1. The 30-second answer: keep the memo; delete the uncompressed log
When people see a doctor post that they “brace themselves” when a patient hands over a long summary, some may hear a different message: “I should bring nothing.” That conclusion goes too far.
The recurring problem is not the medium called a memo. It is a large volume of information with no hierarchy. Asking a clinician to decode several pages of symptoms during a short visit is difficult. A short note that states the purpose of the visit, changes in symptoms, functional impact, preferences, and constraints can move the conversation forward.[R1][R2][R3]
Bring a README.md, not your collected autobiography.
2. Translating “long notes are hard” into “patients with notes are difficult” creates another failure mode
A long, unstructured document can be hard to use. That is different from saying that patients should not prepare.
If the second message dominates, people who forget under stress, struggle to explain themselves verbally, or have several concerns may throw away a useful support tool. There is no evidence that one particular social-media post actually reduces disclosure, so this should be framed as a plausible communication risk, not a proven causal effect.
AHRQ advises patients to think about what they want from an appointment and write down questions and concerns.[R2] Saiseikai’s patient guidance in Japan similarly recommends communicating symptom severity, impact on daily life, time course, aggravating or relieving factors, worries, and concrete preferences such as wanting to discuss tests or a medication change.[R1]
Expressing a preference is not automatically being a “difficult patient.”
3. Oral-only care can become a one-shot memory test
Imagine a patient who comes in for a skin problem. They also want to discuss medication, but never say so. The clinician examines the problem, the visit ends, and only afterward does the patient think, “But I wanted medicine.”
It is easy to say, “Then say it during the visit.” Research on unvoiced agendas shows why that is not always reliable.
A 2000 BMJ qualitative study found that only 4 of 35 patients voiced their complete pre-visit agendas. Unspoken items included worries about diagnosis, patients’ own explanations, side effects, not wanting a prescription, and social context. In all 14 consultations with problem outcomes, at least one problem was related to an unvoiced agenda item.[R4]
A 1999 JAMA analysis of 264 visits found that patients completed their initial statement of concerns in 28.0% of encounters. Physicians redirected the opening statement after a mean of 23.1 seconds. Late-arising concerns occurred in 34.9% of visits with no solicitation of concerns versus 14.9% when concerns were solicited.[R5]
These are old studies in specific settings, not a claim that every modern doctor interrupts at 23 seconds. The useful lesson is narrower: reconstructing every concern and its priority from memory during the visit is a lossy system.
4. Does setting the agenda in advance actually help?
In a 2022 cluster-randomized primary-care trial, patients used a pre-visit agenda tool integrated with the electronic record. The mean number of agenda items discussed rose from 1.94 to 2.25, while consultation time did not significantly differ: 15.45 versus 15.83 minutes.[R3]
A 2018 interview study involving 40 patients with complex chronic conditions and 17 physicians found that many patients prepared written or smartphone lists and home-monitoring data. Common reasons were to avoid forgetting something and to avoid becoming too intimidated to mention it. Visits aligned better when patients raised high-priority items early and negotiated the agenda with the clinician at the start.[R6]
A 2015 review of Question Prompt Lists covered 42 studies and 50 interventions. Findings varied, but an appropriate QPL—especially when endorsed by the physician—could increase question asking and the amount of information provided. Effects on visit length were inconsistent.[R7]
The evidence does not say “write everything.” It supports preparation, prioritization, and early agenda alignment.
5. In psychiatry, symptoms are only part of the payload
Psychiatric visits are not simply a daily mood score.
A Japanese study of 52 psychiatric outpatients and 104 consultations found that, by word count, 37.4% of conversation concerned life, 26.1% symptoms, 16.3% treatment/service use, and 7.3% global state/goals. Daily-life issues, especially work, were major topics.[R8]
Japan’s National Center of Neurology and Psychiatry describes shared decision making (SDM) as a process in which service users and professionals discuss treatment goals, preferences, and responsibilities to find an appropriate course together.[R9] A multicenter cluster-randomized trial of SDM training for depression improved patients’ perceived involvement in decision making after the first visit, although it did not improve every outcome measured.[R10]
A useful visit note therefore includes not only symptoms, but also the main functional problem, what the patient wants to discuss, preferred outcomes, and constraints such as drowsiness, cost, driving, or work requirements.
6. A daily feelings diary is not mandatory; log according to purpose
Useful notes do not imply that everyone should write a long diary every night.
A 2026 systematic review and meta-analysis of mood-monitoring interventions in depression and bipolar disorder included 8 trials and 1,230 participants. It found no robust evidence that mood monitoring broadly improves or worsens mood symptoms.[R11] This does not prove that ordinary journaling is useless; it simply does not support the blanket claim that daily tracking automatically improves outcomes.
Another 2026 systematic review of remote measurement-based mental-health care reported a mean adherence rate of 74.5% across tracking items, and more prompts per day were associated with lower adherence. Evidence for clinical effectiveness remained limited.[R12]
A practical system is therefore:
Routine: update only the visit summary.
Meaningful change: log medication changes, attacks, major sleep disruption, or other events.
Time-series matters: keep detailed sleep, attack, blood-pressure, or similar logs when clinically useful.
What the clinician sees first: a compressed summary, roughly one page.
Keep the database if it helps. Do not dump the SQL file into the appointment.
7. Why do they ask again when you already wrote it on the form?
From the patient side, this is annoying: “I literally wrote that.”
But repeated questions are not always proof that the form was ignored. Different clinics may separate intake collection, nursing verification, and physician confirmation or clarification. A good workflow may even use the written note as a readback: “So these are the main changes, correct?”
Systems vary widely. A memo cannot guarantee that a clinic will integrate information well. Its portable benefit is that the patient has already organized their own information once.
8. The visit README: a copyable template
There is no scientific rule proving that one single-sided A4 page is optimal. It is a practical compression target: roughly one phone screen to one page.
[Main purpose today]
1. I want to discuss ___
2. If possible, I also want to ask about ___
[Symptoms / changes]
- Started:
- Frequency / severity:
- Better or worse since last visit:
[Impact on life]
- Work / sleep / chores / driving:
[Medication / treatment]
- What helped:
- Problems / side effects:
[What I want from this visit]
- Discuss medication
- Ask whether testing is needed
- Confirm whether watchful waiting is reasonable
[Constraints]
- Avoid next-morning drowsiness
- Need to drive
- Limits on cost or visit frequency
[Detailed logs]
- Available if useful
“I want to discuss medication” states the goal. It is not the same as ordering the clinician to prescribe a specific drug.
9. If you use AI, make it compress—not expand
The worst AI workflow is to turn every symptom into polished prose until a one-minute problem becomes a five-page RFC.
Use AI for the opposite task:
- remove duplicates;
- limit today’s priorities to three;
- normalize the timeline;
- separate symptoms from functional impact;
- separate preferences from constraints;
- never invent missing facts;
- move raw logs to an appendix.
Compression, not generation. Indexing, not autobiography.
10. Clinicians can communicate the rule better too
Instead of “don’t bring notes,” a more useful message is:
“Notes are welcome. Put your biggest concern first and, if possible, keep the summary to about one page. If you have a long log, put the summary on top.”
Research on QPLs suggests clinician endorsement can help these tools work as intended.[R7]
The goal is not less information. It is a protocol that moves high-value information to the top.
11. Bottom line: a medical visit is neither an autobiography reading nor a memory exam
You do not need to hand over months of diary entries. You also do not need to abandon preparation because someone dislikes long notes.
A robust sequence is:
purpose → symptoms and changes → functional impact → preferences → constraints.
When it works well, the clinician can say, “So this is what has been happening?” and the patient can confirm and move directly to clarification and decisions.
That is not avoiding conversation. It is what happens when the information architecture was done before entering the room.
A visit memo is not your life story.
It is the README for the version of you that showed up today.
