What kind of AI techniques are used in healthcare?

Healthcare organizations are using a variety of AI techniques to extract insights from structured and unstructured data. Natural language processing (NLP) and machine learning (ML) can be used to identify important pieces of information within medical text for a variety of clinical and administrative applications. Deep learning (DL) can be used to process large datasets and surface common trends across patients populations. Provider organizations are also starting to employ machine vision techniques to analyze imaging for diagnostic purposes.

What’s the difference between natural language processing (NLP) and machine learning?

Natural language processing (NLP) is a set of techniques that help computers process language the way humans do. This includes extracting meaning from written text or speech, and generating readable, grammatically correct original content. Machine learning (ML) is a type of artificial intelligence that focuses on automated knowledge acquisition. The goal of machine learning is to use algorithms to learn, recognize patterns, and solve problems without explicitly programmed instructions.

What kinds of healthcare data can AI analyze?

Artificial intelligence techniques have been largely used to analyze structured data—that is, data that is clearly labeled and tagged in a system so that a computer program can make sense of its composition. However, the vast majority of healthcare data is unstructured—such as the free-form text that makes up doctor’s notes in a clinical record, or scanned documents containing words or images that get uploaded to an EHR. Unstructured data makes up 80% of all healthcare data. While it’s more difficult to analyze with AI, it ultimately yields greater value since it contains more detail and nuance than a typical structured dataset.

Healthcare AI Platform for Risk Adjustment and Quality

The Apixio Platform

More than 1.2 billion clinical documents are generated each year in the U.S., but there is very little analysis of that unstructured information. The Apixio Platform can mine textual data and combine its generated insights with available structured data to craft computable individual health profiles, or phenotypes. We analyze our assembled phenotypes in real-time using a flexible rules engine. This automates the execution of clinical guidelines, quality and risk measures, payment or reimbursement policies, and other operational and administrative rules, to support critical healthcare activities.

Step 1: Data Acquisition

We use a secure extraction process to capture free-text documents, scanned charts, and facsimiles, in addition to clinical structured data, claims and other administrative data, and more. These data are then handled in our Data Loader in preparation for our data and analytics pipelines.

Step 2: Data Processing

Our data pipelines process ingested information, run validation checks, and then assemble the data into Apixio Patient Object Models (APOMs) in preparation for our downstream analytic steps. During this stage, we use our optical character recognition (OCR) pipeline to translate the data in image files into machine-readable text.

Step 3: Data Analysis

Signals are generated from analysis of APOMs, which are then used by proprietary data algorithms and predictive models to generate insights. Processed data stored in APOMs include patient diseases, medications, procedures, encounters, biometric values, other data, and derived information from prior analyses.

Step 4: Phenotype Assembly

Information from AI algorithms are sorted and analyzed for discrepancies and then a robust patient phenotype is created. This phenotype is a rich, longitudinal picture of individual healthcare, which is the basis for our solutions.

Step 5: Application Use

Our solutions use data from phenotypes to execute specific rules to automate clinical, administrative, or operational activities.

View Next Step

Testimonial

The efficiencies Apixio brings to our reviews are just incredible. Their machine learning technology leads you to the right information you need for coding.

Andrew Bloschichak MD, MBA

Senior Medical Director, Government Programs - Capital Blue Cross

Testimonial

The coding team was wonderful to work with, and incorporated our specific coding guidelines when working on our project. When they noticed a trend in what we were rejecting, they shared a Coding Clinic we were unaware of for clarification on how to code the scenarios.

Josh Weisbrod

VP of Risk Adjustment

Testimonial

By using Apixio, we’re improving our auditing bandwidth and enhancing the ability of our coders to focus on other chart audits and other projects that we couldn’t do before.

JoAnn Hayden

Scripps Health Plan Services

Case Study