Medical billsMedical bills

The No Surprises Act (NSA) goes into effect on January 1, 2022. We know what this means for patients—a welcome reprieve from unexpected medical bills arising from circumstances beyond their control. But what about other stakeholders, like insurance companies? One of the provisions of the No Surprises Act requires health plans to update their provider directories more frequently. This gets to the root cause of surprise billing, which is patients’ ability to easily identify which providers are in-network. If payers can make updates to their directories in just 48 hours, as the law mandates, patients will have a much better chance of finding care that’s actually covered by their plans.

Sounds reasonable, right?

Except for the fact that health plans are processing an avalanche of provider data at any given time. And the data are not streaming in seamlessly through a shared platform between payers and providers. Payers are literally receiving Excel spreadsheets from hundreds of different provider groups, each with their own unique and ever-changing templates. The data is so messy that all plans have dedicated staff whose job is to assess, interpret, clean up, and then enter it into the plan’s system. It’s no surprise—pun intended—that these updates can take weeks, cost millions of dollars a year, and still have accuracy rates as low as 60%. That’s hugely problematic, since plans rely on this information not only to update their provider directories, but also to determine which providers should be paid contracted rates, for specialty and licensing updates, and for billing information changes that are necessary to properly pay claims.

While historically it has been challenging for payers to solve for all these issues, recent advances in machine learning and artificial intelligence now make it possible to automate the millions of human hours spent cleansing and keying in this data every year, as well as improve its accuracy. Think of it this way: landscapers have machines that help them shovel, so they can focus on making a yard beautiful. And to give an example from another industry in which human lives are on the line, pilots make all the flying decisions, not the planes.

These are two examples of technology that help humans safely and easily perform large amounts of work. If plans use those technologies, they will not only achieve compliance with the No Surprises Act, they will also be empowered to get employees back to their core business: keeping members healthy.

How payers should be thinking about automation

There are three main things to keep in mind over the next few months as health insurers consider leveraging automation to help them comply with the No Surprise Act.

  1. Learn what automation can—and cannot—do. Automation cannot completely solve the global interoperability problem. What some of the more sophisticated platforms can do, however, is sit between disparate systems and act as a translator. If plans put in up-front effort to codify the knowledge that resides with the staff who have been cleansing and keying in this data, and then feed that knowledge to an algorithm, it can take on the work associated with making sense of the data. Put more simply, if an automation tool is given all the information about how humans have been making decisions about processing this data, then it can apply that same decision-making framework but in a much more efficient and consistent way.
  2. Assess the situation and set realistic goals. Under current manual processes, how long is it taking the payer to make provider updates, on average? What percentage of the updates are accurate? What types of issues does the plan most commonly experience with the data it receives (is it missing column headers and or contain blank fields in Excel files, providers listed at the wrong practice locations, or something else entirely)? The answers to these questions will vary from payer to payer, as will the goals for improvement. National plans obviously receive higher volumes of provider data than those dedicated to single markets. Some plans have already started documenting their data processing decision trees, while others have not even identified the need for this documentation. Health plans that are further along may want to set goals above and beyond what’s required by the No Surprises Act, aiming for turnaround time of less than 24 hours for provider updates (as opposed to the 48 hours that will be required) and accuracy levels of 95% or greater. These are reasonable goals, given the power of the automation tools that are available to tackle this problem today.
  3. Understand the range of automation solutions available. Not all automation solutions are created equal. Some require a “rip and replace” approach that payers may find disruptive to existing IT infrastructure, but other solutions can actually co-exist with current systems. Solutions also vary in terms of the type of data they are able to automate—insurers should seek out those that are sophisticated enough to deal with the inherent messiness of human-generated data. Finally, payers should look for an automation partner that provides human support in addition to technology. It’s not necessary to have a team of data scientists in-house to leverage automation. The right partner can assist with needs assessment and goal-setting, provide guidance on how to go about the process of documenting current manual processes, set expectations around the timeline for putting automation in place, and, of course, expedite the timeline.

January 1, 2022 may seem like it’s still far away, but factoring in upcoming holidays and the time off that people typically take in the fourth quarter of the year, the No Surprises Act deadline is practically around the corner. Payers that want to be in compliance with its provider directory provision must act now.

Photo: fizkes, Getty Images


For researchers who aren’t affiliated with an academic library, finding scientific papers can be time-consuming and expensive—and organizing and sharing them with co-workers can be even harder. A new tool by DeepDyve looks to help researchers address this gap.