The most common reason a treatment change produces no clear answer is that nobody defined what an answer would look like. Six weeks later the question "did it help?" gets "I think so, maybe, a bit" — which cannot be built on, and which sends everybody round the loop again.
The fix takes two minutes at the point the change is made.
Ask what the target is
Not "will this help", but:
"What specifically is this meant to change, and by when?"
A good answer is observable and dated. "The aim is that you can start work in the morning without the usual hour of stalling, and we would expect to see something within two weeks." That is a claim you can check.
A vague answer is worth pushing on, politely. "How will we know if it is working?" is a normal question and clinicians are generally glad to be asked, because the alternative is a review appointment with no evidence in it.
Write the target in your own words
One sentence, before you start. In terms of something you would notice.
Weak targets: feel better, be less anxious, have more focus. Usable targets:
- "Can sit down and start a task in the morning without an hour of stalling."
- "Can read a page and remember what was in it."
- "Wake up without the first hour being unbearable."
- "Get through a week without a crisis over an unopened letter."
- "Fewer than two bad afternoons a week."
The test of a good target is whether someone else could tell if it happened.
Ask the four other questions
"How long before I should expect anything?" This varies enormously — some medications act the same day, others take weeks. Knowing which you have prevents both premature abandonment and pointless waiting.
"What should I watch for that would mean stopping or calling?" The specific side effects that matter, and the ones that are expected and will settle.
"What is the plan if this does not work?" Knowing that there is a next step makes the trial easier to sit through, and it tells you whether the current attempt is one option among several or the last one.
"When are we reviewing this, and how?" A date. If the review is six weeks away and the medication should act within two, ask what to do at two weeks.
Then keep a short record
This is the part that does the work, and it does not need to be elaborate. For the first few weeks, one line a day:
Tue 4th — took it, 8am. Morning fine, started work without stalling. Flat by 3pm. Slept ok. Wed 5th — took it late, 10am. Afternoon worse. Headache from about 4. Thu 6th — missed it. Bad day, could not start anything.
Four things only: whether you took it and when, what happened to the target, anything new, and the obvious confounders — sleep, and whether the day was unusual.
Two weeks of that answers questions that no amount of recollection can: whether the effect arrives and when, whether it wears off and when, whether the bad days were the days you slept badly, whether the missed doses explain the poor week.
Why recollection is not good enough
Three well-known problems, and they compound.
Recency. The last few days dominate the impression. A good fortnight followed by a poor week reads as "it did not work".
Adaptation. Improvement becomes normal quickly. Something that genuinely helped stops feeling like a change, and gets reported as no change.
Attribution. Without notes, everything gets attributed to the medication — including the week you slept five hours a night and the week you had a cold.
A written record does not solve everything, but it distinguishes between these, and they are the three most common ways a working treatment gets abandoned and a non-working one gets continued.
What "partial" means
Frequently the honest answer, and frequently misread as failure.
Partial improvement is a finding, not a dead end. It usually means the direction is right and something needs adjusting — the dose, the timing, the formulation, or something else in the picture. "The mornings are clearly better and the afternoons are unchanged" is one of the most useful things you can report, because it points at a specific adjustment: duration of action.
Compare it with "it sort of helped", which points at nothing.
The thing to resist
Changing two things at once. If a dose changes and you also start sleeping differently, or begin a supplement, or go on holiday, the result is uninterpretable. It is not always avoidable — life does not pause — but when it happens, note it, so that the confound is visible rather than invisible.
At the review
Bring a summary rather than a log:
Three weeks. Mornings clearly better from about day four — I can start work, which I could not before. Afternoons unchanged, and there is a flat patch from about three. Headaches twice a week, always late afternoon. Missed three doses, all weekends. Sleep unchanged.
That is a paragraph somebody can act on immediately, and it probably points at a specific change. Compare it with the same three weeks reported as "I think it helped a bit" — the same experience, and nothing anyone can do with it.