X
April 29, 2021

Prosperdtx taps data science to bring personalized healthcare to your home

By: Sasha Banks-Louie | Brand Journalist

Share

Social distancing mandates of the past year only ramped up the growing interest in delivering healthcare online. And for New Jersey-based startup Prosperdtx, delivering personalized remote healthcare requires two of its core services: Pulling disparate data into one health record and running machine learning analysis on that data, using cloud computing.

Doctoring from a distance

As more clinicians look to care for their patients from remote locations instead of onsite, having all the right information to help prescribe patient-specific treatments has been a perplexing problem to solve.

“Outside of health co-ops and healthcare networks, none of the systems talk to each other,” says Prosperdtx CTO Zachary Wintrob. “The pharmacy doesn’t see what the hospital sees. The primary physician doesn’t see what the cardiologist sees. And because no one shares any information, all of this patient data gets spread out all over the place.”

Integrating multiple sources of healthcare data into what it calls a single “longitudinal patient health record” is a big part of what Prosperdtx has been doing since its founding in 2020. With a newly launched digital therapy platform that runs on Oracle Cloud Infrastructure (OCI), the company not only helps capture data from doctors, pharmacies, wearables, and labs, but also helps analyze it and provide clinicians and patients with recommendations based on available treatment options and a patient’s potential health risks.

“The vast majority of routine health checks can be done virtually, from anywhere, through our platform,” Wintrob says. While patients still need to visit hospitals for emergency services and surgeries, having a single digital record of multiple sources of a person’s healthcare data, accessible by doctors and patients, makes it more practical to do everyday monitoring, such as checking glucose levels or managing medicine dosages remotely. “It also frees up the whole healthcare system to put its main focus on critical care,” he says.

Wintrob discussed his cloud initiatives during an episode of the Built and Deployed video series, which features technology conversations with software architects about OCI.

 

Data science as the secret sauce

To run this new digital therapy platform, Wintrob’s team uses the Python Notebook in OCI Data Science, runs queries in an Oracle Autonomous Data Warehouse, and then structures its analyses using a causal inference model. “That’s where the secret sauce is made,” Wintrob says. Prosperdtx incorporates domain knowledge from published literature, genomic vendors, and electronic medical records (EMRs), and then combines all that with physician input, treatment histories, and decision logic. Finally, the application quickly assesses the validity of various assumptions and provides recommendations to clinicians and their patients.

Patients can log into their EMR system remotely. If the healthcare provider offers Prosperdtx, they can sign up and start pulling in data from their medical records and outside sources.

“After they sync their wearable devices, link to their ancestral DNA profile, and select from a list of current and former healthcare providers, we can then structure all of that data into a longitudinal patient record,” Wintrob says.

After the patient connects all their health data, the Prosperdtx platform pulls it in to Oracle Autonomous Data Warehouse, where it’s structured into a common data model used for EMR data. Even wearable data or ancestral DNA records that don’t comply with medical industry interoperability standards can still be pulled into the Autonomous Data Warehouse and integrated with the patient’s record.

“The combination of OCI Data Science and Oracle Autonomous Data Warehouse gives us the variables we need to really drill down and provide insight at a patient-specific level,” Wintrob says. “This is the true implementation of personalized medicine.”

In addition to using OCI Data Science’s built-in Python tools, such as NumPy for numerical routines and Pandas for data visualization, Wintrob’s team is using Oracle Applications Express (APEX) to help convert its platform to a cloud native application. “The low-code environment on Oracle’s web interface really empowers us,” Wintrob says. “As a small startup with just a handful of people, getting up and running without a Herculean effort or tons of budget is huge.” Wintrob’s team started its cloud ops as the pandemic started. “We had to quickly get off-premises and stand up an environment that could be completely contained in the cloud,” he says.

Wintrob evaluated other cloud vendors, but “the lift for implementing the kind of work we do was too high. It required a much bigger development team than we have,” he says. Beyond Oracle’s low-code, high-performance cloud environment, Wintrob notes that being able to easily provision cloud resources has been key.

“With Oracle, there’s very little to think about in terms of what kind of virtual machine I need or how much RAM or processing cores to provision,” Wintrob says. “I can even provision a single core processor with minimal specs, set Autonomous Data Warehouse to autoscale, and I’m ready to go. If our workloads demand more compute, memory, or storage, I can easily set a cap, and OCI automatically provisions those resources with no input from me.”

References

Brand Journalist

A regenerative hay farmer and writer, Sasha Banks-Louie is a brand journalist at Oracle, covering Oracle Cloud Infrastructure.

More about Sasha Banks-Louie
This is a syndicated post, view the original post here

Share