Non-Epic Smart Tips

Authors: Papova, A.

Synopsis: Tips for optimizing the report-writing process without using EPIC.

Author Disclosures: None.

Overall Description: This committee has provided several tech tips articles specific to EPIC EMR users, but many neuropsychologists do not use EPIC. Whether you are familiar with the EPIC platform and no longer have access to it or EPIC has never been a part of your clinical writing process, this article aims to provide a few similar tips to those who do not use EPIC.

Text Expansion

“Text expansion” refers to tools that can transform a keyword into a significant block of text, such as a paragraph, clinical note, drop-down menu of pre-written options, or even a fillable questionnaire (EPIC equivalent = SmartPhrases). Non-Epic and non-AI alternatives for text expansion include:

• Text Blaze: https://blaze.today/
• Phrase Express: https://www.phraseexpress.com/
• TextExpander: https://textexpander.com/
• Microsoft Word can manage some text expansion via the creation of AutoCorrect key phrases. It’s also possible to insert drop-down menus, similarly to EPIC simple lists. See https://support.microsoft.com/en-us/office/create-reusable-text-snippets-0bc40cab-f49c-4e06-bcb2-cd43c1674d1b and https://support.microsoft.com/en-us/infopath/insert-a-drop-down-list-box

AI Scribes

AI scribes are another option to significantly shorten the report writing process, specifically intake notes. They can record a clinical interview and automatically draft the information into templates. They can handle multiple speakers and accented speech competently, provided the audio equipment (microphone) is adequate. They can also integrate other documents, such as history forms, and notes can be re-generated as needed if the template changes. It is important to keep in mind the numerous limitations of AI, obtain the patient’s informed consent, and manually proofread all output, but these can save a lot of time once appropriate prompts are created, even though they cannot fully replace the note-taking process.

• NovoNote:
  • Developed for mental health professionals
  • APP, HIPAA, GDPR compliant
  • Link: https://novopsych.com/novonote/
• Berries:
  • Developed for mental health professionals
  • HIPAA, PHIPA, SOC2 compliant
  • Link: https://heyberries.com/

Case Study Example:

Here is an example of a prompt that can be included at the beginning of a clinical interview template in NovoNote. These are general instructions, a lot of which were added iteratively to avoid specific issues, and serve as general guidance for the AI on how to handle your template. Keep in mind that different tools may have different formatting requirements.

(This is a clinical intake interview template for comprehensive neuropsychological assessment. Use language consistent with professional neuropsychologists and appropriate medical terminology. Use American English spelling, such as “color” instead of “colour” and “organize” instead of “organise.” Use Oxford commas. Adhere to template closely. Keep all headings and categorize information into them. Keep line spacing and italics/bolding the same. Make it clear when the patient or collateral informant verbally reported something, so as not to imply it was taken directly from the medical record, by using words like “reportedly” or “the patient reported X”. State it when anyone “denied” a symptom, rather than saying it was “not reported”, but do not state that someone denied symptoms that were not asked about. Use direct quotes when prompted to do so in specific parts of the template, otherwise avoid using direct quotes.)

Limitations: Although many of the tools mentioned above are specifically for professional use and are often HIPAA-compliant, data privacy concerns can never be fully mitigated. AI scribes are now very commonly used in medical settings and can significantly shorten the amount of time spent on documentation to allow more time to be spent on patient care. However, for AI in particular, it is imperative to explain data handling, storage, and limitations to patients and obtain informed consent to proceed. It is also important to be aware of AI limitations (e.g., “hallucinations” or omissions), which require careful proofreading of all output. Additionally, the fast-growing field of AI is riddled with ethical concerns such as equity, bias, environmental impact, etc. Ultimately, all ethical, clinical, and professional responsibility rests on the provider.