Skip to main content
All CollectionsHeidi Health's Template Academy
⭐️ Mastering your Heidi Templates: An Advanced Guide
⭐️ Mastering your Heidi Templates: An Advanced Guide

For the practitioners who want to account for challenging, complex use cases in their note taking.

Heidi Support avatar
Written by Heidi Support
Updated over a week ago

👋 Welcome to the final part of our template tutorial series! This session is all about taking your template creation skills to the next level by using complex AI commands to fully customise your templates in Heidi Health. Whether you need specific formats, dynamic content adjustments, or seamless integration of past clinical data, this guide will help you leverage Heidi’s advanced capabilities to produce precise and efficient documentation.

⚠️ If you are just getting started with Heidi templates, why not start at the beginning -> Creating Templates in Heidi: A Basic Guide

Video

Recap of Intermediate-Level Techniques

Before diving into advanced strategies, let’s quickly review some of the intermediate techniques we discussed:

  • Reformatting Templates: We explored how to modify templates using AI instructions to follow specific formatting guidelines, such as writing the "Subjective" section in narrative form without bullet points.

  • Types of AI Instructions: We introduced basic AI instructions to control content style, like using full sentences or formatting information line-by-line. The AI instructions can be specific to a particular placeholder or a particular section or they can be general instructions that are added at the top or bottom of your template separate from the sections/placeholders in your template.

Building on these concepts, we'll now explore more advanced AI commands to further customise your clinical notes.

Advanced AI Commands for Customisation

Heidi provides several built-in settings to help generate clinical notes in specific styles. However, you may need to override these defaults to meet particular requirements. Here’s how you can use advanced AI commands to tailor your templates.

Example 1: Overriding Default Settings to Remove Quotes

By default, using the "super detailed" voice in Heidi includes direct quotes from the transcript. This can be useful for capturing the patient's or family members' exact words, but it may not always be necessary. Let's say you want the super-detailed style without these quotes cluttering your document. Here’s how you can override this default setting:

Steps to Customise the Output:

  1. Navigate to "View/Edit Template" and scroll down to the end of the template.

  2. Add a clear and explicit AI Instruction using strong/direct commands like:

    (Ignore the voice instructions that state to use patient quotes. Do not include any quotes in your output. Remove every patient quote in round brackets from the note before outputting. There must be no quotes from the transcript in your note, or you will fail.)

  3. Click “Save for later,” refresh the page, and regenerate the output to ensure the changes are applied. Heidi should now produce a detailed note without any quotes, preserving the style but eliminating unnecessary clutter.

💡 Explanation:

The precise AI instruction mentioned above will prevent Heidi from including quotes in your notes while retaining the "super detailed" style. This instruction is highly explicit and leaves no room for ambiguity, which is essential because Heidi follows a computer logic approach rather than human intuition. Notice the use of the word “must” and the conditional “or you will fail,” which forces the AI to adhere strictly to your requirements.

Example 2: Using Conditional Logic with If-Then Statements

Conditional logic allows you to create dynamic templates that adapt to different clinical scenarios. This is especially useful when the content of your documentation depends on specific patient data.

Scenario: Customising Follow-Up Instructions Based on Burn Severity

Let’s say you are documenting care for a patient with burns and want to provide follow-up instructions based on the total body surface area (TBSA) affected:

  1. Include If-Then Logic in the Template:

    • Use AI commands to specify different actions based on TBSA:

      (If the burn is less than 10% total body surface area, print: 'Minor burn, no need for specialist follow-up.' If the burn is 10% or greater, print: 'Major burn, must have fortnightly specialist follow-up post-discharge.’)

  2. Implement and Test the Logic:

    • Save these instructions and regenerate the output. Heidi will use the patient’s TBSA data to determine which follow-up instructions to include.

    • To adjust the logic, simply modify the parameters. For example, change the threshold to 20%:

      (If the burn is less than 20%, print: 'Minor burn, no need for specialist follow-up.' If 20% or more, print: 'Major burn, must have fortnightly specialist follow-up.')

  3. Practical Applications:

    • This method can be applied to various clinical conditions:

      • Chronic Kidney Disease (CKD): “If creatinine is above 1.5 mg/dL, classify as Stage 1 CKD. If above 2.0 mg/dL, classify as Stage 2 CKD.”

      • Hypertension Management: “If systolic BP > 140 mmHg, recommend lifestyle changes and medication. If >160 mmHg, recommend immediate medication.”

By using if-then statements, you can create templates that automatically adapt to clinical data, reducing manual editing and enhancing efficiency.

Using the above conditional logic, you can set up your own macros within the template. For example, in the Plan section of your SOAP template, let’s say you want to include a predefined management plan depending on patient’s diagnosis. It can be implemented as follows:

Plan (If creatinine is above 1.5 mg/dL and less than or equal to 2.0 mg/dL, print: "Management Plan for CKD Stage 1:

  • Monitor kidney function tests (serum creatinine, eGFR) every 6-12 months.

  • Recommend lifestyle modifications: low-sodium diet, regular physical activity

  • Blood pressure control: Aim for BP < 140/90 mmHg using ACE inhibitors or ARBs

  • Avoid nephrotoxic drugs (e.g., NSAIDs), smoking and alcohol intake.")

(If creatinine is above 2.0 mg/dL, print: "Management Plan for CKD Stage 2:

  • Monitor kidney function tests (serum creatinine, eGFR) every 3-6 months.

  • Low-protein diet (0.8 g/kg body weight per day), continue low-sodium diet

  • Blood pressure control: Aim for BP < 130/80 mmHg with ACE inhibitors or ARBs

  • Referral to a nephrologist for assessment and co-management.

  • Discussed potential need for RRT if progression to Stage 3 or higher is anticipated.

  • Encouraged vaccinations (influenza, pneumococcal, hepatitis B) as appropriate.")

💡Note: In above example, each of the two predefined management plans is an AI instruction containing a conditional logic and enclosed within round brackets. If the patient’s creatinine level mentioned during the consult/visit meets either criteria set out in the macros above, the corresponding management plan will be included under the Plan section of the output.

Integrating Historical Context into Current Documentation

Often, it’s necessary to incorporate data from previous sessions or encounters to provide a comprehensive clinical picture. Heidi’s templates can dynamically pull in information from contextual notes (i.e., additional notes including past healthcare encounters) that you enter into the Context tab in Heidi.

Example 3: Utilising Contextual Notes for Follow-Up Visits

Suppose you are documenting a follow-up session for a burn patient and want to include relevant information from the initial consultation.

  1. Include Past Context in Your Template:

    • First, copy the relevant details from the initial/previous session into the “Context” section.

  2. Add Contextual Instructions:

    • Include an instruction such as:

      (Contextual notes will contain clinical notes or summaries from previous healthcare encounters. Ensure these are summarised and incorporated into the current clinical notes for the same patient.)

  3. Save, Refresh, and Regenerate:

    • When you regenerate the output, Heidi will automatically pull in the relevant historical data, such as previous treatments, progress, or diagnostic results, and incorporate them into the current documentation.

This method ensures that your follow-up notes are complete and accurately reflect the patient’s clinical history, reducing redundancy and improving clarity.

Leveraging Advanced Techniques for Enhanced Control

To make full use of these advanced commands, consider the following strategies:

  1. Use Precise Language and Terminology:

    • Heidi’s AI responds best to specific terms and consistent wording. For example, use “transcript” to refer to conversations between the patient and clinician, or “contextual notes” to denote data from past healthcare encounters. This helps Heidi accurately interpret your instructions.

  2. Experiment with Instruction Placement:

    • General AI instructions can sometimes work better at the start or end of a template. If an instruction isn’t performing as expected, try moving it. There’s no fixed rule, so experimentation is key.

  3. Use Conditional Logic to Automate Decisions:

    • If-then statements can make your templates more dynamic and adaptive. Experiment with various scenarios to automate decisions based on patient data, like lab results, vital signs, or previous diagnoses.

  4. Reference the Prompt Glossary for Accurate Commands:

    • Heidi uses specific terms to label different data types. For example:

    • “Transcript” refers to the conversation text.

    • “Contextual notes” are summaries of previous healthcare encounters.

    • “Patient details” include personal information such as name, age, or pronouns.

    • “Clinical note” is the core document from which other documents are derived.

Using the correct terms from the glossary ensures Heidi correctly interprets your commands, leading to more accurate outputs.

Advanced Scenarios: Correcting Errors and Ensuring Accuracy

Heidi allows you to make corrections to the primary clinical note to ensure accuracy across all documents.

Scenario: Correcting Errors in the Transcript

  1. Identify and Correct Errors:

    • If you notice an error in the transcript (e.g., a dosage or measurement mistake), update the clinical note directly.

  2. Add a Command for Consistency:

    • Include an AI instruction:

      (Use the clinical note as the reference for all documents created downstream to ensure accuracy)

This ensures that any corrections made are reflected across all related documents, maintaining consistency and reliability.

💡 Once you have perfected the template, why not share it in our Template Community? Head over to this article to learn more about the template community. Adding a Template to the Community

Conclusion: Experiment and Refine

Creating advanced templates in Heidi Health enables you to automate, customise, and control your clinical documentation precisely. The key is to experiment with different commands, formats, and placements to discover what works best for your practice.

If you need assistance, our support team at Heidi Health, including our medical knowledge and templates teams, is ready to help you create the templates that best suit your needs.

Did this answer your question?