Discharge Summary
It seems like no one really knows how to do these.
These are usually dictated or typed into the medical record. The job of the discharge summary is to let people know what happened during that hospitalization and how the patient was treated.
Who actually reads this
Most people write this as though it disappears into a drawer. It has at least five readers, and each wants something different out of the same document.
The primary care physician opens it two weeks later with the patient in the room and eleven minutes on the clock. She wants to know what you changed and what she now owns.
The next hospitalist opens it at two in the morning when the patient bounces back through the emergency department. He wants the story, the trajectory, and why you thought it was safe to send this person home.
The skilled nursing facility builds tonight's care plan from it, so they want functional status, diet, activity and the exact medication list. Your summary is often the only thing traveling with the patient.
The pharmacist doing medication reconciliation reads it as a list of deltas. Give her only a final list and she cannot tell whether the missing lisinopril was a decision or an omission.
And sometimes an attorney, three years later, out loud, to a jury. That reader wants to know whether you documented your reasoning or only your actions. See charting and legal risk for the rest of that argument.
You cannot write five documents. Write one that is complete, organized by problem, explicit about changes, and honest about what is unresolved.
The sections, one by one
The date of admission.
A numbered list of all the acute diagnoses the patient had while hospitalized. This is not their past medical history. List them in order of importance and severity.
List all consultants and their field.
List all major procedures the patient had during the stay. You do not have to list minor things like a urinalysis. We are talking CT scans, stress tests, and the like.
Tells the story of the patient from admission until now: changes over time, lab studies, procedures, results, and what each consultant felt needed to be changed or adjusted.
Include the specifics and the changes each consultant made. Give cardiology their own paragraph, neurology their own paragraph, and so on for each service.
Their last set of vitals, with a date and time.
Their final medication list and dosages. Do not simply write "see medication reconciliation sheet." If you want to refer to it, then also list here which medications were discontinued, which were added, and which dosages changed.
For example: "Please see the medication reconciliation list. Notable changes include his furosemide dose being changed to 40 mg twice daily and..."
What type of diet would you like them on?
Can they resume full activity, or do you want them to avoid lifting heavy objects?
If you want a BMP in three days, give them the script and write it in here.
Follow up with the primary care physician in 1 week, cardiology in 2 weeks. List the names and phone numbers of the physicians you want them to see.
Home? Back to a skilled nursing facility? An assisted living center? Rehab? Be specific.
For a longer template with discharge condition, patient instructions and disposition broken out separately, use the discharge summary note version.
The hospital course The part everyone writes badly
This is where discharge summaries go to die. The standard failure is a day by day recitation copied forward out of the progress notes.
The day by day version, which nobody can use
"Day 1: Patient admitted, started on IV Lasix. Day 2: Continued IV Lasix, cardiology consulted. Day 3: Creatinine up, Lasix held. Potassium repleted. Day 4: Echo done. Lasix restarted. Day 5: Patient improved, discharged home."
Read that at two in the morning. Why did the creatinine rise? Did it come back down? What did the echo show? A day by day course records what happened to the calendar, not to the patient.
Organize by problem instead, in the same order as your diagnosis list, the way you organize an assessment and plan. Each problem gets one paragraph: what it was, what you did, how it responded, where it stands today. Numbers, not adjectives. "Improved" is not data. "Weight down from 94.5 to 88.2 kg" is.
The same admission, organized by problem
1. Acute decompensated heart failure, HFrEF. Three days of orthopnea and leg edema, admission weight 94.5 kg, BNP 1840. Diuresed with furosemide 80 mg IV every 12 hours, plus metolazone 5 mg once on day 2 for a plateau in output. Net negative 6.8 liters, discharge weight 88.2 kg, lungs clear. Echocardiogram showed an ejection fraction of 30 percent, down from 45 percent a year ago.
2. New onset atrial fibrillation with rapid ventricular response. Telemetry the night of admission showed rates of 130 to 145. Rate controlled with metoprolol tartrate 25 mg PO twice daily, converted to metoprolol succinate 50 mg daily, resting rates in the 70s. CHA2DS2-VASc 4, so anticoagulation was started. No cardioversion. Holter placed at discharge to define his burden.
3. Acute kidney injury. Creatinine rose from a baseline of 1.1 to 1.9 on day 3, felt to be cardiorenal and diuretic related. Lisinopril was held and the diuretic reduced for 24 hours. Creatinine 1.4 at discharge, with lisinopril restarted at a reduced 5 mg daily.
Same hospitalization, same number of words, and now the next physician can act on it. Note the last sentence of each paragraph. Every problem ends where it stands today. That is the sentence readers are hunting for.
Medication changes Where the harm happens
Post discharge harm is overwhelmingly a medication problem. Somebody keeps taking the drug you stopped, stops the drug you started, or takes the home dose and the new dose of the same agent under two different names. A final list prevents none of that, because it does not say what is different.
Write the deltas in four buckets: started, stopped, held, dose changed, and give every entry a reason. The reason is what keeps the change from being quietly undone by the next person who sees the patient.
Medication changes this admission
Started: Apixaban 5 mg PO twice daily for new atrial fibrillation, CHA2DS2-VASc 4. Metoprolol succinate 50 mg PO daily for rate control. Spironolactone 12.5 mg PO daily for HFrEF. Empagliflozin 10 mg PO daily for HFrEF.
Stopped: Diltiazem CD 240 mg daily, for the reduced ejection fraction, replaced by the beta blocker above. Naproxen, permanently, since it drives fluid retention and renal injury. Acetaminophen instead.
Held: Nothing at discharge. Lisinopril was held on days 3 and 4 for acute kidney injury and restarted, see below.
Dose changed: Furosemide increased from 40 mg to 80 mg PO daily. Lisinopril decreased from 20 mg to 5 mg PO daily after the acute kidney injury, to be retitrated up by cardiology as the creatinine allows.
Name the drug and dose, not the class. Say why, in six words or fewer. And if you cut a dose because of an event, say so, or the outpatient physician will assume it was a typo and put it back.
Document the indication, agent, dose, intended duration and who is monitoring it. Dose reduction criteria and interactions vary by agent, by renal function and by local policy, so check the current label and your protocol before you write the number. A patient going home on a new anticoagulant with no named owner is the most dangerous discharge in general medicine.
Pending results at discharge
Almost every patient leaves with something unfinalized. A culture with no growth at 48 hours. A biopsy nobody has read. A preliminary imaging read that says "correlate clinically." These come back to a service that has already forgotten the patient exists.
Make the summary the handoff. Give pending results their own heading, never a clause buried in the course, and give every item a named human being who owns it.
Pending results at the time of discharge
- Blood cultures times two from 3/4, no growth at 96 hours, final pending. Responsible: inpatient team, Dr. Alvarez, pager 4471, who will call the PCP and the patient if positive.
- 48 hour Holter placed 3/8. Report routed to cardiology, Dr. Whitfield, reviewed at the 3/22 visit. Responsible: cardiology.
- Preliminary CT chest 3/5 noted a 6 mm right upper lobe nodule, final read pending. Responsible: Dr. Nguyen at the 3/15 visit, with nodule follow up per the radiology recommendation.
Item, date, owner, and when it gets looked at. Then close the loop for real: tell the patient these tests are pending and write in the instructions to ask about them at follow up. A patient who knows to ask is the best safety net you have.
Follow up
"Follow up with PCP in 1 to 2 weeks" is not a plan, it is a wish. A usable follow up line has four parts.
- The name. An actual physician, not a service. If they have no primary care physician, say so and say what was arranged instead.
- The interval. Tied to the risk, not to habit. A heart failure patient on a new diuretic dose is seen in seven days, not a month.
- The phone number. Written in the document. The patient loses the paper and the facility will not know who to call.
- The purpose. What that specific visit must accomplish.
Follow up written so it can be executed
Dr. Nguyen, primary care, 3/15/2026 at 10:00, (555) 214-8890, appointment already confirmed with the office. Purpose: review the metabolic panel drawn that morning, assess volume status and the weight trend, review the final blood culture and CT read, and confirm he is taking the apixaban.
Dr. Whitfield, cardiology, 3/22/2026 at 14:30, (555) 214-9002. Purpose: review the Holter, retitrate lisinopril if the creatinine is stable, and consider a rhythm control strategy.
A complete discharge summary example
Here is the whole thing on one patient, start to finish.
68 year old man, heart failure and new atrial fibrillation
Patient: Robert M., 68 year old man. Admitted: 3/4/2026. Discharged: 3/8/2026. Attending: Dr. Alvarez, internal medicine.
Discharge diagnoses: 1. Acute decompensated heart failure with reduced ejection fraction, EF 30 percent. 2. New onset atrial fibrillation with rapid ventricular response. 3. Acute kidney injury, resolving. 4. Hypokalemia, corrected. 5. Essential hypertension. 6. Type 2 diabetes mellitus.
Consultants: Cardiology, Dr. Whitfield. Nutrition, for sodium and fluid education.
Studies: Transthoracic echocardiogram 3/5. Chest radiograph 3/4. CT chest without contrast 3/5. 48 hour Holter placed 3/8. No invasive procedures.
Hospital course, 1. Acute decompensated heart failure. Three days of orthopnea, two to four pillows, with leg swelling after a week of dietary indiscretion. Admission weight 94.5 kg against a dry weight of 88 kg, BNP 1840, chest radiograph with vascular congestion and small effusions. Troponin mildly elevated at 0.08 and flat on serial testing, felt to be demand ischemia. Diuresed with furosemide 80 mg IV every 12 hours, plus metolazone 5 mg once on 3/6, for a net negative 6.8 liters. Echocardiogram showed an EF of 30 percent, down from 45 percent in 2025, with moderate mitral regurgitation and a dilated left atrium. Transitioned to furosemide 80 mg PO daily on 3/7 and held his weight overnight. Lungs clear at discharge with trace ankle edema.
2. New onset atrial fibrillation. Telemetry the night of admission showed previously undocumented atrial fibrillation at 130 to 145. TSH normal. Rate controlled with metoprolol tartrate 25 mg PO twice daily, converted to metoprolol succinate 50 mg daily on 3/7, resting rates 70 to 80. Cardiology recommended rate control and anticoagulation for a CHA2DS2-VASc of 4. Started on apixaban 5 mg PO twice daily, full dose, as he meets none of the dose reduction criteria. He remains in atrial fibrillation at discharge, rate controlled, with a Holter placed to define the burden before deciding about rhythm control.
3. Acute kidney injury. Creatinine rose from a baseline of 1.1 to 1.9 on 3/6. Lisinopril was held and the diuretic reduced for 24 hours. Creatinine fell to 1.4 by 3/8, with lisinopril restarted 3/7 at 5 mg daily, down from a home dose of 20 mg.
4. Hypokalemia. Potassium fell to 3.1 on 3/6 with diuresis and was repleted orally. Spironolactone added 3/7 for heart failure benefit and potassium stability. Discharge potassium 4.1, magnesium 2.0.
Cardiology recommendations: Rate control rather than rhythm control for now, full dose apixaban, and guideline directed therapy continued with empagliflozin 10 mg daily added 3/7. Retitrate lisinopril up as renal function allows. Repeat echocardiogram in three months. Office follow up in two weeks with the Holter results.
Vitals on discharge, 3/8/2026 at 08:15: BP 118/68, HR 74 and irregular, RR 16, temperature 36.8 C, saturation 96 percent on room air, weight 88.2 kg.
Labs on discharge, 3/8: Sodium 137, potassium 4.1, bicarbonate 27, BUN 32, creatinine 1.4, glucose 142. Hemoglobin 12.8, platelets 214.
Medication changes: Started apixaban 5 mg PO twice daily, metoprolol succinate 50 mg PO daily, spironolactone 12.5 mg PO daily, empagliflozin 10 mg PO daily. Stopped diltiazem CD 240 mg daily for the reduced ejection fraction, naproxen permanently, and aspirin given the new anticoagulant and no known coronary disease. Furosemide increased from 40 mg to 80 mg PO daily, lisinopril decreased from 20 mg to 5 mg PO daily. Metformin 1000 mg twice daily and atorvastatin 40 mg nightly unchanged.
Pending at discharge: Blood cultures times two from 3/4, no growth at 96 hours, final pending, owned by the inpatient team. Final CT chest read, 6 mm right upper lobe nodule on the preliminary, owned by Dr. Nguyen on 3/15. Holter read by cardiology, reviewed 3/22.
Diet: 2 gram sodium, 2 liter daily fluid restriction, carbohydrate modified. Activity: As tolerated. Daily weights on the same scale each morning after voiding. Call the office for a gain of 3 pounds in a day or 5 pounds in a week.
Follow up tests: Basic metabolic panel on 3/15 before the office visit, script given to the patient. No INR monitoring on apixaban.
Follow up appointments: Dr. Nguyen, primary care, 3/15/2026 at 10:00, (555) 214-8890. Dr. Whitfield, cardiology, 3/22/2026 at 14:30, (555) 214-9002.
Discharge to: Home with his wife, independent in all activities of daily living. Home health nursing for two visits in the first week for weight, blood pressure and medication teaching. Condition: Stable and improved, ambulating on room air without dyspnea.
Discharge summaries and readmission risk
The most important thing about a discharge summary is not its contents. It is its arrival time. A perfect summary dictated eleven days out is useless, because the follow up visit already happened and the primary care physician sat through it with nothing but the patient's memory of what you changed.
That is the mechanism behind most preventable bounce backs. Not a missed diagnosis. A patient who kept taking the diltiazem you stopped, or never got the potassium checked, because the person in clinic did not know either was supposed to happen.
Finish it the day the patient leaves, and make sure it reaches the receiving physician instead of sitting in a chart nobody outside your system can open. If they are going to a facility, the summary and the medication list travel in the packet with them.
Remember that this is how the primary care physician who takes care of this patient finds out what you did to them while they were hospitalized, what they need to follow up on, and how their treatment and medications were changed. Write it for that physician, not for the chart.
Before you sign
Run this list. It takes ninety seconds and catches almost everything.
- Do the discharge diagnoses match the problems in the hospital course, in the same order?
- Is the course organized by problem, and does every problem end with where it stands today?
- Are medication changes listed as started, stopped, held and dose changed, each with a reason?
- If they go home on an anticoagulant, are the indication, dose, duration and monitoring owner written down?
- Does every pending result have a named person and a date it gets reviewed?
- Does every appointment have a name, a date, a phone number and a purpose?
- Are discharge vitals and labs in there, with a date and time?
- Are diet, activity and disposition specific enough for a nurse who has never met this patient?
- Did you copy anything forward that is no longer true?
- Would you be comfortable if this were read aloud in a deposition three years from now?
Ten yeses. Sign it and send it today.
Next: the longer discharge summary note template →
This is not medical advice. The drugs and doses in the example below belong to a made up patient. Your formulary, your institution's policy and current references govern what you actually write.
