ElectrifAi
July 8, 2021

Machine Learning Helps Healthcare Providers Identify Missing Charges

Transparency in healthcare pricing has become a big topic. In fact, effective January 1, 2021, the Centers for Medicare and Medicaid Services (CMS) “promulgated a new rule for hospital pricing that would require disclosure of a wide range of hospital prices.”[1]With so many eyes on healthcare billing, it is crucial to ensure there are no errors.

Missing or wrong charges on a claim are one of the biggest problems healthcare providers face today. A lack of resources can result in reduced employees. Reduced employees mean more work is placed on other employees and, as a result, charges can be missed or incorrectly processed in claims.

Wouldn’t it be great if there was a way to ensure no missed or incorrect charges without having to increase employee headcount? How about not having to manually reprocess every claim that is flagged as having the wrong codes? And then you must consider all the lost revenue for missing charges.

Here are a few examples of descriptions that you may see for missing or wrong required claim information:

  • “Claim/service lacks information or has submission/billing error(s) which is needed for adjudication.
  • Missing/incomplete/invalid procedure code(s).
  • Procedure code billed is not correct/valid for the services billed or date of service billed.”[2]

There is an excellent solution to this problem. Machine learning can help identify wrong or missing charges in a provider’s claim. Using machine learning to find missed charges can be used as a complementary technology for:

  • Rules engines
  • Whether the solution is applied to inpatient or outpatient accounts
  • Hospital or professional fee charges

Charge codes change all the time. Healthcare providers have thousands of codes to indicate a variety of options (e.g., surgery, anesthesiology, room charges, etc.) and it can be difficult to manually keep track of so many codes.

“Hospitals deal with more than 1,300 insurers. Each has different plans and multiple and often unique requirements for hospital bills. Add to that decades of government regulations, which have made a complex billing system even more complex and frustrating for everyone involved. In fact, Medicare rules and regulations alone top more than 130,000 pages, much of which is devoted to submitting bills for payment.”[3]

With machine learning, you don’t have to worry about not providing the right code or missing a charge that usually accompanies a code. ElectrifAi’s pre-built machine learning model, Missing Charges, is trained on each facility’s data to produce the best insights specific for each facility.

Moreover, a second application of the Missing Charges machine learning model can be trained on all facilities within a healthcare system to make use of a larger dataset to improve performance predicting common charges.

What makes this machine learning model so powerful? Some technical highlights include:

  • Semi-supervised machine learning (less manual involvement to save time).
  • Specific data mapping is not required.
  • Advanced pattern-based system continually learns.
  • Predictive analytics identifies complex relationships that are difficult or impossible to capture through a rules-based system.

In environments like healthcare—with changing regulations, contract modifications, and payer alterations—rules always need to be adjusted. These adjustments require manual intervention, and that need for constant updating makes rules engines less effective over time.

The dynamic intelligence of artificial intelligence (Ai) and machine learning combats limitations of traditional revenue cycles and helps to track, analyze, and identify missed charges. ElectrifAi’s machine learning model identifies patterns using machine learning algorithms and predicts with amazing accuracy missed charges.

Furthermore, this machine learning model leverages feedback models to learn from auditor’s responses to make more intelligent predictions by continually learning.

There are many data sources that make this machine learning model extremely effective for healthcare providers, such as:

  • Medical charges
  • Charge data: HCPCS/CPT codes and department charge codes
  • Patient information
  • Demographic information, financial class, admit type, diagnosis codes, procedure codes (ICD10, HCPCS, etc.)
  • Insurance data
  • Patient’s insurance plans
  • Other data
  • Hospital chargemaster

Altogether, the data outputs a score of every possible charge code. High scores mean the charge should be on the claim and lower scores are a recommendation for review.

ElectrifAi’s Missing Charges machine learning model has been used in the real world, not just in a theoretical environment. The business impact for past use of this model has:

  • Improved a hospital’s billing accuracy, thus increasing revenue by not missing charges.
  • Increased efficiency to reduce auditor costs by 50%.
  • Yielded up to 0.5% net recovered of additional net patient revenue capture.

Machine learning is a highly effective tool that can help bring healthcare providers increased revenue, lower risk, and drive operational efficiencies.

To get started, you need to consider how to bring the benefits of machine learning to your facility. Do you start your own team of data scientists and all the overhead costs that incur? Or do you use a tried-and-true machine learning model like ElectrifAi’s Missing Charges model and quickly get actionable insights?

The choice is yours to make, but make sure you decide now. After all, building your own machine learning model can take months if not years to start producing results. Start reaping the benefits right away with ElectrifAi's experienced data scientists.

To find out more about how ElectrifAi can help, contact us today for a custom demo!

[1]Wheeler, C., & Taylor, R. (2021, January 19). New Year, New CMS Price Transparency Rule For Hospitals. Health Affairs Blog. https://www.healthaffairs.org/do/10.1377/hblog20210112.545531/full/.

[2]Noridian Healthcare Solutions. (2020, October 20). Missing/Incorrect Required Claim Information. Denial Code Resolution. https://med.noridianmedicare.com/web/jeb/topics/claim-submission/reason-code-guidance/missing-incorrect-req-claim-info.

[3]American Hospital Association. (2017, September). Fact Sheet: Hospital Billing Explained. https://www.aha.org/system/files/2018-01/factsheet-hospital-billing-explained-9-2017.pdf

You have data, we have solutions!

Find out what ElectrifAi has to offer by filling out the information below.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.