Feature | Radiology Business | October 17, 2019 | By April Wilson

Combining Data Science, Machine Learning and Frontline Expertise

Using big data and AI to improve imaging workflows and the revenue cycle

An illustration of radiology department analytics data showing GE Healthcare’s business analytics software.

An illustration of radiology department analytics data showing GE Healthcare’s business analytics software.

April Wilson

April Wilson

According to IBM, the world creates 2.5 quintillion bytes of data daily. A large chunk of this data is healthcare information. Medical data continues to expand whether from genomic testing and large imaging studies, to billions of payment transactions. All of this data represents a strong match for advanced systems analysis.

Bill Gates previously said, “A breakthrough in machine learning would be worth 10 Microsofts.” Both the front-end of imaging workflows and the back-end of healthcare revenue cycle management (RCM) stand to reap the rewards of big data analytics and artificial intelligence (AI).

Imaging Workflow Ecosystem Challenges

Imaging clinicians must carefully balance competing time and resource demands. Common challenges include:

   • Increasing study volume per reading radiologist

   • Decreasing misdiagnoses

   • Dealing with large increases in secondary study volumes

Healthcare Revenue Cycle Challenges

Today, health systems are also examining how to tackle an expanding amount of imaging data contained in electronic medical records and billing systems. Data issues related to RCM include:

   • Missing or inaccurate charges

   • Data consolidation across multiple payment vendors

   • Overburdened financial support service staffs

   • Manual data entry errors

   • Disconnected silos of care and information

   • Data transfer timing delays

 

Thankfully, machine learning presents a way to determine and analyze data patterns, which in turn can enhance imaging workflows and the healthcare revenue cycle. Hospitals already use artificial intelligence applications and big data analysis for insurance pre-certifications, denial prediction and ICD-10 billing code verification.

Using AI to Solve Imaging Workflow Challenges

The last challenge in implementing a successful enterprise imaging solution, after leaping connectivity and various interoperability hurdles, is adding intelligence and machine learning applications.

Companies are developing AI intelligent assistants using machine vision, deep learning, computer-aided diagnosis (CAD) and other advanced proprietary algorithms, on the clinician-facing front end, and for imaging workflow and storage needs. AI applications span across multiple disease states from cardiovascular disease monitoring to lung and prostate cancer detection.

For example, studies show that false negative diagnosis rates are at 30 percent and false positive diagnoses are at 2 percent.1 AI can help proactively prioritize studies that need immediate attention, as well as go back and suggest prior studies that may require re-examination.

“Advances in AI imaging tools, and the workflow engines that power them, enable clinicians to make better diagnoses and provide efficacious treatments across a wide array of disease states,” said Beau Jones, COO of DataFirst, developer of the SilverBack workflow engine.

Big Data Can Help Solve Healthcare Revenue Leakage Issues

It is true that health systems have processes to pinpoint revenue leakage, but many are not efficient. For example, manual audit reviews are cumbersome, time-consuming and can even be inconsistent because of data silos. And, rules-based systems, such as bill scrubbers, often can only find pre-identified errors.

Healthcare systems and imaging centers typically lose 1 to 5 percent of net revenue due to leakage and this will likely grow with ICD-10, the Accountable Care Act (ACA) and more out-of-pocket payments. To combat revenue leakage, a closer look must be given to analytics-based systems using big data and machine learning.

Billing data (such as diagnostics attributes, procedure attributes, billing code attributes, charge code attributes and hospital history) can be synthesized to build a database. Present cases can be aligned with past cases, so any variances can be examined.

Tackling Vendor Management

When thinking about using big data, back-end processes, such as billing, cannot be ignored. Most healthcare organizations engage in a complicated receivables process involving multiple vendors. These vendors work with healthcare organizations on boutique processes ranging from payer denials to patient collections.

Big data can help motivate employees, enhance the patient financial experience and enrich a hospital’s financial health. Data-driven decision models can streamline revenue cycle processes and financial operations to improve patient communications and profits.

Vendor management of a variety of providers is a challenge because each as its own methodology, interface and data types. How do healthcare organizations ensure that their vendors are identifying incorrect statements and charging the right rates? The answer again is big data.

Pulling data from each vendor (instead of utilizing vendor spreadsheets and reports) offers a tremendous benefit allowing hospitals to comb through information about vendors and realize trends. This includes vendor performance data, such as screening rates, cycle times, rejection reason summaries and in-house application summaries.

This provides an opportunity for health systems to provide vendors with valuable feedback that can lead to corrective actions for improved patient satisfaction ratings. Data science and machine learning in combination with frontline expertise helps receivables operations run smoother and even more profitably.

Shifting to Value-based Imaging

Recent advances in front-end imaging workflows, enterprise imaging engines and back-end revenue cycle process all stand to support the American College of Radiology’s (ACR) Imaging 3.0 Initiative by delivering systems that enable AI applications to work across modalities, optimizing practice management and patient care.  itn

 

Reference:

1.    www.siemens.com/press/pool/de/events/2017/healthineers/2017-11-rsna/white-paper-medical-imaging-in-the-age-of-artificial-intelligence.pdf

 

April Wilson is the vice president of marketing and analytics for RevSpring. The company predicts payment outcomes so they can be improved, transforming complex financial workflows into intuitive financial pathways.

 

Related content:

Data Driven Workflows on the Rise as the User Changes

Using Big Data and AI to Improve Imaging Workflows and the Revenue Cycle

AI in Medicine: On the Way to Growth

Related Content

Sponsored Content | Videos | Artificial Intelligence | February 21, 2020
In Artificial Intelligence at RSNA 2019, ITN Contributing Editor Greg Freiherr offers an overview of artificial intel
An axial CT image obtained without intravenous contrast in a 36‐year‐old male (Panel A) shows bilateral ground‐glass opacities in the upper lobes with a rounded morphology (arrows). #coronavirus #nCoV2019 #2019nCoV #COVID19

An axial CT image obtained without intravenous contrast in a 36‐year‐old male (Panel A) shows bilateral ground‐glass opacities in the upper lobes with a rounded morphology (arrows). Image courtesy of Radiology Online.

News | Computed Tomography (CT) | February 20, 2020
February 20, 2020 — In new research
Altamont’s zero-footprint solution, CaptureWare, allows Mach7’s Enterprise Imaging Platform (EIP) to ingest more DICOM and/or non-DICOM data from various sources in a facility
News | PACS | February 20, 2020
February 20, 2020 — Mach7 announced its partnership with Altamont
Two magnetic resonance imaging (MRI) findings — joint capsule edema and thickness at the axillary recess, specifically — proved useful in predicting stiff shoulder in patients with rotator cuff tears, according to an ahead-of-print article in the May issue of the American Journal of Roentgenology (AJR)

A: Oblique coronal fat-suppressed T2-weighted MR image shows normal hypointense joint capsule at axillary recess (arrow). Note full-thickness tear of supraspinatus tendon (arrowheads) B: Oblique sagittal proton density MR image shows preserved subcoracoid fat triangle (asterisk). Image courtesy of the American Journal of Roentgenology (AJR)

News | Magnetic Resonance Imaging (MRI) | February 20, 2020
February 20, 2020 — Two ma...
Chest CT imaging of patient. #coronavirus #nCoV2019 #2019nCoV #COVID19

Examples of typical chest CT findings compatible with COVID-19 pneumonia in patients with epidemiological and clinical presentation suspicious for COVID-19 infection. This image is part of the original research, Sensitivity of Chest CT for COVID-19: Comparison to RT-PCR, published Feb. 19, 2020, in Radiology Online.

News | Computed Tomography (CT) | February 19, 2020
February 19, 2020 — In new research
Sponsored Content | Videos | Enterprise Imaging | February 19, 2020
Bill Lacy, vice president, Medical Informatics at FUJIFILM Medic...
Recognized as the “Pulitzer Prize of the business press,” the Jesse H. Neal Award finalists are selected for exhibiting journalistic enterprise, service to the industry and editorial craftsmanship
News | Radiology Business | February 19, 2020
February 19, 2020 — Connectiv, a division of The Software and Information Industry Association (SIIA), has announced