NoHarm modernizes clinical pharmacy and boosts patient safety with Oracle
NoHarm improves the accuracy of its AI algorithms up to 95% on Oracle Cloud Infrastructure for better clinical pharmacy reviews.
“NoHarm uses OCI Compute instances and NVIDIA GPUs to screen over 5 million prescriptions at 70 hospitals, realizing savings of more than $2 million and increasing the volume of prescription evaluations up to 8X.”
In many hospitals, clinical pharmacy reviews are a time-consuming and error-prone process that can lead to medical errors, partly because of the sheer number of medications prescribed every day. Each prescription has to be read and evaluated for the exact drug information, dosage, and frequency of application, among other data. The relationships between prescribed medications and other drugs and procedures must be identified—by a minimal number of clinical pharmacists working in this process.
At Conceição Public Hospital in Brazil, where NoHarm began its development, there are more than 1,000 patients—each getting 10 to 20 medications per day—and only 3 pharmacists, making it impossible to serve patients with quality and efficiency.
NoHarm’s goal is to systemize clinical pharmacies by integrating patient data to validate medical prescriptions. Its aim is to allow providers to make artificial intelligence (AI)-based decisions to prioritize non-standard cases as well as indicate potential risks in drug prescriptions.
The Oracle for Startups partnership empowers us to develop algorithms for patient safety. We want technology to be at the service of healthcare professionals.
Why NoHarm chose Oracle
NoHarm selected Oracle because it could partner with the Oracle for Startups program to develop the algorithms it needed to pursue its work to improve patient outcomes at hospitals. Not only did Oracle give NoHarm access to the computing power it needed, but it also provided the necessary cloud credits to fund further research.
NoHarm required graphics processing units to develop its algorithms with neural networks. NoHarm Care, based on BERT, extracts patient risk information from clinical notes. NoHarm Gehn (Generative Electronic Health Notes), based on GPT, generates discharge summaries.
The organization used Oracle Cloud Infrastructure (OCI) and Oracle Analytics Cloud to develop studies, analyze clinical data, and extract information. Oracle provided NoHarm with more powerful machines than any other cloud provider offered.
Furthermore, more than half of the hospital units assisted by NoHarm belong to Brazil’s public healthcare system, and they do not have the budget to deploy advanced technology or train users. Cost was another reason why NoHarm selected Oracle, because it offered to host NoHarm’s systems in OCI for free.
NoHarm uses Oracle Cloud Infrastructure to screen over 5 million prescriptions at 70 hospitals, realizing savings of more than $2 million and increasing the volume of prescription evaluations up to 8X.
After joining the Oracle for Startups program, NoHarm was able to support hospital pharmacists in validating medical prescriptions by using AI. Oracle for Startups gave NoHarm the necessary infrastructure in OCI to deliver the computing power it needed to develop its resource-heavy No Harm Care and NoHarm Gehn AI models.
Using Oracle Cloud Infrastructure and Oracle Analytics Cloud, the organization’s AI models showed a substantial improvement. Before it partnered with Oracle, these models had 72% accuracy. When NoHarm started deploying its application on Oracle machines, with increased flexibility to test neural network parameters, it arrived at an algorithm that identifies keywords with 95% accuracy.
This exceptional accuracy in AI-powered clinical pharmacy reviews is paramount because the NoHarm application alerts users about duplications, drug interactions, allergies, and cross-reactivity, among other issues. It helps qualify the pharmacist’s evaluation—decreasing the risk of errors, and reducing the time spent on this task by integrating all related information in a single system.
The company regularly receives credits as part of the Oracle for Startups program, which allows it to expand its research and add value to its AI models by automating database routines.
NoHarm’s principal mission statement is focused on patient safety. However, according to the hospitals it serves, there is a cost reduction in the price of drugs as a result of using the system, because the inherent analytics help hospitals source more economical medications.
Today NoHarm works at 45 healthcare facilities and has screened more than 2.3 million prescriptions since deploying its solution on OCI. The organization impacted more than 250,000 patients, and enabled hospitals to realize savings of US$300,000. At the same time, hospitals using NoHarm see a 4X to 8X increase in the volume of clinical pharmacy evaluations.
NoHarm is currently working on developing a new discharge prediction algorithm. This model will act on patient flow predictions, integrated with the management of bed count as well as medical procedures and surgeries.
About the customer
NoHarm, a nonprofit startup, uses AI algorithms to efficiently systemize the prescription review procedures of clinical pharmacies. It operates in 45 hospitals across Brazil, of which 70% are free state clinics run by Brazil’s Sistema Único de Saúde, the world’s largest public healthcare system based on the number of beneficiaries.