Large Language Models (LLMs) and Generative Artificial Intelligence (AI) models, such as OpenAI's ChatGPT, have caused a stir online and sparked debate about the potential of AI replacing or facilitating humans. LLMs and Generative AI models existed even before 2022, but ChatGPT and Stable Diffusion caught the world by storm. Text creation is one of the dimensions of Generative AI models that is now gaining prominence.
Generative AI models have a broad range of applications and uses, enabling businesses to automate intelligence and create a comprehensive knowledge base that spans multiple domains. These models can aid in accelerating innovation in AI development across sectors. Healthcare is set to be one such sector that will be significantly impacted.
Care provider challenges that threaten patient experience
Electronic medical records or EMRs (also known as Electronic Health Records or EHRs) were developed to raise the standard of care and care coordination. Clinical workflow, physician and patient communication, billing process, and other areas have improved by deploying EMRs.
However, government mandates and regulations regarding EMRs have made them considerably more complicated over the last ten years. It causes additional stress to medical care providers when it comes to documentation since they may not have enough time during patient care. Unsurprisingly, unfamiliarity with system usage adds a degree of difficulty to their work. As a result, patient experience takes a hit.
According to a study by Tech Target, there is a high correlation between provider stress caused by using health IT-related products such as EMRs, and provider burnout, with 40% of their time spent on EMRs.
Generative AI and LLMs in EMR can revive the care provider experience
Traditionally, care providers and care managers spend a lot of time documenting reports, plans, and summaries as well as structuring the data for better readability. Generative AI models can benefit care providers by summarising clinical data into patient records, generating detailed treatment notes and care summaries, interpreting lab results, developing tailored patient education materials in the patient’s preferred language, and more. These models can also help care managers create comprehensive diet plans, care plans, and wellness plans based on the patient's demographics, medical problems, and medications.
Large Language Models (LLMs) and Generative AI models provide the output in natural language, making them suitable for developing reports based on data analytics and generating insights. If EMRs are coupled with such LLMs, they have a great chance of rejuvenating providers' experience.
OpenEMR and LLM integration
OpenEMR is a popular free open-source-based EMR with a strong market and community presence. Customised OpenEMR can be integrated with Generative AI APIs to produce provider documents such as generating diet plans, care plans, wellness plans, lab result analysis, patient education materials, and more. The model's response to the pre-defined prompts and the output generated can be astonishing.
For instance, such a Generative AI platform can create a wellness plan which can be integrated into patient records within just a fraction of a second. It can also generate a first-level lab result analysis making it easier for doctors to complete their analysis and reporting. Providers only need to examine the generated plans and reports and finalise them as the model has already generated 80% of the analysis.
A recent survey of data scientists published by Forbes Innovation found that Generative AI has the biggest spike in adoption among all AI techniques most commonly in use today in the enterprise, with a 20% difference between respondents currently implementing generative AI vs those anticipating implementation in 2023, according to a survey by Hubspot.
That said, with these technological adoptions, care professionals will face fewer obstacles and challenges, allowing them to go above and beyond to improve the patient-provider relationship.
Fine-tuning LLMs for healthcare
Most of all the Large Language Models are generic pre-trained models and may not be tailored specifically for healthcare use cases. To address this, the model must be fine-tuned with a domain-specific data set and evaluated for performance. Organisations must have a plan of action to fine-tune the model for their needs.
Being compliant with HIPAA
All the Large Language Models used in the healthcare industry must adhere to the HIPAA (Health Insurance Portability and Accountability Act) regulations. It is essential that these models are hosted on a secure and compliant IT infrastructure, such as HIPAA-compliant cloud storage, virtual private clouds, or on-premises private cloud. Furthermore, it is important to program the models so that all personal health information in the prompt is automatically de-identified in the request and re-identified in the output, thus reducing the chance of any breach. Businesses must also make sure that these models comply with data-sharing agreements and patients’ consent.
Generative AI, like Large Language Models (LLMs), has already kickstarted a revolution, although it is still in the nascent stage. However, there is no doubt that healthcare will harness the potential of AI language models to improve patient care and population health. As AI technology continues to evolve, it is important to keep up with the changing landscape and make it an integral part of success.
(The article is authored by Devi G and Sourav Ghosh)
Devi is a thorough professional with over 16 years of experience across various verticals in the Healthcare industry. She is a seasoned Digital Transformation Consultant with the proven ability to translate transformation objectives into measurable outcomes. She has made significant contributions to Healthcare consulting, Product Development, and Market Research. Devi is experienced in Provider & Payer solutions, Care Management, and Population Health with an emphasis on analytics, product & process reengineering, process automation, FHIR interoperability, Business Intelligence reporting, and modernization. She is certified in PMP, PBA, and CSM, and holds a Master’s degree in Business Administration.
Sourav advises organizations on their operations strategy, assists in improving profitability, efficiency of business processes and in executing business transformation through calibration of operating model and technology. In his current role with Infosys BPM, Sourav oversees Digital Solution Design and Delivery across a number of industries – Banking, Insurance, Healthcare, Retail and Telecom.
(Articles under 'Fortune India Exchange' are either advertorials or advertisements. Fortune India's edit team or journalists are not involved in writing or producing these pieces.)