Oracle outlined a broader strategy for AI across its healthcare portfolio at the Oracle Health and Life Sciences Summit in Orlando last week – where it announced new capabilities spanning revenue cycle management, clinical research and oncology.
Speaking to Pulse+IT, Oracle Health and Life Sciences executive vice president and general manager Seema Verma said Oracle products could be localised for different markets, with the company aiming to take its new Oracle EHR and other products into markets where it already operates as well as some new ones.
“We already have that with Millennium, but we’ll be with the Oracle EHR and every Oracle product,” Ms Verma said.
“We will bring them to all the different markets that we are in today and also some new ones as well.”
She said organisations would also not necessarily need to use Oracle’s EHR or other applications to adopt parts of its AI strategy.
Oracle’s data platform is designed to bring together information from different systems, while a number of its products are intended to work with different EHRs, revenue cycle and clinical trial systems.
“We want to create that flexibility to be able to meet our customers where they are,” she said.
Ms Verma referenced the UK during her opening keynote at the summit, telling attendees Oracle continued to support Australia and the UK with data interoperability.
The company wanted to move AI beyond standalone tools and embed it across clinical, operational, financial and research workflows, supported by connected data.
“If we simply layer AI onto fragmented systems, disconnected data, manual processes and technology architectures built for another era, we risk amplifying the discord rather than creating harmony,” Ms Verma told the summit.
She said healthcare technology could no longer operate solely as a system of record, but needed to bring together information from across organisations and external sources, understand its context and help turn insights into action.
‘Less bureaucracy’
Ms Verma said Oracle’s existing healthcare AI agents were being used to reduce documentation burden and improve coding, with the company now expanding into nursing, discharge planning, care pathways, care coordination and prior authorisation.
In the US, prior authorisation is the process under which insurers approve coverage for certain treatments or services.
Oracle is also looking to apply AI further into administrative processes including billing, claims and contract management.
“The goal is not to automate every step of a broken process,” Ms Verma said during her keynote, noting that Oracle wanted to connect clinical, financial, research, clinical trial and claims data with analytics and AI, to allow routine transactions to be automated while retaining human involvement for decisions requiring judgement.
“So this isn’t faster bureaucracy, it’s less bureaucracy,” she said.
Revenue cycle management is one of the areas where Oracle was aiming to expand that approach.
At the Summit the company announced new AI capabilities across its revenue cycle portfolio, which it said were designed to identify problems earlier in the patient and payment journey as well as to connect information across scheduling, financial clearance, clinical documentation, charge capture, billing and payment.
Other planned capabilities include AI support for prior authorisation, clinical documentation integrity, charge capture, coding and appeals management.
The prior authorisation capability seeks to retrieve payer requirements, gather relevant clinical information and draft submissions, while AI-supported appeals workflows are designed to help identify and organise information needed to challenge denied claims.
The new capabilities are initially planned for healthcare organisations in the US, with Oracle saying they are expected to be released in the coming months.
Asked how far automation of revenue cycle processes could ultimately go, Ms Verma said there was a “tremendous opportunity” to automate much of the work, while retaining human oversight.
“There’s always going to be these outliers that are going to require human judgment,” she told Pulse+IT.
“When we build our AI, it’s always human in the loop. So whether it’s rev cycle or clinical care, we need to have humans that are overseeing the work that the AI is doing.”
Ms Verma said Oracle’s approach spanned from patient registration through to claims and into enterprise financial systems – with clinical information from the EHR available to support processes such as prior authorisation and billing.
She said the company was also looking to demonstrate measurable outcomes from its AI products rather than simply their technical capabilities.
For existing clinical AI agents, Oracle was measuring indicators including time spent in the EHR and coding accuracy, with similar KPIs to be collected for other AI agents.
“When it comes to health care, resources are very scarce and we have to make sure that our customers, when they’re making an investment, that they can see the outcome that they hope to see.”
AI for research
Also featured at the Summit was Oracle’s move into clinical research – where it was aiming to combine large real-world datasets with AI tools intended to reduce some of the manual work involved in designing and analysing studies.
Oracle says Life Sciences Data Intelligence combines real-world data, domain-trained AI and analytics – with tools allowing researchers to use natural language to explore and refine patient cohorts, analyse outcomes and automate multistep research workflows.
The company says its real-world data foundation contains more than 130 million de-identified patient records – with organisations able to combine Oracle data with their own and third-party datasets.
Ms Verma said bringing together longitudinal patient records with other information including claims, genomic and prescription data could reduce the work researchers currently undertake to aggregate information from multiple sources.
“The second thing is they can bring their own data onto the platform, and they can combine the data,” she told Pulse+IT.
“So you can imagine the type of resources and the type of research that you’re able to do when you have all these different types of data.”
She said AI could then be applied to tasks including study design, modelling clinical trial scenarios, developing patient cohorts and investigating potential additional indications for drugs.
“If you think about setting up a study, it would take you months to even figure out the study design,” she said.
“With AI, we can do that very easily and fast.”
Oracle has also introduced an AI clinical trial matching agent designed to identify potentially eligible patients from connected data and match them with appropriate trials.
During her keynote, Ms Verma described the objective as shifting from relying on patients finding trials towards enabling “the trial to find the patient.”
The company’s collaboration with the Buck Institute for Research on Aging was given as one example of how it expects the data and AI infrastructure to be used.
The institute is using Oracle technology as part of the US Advanced Research Projects Agency for Health-funded THRIVE initiative, which is investigating biomarkers associated with ageing and age-related disease.
Oncology AI
Meanwhile, another development featured at the Summit was the extension of the company’s AI-focused EHR strategy into specialist clinical care, with the announcement of the Oracle Health Oncology EHR – a cancer-specific system built on its newer ambulatory EHR and designed to bring together oncology workflows and information from multiple sources.
The system is intended to connect genomic, radiology, pathology, laboratory, pharmacy and other patient information across diagnosis, treatment and follow-up.
Planned capabilities include AI assistance for tumour board preparation and treatment planning, alongside oncology-specific workflows such as chemotherapy and immunotherapy regimen planning and support for nurse navigators.
In her keynote, Ms Verma described the oncology EHR and a new pharmacy system as examples of bringing intelligence closer to clinicians working to detect disease and select treatments.
“These are not isolated product announcements. They are pieces of a much larger transformation across payers, providers, pharma and for all patients.”
Humans in the loop
With the expansion of AI capabilities across the company’s product portfolio, Ms Verma said human oversight and the ability to understand how AI reaches an output would remain important.
Asked by Pulse+IT how researchers could verify and trust evidence produced with AI, she pointed to Oracle’s framework for assessing healthcare AI.
“Every single AI model that we bring to the market in healthcare is thoroughly tested. It’s monitored. It’s transparent. So when the AI is giving conclusions, it’s very clear of where did that conclusion come from and why.”
She said the framework included transparency, traceability and continuing monitoring after deployment, as well as testing for issues including bias.
The company noted its AI Review for Health assesses healthcare AI features before release against standards covering areas including safety, performance, transparency, privacy and human oversight.
The company has also published a responsible AI framework covering human oversight, transparency and traceability, model governance, privacy, cybersecurity, fairness, accountability and auditability.
Ms Verma told the summit healthcare required AI with humans “always in the driver’s seat,” noting that more AI would not necessarily lead to better healthcare without appropriate safeguards and the data and infrastructure to support it.
The author was a guest of Oracle.
Add Comment