metamorworks - stock.adobe.com
Independent laboratory Ascend Clinical is using AI to try to improve patient care and provide its customers with the best and most efficient forms of treatment for renal disease.
The tech-oriented lab, with roots in Silicon Valley, performs blood testing and other services for kidney dialysis clinics across the country.
According to the National Center for Biotechnology Information, more than 500,000 people in the U.S. live with end-stage renal disease.
Using a new product released by AI hardware and software vendor SambaNova Systems and AI sentiment analysis vendor DeLorean AI, Ascend is moving toward its goal of using AI technology to make recommendations for providing more efficient renal care.
On March 16, SambaNova and DeLorean AI revealed that Ascend is their first customer to use DeLorean's new Medical Renal Model running on the SambaNova AI platform.
The lab company said it has been using and evaluating the system since February during a proof-of-concept phase -- during which it paid for a software license -- but is working out cost details with the vendors for the long term.
The Medical Renal Model ingests structured and unstructured data from different sources, including test results from labs such as Ascend, health records, data from medical procedures and insurance claims. The AI model uses the data to predict if a patient is high or low risk, and recommends the next best action for healthcare professionals, according to the vendors.
Ascend was first introduced to the AI product during a HIMSS -- the Healthcare Information and Management Systems Society -- health IT conference last year, Ascend CIO Helmut Oehring said.
Ascend Clinical and the AI product
As the first customer of the AI product, Ascend gathered data from patients' blood test results and information from patients' electronic health records, including hospitalizations, medication lists and treatments patients were receiving at dialysis centers.
The lab prepared data to feed into the AI model in December 2021. Then, in February, Ascend performed its first run in the AI model.
Helmut OehringCIO, Ascend Clinical
The model crunches the data to find patterns that reveal best practices for caring for patients. Clinicians are then able to evaluate those recommendations and determine if they truly indicate the best course of action to take, Oehring said.
"This also allows for our ... clinics to look at populations of patients," he continued, referring to the practice of population health management, which involves using health data to improve the overall health of communities. "This definitely feeds into value-based care in healthcare where we're trying to keep people populations as healthy as possible, not just treat any individual patients and help them get healthier."
Using the AI technology from the two vendors will also help healthcare providers predict when patients will move from one stage of renal disease to another, he said.
Up to now, Ascend has only used the AI model to bring all its data together and do an internal review of the recommended best steps for clinicians it works with, Oehring said. The lab has yet to validate whether clinicians will feel comfortable with the recommendations made by the AI model.
"There have been a lot of circumstances where clinicians are used to seeing AI analyze X-rays and visualizations, and seeing the patterns and the nuances," Oehring said. "This is one of the next steps for them to apply the similar kind of capabilities on data points as it relates to test results."
Challenges and costs
One challenge Ascend has encountered is using all the data it has and making sure it's important enough for the AI model to learn and recommend the best course of action, Oehring said.
"Not all data points are as important as some data points," he said. "Making sure there is a comprehensive set of data is key."
Pricing for the AI product is subscription-based, and organizations can use it in the cloud or on premises. Healthcare providers can also use the SambaNova-DeLorean system to track other types of diseases, including diabetes, heart disease and cancer.