This article explores how Apixio leverages Big Data, AI, machine learning and deep learning algorithms to analyze structured and unstructured sources of patient records. It’s mission is to use Artificial Intelligence (AI) to change the way healthcare is measured, care is delivered, and discoveries are made.
Apixio at a Glance
- A private company founded in 2009
- Headquartered in San Mateo, CA
- 27 active products
- $43M in funding (Series D)
Apixio’s Value Proposition
Apixio’s value proposition is to learn how to make sense of clinical information at scale to provide better individual care and gain insights on overall population health by mining Big Data. It pitched the same to stakeholders to gain data access so holistically coded data and unstructured data can be analyzed across healthcare plans and providers to lower costs, provide effective care delivery, and streamline processes.
Apixio transforms disparate data into actionable knowledge that can be leveraged by entities like healthcare systems, clinics, hospitals, medical insurers, the U.S. Medicare program, medical practitioners, and patients in areas of risk adjustment, interoperability, clinical guidance, and quality.
Impediments to leveraging Big Data
While volumes of health data exist, much of it is underused. Few impediments include:
- 80% of medical and clinical information is unstructured. It comes in many different formats and templates and from numerous systems. E.g. physician notes, consulting notes, pathology results, handwritten notes, Medicare records, hospital records, PDF scans, etc.
- Sometimes getting healthcare providers and healthcare insurance plans to share data is hard.
- Data security and legal requirements are stringent with the need to encrypt data in rest and in transit. Apixio demonstrated its security governance and interoperability standards.
Potential Use Cases
- User-friendly interfaces that coders and plan operators can understand.
- Improving quality and frequency of data extraction by enabling direct access to clinical and administrative data from leading EHR platforms such as Epic, Cerner, athenahealth, and Allscripts.
- Mining clinical trial data to determine which treatments are most effective across vast populations takes us a step closer to the “learning healthcare system”.
- It has implemented a cognitive computing platform to reduces manual aspects of workflows, automate key administrative and operational tasks where possible, and streamline processes across the industry. Lesser administrative burden means affordable healthcare costs.
- Rather than traditional risk adjustment done by coders combing through data, Apixio has automated data reading process and the ability to gain anecdotal insights resulting in risk adjustment coding being more productive, accurate, efficient, transparent, and predictive.
- With machine learning and NLP, it provides insights to health plans to manage individual value-based care, creating accurate risk and quality measurement and actionable care decision guidance.
- While individual patient data models are available, Apixio aggregates data for pattern analysis across the patient population.
- Gaps inpatient documentation can be easily identified to provide a clear picture of diseases, treatment, and aid in coordination and management of patient care.
- Presently only 5% of the Medicare data is audited. AI capabilities could help audit 100% of the data, especially important as our population increases.