How Machine Learning Can Improve Patient Eligibility Verification

By PokitDok Team,


The current state of the consumer healthcare experience has significant room for improvement. According to the Accolade Consumer Healthcare Experience Index, nearly a third of Americans say that navigating a healthcare system makes them more uncomfortable than does buying a home, a car, or some expensive technology.

Sarah Buhr at TechCrunch describes medical billing as "murky" since "most of the time it's not clear how much something will cost and sometimes you don't even get the (possibly whopping) bill until months down the road." Clearly, medical billing and electronic claims processing can be improved to be more transparent and intuitive.

A report by Meritalk highlights how data integration issues with healthcare benefits verification led to a staggering $343 billion in economic costs annually, which are borne by government health and human services agencies. The lack of interoperability between different healthcare systems causes government agencies to bear the healthcare costs of people who aren't actually eligible. Difficulties in determining patient eligibility also affect the quality of patient care.

In recent years, healthcare organizations have looked to technology to improve medical billing and transform the customer experience. Instead of allocating resources to manually verify insurance eligibility, a growing number of companies are leveraging healthcare APIs that automate the checking of patient eligibility and their payment responsibilities.

So, how can technology further streamline the patient eligibility and payment processes? Here are some ways where machine learning is already making a difference.


Accelerate Insurance Preauthorization

Health insurers frequently require preauthorization to verify that a patient is entitled to receive medical services before he or she actually receives them. While often necessary, preauthorization requests can disrupt and slow a patient's access to treatment and add unnecessary busy work for providers. Recondo Technology's Heather Kawamoto writes at Digital Commerce 360, that the manual activity required for one preauthorization request takes about 15 minutes. This hampers the productivity of the clinical and administrative staff, who could instead be involved in more meaningful aspects of patient care.

With machine learning, the preauthorization process is automated and requires very little human intervention. As Kawamoto explains, authorized web crawlers can conduct preauthorization requests quickly and easily. Once the request results are in, web crawlers can automatically sort the results and route the information to the providers' system for the next steps. Approved preauthorizations can lead to registration and scheduling systems, whereas denied requests can be routed to staff for follow up.

Save Time and Resources -- For Both Patients and Providers

It's not just healthcare providers that are wasting resources on patient eligibility; verifying  insurance benefits is also a time-consuming process for patients. Christina LaMontagne at Huffington Post notes that savvy patients who want to make the most of their health plans and minimize their own out-of-pocket expenses have to do their own homework.

Kate Sahnow at Health Partners has a detailed guide for patients to find out what their health plan covers, and it is a tedious process. It involves checking various sources for information, verifying with customer service representatives, and keeping track of this information over time. Insurance companies typically have a database of in-network doctors on their websites that patients can refer to, but this data can be outdated, so patients often have to call their healthcare providers and the insurance companies in order to verify eligibility prior to their visits. This, of course, is not the ideal patient experience.

Even when patients call a healthcare provider to check on their insurance eligibility, providers face the problem of not having complete information. Even if they are an in-network provider, they still may not know what services are covered by insurance and what costs the patient must pay out of pocket.

This is where machine learning and patient eligibility APIs can make a difference. On the front-end, self-service technology such as virtual avatars and chatbots can be used to handle queries on insurance eligibility through the integration of healthcare systems and machine learning. Rowland Manthorpe at Raconteur examines how this type of technology can bring speed and awareness to healthcare processes. Patient eligibility is certainly one area that can benefit from this.

In addition to chatbots and virtual avatars on the front-end, there are currently APIs -- the things that let distinct systems and software communicate and work together seamlessly in the background -- that can help healthcare providers on the back-end. APIs can handle patient eligibility checks in minutes, if not seconds, freeing up a care provider's admin team to focus on more productive work.  


Reduce Billing Inaccuracies and Provide Faster Billing

Medical billing can be prone to inaccuracy, especially when manual input opens the door to human error. Even a typo in a patient's address can throw a wrench into a billing cycle. Improving the benefits verification process could reduce medical billing inaccuracies and thereby streamline billing.

Further, a smarter billing system could help estimate patient payment responsibilities. This is especially pertinent today as patients bear more medical costs due to the prevalence of high-deductible health plans. This was reported in a 2016 InstaMed survey which found that about 74 percent of healthcare providers reported a significant increase in patient financial responsibility, compared to the previous year.

Electronic healthcare information that leverages machine learning can more accurately predict patient payment responsibility and a patient's propensity to pay in real time. This allows providers to create new payment options and to move payments forward in the revenue cycle by enabling more reimbursements to be collected at the time of care. In doing so, providers can increase revenue and reduce bad debt.

This can have a significant financial impact. According to McKinsey, "about $300 billion a year -- 15 cents of every dollar spent on healthcare -- is lost on claims processing, payments, billing and revenue cycle management, and bad debt."

Muxi Li and Danielle Dean at the Cortana Intelligence and Machine Learning Blog write about how machine learning helps healthcare providers to better manage their revenue cycles and cash flow. Machine learning is able to forecast:

  • Whether human intervention is needed to speed up the claim payment process
  • The amount of time required for an insurance company to pay a claim

This allows for greater operational efficiency and better resource management, thanks to the machine learning predictions that prioritize which claims need additional work.

Reduce Fraudulent Claims

Hugh Anderson at Digitalist Magazine believes that machine learning can help lower risk for providers by reducing fraudulent claims.

Tom Lawry at Microsoft Corporation shares the case study of Fullerton Health and KenSci, who utilized machine learning to more effectively manage fraudulent claims. By implementing machine learning to data sets that had already been reviewed by 20 claims specialists, they uncovered more than $1 million in questionable and fraudulent claims.

The capacity of machine learning to process large data sets in a short amount of time means that providers can identify data patterns and information that might otherwise have remained cloaked.


Empower Patients

Centric Digital argues that the true value of using technology such as machine learning in healthcare, is that it empowers patients. No longer are patients "moved around like a passive chess piece between providers and insurers."

Instead, machine learning can provide patients with better information that allows them to choose the most cost-effective and appropriate healthcare options. Patient eligibility APIs provide patients with key data on their health coverage, plans, deductibles, and more -- and all of this is information that can help them better budget their out-of-pocket healthcare expenses.

This, Centric Digital says, will lead to higher patient satisfaction since patients will now be able to make more active choices about which provider, service, or treatment is the best choice for them. Even those with more limited healthcare choices will still be able to benefit from "the increased coordination between providers and payers [that] enables the patient to be directed automatically to the most appropriate, effective and affordable care option."

By improving just the patient eligibility verification process, a whole set of benefits is introduced to the healthcare system: more efficient billing cycles, fraud detection, and patient empowerment are some of the big ones. We believe the opportunities are there for machine learning technologies to step in, create those kinds of efficiencies, and help all of us work toward a better, more value-centric approach to healthcare.

Images by: Meditations, Cobanams, Pexels, valelopardo

The opinions expressed in this blog are of the authors and not of PokitDok's. The posts on this blog are for information only, and are not intended to substitute for a doctor-patient or other healthcare professional-patient relationship nor do they constitute medical or healthcare advice.

  Tags: API, Consumer, Health Innovation


  Comments: 1

Your feedback