Medical chart summary: the 5 formats clinicians use
Answer in brief: a medical chart summary is a short document that pulls from the chart what the next clinician needs to act. There are 5 working formats, and each has a different reader:
- Pre-visit summary: for you, before you see the patient. What changed since the last visit and what is due.
- Consult summary: for the team that asked. The question, your answer, specific recommendations.
- Transfer summary: for the team taking over today. Why the patient is moving, current status, what is pending.
- Discharge summary: for the primary care physician or facility. What happened, what changed, who follows up on what.
- Problem-list summary: for everyone. The running list of active problems the other 4 draw from.
The items most often missing from discharge summaries are what the patient was told, tests still pending and key test results. The numbers are below.
Limits: Vendor facts come from each vendor's own website on 2026-11-12 and were not tested by us. Rules are summarized from the cited text; your hospital's bylaws, payer contracts and state law can ask for more.
What is a medical chart summary?
A medical chart summary is a selection from the chart, written for 1 reader and 1 decision. A discharge summary tells a primary care physician what to do next week. A transfer summary tells the receiving team what to watch tonight. Same chart, different facts on top. A summary that tries to serve every reader turns into a copy of the chart, and the facts that matter sink under the ones that don't.
This site covers summaries written for patient care: in clinics, emergency departments and hospitals. Summaries written for insurance claims or lawsuits follow other rules, and most of what ranks for this phrase on Google is written for them.
The 5 chart summary formats at a glance
Each format answers 1 question for 1 reader. The last 2 columns show the rule that asks for it and the LOINC document code your EHR uses when it sends it to another system.
| Format | Reader | The question it answers | Rule or standard behind it | LOINC code |
|---|---|---|---|---|
| Pre-visit | The clinician about to see the patient | What changed, and what is due today? | No federal rule; AMA team-care guidance says plan 2 to 3 days ahead | None of its own |
| Consult | The requesting physician | What is the answer to my question? | Medical staff bylaws; C-CDA Consultation Note | 11488-4 Consult Note |
| Transfer | The receiving team | What do I need to keep this patient safe tonight? | 42 CFR 482.43(b) and (c); EMTALA, 42 CFR 489.24(e)(2)(iii) | 18761-7 Transfer Summary Note |
| Discharge | Primary care physician, facility, patient | What happened, what changed, what do I do next? | 42 CFR 482.24(c)(4)(vii); 42 CFR 482.43(b) | 18842-5 Discharge Summary |
| Problem list | Every clinician who opens the chart | What is active, and what is being done about each? | Weed's problem-oriented record (NEJM, 1968) | 11450-4 Problem list |
Rules read from eCFR on 2026-11-12. LOINC codes looked up through the NLM Clinical Tables service on 2026-11-12. Field-by-field lists for each format are on the formats page.
What the top Google results for this phrase leave out
On 2026-11-12 we pulled the Google top 10 for "medical chart summary" twice, plus "patient chart summary", "clinical summary of patient chart" and "chart summary template", and read every result. 14 distinct pages came back for the exact phrase. What we found:
- 8 of the 14 come from companies that sell summaries for claims, litigation or medico-legal review. They describe a claim-file digest for adjusters, law firms and examiners, not a document used to care for a patient.
- 4 are for patients: a patient portal, a hospital blog on what a chart is, a personal health summary guide and a 1-page registry form for a bleeding-disorder database.
- None lists the fields a pre-visit, transfer, discharge, consult or problem-list summary must contain.
- None cites a rule. Not the CMS hospital rules on discharge summaries, not EMTALA's list of what goes with a transferred patient, not the HL7 document types EHRs send.
- None gives omission rates, though the research on what discharge summaries leave out goes back to 2007.
- "Chart summary template" returns mostly data-chart templates (Pinterest, stock images), and plain "chart summary" returns only data-visualization pages.
So this site does 5 things those pages don't: field lists for each clinical format tied to their source, copyable templates, 1 synthetic chart summarized 4 ways, an omission checklist built on the published numbers, and a dated table of 19 tools.
The 8 fields every chart summary needs
The 5 formats differ in emphasis, but 8 fields show up in all of them. If a summary is missing 1 of these, the reader has to open the chart, and the summary has failed at its 1 job.
- Who, when, from where. Patient identifiers, the date range the summary covers, and the source documents it was built from. A summary without a date range invites the reader to assume it is current.
- The reason. The visit reason, the consult question, the reason for transfer or the reason for admission. Put it first.
- Active problems, with status. Not the whole problem list: the problems that matter for this reader, each with "better, worse or unchanged".
- Medications, with changes marked. What was started, stopped, held or changed, and why. Cornish and colleagues found that 53.6% of admitted patients on 4 or more medications had at least 1 unintended medication discrepancy, and the most common error was a regular medication left out (Arch Intern Med, 2005).
- Allergies and reactions. The reaction, not just the drug. "Penicillin: hives" and "penicillin: anaphylaxis" lead to different choices.
- Key results, dated. The results that drive decisions, each with its date, plus every result still pending and who will act on it.
- What happened. The course of care in a few sentences: what was tried, what worked, what didn't.
- The plan, with an owner. Each follow-up item names who does it and by when. "Follow up with cardiology" is not a plan. "Cardiology clinic visit within 7 days, booked for 2026-10-19, to recheck potassium and titrate metoprolol" is.
For transitions (transfer and discharge), add code status and goals of care. The receiving team can't safely assume either.
Where chart summaries fail: the published numbers
The research we found is mostly on discharge summaries, because that is where a missing fact travels furthest. The results are consistent: summaries arrive late and leave out the same things.
| Finding | Number | Source |
|---|---|---|
| Discharge summary available at the first post-discharge visit | 12% to 34% | Kripalani et al., JAMA 2007, systematic review |
| Still not available at 4 weeks | 23% to 49% (available in 51% to 77%) | Kripalani et al., JAMA 2007 |
| Summaries missing test results pending at discharge | 65% | Kripalani et al., JAMA 2007 |
| Summaries missing diagnostic test results | 33% to 63% | Kripalani et al., JAMA 2007 |
| Summaries missing follow-up plans | 2% to 43% | Kripalani et al., JAMA 2007 |
| Summaries missing discharge medications | 2% to 40% | Kripalani et al., JAMA 2007 |
| Patients discharged with results still pending | 41% (1,095 of 2,644) | Roy et al., Ann Intern Med 2005 |
| Of those pending results surveyed, physicians unaware of | 61.6% | Roy et al., Ann Intern Med 2005 |
| Admitted patients with at least 1 unintended medication discrepancy | 53.6% | Cornish et al., Arch Intern Med 2005 |
Abstracts read on PubMed on 2026-11-12. The ranges are across the studies each review pooled, so treat them as a spread, not a single rate. The omissions page turns these into a checklist by format.
Structure helps. The I-PASS handoff program, which standardized oral and written handoffs across 9 hospitals, saw the medical-error rate fall 23% (24.5 to 18.8 per 100 admissions) and preventable adverse events fall 30%, with no change in how long a handoff took: 2.4 minutes per patient before, 2.5 after (Starmer et al., NEJM 2014). A fixed template is the cheapest version of that structure, and it is why every format on this site comes with one.
AI chart summaries: what the 2025 evidence says
The large EHR vendors now offer AI summaries of the chart. Epic lists AI summaries for admitted patients, referrals, the ED problem list and a draft hospital course. Oracle Health announced acute nursing summaries in September 2026. athenaOne summarizes outside documents. The tools page lists 19 products with a source for each.
The best evidence so far is mixed in a specific way. In a 2025 study at NYU Langone, residents edited AI-drafted hospital courses from an EHR-embedded model and physician-written ones, blind to which was which. They changed 31.5% of the AI text versus 44.8% of the physician text. Attending hospitalists rated the edited AI drafts more complete, but with more confabulations: statements the chart doesn't support (Small et al., JAMA Netw Open 2025). An accompanying commentary noted that the AI drafts still held confabulations after human editing.
That gives 3 working rules for any AI summary you sign or rely on:
- Every line needs a source you can open. Epic says its AI Insights link back to the source documentation. If a tool can't show where a sentence came from, check that sentence by hand or delete it.
- Check what is missing, not just what is wrong. A summary that reads well can still drop the pending result or the held medication. Run the omission checklist against it.
- Treat accuracy claims as claims. In 2024 the Texas Attorney General settled with Pieces Technologies over what the state called deceptive claims about the accuracy of an AI product that summarized patients' condition and treatment for hospital staff. Ask a vendor how it measured, on what charts, and whether you can test it on yours.
How to pick the right format in 30 seconds
Name the reader and the decision they will make. The format follows.
| If the reader is... | ...and they need to... | Write a |
|---|---|---|
| You, tomorrow morning | Walk into the room knowing what changed | Pre-visit summary |
| A physician who asked you a question | Act on your answer | Consult summary |
| A team taking over care today | Keep the patient safe through the next shift | Transfer summary |
| A primary care physician or nursing facility | Pick up care after the hospital | Discharge summary |
| Anyone opening the chart cold | See what is active in 1 screen | Problem-list summary |
| A quality or peer review committee | Reconstruct what happened and when | A dated, cited chronology (outside the scope of this site) |
2 shortcuts. If the patient is moving between teams or buildings today, write a transfer summary even if a discharge summary will follow; the receiving team can't wait for it. And if you are writing a consult, put the answer in the first 2 lines. Goldman, Lee and Rudd's 1983 rules for consultations start with pinning down the question being asked, and include keeping the note as brief as the case allows and making specific recommendations (Arch Intern Med, 1983).
Questions about medical chart summaries
What should a medical chart summary include?
A medical chart summary should include the patient and date range, the reason for the summary, active problems with status, medications with changes marked, allergies with reactions, key dated results including anything pending, a short course of care, and a plan where each item has an owner. Transfer and discharge summaries also need code status and goals of care.
How is a chart summary different from a discharge summary?
A discharge summary is 1 type of chart summary. It is written when a patient leaves the hospital, for the clinician who picks up care, and CMS requires it to cover the outcome of the hospitalization, the disposition and follow-up care (42 CFR 482.24(c)(4)(vii)). A chart summary can also be a pre-visit, transfer, consult or problem-list summary.
How long should a chart summary be?
As short as the reader's decision allows. A pre-visit summary should fit on 1 screen. A transfer summary for an ICU patient can run longer because the receiving team needs current status by system. If a section doesn't change what the reader does next, cut it or move it to an appendix.
What do chart summaries leave out most often?
Test results pending at discharge, patient and family counseling, diagnostic results, follow-up plans and medication changes. A 2007 JAMA systematic review found pending tests missing from 65% of discharge summaries and follow-up plans missing from 2% to 43%. The omissions page has the full list by format.
Can AI write a medical chart summary?
AI can draft one, and the large EHR vendors now offer AI summaries. A 2025 JAMA Network Open study found AI-drafted hospital courses needed less editing than physician drafts but held more confabulations. Use AI drafts only with line-level sources, and check them for what is missing as well as what is wrong. The clinician who signs is responsible for the content.
Is there a standard format for chart summaries?
For electronic exchange, yes: HL7's Consolidated CDA (C-CDA) defines document templates including the Discharge Summary, Transfer Summary, Consultation Note and Referral Note, each with required sections. For the text a clinician reads, there is no single national format; hospitals set their own templates, and the templates page gives you a starting point for each.