The Power of AI: Transforming Healthcare for Providers, Payers, and Biopharma
Written by: Auren Weinberg M.D., M.B.A., Scott Stratton, and Saima Qasim
Artificial intelligence (AI), machine learning (ML), and natural language processing (NLP) are increasingly becoming integral components of the healthcare landscape. Across the healthcare industry, AI has already demonstrated remarkable achievements. Deep learning algorithms have achieved accuracy rates comparable to human experts in diagnosing certain diseases from medical images. Natural Language Processing algorithms have transformed the way medical records are analyzed, enabling researchers to extract valuable insights and identify patterns for research and population health management. Furthermore, AI-powered chatbots have shown great promise in improving patient engagement and adherence to treatment plans.
At Veradigm, we recognized the transformative potential of these technologies, applied them in key products, and are continuing to incorporate them into our current solutions and those in development.
In this article, we provide a high-level overview of key AI-related concepts and share Veradigm’s notable AI-achievements thus far our vision for the future, and the significance this technology holds for providers, payers, biopharma, and the patients they serve.
What are some of the most common AI-related terms?
To understand this transformation, it’s important to first define these terms:
- Artificial Intelligence (AI) refers to the simulation of human intelligence in machines that are programmed to think like humans and mimic their actions.
- Generative AI is a subfield of AI that focuses on the creation of new content. It’s a type of AI that takes in data and tries to generate similar data. This can include creating text, images, sound, or any other type of output.
- Machine Learning (ML) is a subset of AI and refers to the concept that a machine can learn and improve from experience without being explicitly programmed.
- Natural Language Processing (NLP) is a branch of AI that focuses on the interaction between computers and humans using natural language. The ultimate objective of NLP is to read, decipher, understand, and make sense of human language in a valuable way.
- Optical Character Recognition (OCR) is another form of AI that involves converting documents into searchable and editable data. The purpose of OCR is to digitize printed texts so they can be edited, searched, stored more compactly, as well as to allow machine processes like data mining, text-to-speech, machine translation and cognitive computing.
How is AI transforming the healthcare landscape?
These technologies are playing a significant role in reshaping the provider and payer landscape. AI is being used to predict patient outcomes, enhance diagnostic accuracy, optimize workflows, and personalize patient care.
ML, which is a subtype of AI, is helping in the analysis of vast amounts of data to identify patterns and make predictions, while NLP is being utilized to extract meaningful information from unstructured data, such as clinical notes.
When referring to AI, people may be including ML and sometimes even just computer programming. Understanding these differences and how they relate to healthcare is important when considering their impact. Computer programming, like in some forms of clinical decision support, can be used to help support healthcare providers in diagnosing and treating conditions. For example, if a patient has three recorded blood pressures that are elevated and hypertension has not been diagnosed, the system can be programmed to highlight this for the provider. It can then ask whether hypertension should be diagnosed and ask whether the provider wants to prescribe lifestyle changes and/or medication.
ML, on the other hand, can involve reviewing large data sets and drawing inferences from patterns. For example, a pharmaceutical company may want to use ML to review medical registries which contain vast amounts of data to find which patients are likely to have the best blood pressure control on which of their medications. While a diabetic patient may get better blood pressure control on one medication, a patient with poor kidney function may benefit more from a different one. Even better, ML could find a combination of blood pressure medications that are most likely to prevent heart attacks in patients with a family history of early heart attack.
AI involves computers emulating human thought and planning a future action based on a previous result. Using the blood pressure medication example, a pharmaceutical company may want to find the best chemical structure for their next medication. The AI would have to analyze many potential factors including chemical stability, body receptor targets, side effect mechanisms, body clearance pathways, clinical trials and patient outcomes from similar medications, and potential cost of production. Moreover, the AI may suggest other potential medications for other conditions as a result of the exercise.
Veradigm’s ongoing commitment to AI
AI is creating a paradigm shift in healthcare, and Veradigm is proud to be part of this transformation. We have embraced AI and its subsets in various aspects of our offerings.
Veradigm Network EHR Data
Veradigm applies artificial intelligence (AI) to electronic health record (EHR) data to create a diverse, research-ready real-world dataset. Veradigm Network EHR Data is one of the largest real-world databases and is derived directly from community-based physician EHRs, making it representative of a wide, nationally distributed patient population. Via the Veradigm Network, Veradigm uses multiple EHR data sources to populate this database and has access to both structured and unstructured de-identified data through direct access to clinicians’ EHRs. Unstructured data includes clinical notes, attachments, images, and more. A sizable amount of important text still resides in images like faxes.
Veradigm Data Enrichment Service
Via Natural Language Processing (NLP), our data enrichment service exemplifies our commitment to AI. This service utilizes our analytic tools to extract and de-identify patient data from unstructured clinical notes available in our ambulatory patient EHRs and registries. The service mines structured facts and generates new insights, enhancing our ability to answer challenging research questions, optimize clinical outcomes, and advance patient care.
In 2021, researchers from Veradigm used NLP and ML capabilities to create a framework to extract and report social risk factors from ambulatory EHR data, which was published at the American Medical Informatics Association (AMIA)’s 2021 Symposium. This research resulted in an open-source Python dictionary to facilitate the reporting and extraction of social risk factors.
Additionally, Veradigm’s researchers worked with a biopharma client to analyze unstructured qualitative provider notes recorded in the EHR during visits with patients with atopic dermatitis (AD). Through ML and NLP capabilities, they demonstrated that these encounters primarily focused on patients’ symptoms and symptom relief. However, these encounters rarely documented the burden of AD on daily functioning and quality of life. This study showed that combining defined data from EHR structured fields with NLP-extracted information from provider notes captured in unstructured fields could potentially broaden the understanding of AD impact and management.
Veradigm eChart Coder
Veradigm’s eChart Coder uses Optical Character Recognition (OCR) technology to extract text from them, and then applies machine learning (ML) and natural language processing (NLP) to organize and extract meaning from the unstructured text. For those using Veradigm coding services, an inference engine based on human-curated machine learning (ML) identifies diagnoses yet to be coded and can show the human reviewer each piece of evidence supporting that inference – something many ML applications do not offer.
As these health information professionals interact with the tool, adding and removing bits of evidence, the engine collects both the changes and the rationale stated by the coder, which it incorporates into the next learning cycle. This markedly improves both speed and precision of decisions, enabling a more complete and insightful view of the patient’s care status and journey.
Dynamic Intervention Planning
There has been much (overdue) attention recently about health equity and diversity, and in socioeconomic and demographic factors that can impede a patient’s ability to seek and access care, and which are often referred to as Social Determinants of Health (SDOH). From its early days, Veradigm’s Dynamic Intervention Planning analytics considered such factors to highlight patients needing intervention to close care gaps and realize opportunities.
The future of AI and Veradigm
As Veradigm continues to invest in AI and its subsets, our goal remains clear: to empower healthcare stakeholders with advanced, AI-powered tools to deliver better patient outcomes and improve the overall healthcare system.
We are excited about the potential of AI in healthcare and are committed to keeping payers, providers, and biopharma stakeholders informed and prepared for the future of AI-driven healthcare solutions. We believe that our investment in AI today will drive the healthcare innovations of tomorrow.
Through collaborations, research initiatives, and our unwavering commitment to innovation, we are well-positioned to harness the power of AI and machine learning to develop groundbreaking solutions that will empower providers and payers, ultimately improving patient care, operational efficiency, and health outcomes. Together, let’s transform healthcare, insightfully, with AI, machine learning, and natural language processing at the forefront.