Start with a phone call. Your doctor’s office has a triage line and a nurse will talk it through with you. They are genuinely good at this, they do it all day, and it costs nothing. Most practices have a phone line where a nurse takes calls about whether something needs to be seen, and how soon. It is often not advertised — you ring the main number and ask to speak to the triage nurse. There is no charge and you do not need an appointment to use it.
If it feels more serious than that, go to urgent care. You will usually be seen faster than at a hospital, and the care is good.
If it feels life-threatening, call 911 or go to the emergency room. Do not wait until you are sure.
Which one is yours to decide, not ours. But our advice is to assume it is worse than it looks and take the more cautious road. And trust your body — if it is telling you something is wrong, it is probably right. The best possible outcome of a trip to the emergency room is walking out saying well, that was a waste of an evening — but I feel a lot better knowing it is nothing serious.
Before you call, have these ready. They are what the nurse will ask, and the call goes better when you are not working them out on the phone.
You are not diagnosing yourself by having this ready. You are handing them the things they would otherwise spend the call extracting — and the decision stays entirely theirs.
Writing to the portal instead of calling? The same list works, in that order, in one message. Put the direction it is going and the immune-system line near the top — portal messages get read quickly, and those two change how the rest is read.
The appointment is short and it is not your doctor’s fault that it is. Walking in prepared is how you both get more out of the time there is. Print this, or copy it into your phone.
None of this is a script and none of it needs to be done perfectly. Even the first two lines, on the back of an envelope, change the appointment.
The short answer, before anything else.
Almost none of this is a doctor deciding not to answer you. A message you send arrives in a queue that somebody else sorts. A portal message can sit unread for days by design, and nobody tells you that.
→ where this is explained: What usually goes wrong, and where
The appointment is short, and that is not your doctor’s doing either. The same pressure that swallows your message is the one that ends the visit early.
→ where this is explained: What usually goes wrong, and where
So the useful question is not how do I make them care. It is what makes a message answerable — and there are specific things that do, below.
→ where this is explained: What makes a message answerable
The one thing to do next: Send one thing, not five, and put the direction it is going near the top. If nothing comes back, there are five rungs on this page to climb, in order, and the first is not a complaint.
Everything below explains each of those, in whatever order suits you.
A portal message can sit unread for days by design. None of the above can wait for that. Call emergency services or go in. Nothing further down this page applies to any of them.
Almost none of it is a doctor deciding not to answer you.
A message you send arrives in a queue that somebody else sorts. It may be read by a nurse, an assistant or a scheduler before it reaches the person whose name is on it — and a large share of messages are handled without that person ever seeing them. That is not a scandal; it is the only way the volume gets managed at all. But it has a consequence: the message that is easy to route gets routed, and the message that is hard to route waits.
The seven-minute visit is the same problem inside the room. It is not a problem the physicians created. It is one the system imposed on them, and the ones who find a way to do right inside it — who take the time to admit uncertainty, who make the referral that gains them nothing — are doing the hard thing in a system that rewards the easy one.
So the useful question is not how do I make them care. It is how do I make this answerable in the ninety seconds it is actually going to get.
Four things, and the fourth is the one almost everybody leaves out.
Seventy-two words. One subject, one date, one ask, one exit.
This is an escalation ladder, and the order is the whole of it. Each rung is used only when the one before it produced nothing. Starting at the bottom is the commonest mistake and the most expensive: it turns a question into a challenge, and a challenge is far harder to answer than a question.
The plain question, in the shape above. Most things end here, and that is the point — the ladder exists so that you rarely need the rest of it.
Same thread, naming the date. Not a fresh message: the same one, so the whole timeline stays in one place where anybody can see it.
Now name the specific thing, as a suggestion rather than a demand. “Would a urine protein test be reasonable here?” Naming the test or the referral gives the answer somewhere to land; without it, replying means first working out what you are asking for.
The rung that does the most work. Set out your reasoning in three or four sentences and invite it to be knocked down.
It is not a rhetorical trick, and it should not be used as one. It works because it asks for the thing a clinician can do quickly and does well — correct an error — instead of asking them to build a case from nothing. People correct far more readily than they compose. And if the answer is that you are wrong, you have got what you came for.
Last, and once. The written summary with the dates, the results and the citations. It carries weight because the four lighter rungs came first. Sent cold it reads as an accusation, and it will be handled as one.
One line per message. It takes a minute, and it is the only thing that will still exist in a year.
| Date sent | Subject | Reply? | Reply date | Adequate? | What I did next |
|---|---|---|---|---|---|
| 12 Aug | Ankle swelling since 4 Aug | N | — | — | Followed up 19 Aug |
| 19 Aug | Follow-up, same thread | Y | 20 Aug | N | Asked about a specific test |
“Adequate” is the entire reason for the table. A reply is not an answer. “We will discuss this at your next visit” is a reply. Whether it answered the question you asked is a separate fact, and nothing else anywhere in medicine writes that fact down.
This is not for building a case against anybody. It is so that six months from now, when somebody asks when you first mentioned it, you know — instead of guessing.
Under the information-blocking rules of the 21st Century Cures Act you have a right of access to your own electronic health information, and results are generally released to you as they become available rather than held back until an appointment. You can ask for your records, and a refusal needs a reason.
Second, and less well known: results held by one institution do not travel to another on their own. If a test was run somewhere else, somebody has to request it — and sometimes nobody has. “Has anyone asked for the results from [place]?” is a short question that occasionally finds a great deal.
Everything above is written from the patient’s side, because that is the side we know. The same breakdown frustrates the person at the other end of it, and about their half we would only be guessing.
So instead of describing a working day we have not had, here is what we currently assume. Each one is written to be answerable in a single line, and each one turning out to be wrong is worth more to us than being right.
Grade U is our own evidence scale: asserted, not established. These are assumptions, labelled as assumptions, on a site that grades everything else it says.
And the big one. We assume communication between specialties is minimal, and that it is correctable rather than inherent in the way medicine is organised. That assumption sits underneath most of what this site argues. If it is wrong, a great deal of what we are building is wrong with it.
Perhaps our interpretation and our methodology are wrong. How do you see it?
There is no form and no sign-up. Use the comment rail in the margin — it opens where you are and keeps your place. One sentence is a complete answer, and correcting one of these is more use to us than agreeing with all five.
Where this comes from. The five rungs are Fred Schwacke’s framework,
recorded on 2 April 2026 as “question → follow-up → polite suggestion
→ ‘show me where I am wrong’ → the full document with thorough
citations”, and built here for the first time. The message log, including the
adequate response column, comes from the memory-clinic intake work of January 2026.
The reasoning about escalation and about records is drawn from an advocacy record kept over
two years; the particulars of that case are not published here.
What is not settled. The title of this page is provisional and will be
rewritten from what people actually search for rather than from what we would like to call
it. The five assumptions above carry Grade U and stay at Grade U until clinicians correct
them — at which point the correction goes on the page and the assumption that produced
it stays visible underneath. That is deliberate. The places where this turns out to be wrong
are where the work gets done.