How to Write a Clinical Summary

Patient summaries play a crucial role in improving healthcare efficiency and the continuity of care, as they provide an accurate picture of who the patient is and how they have been doing. A good patient summary should be a narrative that synthesizes the information, provides context, and alerts downstream clinicians about any follow-ups needed by the patient.However, there is a great potential for generative AI technology to automate narrative summaries and save doctors time. By leveraging AI, healthcare providers can streamline their workflow, improve continuity of care and provide better patient outcomes.

The average clinic visit is only about 15 minutes. There just isn’t enough time nowadays to dig through a patient’s chart and thoroughly understand their medical history. Then, add on top of that two main issues: (1) the information is most likely incomplete due to interoperability issues, and (2) there is a lot of clinical data to sift through that is not as important. So primary care doctors rely on other doctors’ summary notes: notes such as Progress Notes, Emergency Department Provider Note, and the Discharge Summary. All of these notes are required and regulated by the Joint Commission because of the importance of patient summaries for a proper transition of care. A lot of research studies have shown a patient summary significantly improves patient outcomes. So, given the importance, what are the best strategies to write a good patient summary?

Do not just copy and paste previous information into your notes.

Downstream doctors are not interested in reading the entire MRI results regurgitated again. It’s not helpful. The primary goal of a summary is to synthesize the information. For example with LDL cholesterol levels, your note shouldn’t just aggregate the minimum and maximums outside of the normal ranges; you need to interpret the results and combine them with treatment information such as “in response to elevated LDL levels, a statin was initiated and LDL levels decreased to normal.”

A typical summary will first start with a brief synopsis of the patient’s medical history to date such as their current age, important diagnoses and conditions, pertinent family and social history, and relevant allergies.

Here is an example:

Ms. Jane Doe is a 72 year old female with a past medical history of hypertension (HTN), hyperlipidemia (HLD), migraines, gastroesophageal reflux (GERD), depression, psoriasis, left frontal arteriovenous malformation (AVM), an 8mm wide nicked anterior communicating artery aneurysm, and a fusiform 9mm ectasia of the left anterior cavernous carotid artery.

We can quickly understand that our patient most likely needs something like an angiogram and stent for this visit. And we know our patient has a more complicated clinical profile with comorbidities such as hypertension and high cholesterol.

Stating the patient’s medical diagnoses though is not enough.

We need to know the context of how the patient has managed their conditions. As you may be aware, 10% of the US has Type 2 diabetes. While in its early stage, it’s a disease manageable with regular exercises and a healthy diet, if left untreated, you can end up in the hospital with serious complications as your blood sugar regulation gets out of control. It would be exceedingly useful to know that a patient has not been managing their diabetes appropriately and we could therefore alert them to alter their behaviors to prevent hospitalization. So summarizing a patient’s course of treatment is essential for understanding how the patient has been managing their health: what has worked well for them and what procedures or interventions have not worked so well. It helps us understand the underlying story of why our patients are now here for their visit and more broadly what challenges we are dealing with as a whole.

Lastly, we need to alert downstream clinicians about any remaining follow-ups needed by the patient, such as lab tests or outpatient or specialty clinic visits. Most patients will need ongoing treatment and follow-ups after their visit. Timely follow-ups help improve the likelihood of positive patient outcomes. While your EHR system might be able to alert you in a timely manner of follow-ups, it’s likely the downstream clinicians who might receive the patient and read your summary won’t have the same EHR as you, so it’s essential to put any remaining follow-ups of care at the end of your summary.

For example:

• Patient is to follow-up with neurosurgery for ventricular drain placement.
• He may need a colectomy in the future if bleeding recurs.
• Patient is to obtain basic chemistry panel including BUN and Creatinine before visit with Dr. Smith.
• Enalapril was discontinued during her hospitalization. Patient is advised to follow up with this medication with his PCP.

Generative AI holds great promises to help automate a narrative summary so doctors don’t have to spend 15–30 minutes with this process manually. Especially if a patient has been in the hospital for many months with a complicated course, summarizing all that care can be taxing. A tool using large language models, similar to ChatGPT, would greatly assist doctors. Our company, Abstractive Health, is focused on solving this problem to make medical summarization easier and faster for clinicians. Check out how we are automating the Discharge Summary, ED Provider Note, and Progress Notes, so that physicians can deliver the best care that patients deserve.