top of page

Rise of Generative AI in Healthcare: A Comparison of BARD, LLAMA 2, CHatGPT

In the rapidly evolving landscape of healthcare technology, Generative AI has been making significant strides, transforming the way diagnoses, treatment plans, and predictive analysis are conducted. This article delves into the rise of Generative AI in healthcare, particularly focusing on the comparison of three prominent models: Google's PaLM 2 (Bard), Meta’s LLAMA 2, and OpenAI’s GPT-4 (CHatGPT). 

Comparison of Bard, Llama and Chatgpt

The progression of these models has undeniably transformed the healthcare sector, offering innovative solutions for complex challenges. Each of these models comes with its unique set of features and capabilities, contributing to a more efficient, accurate, and personalized healthcare system. The following sections will provide an intricate overview of each model, their key features, and a comparative analysis focusing on ICD-10-CM and HCC coding. 

OpenAI GPT-3.5/4 (CHatGPT):

A state-of-the-art language model known for its ability to generate human-like text based on the input provided.

Google Bard:

Google's cutting-edge language model designed to generate medical text, aiding healthcare professionals with diagnosis and treatment.

Meta LLAMA 2:

Meta's advanced language model that is specifically designed to handle healthcare-related queries and generate corresponding responses.

"The intricate comparison among these models helps in better understanding the potential and performance of these AI models in the healthcare industry, especially when it comes to ICD-10-CM and HCC coding."

In the subsequent sections, we will delve deeper into the respective key features of CHatGPT, Bard, and LLAMA Chat, offering insights into their capabilities and contributions to the healthcare industry. 

Following this, a comparative analysis will be conducted focusing on their performance and use in ICD-10-CM and HCC coding, providing a comprehensive understanding of their practical implications in medical coding and healthcare informatics.

The biggest challenge in healthcare is not diagnosing diseases, it's managing patients' health between visits.
- Eric Topol


In the burgeoning field of artificial intelligence, three models have stood out for their innovative applications in healthcare: OpenAI’s GPT-3.5/4 (ChatGPT), Google's Bard, and Meta’s LLAMA 2. 


ChatGPT, developed by OpenAI, is a genitive AI model based on the GPT-3.5/4 architecture. GPT-4, the latest version, exhibits exceptional language understanding capabilities due to its transformer-based model, which leverages unsupervised learning to generate human-like text. It's a conversation AI model that's been utilized for a plethora of applications, including drafting emails, writing code, creating articles, and providing tutoring in a wide range of subjects. In the healthcare sector, ChatGPT has demonstrated potential for use in medical transcription and preliminary patient interaction.


Google's Bard, also known as the PaLM 2, is another remarkable contribution to AI in healthcare. Bard is designed to predict clinical events and make sense of unstructured electronic health records (EHRs). This machine learning model can process vast amounts of healthcare data, identify patterns, and predict possible outcomes. Bard's prowess lies in its ability to provide meaningful clinical insights and aid in decision-making processes for healthcare professionals.  


Meta's LLAMA 2 is an AI model developed with the specific intent of tackling healthcare-related tasks. This model focuses on understanding medical language and knowledge extraction from medical text. It's known for its proficiency in ICD-10-CM and HCC coding, which helps in categorizing patient diagnoses and risk adjustment factors in healthcare facilities. LLAMA 2 stands as a testament of AI's potential in the realm of medical coding and language understanding.


In the rapidly evolving landscape of healthcare and Artificial Intelligence (AI), three prominent systems have emerged as leading contenders in the application of generative AI technology: OpenAI's ChatGPT, Google's Bard, and Meta's LLAMA2. Each system boasts a unique set of features that are designed to enhance the efficiency, accuracy, and overall quality of healthcare services. Below, we delve into the key features of each system. 

ChatGPT – Key Features 

OpenAI's ChatGPT is renowned for its advanced language processing capabilities, built on the foundation of the robust GPT-3 model. The system is designed to understand, generate, and engage in human-like text-based conversation, offering significant utility in patient interaction and information extraction. The key features of ChatGPT include: highly nuanced language understanding, ability to generate detailed and coherent responses, advanced dialogue management, impressive contextual understanding, and an extensive knowledge base.

Application ChatGPT in healthcare

This helps in providing a high degree of accuracy in interpreting patient symptoms, their medical history, and other pertinent information during healthcare interactions. Furthermore, its application in healthcare extends to providing health education, aiding diagnosis, and facilitating patient-provider communication.Continuing this technological marvel, the GPT-4 (ChatGPT) has further improved on these aspects and has demonstrated remarkable efficiency in processing and understanding complex medical terminologies.

It's sophisticated algorithms have shown to improve healthcare delivery by assisting in medical decision-making and enhancing patient engagement. The system's contribution to ICD-10-CM and HCC coding has further solidified its standing in the realm of healthcare AI.In comparison to its predecessors, GPT-4 has shown a heightened ability to comprehend and generate contextually relevant responses. Its proficiency in interpreting multi-modal tasks - a combination of textual and image-based tasks, make it an advanced tool in the field of healthcare.

Furthermore, GPT-4's understanding of medical jargon, its ability to predict likely diagnoses based on given symptoms, and its potential to aid in the creation of personalized treatment plans, have revolutionized the way professionals approach healthcare challenges.From sifting through vast amounts of medical literature to providing contextually relevant patient education, its capabilities are multifarious.

Its proficiency in understanding and generating nuanced, natural language responses has marked it as a vital tool in telemedicine and remote patient monitoring systems, thereby reinventing traditional medical consultation processes.


Google Bard, also known as PaLM 2 (Predictive Language Model), is a generative AI developed by Google Health. It is designed specifically for the healthcare industry, focusing on enhancing the understanding and generation of medical language. Here are some of its key features: 

Firstly, Bard uses a bidirectional transformer, enabling it to understand context from both previous and subsequent text for better prediction of words and phrases.

Secondly, it has been trained on a vast medical corpus, making it highly tailored for healthcare applications. It can generate clinically relevant sentences and even full notes, hence facilitating effective communication within the healthcare field.

Lastly, Bard has the ability to generate ICD-10-CM and Hierarchical Condition Category (HCC) codes, which are essential in the billing and insurance processes in healthcare.

OpenAI's GPT-4, commonly referred to as ChatGPT, is another transformative force in the realm of healthcare. It operates on the principle of a unidirectional transformer, inherently different from Bard's bidirectional model.

The considerable size of its training data, including a diverse range of internet text, equips it with a rich understanding of human language. While not specifically healthcare-oriented like Bard, ChatGPT's language versatility allows it to generate coherent and contextually accurate medical text. However, it lacks the specific ability to generate ICD-10-CM and HCC codes that Bard possesses. 

Meta's LLAMA 2 operates on a similar bidirectional transformer model as Bard. This model, enriched with a significant corpus of medical texts, makes LLAMA 2 another potent AI tool specifically designed for healthcare applications. Beyond the generation of relevant medical text, it also has the capability to generate accurate ICD-10-CM and HCC codes, making it a direct competitor to Bard in the field of medical coding and billing. Google's BERT has 340 million parameters


The LLAMA Chat, developed by Meta, is a revolutionary application of generative AI in healthcare. Its key features are particularly tailored towards enhancing the experience of healthcare professionals and improving patient outcomes. 

Natural Language Processing:

LLAMA Chat showcases superior natural language processing capabilities. It can effectively understand and respond to complex medical queries, thereby facilitating better doctor-patient communication.

ICD-10-CM and HCC Coding:

A cardinal feature of the LLAMA Chat is its proficiency in ICD-10-CM and Hierarchical Condition Category (HCC) coding. It can accurately identify, process, and present relevant codes, assisting healthcare providers with accurate diagnosis and treatment planning.

Data Security:

Data privacy and protection is a critical aspect of any AI application in healthcare. LLAMA Chat offers robust security features, ensuring that sensitive patient data is secure and compliant with healthcare regulations.


LLAMA Chat is designed to be scalable, implying that it can expand its services as per the growing needs of the healthcare institution.

The LLAMA Chat has been designed to address the complex needs of the healthcare industry. Its superior features make it a potent tool in the hands of healthcare professionals, aiding them in providing better patient care. 


ICD-10-CM and HCC coding are integral to the healthcare industry, with their use being essential for accurate diagnosis and treatment. A comparison of the proficiency of LLAMA Chat, CHATGPT, and Google BARD in dealing with these codes provides a clear picture of their effectiveness.


When it comes to ICD-10-CM and HCC coding, LLAMA Chat has demonstrated superior proficiency. While CHATGPT and Google BARD also exhibit capabilities in handling these codes, LLAMA Chat's dedicated features offer a more precise and comprehensive solution. The ability of LLAMA Chat to correctly identify and process relevant codes has been lauded by healthcare professionals, marking it as a preferred choice for handling ICD-10-CM and HCC coding.On the other hand, CHATGPT and Google BARD, while capable, are more generalized AI platforms, with a broader focus beyond healthcare coding.

Their effectiveness in this domain, although commendable, is not as specialized as LLAMA Chat's. Therefore, for tasks specifically related to ICD-10-CM and HCC coding, LLAMA Chat maintains a competitive edge. However, it is critical to note that the performance of these AI models is contingent on the specific requirements and nuances of each task at hand. CHATGPT's strengths lie in its advanced natural language processing capabilities, which can prove invaluable in patient interaction contexts.

In comparison, Google BARD's key strength is its powerful and highly accurate search functionality, which could be instrumental in medical research and information retrieval. Ultimately, the choice between these AI platforms depends on the precise needs of the healthcare institution and the tasks they wish to automate or augment. On the other hand, Meta's LLAMA 2 stands out for its ability to understand and generate natural language text, with a particular emphasis on medical dialogue. It also excels in handling complex medical queries, making it a valuable asset in the diagnostic process.

All three models have the potential to dramatically transform healthcare delivery, but their effectiveness can only be ascertained through rigorous testing and iterative implementation in real-world clinical settings.  Furthermore, the comparison of these models in ICD-10-CM and HCC coding exhibits distinct capabilities. In this aspect, LLAMA 2 demonstrates proficiency in understanding complex medical terminologies and coding, offering a significant edge.

However, the comparative performance of CHatGPT and Bard in this area requires extensive examination. It is paramount to remember that these AI models, while promising, are tools meant to augment and not replace the critical role of healthcare professionals.

Generative AI in Emedlogix NLP:

Emedlogix's NLP tool is revolutionizing the medical field by leveraging advanced technologies such as AI, ML, and NLP, combined with a custom rules engine. This tool specializes in extracting ICD-10-CM and HCC codes from medical charts, primarily for Medicare providers. Recognizing the potential of Generative AI, Emedlogix has integrated it into their system. The results have been astounding, with significant improvements in code extraction accuracy and efficiency. The company is excited about the advancements and is gearing up to introduce this innovative feature by the end of August 2023 or early September 2023. This enhancement is expected to further streamline the coding process, ensuring more accurate billing and improved patient care. As the medical industry continues to evolve, tools like Emedlogix's NLP are setting the standard for technological integration and excellence.

AI will transform healthcare from reactive to proactive, from disease-focused to wellness-focused, and from population-based to individualized care.
- Eric Topol
568 views0 comments


Rated 0 out of 5 stars.
No ratings yet

Add a rating
bottom of page