Artificial Intelligence (AI) Tools in Primary Care

Last Updated: July 15, 2024

No Results Found 0/0

The opportunities for AI tools to transform primary care and reduce clinician burnout are immense. Check back often for information about high-value AI tools as they emerge. 

See the AI Primer for basic information about design, capabilities, and limitations of emerging AI-powered tools with potential to benefit primary care practitioners.

Topics under development

  • AI chatbots for back-office support  
  • AI-supported clinical decision support systems   

Key messages

When adopting AI tools into clinical practice, practitioners should: 

Prioritize use of applications designed for healthcare

These tools will be more accurate and perform better than those not tailored for use in medicine. See the AI Primer to learn how design impacts performance of AI-powered tools.

AI scribes New

See the AI Primer for basic information about design, capabilities, and limitations of the models that underpin AI-powered tools. 

Jump to:

Value of AI medical scribes in clinical workflow

While data on the effectiveness of AI tools in healthcare is limited, early research suggests that AI medical scribes could significantly decrease after-hours documentation time and increase time for patient visits during work hours (Government of Ontario, April 2024). Results from a 2023 preliminary pilot conducted by OntarioMD (OMD) and supported by the Ontario Medical Association (OMA) showed that use of an AI medical scribe product: 

  • Reduced cognitive load and administrative burden 
  • Increased interaction time with patients and enhanced engagement 
  • Improved accuracy of documentation details during patient visits

In 2024, the Ontario Ministry of Health and Ontario Health funded an AI scribe evaluation study by OMD, the eHealth Centre of Excellence (eCE), and the Women’s College Hospital Institute for Health System Solutions and Virtual Care (WIHV). Including patients, clinicians, and AI scribe vendors, the study aimed to help clinicians and patients adopt AI scribes efficiently, enhancing practice without adding burden. Detailed results are expected for release later in 2024 (OMD, 2024; eCE, 2024). 

Examples of AI medical scribe outputs 

Depending on the product, AI medical scribes generate valuable outputs for clinical workflow, including: 

  • SOAP notes for EHRs 
  • Referral letters and request for consult 
  • Insurance documentation

How AI scribes work

While there will be differences between specific products, most AI scribes will use Automatic Speech Recognition (ASR) to process human speech into readable text and employ other models, such as Large Language Models (LLM) and Natural Language Processing (NLP), to interpret and correct transcriptions and generate an output.

For definitions of common AI terms, see the AI Primer

How design impacts performance of AI scribes

Hear conversation
Hear conversation

Medical scribes are optimized to operate in a busy or noisy environment, whereas generalized AI tools are susceptible to suboptimal environmental conditions or background noise. 

Identify people and terms
Identify people and terms

Medical scribes are trained to recognize specialized healthcare lexicon, including non-English languages and accents, while generalized AI tools may struggle with terms, accents, dialects, or ways of speaking not part of their training data. 

Follow conversation
Follow conversation

Medical scribes are trained to interpret the conversational structure of clinic visits, but generalized AI tools may struggle with fragmented or non-linear conversations. 

Optimize performance
Optimize performance

Medical scribes prioritize reducing errors that impact clinical accuracy, whereas generalized AI tools treat all errors equally. 

Gauging accuracy of AI scribes

Reliable, up-to-date information about accuracy rates for specific AI medical scribes is scarce. Often accuracy rate information is included within promotional materials produced by the product developers. However, lack of transparency makes it difficult to assess validity of advertised performance.

At this time, the best way to assess the appropriateness of a scribe for clinical use is to: 

  • Use a well-known AI medical scribe advertised specifically for use in clinical documentation.  
  • Solicit information on accuracy from colleagues who are using scribes.   
  • Try the scribe in practice and do your own assessment of accuracy.  
    • How often do you have to correct documents that are created?  
    • Are corrections more often generic parts of conversation, or are there errors in critical aspects of the conversation about health? 

Effective use: practical guidance from clinicians

About the AI Learning Centre New

The CEP’s AI Learning Centre is a resource for clinicians who want to better understand how AI will impact their work and build confidence in the effective and ethical use of AI to support clinical practice. 

Building off the success of the CEP’s COVID-19 Resource Centre and our proven rapid knowledge translation approach, the AI Learning Centre will prioritize collaboration with external partners and experts to develop responsive, iterative guidance for primary care clinicians while reducing duplication and noise in the system. 

Aims

The aims of the AI Learning Centre are to: 

  • Provide trusted, practical information to help Ontario’s primary care clinicians make informed decisions about using AI in clinical practice. 
  • Promote complementary resources and partner with other trusted organizations to align messaging and reduce noise in the system. 
  • Monitor the space where AI meets primary care in Ontario, keeping clinicians updated with critical information and insights as they develop. 
  • Fill information gaps, demystify concepts, and debunk misconceptions where possible. 

Why

A major AI revolution is here. Its effects are already profound and will continue to shape our world. Current foundational AI models are being tested to solve critical problems in every aspect of our professional and personal lives. 

The opportunities for AI tools to transform primary care and reduce clinician burnout are immense. Early research suggests that AI medical scribes could significantly decrease after-hours documentation time and increase time for patient visits during work hours (Government of Ontario, April 2024). Tools for all other areas of primary care provision are available or in development, ranging from clinical workflow, diagnosis, management, treatment, and beyond.  

How can we distinguish between the performance and hype of emerging technologies? The sudden need to assess the utility, efficacy, appropriateness, and ethics of using AI tools in specific contexts can feel overwhelming.  

We don’t need to become computer scientists to understand the implications. We do need to become literate in this new field, to distinguish between truth and hype and critically evaluate products and services. AI literacy is essential for making adoption decisions and feeling confident in those decisions. 

Share your thoughts

We want to hear from you

Are you a practicing primary care clinician?