The money is the least of it. £389.40 over nineteen months. Fine. I have wasted more than that on a conservatory. What it cost was that I stopped looking. Once you believe your own body is the exception you close the whole category, because you have already run the experiment and you have the spreadsheet to prove it.

I went back to my GP in April. 154 over 96, taken twice in her room. She was kind and completely reasonable and she started me on amlodipine 5mg. I still take it every morning, and I am not telling one single person to change anything about their prescription — the internet is full of people who will, and they are dangerous.

But she also said a sentence at the end of that appointment which I have thought about nearly every day since. It was not unkind. It was arithmetic. We will start with the one and see whether it holds it; if it does not hold it, we add a second.

That is the thing I am actually frightened of. Not 154. Not the drawer. The staircase. One tablet at sixty-one, two at sixty-four, three at sixty-eight, and a review every six months for the rest of my life where somebody kind tells me the number has crept again.

And here is what nineteen months of receipts and 574 readings had failed to teach me. Every single thing on that staircase — the reading in her room, the reading on my kitchen table, the decision about a second tablet — comes off one piece of my body: the wall inside the vessel. I had a column for the output of that wall twice a day for nineteen months. Nobody had ever given me a column for the wall itself. And I did not find that out from a doctor or from a label. I found it out next to a plate of sausage rolls at my mother's eightieth.

My nephew Sam is a technician at a contract analysis laboratory in the Midlands — the sort of place that does third-party testing for companies who want a certificate they can print. At some point in the evening I told him the drawer story. As a joke. It is my party piece: the woman with the spreadsheet and no sense.

He did not laugh. He put his plate down and said, "Auntie Dawn, what did the garlic one say the ageing time was?"

I said I had no idea what he meant. So he asked for a pen, and because I am who I am there was a receipt in my handbag, and he drew on the back of it: a tube, then a line inside the tube hard up against the inside of it. Everything you have been measuring for nineteen months, he said, is happening in that line.

Every artery I own is lined with a layer one cell thick. It is called the endothelium, and it is not lining in the plumbing sense — it is an organ, and altogether it weighs about as much as my liver. Its job is to make nitric oxide, the signal that tells the muscle wrapped around the vessel to let go.

Production of that signal starts falling from about forty and keeps falling every decade after. Quietly. Nobody measures it. There is no line for it on any printout, and there was no line for it in nineteen months of a spreadsheet built by a woman who measures things for a living.

A vessel that stops getting the let-go signal stays tight, and a wall that stays tight stiffens. A heart pushing through tubes that will not yield has to push harder, and the gauge on my arm reads higher. That is my systolic — the number I have been logging twice a day with a date stamp. It is a reading taken off the wall.

The same stiff, irritated wall is the surface cholesterol gets into, which is the other reading they take off me once a year; and the pressure inside those same vessels is what the kidney filters against, which is the third.

Three readings. One wall.

I am not about to blame anybody for that. There are perfectly good reasons a clinician talks about the reading rather than the wall: the reading is what can be measured in ten minutes, and the tablet is aimed at the reading. I would have run it the same way myself.

But it leaves a hole. The tablet is aimed at the gauge, the staircase is a series of decisions about the gauge, and nothing anybody had handed me was aimed at the wall the gauge is measuring. That is the hole I had spent £389.40 trying to fill, badly, with six bottles that could not reach it.

Twice a day for nineteen months without missing one. Date-stamped, so I cannot even flatter myself about it. If adherence were the problem my spreadsheet would say so, and it does not. Whatever failed in that drawer, it was not the woman holding the bottle.

Six bottles, and not one of them could put its working compound into my bloodstream, which is the only route to the wall. That is not a small effect. It is an absent one, and I will show you the line on the label that proves it.

A trial with no active ingredient is not a test of the patient. Nineteen months of my data says nothing about my body; it says six things about six bottles.

You cannot feel it working, the label will not tell you what is inside, so the only feedback you have is your own monitor against a compound that was never in the capsule. Everybody quits at about twelve weeks. That is not an accident.

This is the one that cost me. I concluded that supplements do not work on me. What the data actually supports is narrower and much more useful: six specific products, none of which could reach the wall, did nothing to a reading made by the wall. Those are not the same conclusion, and only one of them leaves you with somewhere to go.

Raw garlic's famous compound is allicin, and it is what nearly every cheap tablet on the high street is built around — that is what "high allicin yield" means on the front of a tub. Allicin is not a delivery vehicle. It is a firework: the bulb's own defence chemical, created the moment the clove is crushed and falling apart from that moment on. Put it in stomach acid and it is essentially gone inside sixty seconds, so it never arrives in the bloodstream in any meaningful amount, which means it never arrives at the wall.

Sam put it in a way I understood immediately, because it is my language: you have been buying a material with a shelf life measured in seconds and expecting it on site three days later.

So how does the research work, I asked him, because the trials are real — I had read them, and that was the whole reason I bought the first bottle. That is the other half of it. There is a process, Japanese in origin, where whole garlic is aged. Not cooked, not dried and powdered: held in controlled darkness, no heat, no solvents, for twenty-four months, over which the unstable allicin converts into a completely different compound, S-allyl cysteine.

S-allyl cysteine is the opposite of allicin in every way that matters. Stable, water-soluble, it survives stomach acid and turns up measurable in blood — the one claim raw garlic powder can never make. And it is the compound the cardiovascular research is built on, studied specifically against the wall.

Then Sam said the sentence that made me go home and open the drawer. "Aged garlic takes two years to make. Nobody selling you a tub for £8.99 has waited two years for anything."

It is also why nobody can make more of it when it runs low. Harvested once a year, then two years in the dark; you cannot order extra in July for Christmas. Whatever is on sale today went into the dark two years ago, which is arithmetic rather than a sales tactic, and it is the one part nobody can fake with a marketing budget.

I did what I should have done on day one: stop reading the front and start reading the back. Six bottles, kitchen table, glasses on, pen in hand, exactly the way I used to go through a tender. Not one of them stated an ageing time. Zero out of six.

The garlic one said 15,000mg on the front in silver foil. Read the small print and that is not fifteen thousand milligrams of anything in the capsule — it is an equivalence claim, powder "stated as equivalent to" that weight of fresh bulb. What is in there is a few hundred milligrams of dried, heat-processed powder, plus rice flour, magnesium stearate and a gelatin shell. The big silver number describes a vegetable that no longer exists.

Then there was "odourless", which is the one that genuinely angered me, because I had been treating it as a courtesy — a coating, so you do not clear a lift. It usually is not. The smell of garlic is the sulphur chemistry; they are the same molecules. So on a cheap tub, odourless very often means they drove the smell off with heat, which is to say they removed the smell by removing the thing you came for. It is a confession printed on the front as though it were a feature, and I paid £8.99 for it and thanked them.

A properly aged extract is odourless for the opposite reason: there is nothing left to strip out, because two years in the dark already converted the volatile compounds into the stable one. No breath, no next-day sweat, no taste coming back at four in the afternoon — not cooked out, converted. Same word on two labels, opposite meanings, and with the bottle in your hand the fastest way to tell them apart is to ask why it is odourless.

And the £34 "cardio blend" that smelled like a hedge? Third ingredient rice flour, filler list longer than the active list, and the smell was the only honest thing about it.

Six bottles, nineteen months, twice a day without missing one — and the compound the research measures was never inside any of them, so nothing I did was ever going to reach the wall. I sat at that table angrier than I have been in years, and the anger went at me first, for being a woman with a spreadsheet who never turned a bottle round. It took me a week to work out that was the wrong address.

Old habits: I wrote it the way I would have written a tender requirement. Three lines, and I am giving them to you whole, because they cost nothing and you can run them in ninety seconds on any product page in Britain, including this one.

One — a stated ageing time, in months, printed on the pack. Not "aged", not "matured", not "traditional method". A number. If a company will not put a number on the one step that costs them two years, it is because the number would embarrass them.

Two — why is it odourless? Because of the two-year conversion, or because they cooked the actives out? Same word on two labels, opposite meanings. That is the cheapest test on the shelf and I wish somebody had told me nineteen months ago.

Three — how long is the ingredients list? If I am reading half a page of cellulose, stearates and rice flour before I reach the garlic, the garlic is a guest in somebody else’s capsule and I can put the bottle back.

I assumed nothing would pass; after nineteen months I half-wanted to be proved right about the whole category, because being right is more comfortable than starting again. BioZentra Odourless Aged Garlic Extract. Twenty-four months of cold ageing in darkness, stated on the pack. S-allyl cysteine named on the label. One softgel a day, short ingredients list, no smell and no repeat — because there is nothing left to cook out. Sixty days, refunded in full, no questions, and you do not send the bottle back. No subscription, which mattered to me, because I have spent enough of my life on the phone cancelling things.

And the part that finally made me trust it: garlic is harvested once a year, then sits in a controlled room for twenty-four months. Nobody can order more in July for Christmas. Whatever is on sale today went into that room two years ago. That is real scarcity — arithmetic, not a countdown timer — and it is the one thing you cannot fake with a bigger production run.

And I want to be plain about what it is not. It is not a replacement for the amlodipine and it is not aimed at the same thing: the tablet is aimed at the gauge, this is aimed at the wall the gauge reads. It sits alongside, one a day, and the person deciding about my medication is my GP with my readings in front of her.

No. It is a one-off purchase, there is nothing to cancel, and nobody is going to take money off your card next month. I checked that before I ordered, because I have read the reviews of the ones that do it.

Because of one line. Get your bottle out of the drawer and look on the back for a stated number of months. If it is not there, nothing was aged, and you have not tried aged garlic extract — you tried heat-dried powder with a big silver number on the front. I ran that experiment for nineteen months with 574 readings, and it is not a test of you.

Run the three lines on us. Ageing time in months on the pack — it is there. Why it is odourless — the two-year conversion, not heat. And the length of the ingredients list — short. If we failed any of the three you would be entitled to shut the drawer on us too.

You are in the best position of anybody reading this, and I say that as somebody with 574 rows. Do not start a new system. Use the log you already have as your control. Take a fortnight of baseline first, same arm, same time of day, sitting still for five minutes, then change exactly one variable and change nothing else for sixty days. That is the whole protocol. Nineteen months of my data was useless because I never isolated a variable, not because I did not measure.

No. Not one thing, not the dose, not the timing, and I would be appalled if anything here read that way. I take amlodipine every morning and I intend to keep taking it. This sits alongside. Take your written readings to your next review and let the person with your notes tell you what she sees in them.

Yes, and not because it was cooked. The two years already converted the volatile compounds, so there is nothing left to strip out with heat. No breath, no next-day sweat, no taste coming back at four in the afternoon. My £34 blend smelled like a hedge and did nothing; the smell was never the point either way.

Here is what a woman who files receipts does with a drawer full of failures: she does not throw the data away. I kept the old spreadsheet. 574 rows, nineteen months, two points. It sits on the shelf above the kettle next to the new one, and not out of bitterness. It is the most valuable thing I got out of £389.40, because it is a nineteen-month baseline of my own body taken at the same time of day in the same chair, and almost nobody has that.

So I did it properly, the way I would have run any trial that crossed my desk. A fortnight of baseline. Then one variable changed and nothing else — no new diet, no new exercise, no second product, nothing that would muddy the column. One softgel a day, sixty days, which is long enough for a fortnight of baseline and six weeks of readings, which is long enough to know something. Thirty days is not, and that is precisely why thirty is the industry default.

I am not going to give you my numbers, and the reason is that anything I quote becomes a promise I am not able to make. I am one sixty-one-year-old woman in Bristol on a tablet. What I will tell you is that the first fortnight was completely undramatic — no lift, no rush, no signal of any kind, which is the exact point at which I quit six times before.

Then I did the part that nineteen months of bottles never once led to. I printed both columns, the old and the new, and took them to my review.

She read them properly, for longer than I expected. What she and I agreed about my medication is between me and her, because the entire point of this article is that it should be a conversation with somebody who has your notes, not a decision a woman makes alone at a kitchen table with six bottles and a bulldog clip.

What I will say is that I walked out of that room without hearing about a second tablet, and that for the first time since April the person deciding about my body and the person living in it were looking at the same sheet of paper.

So if you have got that drawer — and if you have read this far, you have got that drawer — here is what I would do, and the first two parts of it are free.

Go and get one bottle out of it, turn it round and look for a number of months. You will not find one, and that absence is the whole story, and it is not about you.

Then, whatever else you take from a stranger on the internet, take this. Write down two numbers tomorrow morning — same arm, same time, five minutes sitting still first — and do it for a fortnight before you change anything at all. If you already log them you are further ahead than I was for nineteen months: you have the control column, and all you have been missing is a single variable worth testing against it.

And then, if you want to test something, change one thing, aim it at the wall rather than at the gauge, and give it sixty days with your own handwriting as the judge. That is not faith. It is a control, a variable and a deadline, which is the only way I have ever known how to decide anything. I was disciplined for nineteen months about six things that could not work, and I would rather you did not spend yours the way I spent mine.

Dawn Petrie, Bristol