How AI-based tools can help manage COVID-19’s aftermath

New AI features like natural language processing can make documentation less of a chore.

Payers and providers have traditionally experienced competing interests, but today they share the same reality of an uncertain business environment caused by COVID-19 – and in many cases they are using the same technology to overcome operating challenges created by the pandemic.

As the pandemic began to unfold, many providers paused elective surgeries and others experienced significant declines in office visits, screenings, and routine care. The pause on elective surgeries alone created an estimated $200 billion in financial losses for hospitals and health systems between March and June 2020 alone, according to a McKinsey report.

Then the pandemic eased for a little while and volumes and procedures began to return to usual rates until the omicron variant led many providers to again postpone nonemergent procedures in late 2021 and early 2022.

Now, payers and providers both must deal with a backlog of elective procedures and it’s anyone’s guess how long it will take to clear. One executive with an endoscopy center said the backlog of colonoscopies could take “years to recover from.” In reports cited by McKinsey, the Journal of Bone and Joint Surgery estimated a cumulative backlog of more than 1 million joint and spine surgeries by 2022 and the Journal of Cataract & Refractive Surgery estimated a backlog of between 1.1 and 1.6 million cataract surgeries by the end of this year.

At the same time, patients have faced significant financial challenges from the pandemic. For example, a Kaiser Family Foundation analysis found that over a three-month period last year, the cost of treating unvaccinated COVID-19 patients in hospitals was a staggering $5.7 billion. Additionally, a study published in JAMA Network Open estimates that Americans who get seriously ill from COVID-19 might be responsible for thousands of dollars in medical bills from hospitals, doctors, and ambulance companies.

For providers, postponements have equated to delayed care and declining revenues. For payers, delays in care, testing and procedures often translate to more expensive care needs for patients whose conditions were complicated and worsened when timely treatment was not available.

A better alternative for chart reviews
To address patients’ care needs, payers and providers must identify gaps in care and prioritize the highest risk individuals. To do this effectively and efficiently, many payers and providers will turn to artificial-intelligence-based technologies such as natural language processing (NLP). By enabling computers to “read” and understand text by simulating humans’ ability to interpret language, NLP helps payers and providers extract key insights from colossal amounts of medical records data – without the bias or fatigue that is inherent to humans. NLP automates – with more accuracy - expensive, manual chart reviews, without requiring clinicians to wade through thousands of pages of documentation to pick out tiny pieces of data.

In giving computers the ability to read, understand and interpret language, NLP does far more than merely identify the presence of words or elements within text. By leveraging NLP, clinicians and researchers can organize data from an individual’s health journey, an entire patient population, or an enterprise's data warehouse. Additionally, NLP gives organizations the ability to retrospectively analyze longitudinal health data to find one particular piece of information about one particular patient or identify populations that require further exploration.

Following are three ways that NLP can help payers and providers manage in the aftermath of the COVID-19 pandemic:

  1. Prioritizing patients for care: In the wake of the pandemic, it has become critical for both providers and payers to triage care to determine which patients are in the most urgent need and, alternatively, which can effectively be placed at the back of the line. NLP can help payers and providers sort through mountains of patient records to pinpoint the information that reveals which patients are most in need of urgent interventions.
  2. Closing care gaps: As healthcare workers across the nation endured disruptions to usual operations and pandemic-related stress, it’s inevitable that some vulnerable patients fell through the cracks – and in some cases likely saw chronic conditions deteriorate. By analyzing longitudinal patient records to identify care gaps, payers and providers can generate outreach lists of patients to target for much-needed follow-up care.
  3. Ensuring payment integrity: With shifting regulations and policies in flux as a result of the pandemic, it has been challenging for payers and providers to ensure proper payment for COVID-19-related tests and procedures. NLP can automate audits of medical records to identify potential fraud or improper payments, billing inaccuracies, or clerical errors, and to assess the accuracy and completeness of clinical documentation. Records are reviewed with more efficiency and accuracy, which also facilitates faster claims processing.

Although it’s unclear when the pandemic will transition to an endemic, payers and providers need to be ready to manage a backlog of delayed care whenever it happens. Once it does, AI-based tools such as NLP will represent a critical means of prioritizing care, closing care gaps and ensuring payment integrity.

Ketan Patel, MD is chief medical officer for SyTrue. He is also an emergency medicine physician with a focus on healthcare information technologies aimed at improving clinical data workflow and improved detection of disease at the point of care.