the first time anyone has treated the list as a single object with an owner, which is the only way a table covered in eleven bottles ever gets shorter.