ARTICLE | AI TECHNOLOGY
ARTICLE | AI TECHNOLOGY
Steven Lee is co-founder and CEO of ianacare, and a seasoned entrepreneur who has scaled companies from the garage stage to IPO. After watching his family confront his grandfather’s Parkinson’s diagnosis, he became passionate about using technology to make a meaningful social impact and address gaps in caregiver support.
Key Takeaways:
In the post-pandemic world, we’re realizing the importance of providing care in the home and relying more on community support. This becomes even more important with the silver tsunami, where our 65+ population in the US doubles from 2000 to 2040. More than ever, we must understand that non-medical wraparound care and community resources that enable care in the home are crucial for patient well-being and outcomes.
Family caregivers are a critical component of that community support. For caregivers, finding resources such as home help and local community-based organizations to support their loved one’s care needs and ensure high-quality care is a daunting task. There are numerous challenges, including the difficulty of finding the right resources quickly, especially for caregivers who are already juggling multiple responsibilities. This is why our caregiver navigators have historically spent the most time curating appropriate resources for our caregivers.
This program is the first of its kind and marks a shift toward recognizing the influence and impact informal/family caregivers can have on the health & medical outcomes of patients in their care.
Unlike past and existing CMS programs which have historically been aimed at reducing the cost of care, the intent of the GUIDE program is not to create a shared savings or evaluate reduction in total cost of care. It is purely focused on prioritizing and financing caregiver support.
The development of the GUIDE model underscores CMS’s realization that family caregivers play a concrete and crucial role in shaping health outcomes. By offering enhanced support to family caregivers, the program aims to alleviate some of the burdens of caring for patients with Alzheimer’s, resulting in better care for patients.
Lifting these burdens improves the physical and emotional health of caregivers, which has translated to improved health outcomes. Patients whose caregivers are depressed, burned out, or fatigued experience a 73% increase in ED utilization & $1,937 higher medical costs.
Today, when we can think of something we need, and in an instant, have it delivered to our doorstep, why and how have we not found a way to translate this technology to the moments that matter the most? To the moments we try our best to provide our loved ones with the care they need and with the support we, as caregivers, deserve?
While there has been some progress in this area, overall innovation has been slow. With the emergence of generative AI technology, our team at ianacare is making significant strides in leveraging this technology to provide accurate and reliable information quickly, both to aid our navigators and to provide on-demand access to our caregivers.
We recognize the revolutionary potential of using AI in care management and coordination, but it’s essential to approach any use cases with careful consideration and understanding of the current limitations of AI. We must be honest about what is working well and what needs improvement.
With all of this in mind, our team has been hard at work fine-tuning Large Language Models (LLMs) to understand the nuances of care situations.
It’s important to acknowledge that we’re not on the brink of achieving Artificial General Intelligence (AGI) just yet. While the widespread use of LLMs like ChatGPT may give the impression that AGI is within reach, these models serve as foundational building blocks rather than fully realized AGI. These models are generally good at grasping concepts from language, but they are not perfect, and need a lot of tuning to make them work well for a specific domain. We see this come to life with “hallucinations” or mistakes and errors AI often makes by unknowingly feeding a user incorrect or nonsensical information. And when we’re applying AI to areas like healthcare & care coordination, these types of errors are unacceptable.
Our AI team at ianacare has been working hard to make LLMs better at understanding really specific situations, like those involving the care of a loved one with a specific condition. We started with a generic state-of-the-art foundation model on the left that we soon discovered wasn’t very good at discerning relevant vs. irrelevant resources based on care situations.
Each week, we see the trends on the right where relevant vs. irrelevant resources separate more and more. Through training, we see improved pattern recognition driving better decisions and giving responses and resources that are highly accurate, building trust with users.
Whether it’s connecting someone with a meal delivery service, helping them access financial assistance, or simply providing information and guidance when it’s needed most, this type of technology will completely revolutionize the way caregivers search for and receive support. Our navigators are already leveraging this system to become more efficient in resource curation. We can’t wait to make this directly available to caregivers later this year.
In a world where technology can feel distant and impersonal, it’s refreshing to harness it in a way that has a deeply personal impact on someone’s life and well-being. We’re committed to innovation through compassion, building technology as a force for good that can allow us to care for one another in times of need.
