Here’s how AI will help hospitals take back control of their revenue cycles

Hospitals are falling behind in the revenue cycle race in part through a perfect storm of disadvantages — budget limitations, outdated systems, staffing struggles and crushing workloads on their revenue cycle teams.

Meanwhile, payers have widened their lead with AI that can analyze and deny claims in seconds – leaving hospitals falling further behind while spending more than $25 billion and countless working hours fighting denials. As denials mount, payers’ technology advantage seems starker, particularly when it comes to outcomes of denied or delayed payments on claims that are hospitals’ financial sustenance.

Unless hospitals find a way to recover those losses — and fast — their financial situation will become more precarious. How can they compete in the AI battle to take back control of their revenue cycles while advancing care for their patients and recapturing the hundreds of millions of dollars being left on the table?

Fortunately, hospitals are sitting on a goldmine: their rich patient datasets. These contain all the information needed to supercharge their revenue cycles. Clinical documentation integrity (CDI) teams are the workhorses behind this effort, carefully analyzing patient charts against billing claims to ensure they accurately detail the care provided. But just one hospitalization can generate tens of thousands of data points, and more for clinically complex diagnoses.

The problem with massive amounts of complex, unstructured data and second-level chart reviews can be illustrated by the experience of a medical director for transformation of one large health system. After running queries to determine how many of the health system’s patients had heart failure, the director found that while numerous patients had been given treatment for the disease — a specific, high dose of diuretics — the diagnosis and code for heart failure had somehow never made its way into their charts because of the sheer amount and format of the complex data.

Despite CDI teams’ impressive accuracy, a single person couldn’t possibly scour each line in the structured fields of an EHR to identify new revenue opportunities in any reasonable timeframe. Then there’s unstructured data: the doctors’ notes and discharge summaries that hold additional valuable information about care provided. Hospitals need a way to help their teams find the needles in massive hospital data haystacks to ensure nothing is missed. They need technology that is both scalable for industry-wide impact and attuned to the complexities of real-world healthcare data and operations. That’s where AI comes in.

To fully take advantage of their data, hospitals can combine clinical documentation efforts with AI. This means analyzing all patient data, structured and unstructured, to strengthen claims before they’re submitted to payers. These bills have all the accuracy and granularity needed to capture every bit of revenue they’ve earned the first time, before the bill ever goes out the door.

This AI-enabled edge isn’t just nice to have; it is essential to hospitals’ financial futures. Choosing the right AI tools: for example, those with the baked-in clinical acuity needed to handle complex healthcare data, can help hospitals capture millions of dollars in net new revenue while improving quality scores.

Enhancing prebill CDI analyses to boost hospital revenue

Clinical AI that’s specially designed to enhance CDI teams’ work can help hospitals rewrite their financial and technology stories. This means introducing AI early in the revenue cycle to ensure all claims are complete and accurate before they’re submitted to payers.

In the absence of an easy way to analyze complex EHR data, hospitals typically rely on discharge summaries for billing. There’s a problem with this strategy: discharge summaries are written as notes, and unclear documentation means up to 25% of diagnosis codes could be missed. For example, in one health system, 40% of unstructured notes did not mention the vital signs that led to a sepsis diagnosis, although nearly all the EHRs did. CDI teams must go back and forth with clinicians to fill in the blanks to complete the claim. However, AI models can rapidly and completely analyze 100% of patients’ EHR data against unstructured notes, instantly revealing opportunities to strengthen the documentation.

Whether the claim includes missed codes or inaccuracies, the AI flags opportunities for improvement to human CDI reviewers alongside the evidence it used to make its recommendations. CDI teams and hospital leaders can use these insights to decide whether to adjust the claim. By the time the claim is submitted, the hospital is in the best position to capture all of the earnings from the care it provided. At scale, that can mean millions of dollars in net new revenue and better quality scores. It’s a game changer that hospitals can employ today to improve their revenue yield.

How much should it cost?

Advising cash-strapped hospitals to invest in new technology may sound contradictory. Some AI tools come with high price tags and little evidence to back up their claims of efficiency or cost savings, and there’s an undeniable history of over-promised and under-delivered tech solutions.

CDI solutions are different, especially those with transparent reporting that clearly detail costs, savings and ROI. Advances in machine learning have enabled AI to both think like a doctor and bridge the billing gaps that often leave revenue untapped. And that recovered revenue can mean millions of dollars back in a hospital’s budget, revenue cycle improvements quickly eclipse the cost of the tech. 

In the face of hospitals’ financial crisis, AI has the power to drive real change in the form of dollars and workflow efficiencies. By deploying AI technology proactively within their own datasets, hospitals have an opportunity to flip the script. Not only will that put hospitals on equal revenue-cycle footing with payers, but it will also shift their position to technology innovators, putting them at the forefront of change.