Advancing IDNs’ Comprehensive Care Goals With Clinical NLP

55

Photo credit: Depositphotos

By Calum Yacoubian, Associate Director NLP Healthcare Strategy, Linguamatics an IQVIA company

With the shift to value-based care, providers are increasingly focused on holistic and preventative care that optimizes patient health and decreases the cost of unmonitored disease progression. The emphasis on long-term monitoring and management of patients has fueled industry consolidation and the growth of Integrated Delivery Networks (IDNs) and similar organizations that seek to manage patients across the continuum of care. 

The IDN market is expected to grow at a compound annual growth rate of 10.1% until 2027 as providers work to deliver more effective, efficient, and quality healthcare at a lower cost. Because clinicians work within the same network, they theoretically have access to a more complete view of a patient’s data, enabling them to deliver comprehensive, patient-centered care. 

Yet even within an IDN, clinicians may struggle to obtain a 360 view of a patient’s health because care is often delivered by multiple specialists across different clinics that may not share a single electronic health record. An additional challenge is that approximately 80% of healthcare data is trapped within unstructured sources, such as nurse notes, discharge summaries, radiology reports, pathology, etc. 

However, using AI technologies such as natural language processing (NLP), organizations can gain rapid and effective access to the knowledge buried in such documents to maximize patient benefits, leverage clinical analytics, and advance value-based care objectives. With NLP, for example, organizations can automate the capture of particular insights and terms within clinical documentation, reducing the time required for time-consuming manual searches. In addition, such tools can also normalize the surfaced data to make it useful and ready for downstream analysis. 

To better understand how clinical NLP can help IDNs advance comprehensive care goals, consider the following use cases. 

Social determinants of health (SDoH) identification

Social determinants of health (SDoH) such as education, income level, access to healthcare and other factors are known to impact an individual’s health and health outcomes. SDoH can provide a wealth of information about non-clinical factors impacting a patient’s overall well-being, but critical details are often trapped as unstructured text within clinical notes such as nursing or admission notes, and are therefore unstandardized, difficult to access and thus not used.  

To ensure positive outcomes, IDNs must consider the SDoH of their patient population. With AI-based tools such as NLP, providers can extract specific SDoH details that can be combined with structured data to identify at-risk patients, create a more complete 360-view of patients, and take early action to mitigate disease complications. Take for example social isolation, which may be noted at nursing admission “Mr. Henry, this 73-year-old man, is recently widowed.” This seemingly innocuous piece of information can actually tell us that the patient may be at increased risk of missed appointments, poor medication compliance and depression. By using NLP to surface these social determinants of health – practices can be put in place to reach out to those more likely to need additional support to prevent unmonitored disease progression – improving patient outcomes, and reducing overall costs associated with worsened disease.

Assessing co-morbidity status

Patients with multiple chronic conditions account for a disproportionate amount of total healthcare costs. In fact, 50% of all medical costs are related to care for the sickest 5% of the population. 

To control total healthcare costs, IDNs must identify patients that have earlier stages of their disease or have multiple comorbid conditions. By targeting these individuals, organizations can take proactive preventative measures to keep this population healthier. 

Traditionally providers have relied on billing codes to identify patients with particular medical conditions. However, to understand a patient’s true health status, clinicians need more than billing codes from a visit encounter; they also require subtle details about a patient’s health status to identify true disease status and likelihood of progression. For example, if a patient’s A1C levels, which are typically in lab reports, are gradually increasing over time, this could indicate poor diabetic control. Similarly, if a patient has periodically complained of shortness of breath and fatigue – which are often noted in the physician narrative – this could indicate early coronary heart disease. 

With NLP tools, IDNs can uncover hidden health details that might be easily overlooked if relying solely on billing information. By surfacing a more complete health picture, clinicians can identify disease earlier and take action to slow disease progress, improve health outcomes, and reduce the economic toll of advanced disease. 

Connecting patients with clinical trials

Enterprises can also enhance healthcare outcomes by using clinical NLP to connect patients with clinical trials or to identify emerging therapeutic treatments. Identifying patients who are eligible for clinical trials is time consuming and challenging – with protocols containing complex eligibility criteria – making it hard to pinpoint the right patient for the right trial. The majority of the disease specific information that needs to be identified for enrollment in clinical trials exists only in unstructured data. With NLP, eligible patients can be identified, and offered these innovative treatments, which in areas like oncology represent gold standard care. 

Other areas of clinical research such as observational studies can also make use of NLP, where the real world data (non—identifiable to a particular patient) that IDNs house in their data lakes or data warehouses can be used to generate real world evidence that can shape treatment guidelines and pathways.   

Driving better care with technology

As IDNs seek to improve outcomes across their patient population, NLP technologies can help organizations advance their clinical and financial objectives by giving clinicians rapid insights into the health of their patients. With the right tools, IDNs are empowered to reduce time-consuming manual search efforts, identify at-risk patients sooner, and connect patients with resources that drive optimal health outcomes.  

LEAVE A REPLY

Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.