Healthcare’s Real Challenge Isn’t Claims — It’s Intelligence
Healthcare has spent years investing in digital systems, automation and data collection, yet one of the sector’s most expensive operational challenges remains stubbornly difficult to solve: getting healthcare providers paid accurately and efficiently.
The problem may no longer be a shortage of technology or information. Instead, a growing argument within healthcare technology is that providers have enormous quantities of data but lack the connected intelligence required to interpret that information early enough to influence decisions.
Mark Sithi, Senior Vice President of Product at healthcare revenue-cycle company R1, has argued that the industry needs to look beyond claims processing itself and examine the fragmented information environment behind it. R1 says Sithi’s work focuses on areas including AI-driven automation, payer connectivity and revenue-cycle intelligence.
With billions of transactions taking place each year between the organizations paying for healthcare and those delivering it, improving how those signals are understood could have significant financial consequences.
Healthcare’s $200 Billion Administrative Challenge
The scale of the issue is considerable.
Research published in JAMA estimated that the financial transactions ecosystem — including claims processing, revenue-cycle management and prior authorization — accounts for approximately $200 billion annually in U.S. healthcare spending.
A subsequent analysis published in Health Affairs Scholar similarly placed annual spending on interactions between patients, providers and payers involving areas such as claims processing, payments, collections and prior authorization at approximately $200 billion annually. It also estimated that more than 9 billion claims are processed each year.
Denials add another major layer of expense.
Healthcare providers spend an estimated $20 billion managing and appealing denied claims. American Hospital Association data citing research from Premier reported that hospitals were spending just under that amount annually on denial appeals, with a significant proportion of those costs associated with claims that ultimately should have been paid.
These costs highlight a broader problem: much of healthcare revenue management still responds to problems after they have already occurred.
Plenty of Data, But Limited Visibility
Healthcare providers are certainly not lacking information.
Payment policies, contracts, coding requirements, authorization rules, fee schedules, claim histories and previous reimbursement decisions can generate enormous amounts of data.
The difficulty is that this information often exists across separate systems.
A healthcare organization may know what an insurer’s published reimbursement policy says, for example, while its historical claims data suggests that claims are being treated differently in practice.
When these datasets are not connected, teams can struggle to recognise changing behaviour until a payment is delayed or a claim has already been rejected.
That puts revenue-cycle teams into a reactive position. Employees investigate the denial, determine what happened, collect supporting documentation and begin the appeal process.
By the time a recurring pattern becomes obvious, the provider may already have experienced multiple similar cases.
Patients can also feel the consequences. Claims disputes and reimbursement delays can create uncertainty surrounding coverage and financial responsibility at a point when the patient’s primary concern should be receiving care.
AI Could Make Payer Behaviour More Visible
Artificial intelligence is increasingly being introduced into healthcare revenue management as a way to automate administrative work.
However, automation may represent only part of its potential.
One of AI’s more significant capabilities is its ability to analyse large quantities of previously disconnected information and identify patterns across millions of interactions.
That could allow healthcare organizations to examine reimbursement behaviour in a fundamentally different way.
Instead of analysing individual claims independently, AI systems can potentially compare payment outcomes with policies, authorization requirements, coding decisions and previous payer responses.
Patterns that would be difficult for employees to identify manually could therefore become visible much earlier.
This introduces a concept already familiar in areas such as software engineering: observability.
Technology companies continuously monitor system behaviour to identify anomalies before they develop into major failures. Applying a similar approach to healthcare revenue management could allow organizations to detect changes in reimbursement behaviour before they develop into widespread financial problems.
Recent R1 analysis has similarly argued that understanding how payers actually make reimbursement decisions is becoming increasingly important as requirements change across policies affecting documentation, authorization, coding and billing.
Moving From Denial Recovery to Denial Prevention
Traditional revenue-cycle management has frequently focused on what happens after a problem appears.
A provider submits a claim. The payer denies or delays it. Staff investigate the reason. Additional documentation may be collected, corrections made and an appeal submitted.
Technology can make each stage faster, but that still leaves the underlying model largely reactive.
Connected intelligence creates the possibility of moving decision-making earlier.
If an organization can analyse previous outcomes and recognise how a particular payer is likely to respond, those insights could potentially be introduced during authorization, documentation, coding or billing.
Rather than simply accelerating an appeal, the goal becomes preventing an avoidable denial from occurring.
The financial implications could be substantial.
Research on healthcare financial transactions has estimated that improvements including greater automation and other structural changes could reduce spending within the U.S. financial transactions ecosystem by $40 billion to $60 billion.
For providers, reducing preventable denials could also mean more predictable cash flow and fewer employees spending time investigating problems that could potentially have been identified earlier.
Intelligence Could Also Change the Payer-Provider Relationship
Discussions surrounding AI in healthcare payments can easily become focused on competition.
Providers want to maximise appropriate reimbursement, while insurers are responsible for managing utilization and healthcare expenditure.
AI could consequently be viewed simply as another technology being deployed by both sides of that relationship.
However, inefficient information exchange creates costs for everyone involved.
Duplicate work, repeated requests for documentation, unclear requirements and multiple rounds of appeals require resources from both healthcare providers and payers.
Better visibility into where these problems repeatedly occur could provide another benefit: identifying processes that create substantial administrative work without providing equivalent value.
Rather than using AI exclusively to process disputes faster, data could therefore help organizations understand why those disputes are occurring repeatedly in the first place.
This could ultimately create opportunities to simplify processes and reduce unnecessary interaction between providers and insurers.
From Revenue-Cycle Data to Predictive Intelligence
Healthcare has already completed much of the first stage of digital transformation.
Medical records have become digital. Claims are processed electronically. Revenue-cycle systems capture enormous quantities of information about authorizations, payments, denials and reimbursement.
The next challenge is connecting those digital signals.
For healthcare organizations, competitive advantage may increasingly come from the ability to transform historical data into foresight — identifying unusual payer behaviour, recognising emerging reimbursement risks and taking action before those issues materially affect revenue.
AI and autonomous systems could accelerate that transition, but simply introducing more technology will not automatically resolve the problem.
The underlying information must first be accessible, connected and interpretable.
Without those foundations, automation risks making fragmented processes faster rather than making them fundamentally better.
With them, however, healthcare revenue management could gradually move away from a system built around recovering money after problems occur towards one capable of predicting and preventing many of those problems beforehand.
And that distinction could ultimately prove more important than claims automation itself.
References
JAMA — Administrative Simplification and the Potential for Saving a Quarter-Trillion Dollars in Health Care
Research examining U.S. healthcare administrative spending, including the approximately $200 billion financial transactions ecosystem. View the JAMA research
Health Affairs Scholar — Active Steps to Reduce Administrative Spending Associated With Financial Transactions in US Health Care
Analysis of approximately $200 billion in annual financial-transaction spending and the potential for $40 billion to $60 billion in savings. View the Health Affairs Scholar study
American Hospital Association — Payer Denial Tactics: How to Confront a $20 Billion Problem
Analysis of denial rates, appeals and the financial burden placed on healthcare providers. View the AHA analysis
R1 — Mark Sithi
Background on Mark Sithi and R1’s work in healthcare revenue-cycle automation and intelligence. View Mark Sithi’s R1 profile
0 Comments