Discharge summary: what it contains and how to check it against the stay

A discharge summary is written when a patient leaves the hospital. It tells the next clinician, usually a primary care doctor, a nursing home or a home health agency, what happened during the stay and what needs to happen next. Every hospital record must have one, and outpatient surgery and emergency visits have a shorter equivalent.

It is not the same as the discharge instructions. Instructions are written for the patient in plain language: medications, activity, warning signs, appointments. The summary is written for clinicians. A file usually contains both, and they should agree.

For a reviewer, the summary is a map, not the territory. It points to the admission, the key events and the diagnoses, and it saves time on a first pass. But it is written at the end of the stay by 1 physician, often one who did not see the patient every day, and it compresses days or weeks into a few paragraphs. Every important statement in it should be traced back to the record it summarizes.

Terms you will meet

Hospital course

The narrative of the stay, usually by problem or by day. The longest and most compressed section.

Principal diagnosis

The condition established after study to be chiefly responsible for the admission. It drives coding and billing.

Secondary diagnoses

Other conditions present or arising during the stay. Complications appear here, or do not.

Disposition

Where the patient went: home, home with services, skilled nursing, rehabilitation, hospice, another hospital, against medical advice, or death.

Discharge medications

The list the patient goes home on, ideally marked as new, changed, stopped or continued.

Pending results

Tests sent during the stay whose results were not back at discharge, such as cultures, pathology or send-out labs.

Transfer summary

The equivalent document when a patient moves to another facility. Death summaries cover patients who died in hospital.

Co-signature

When a resident, nurse practitioner or physician assistant writes the summary, the attending physician co-signs to verify it.

The mock summary below closes a 4-day admission for pneumonia. The layout is invented; every hospital's template differs. The 9 marked parts appear in some form in almost all discharge summaries.

What each section should be checked against

The best evidence on discharge summaries comes from studies of what reached the next doctor. Kripalani and colleagues reviewed 55 observational studies. Direct communication between hospital and primary care physicians happened in 3% to 20% of cases. The discharge summary was available at the first follow-up visit in 12% to 34% of cases, and still only in 51% to 77% at 4 weeks. When it did arrive, it often lacked diagnostic test results (missing from 33% to 63%), the hospital course (7% to 22%), discharge medications (2% to 40%), test results pending at discharge (65%) and follow-up plans (2% to 43%).

Pending results are the most studied gap. Were and colleagues reviewed discharge summaries for 668 patients who all had tests pending at discharge:

Those results do come back, and some matter. Roy and colleagues followed 2,644 patients discharged from hospitalist services: 41% had test results return after discharge, and 9.4% of those results were potentially actionable. Of the potentially actionable results they surveyed physicians about, the physicians had been unaware of 61.6%.

The period after discharge is also when harm shows up. Forster and colleagues followed 400 patients discharged home from a general medical service. 19% had an adverse event within about 3 weeks, and 66% of those were adverse drug events. Some were judged preventable. That is why the medication list in the summary gets the closest reading.

Federal hospital rules require a discharge summary in every record and set a deadline for completing the record. A separate rule requires the hospital to send the information the next providers need.

CMS guidance for surveyors adds detail. Follow-up care provisions include post-hospital appointments and plans for care by home health, hospice, nursing homes or assisted living. The admitting physician is responsible for the summary; it can be delegated to a nurse practitioner or physician assistant where state law allows, and then:

"For delegated discharge summaries we would expect the MD/DO responsible for the patient during his/her hospital stay to co-authenticate and date the discharge summary to verify its content."

Copied text and late summaries

Many hospital course sections are built by copying earlier notes forward. That is not improper by itself, but copied text can carry forward findings that changed, or leave out events that happened between the copied note and discharge. A summary dictated weeks after discharge was written from the chart, and it was written after any complaint or readmission. Our guides to altered medical records and EHR audit trails cover how to test when and how a summary was written.

Some problems recur often enough to check for every time. Most are differences between the summary and the records it summarizes.

The medication pattern is the one to check first. The summary's list, the discharge orders, the last day of the medication administration record and the patient instructions are 4 versions of the same list. They often differ, and the difference is where adverse drug events after discharge start.

What each flag means

1. A new drug missing from the summary. The anticoagulant was started on day 3 and ordered at discharge. The summary and the patient instructions say home medications are unchanged. Which list the patient followed is a question for the patient and the pharmacy records.
2. A fever the day before discharge. The hospital course says afebrile. The lab log shows a culture drawn for a fever of 38.6.
3. A pending culture not listed. "Pending: none," though the culture was in progress at discharge. The result came back the next morning.
4. No record of who acted on the result. The file needs the lab's notification log and any call to the patient or the primary care office between 05/07 and 05/12.
5. A summary written after the readmission. Signed on 05/20, 8 days after the patient came back, and faxed the next day. The primary care doctor could not have had it before the readmission.

The same stay on 1 timeline

The method works for a 2-day stay or a 2-month one. Read the summary first for orientation, then set it aside and test it.

1. Note the dates. Admission, discharge, dictation, signature, co-signature and date sent.
2. List the claims. Every diagnosis, procedure, event, result and medication the summary states.
3. Trace each claim. Find the source note, result or order for each. Mark the ones with no source.
4. Read the chart for omissions. Rapid responses, falls, transfers, new diagnoses, consults and complications that the summary does not mention.
5. Find pending results. Every test ordered during the stay with a result time after the discharge time.
6. Compare the 4 medication lists. Summary, discharge orders, last-day MAR and patient instructions.
7. Check the diagnoses against the codes. Summary diagnoses vs UB-04 diagnosis codes and present-on-admission flags.
8. Follow the handoff. Who was named for follow-up, whether an appointment was made, and when the summary reached them.

Can this discharge summary be relied on as written?

DISCHARGE SUMMARY
Patient: [NAME / ID]
Admitted: [DATE]    Discharged: [DATE, TIME]
Attending: [NAME]    Author: [NAME, ROLE]

Reason for admission: [ ]
Principal diagnosis: [ ]
Secondary diagnoses (including complications): [ ]
Procedures, with dates: [ ]
Consultants: [ ]

Hospital course, by problem:
  [Problem]: [what was found, done, and the result]
  Significant events (falls, rapid responses, transfers): [ ]

Key results: [labs, imaging, pathology, with dates]
RESULTS PENDING AT DISCHARGE: [test, date sent, who will follow up]

Discharge medications:
  NEW: [ ]   CHANGED: [ ]   STOPPED: [ ]   CONTINUED: [ ]
Allergies: [ ]

Condition at discharge: [ ]
Disposition: [home / home health / SNF / rehab / hospice / other]
Follow-up: [NAME, specialty, DATE or timeframe]
Patient and family informed of: [ ]

Dictated: [DATE, TIME]    Signed: [NAME, DATE, TIME]
Co-signed (if delegated): [NAME, DATE]
Sent to: [PROVIDERS, DATE]

DISCHARGE SUMMARY CROSS-CHECK
Patient: [ID]    Admission: [DATES]
Summary signed: [DATE]  co-signed: [DATE]  sent: [DATE, TO]

Type (claim / omission / pending / medication)
  | Summary says (page)
  | Source record says (page)
  | Match? | Question

Examples:
Pending | "None" (p. 98)
        | Blood culture drawn 05/05 18:40, resulted 05/07 (p. 74, 76)
        | No | Who was notified of the result?
Medication | "Home medications unchanged" (p. 99)
           | New anticoagulant 05/04 to discharge (MAR p. 86)
           | No | Which list did the patient follow?

Before you rely on a discharge summary

• Signature and send dates are knownIncluding any co-signature.
• Progress and nursing notes are in handTo find events the summary omits.
• Result times for every test are in handTo find results pending at discharge.
• All 4 medication lists are in handSummary, orders, last-day MAR, instructions.
• The UB-04 or coded diagnoses are in handWith present-on-admission flags.
• Case management notes are in handFor the discharge plan and destination.

0 of 6 checked.

For diagnosis codes and present-on-admission flags, see the UB-04 and CMS-1500 guide. For the notes the hospital course summarizes, see the SOAP note and nursing notes guides, and for surgical stays, the operative report guide.

A discharge summary is 2 pages. The stay it summarizes can be 2,000. Chapter 7's method is tracing each claim in the summary back to the chart, and scanning the chart for what the summary left out. Doing that by hand is slow; it is the kind of matching software does well.

AI medical record review can build the timeline from the source records rather than from the summary, put every lab result, medication and note on it with a page-level citation, and show the summary's claims beside what the chart says. That turns an admission into an AI medical chronology where a pending result, a medication that differs between lists, or an event the summary omits is visible at a glance and can be checked against the page in seconds.

There is a specific trap here. A large language model (LLM) asked to summarize a hospital stay will lean on the discharge summary, because it is already a summary, and repeat its omissions. Hallucination and omission are both risks, which is why every line needs its page and why the timeline should be built from the underlying records. Whether a discharge was premature or a result should have been acted on is a clinical judgment. The human in the loop reads the flagged pages.

Medrecords AI works this way on the files you upload. It builds a cited medical chronology from the notes, results, MAR and orders, each entry linked to its page. It answers questions like "every test resulted after the discharge time, and whether the summary mentions it" through cited Q&A and record search, writes summaries with a citation on every line, and flags records that should exist but were not produced. Flags are signals, not verdicts. It does not decide whether care met the standard, and it does not retrieve records from providers. It runs under SOC 2 and HIPAA with a signed business associate agreement (see security and HIPAA). Self-Service bills 10 cents a deduplicated page, down to 5 cents at volume, duplicates free; Enterprise On-Prem is an annual license.

If you are comparing tools, our guide to medical record review covers what to ask any medical chronology software vendor. See how discharge review fits medical malpractice, wrongful death, hospital risk management and legal nurse consultant work.

What should a discharge summary include?

The reason for admission, diagnoses, procedures, the hospital course, key results, results pending at discharge, discharge medications marked as new, changed or stopped, condition and disposition, follow-up plans with names, and the author's signature with date and time.

Is a discharge summary required?

Yes. Federal hospital rules require every record to have a discharge summary covering the outcome of the stay, the disposition and follow-up care, and the record must be completed within 30 days after discharge.

Who writes the discharge summary?

The physician responsible for the patient during the stay. It can be delegated to a resident, nurse practitioner or physician assistant under hospital policy and state law, and the responsible physician then co-signs it.

What is the difference between a discharge summary and discharge instructions?

The summary is written for clinicians and covers the whole stay. The instructions are written for the patient: medications, activity, warning signs and appointments. Both belong in the file and should agree.

What are pending results at discharge?

Tests sent during the stay whose results were not back when the patient left, such as cultures or pathology. Studies found discharge summaries mention only a small share of them.

Can I rely on a discharge summary in a legal case?

As a guide to the stay, yes. As proof of what happened, check each statement against the notes, results and orders it summarizes. Summaries often omit events and results.

Why does the date a discharge summary was signed matter?

A summary signed after a readmission or complaint was written with knowledge of the outcome, and it could not have guided the care that followed discharge.

Can AI write or review a discharge summary accurately?

It can build a timeline from the underlying records and show where the summary differs from them, with a page cite on every line. An AI summary that relies on the existing discharge summary will repeat its gaps.

Is it HIPAA compliant to upload hospital records to an AI tool?

It can be, with a vendor that signs a business associate agreement, holds a SOC 2 report and commits in writing not to train on your data. A consumer chatbot without a BAA is a different matter.

Regulations and the CMS State Operations Manual (Appendix A, revision 248) were read in full text from primary sources in September 2026, and quoted text is verbatim. Study figures come from the published abstracts. The mock summary, patient, physicians, dates and pages in chapters 2, 6 and 7 are hypothetical. Product facts come from this site's product pages. Nothing here is legal or medical advice.

• 42 CFR 482.24: medical records, the discharge summary and the 30-day completion rule.
• 42 CFR 482.43: discharge planning and transmission of necessary medical information.
• CMS State Operations Manual, Appendix A: interpretive guidance at tags A-0468 (discharge summary) and A-0469 (final diagnosis and record completion).
• Kripalani S, LeFevre F, Phillips CO, et al., "Deficits in communication and information transfer between hospital-based and primary care physicians: implications for patient safety and continuity of care", JAMA 297(8):831-841, 2007.
• Roy CL, Poon EG, Karson AS, et al., "Patient safety concerns arising from test results that return after hospital discharge", Annals of Internal Medicine 143(2):121-128, 2005.
• Were MC, Li X, Kesterson J, et al., "Adequacy of hospital discharge summaries in documenting tests with pending results and outpatient follow-up providers", Journal of General Internal Medicine 24(9):1002-1006, 2009.
• Forster AJ, Murff HJ, Peterson JF, et al., "The incidence and severity of adverse events affecting patients after discharge from the hospital", Annals of Internal Medicine 138(3):161-167, 2003.