Keep what your doctor told you, not just the summary

What should you keep after a doctor's visit?

Longetic, MyChart, CareClinic, KeepMD and Guava Health all store records from a doctor's visit, and none of them changes what the clinic actually handed you. The after-visit summary printed at checkout is a courtesy document, not a legal one. What United States federal rules guarantee is electronic access to the underlying data, which is a different and much larger thing.

There is no definition of an after-visit summary anywhere in the Code of Federal Regulations. The phrase is an electronic health record vendor's term, and what lands in the document depends on what the clinic's software auto-filled and what the clinician typed before you stood up. It is the most visible thing you leave with and the least authoritative. The other half of this is what you write down before the appointment.

The paperwork that does have legal weight sits behind it. Under United States rules your clinical notes, test results, medication list and problem list are electronic health information, and a clinic that delays your access to them without a good reason is doing something the regulations have a name for. The practical consequence is simple: treat the printout as a receipt, then go get the record.

This page is written for people whose care is spread across several clinicians, and for anyone keeping records on behalf of a parent or partner. It covers the United States in detail because that is where the rules are most specific, and it gives the England, Germany and Poland equivalents in their own row rather than letting the American answer stand for everyone.

Is the after-visit summary the same as your medical record?

No: the after-visit summary is a recap the clinic's software assembles at checkout, while the clinical note is what the clinician wrote about the encounter and is a separate record. Federal certification rules describe a downloadable summary your portal must be able to produce, which is a third thing again, and all three can disagree with each other.

The distinction matters the moment something is wrong. A summary that lists the wrong diagnosis is a printing problem. A note that lists the wrong diagnosis follows you to the next clinician, the next insurer and the next referral. AHRQ's Patient Safety Network published a case where exactly this happened: a teenager treated for herpetic gingivostomatitis left with a summary saying thrush, because the document had been populated from triage and never corrected.

In that write-up the commentary author put the general lesson plainly: "Information alone does not necessarily imply understanding, nor does it decrease patient anxiety in the face of illness."

The third row is the one almost nobody uses. Certification criteria require a patient portal to let you view, download and transmit your record, in human readable form and in a structured format, and to let you "Select data within an identified date range". That means you can pull a machine readable summary for the day of your appointment, not just the paper version of it.

How often is the after-visit summary wrong or incomplete?

Often enough that it should never be your only copy. A review of 250 after-visit summaries in one primary care clinic found 16 percent were not meaningful, and among the reasons they were judged not meaningful, 39.2 percent lacked any information on the presenting problem. Separately, 21.1 percent of patients who read their own clinical notes reported finding a mistake.

The 2019 review of 250 randomly selected summaries from a family medicine residency clinic broke the failures down. Among the reasons summaries were judged not meaningful, 39.2 percent lacked any information on the presenting problem, 35 percent lacked the physician intervention, and 18.4 percent lacked a plan of care. Of the patients who did not speak English in that sample, 15.2 percent received their summary in English.

The medication section drifts hardest, which is its own reason for keeping one medication list across several prescribers rather than trusting whichever printout is newest. A 2026 study of 151 hospitalised older adults covering 319 medication changes found that patients' stated reasons for a change matched the discharge summary for only 38 percent of changes. Where the discharge summary and the discharge instructions agreed with each other, patient understanding was better, 42 percent against 23 percent.

Reading level is the quieter failure. A 2016 intervention at one residency programme found only 30 percent of physicians were writing summaries at a sixth grade level or below before training, rising to 68 percent after. A 2026 analysis of 982 AI generated patient visit summaries in an orthopaedic clinic found a mean Flesch-Kincaid grade level of 9.3, with only 0.4 percent meeting the sixth grade benchmark.

None of which matters if the document is never opened. In a survey of 355 patients at a primary care internal medicine practice, 82.8 percent recalled receiving a summary and 67.4 percent consulted it, so roughly one in three people who were handed one never looked at it again. The same study found that among the 133 patients who recalled personalised free text instructions, 96 percent found them easy to understand. The part a human wrote is the part that works.

What are you entitled to after a visit, and how fast?

In the United States two separate rights apply: HIPAA gives you a copy of your designated record set, and a covered entity "must act on a request for access no later than 30 days after receipt of the request", while the information blocking rules ban unreasonable delay in electronic access, with no fixed clock. Europe answers this country by country.

The second row is the one clinics most often get wrong. ASTP/ONC has answered this directly: "While the information blocking regulations do not require actors to proactively make electronic health information (EHI) available, once a request to access, exchange or use EHI is made actors must timely respond to the request". The agency's own guidance treats a policy of holding lab results back so a clinician can review them first as likely interference, and says access can run "in parallel to the availability of the test results to the ordering clinician".

Two corrections worth carrying, because older articles still get both wrong. There are now ten information blocking exceptions, not eight: preventing harm, privacy, security, infeasibility, health IT performance, protecting care access, manner, fees, licensing and TEFCA manner. And the rules' scope excludes psychotherapy notes and anything "compiled in reasonable anticipation of, or for use in, a civil, criminal, or administrative action or proceeding". Everything else in your designated record set is in scope.

One change in 2026 matters for what your portal has to give you: the adoption of the first version of the United States Core Data for Interoperability "expires on January 1, 2026", leaving version 3 as the baseline your portal has to support. Version 3 carries 19 data classes, including Clinical Notes, Medications, Problems, Procedures, Care Team Members and Goals. The wider expansion many people expected did not arrive: the patient engagement proposals were formally withdrawn in December 2025, and the proposed rule that would adopt version 3.1 closed comments on 27 February 2026 and has not been finalised.

Outside the United States the vocabulary changes. England has no after-visit summary, it has prospective access to the GP record, and NHS England Digital states that "All GP surgeries in England are required to give patients online access to new information as it is added to their GP health record". Germany's nearest equivalent is the Arztbrief, a letter written from one doctor to another that the patient can now read in the ePA. Poland's is the specialist's letter to the referring doctor, which carries the diagnosis, treatment, prognosis and prescribed medicines, and which you read in your Internetowe Konto Pacjenta.

What should you do in the first week after a visit?

Six things, in order, and none of them takes long. Read the summary the same day, check the medication list against the bottles at home, open the portal for the clinical note, write down the reason for every medication change, put every ordered test and referral on a date you control, then file the document where the next clinician can reach it.

Step 1. Read the summary the same day, before the visit fades

Check three things only: the diagnosis, any medication that was started, stopped or changed, and anything that was ordered. This is worth doing while you still remember the room. A narrative review in the Journal of the Royal Society of Medicine reports that "40-80% of medical information provided by healthcare practitioners is forgotten immediately" and that "almost half of the information that is remembered is incorrect".

Step 2. Check the medication list against what is in your cupboard

Printed lists go stale in a specific direction. In a study of 85 medication reconciliation phone visits, 233 discrepancies were found, and the single largest type, 70.4 percent of them, was a medication recorded in the record that the patient had already stopped taking. Roughly 80 percent of the discrepancies were traced to the system rather than to the patient.

Step 3. Open the portal and look for the note, not just the summary

The note is a separate document from the summary and it is usually longer and more specific. If your clinician takes part in the CMS Promoting Interoperability measure, the standard they are scored against is access "Within 4 business days of the information being available to the MIPS eligible clinician". If nothing has appeared after that, ask, and keep the request in writing.

Step 4. Write down why each medication changed

Not just what changed. The reason is the thing that goes missing first, and it is the thing the next clinician will ask for. One line per change is enough: the drug, the change, the reason you were given, and the date.

Step 5. Give every ordered test and referral a date you own

Do not assume the system will chase you. A systematic review of 19 studies found tests not followed up ranging from 6.8 percent to 62 percent for laboratory tests and from 1.0 percent to 35.7 percent for radiology. Put the date in your own calendar, with the name of the test and who ordered it.

Step 6. File it where the next clinician can actually reach it

A summary in a kitchen drawer helps nobody at the next appointment, and a summary in a portal you cannot log into from abroad helps nobody either. Keep a copy you control, in the same place as your lab results and medication list, so that the next time you have to assemble a record a new doctor can read the material is already together.

Which apps keep the visit, the instructions and the results together?

Most of them keep the visit and almost none of them keep the instructions. Doctor visits appear as records in every product below, and several will take a photo or PDF of a document you were handed, but only CareClinic publishes follow-up actions attached to a visit record. Reminder types differ sharply, so read that column rather than the word reminder.

Read that table for what is missing as much as for what is there. Apple Health Records is not in it because Apple does not publish an encounter or visit record type and publishes nothing about uploading your own document, and Healthmatters.io is not in it because it handles laboratory results only. Where a vendor publishes nothing, the cell says so rather than guessing.

As a disclosure, Longetic is our product, so weigh this accordingly. It stores blood tests, medical records, doctor visits and medications in a single searchable timeline, and it reads biomarker values out of a lab report you upload as a PDF or photo, which is the part it is actually built for and the reason people use it to track a biomarker across every report you have. It does not publish an after-visit summary import, it does not keep your doctor's instructions as tasks, and the app is iPhone only, with no Android app. If what you need is a scanner for paper and a transcript of the appointment itself, KeepMD is built for that and we are not. If you want the clinic's own copy, your patient portal already has it and it is free.

What happens to the tests and referrals your doctor ordered?

A measurable share of them quietly stop. Published rates vary by setting and by how completion was recorded, but the direction is consistent across every study: results that are never communicated, referrals with no documented appointment, and abnormal values that nobody follows up. The gap is a systems problem, and the only part of it you control is your own calendar.

The referral number needs its caveat stated rather than buried. In that single health system, 34.8 percent of referral scheduling attempts ended in a documented complete appointment, but 38.9 percent lacked a documented appointment date. That is a record keeping gap as much as a care gap, and it is not evidence that two thirds of referrals never happen. What it does show is that the loop is not reliably closed on paper, which is exactly why your own date matters.

The comprehension research points the same way. A study of 140 patients discharged from emergency departments in two health systems found that 78 percent showed deficient comprehension in at least one area, while patients "perceived difficulty with comprehension only 20% of the time when they demonstrated deficient comprehension". Feeling clear about your instructions is not evidence that you are.

And when patients do read the underlying notes, they find things. In a survey of 22889 people who had read their own ambulatory care notes, 21.1 percent perceived a mistake, and of those, 42.3 percent judged the mistake serious. The most common very serious errors involved a current or past diagnosis, then inaccurate medical history, then medications or allergies. Reading your note is the cheapest error check available to you.

Frequently asked questions