The Riddle of the Middle: Why Middle Rev Cycle Is Healthcare's Most Under-Invested, Highest-ROI Opportunity
Coding, documentation, and auditing already anchor your revenue integrity. Almost nobody owns all three.
Part 1 of a 3-part series on the future of coding, auditing, and revenue integrity.
Every health system has a department for the front of the revenue cycle. Registration, eligibility, prior auth. Someone owns it. Every health system has a department for the back end, too. Collections, denials, appeals, and patient billing. Also owned.
Then there's the middle. Coding. Clinical documentation integrity. CDM and pricing. Charge capture. This is the translation layer between what a clinician documents and what a payer will reimburse. It's arguably the most technically demanding, highest-dollar-impact part of the revenue cycle, and it's also the part that tends to be the least clearly owned, or even understood.
Ask a CFO who's accountable for a DNFB spike, and the answer often depends on who you ask. HIM may point to documentation. CDI may point to coding capacity. Finance may see it as a shared issue without a clear owner. That ambiguity isn't accidental; it's structural. Middle rev cycle sits at the intersection of clinical, informatics, and finance: three functions with different reporting lines, different KPIs, and, historically, different incentives to take the lead.
That's also where we see the biggest opportunity.
A workforce and complexity problem that's converging
The workforce numbers are fairly well established at this point. The American Medical Association estimates a national shortage of certified medical coders at roughly 30%, and the Bureau of Labor Statistics projects about 14,200 coding positions will open each year through 2033, as experienced coders retire faster than new ones are certified. Many organizations are closing that gap with flexible coding capacity that scales with demand rather than headcount alone. At the same time, coding complexity continues to climb. The most recent ICD-10-CM update alone added nearly 500 new billable codes, almost double the prior year's total, with much of the growth concentrated in more granular documentation requirements around social risk factors and advanced treatment pathways.
In other words, the work is getting harder at the same time the pool of experienced people available to do it is shrinking. And the layer of the revenue cycle absorbing most of that pressure isn't the front end or the back end. It's the middle.
Automation helps, but it's not the whole answer
The most common response to this pressure right now is autonomous coding software, and there's good reason for that. It's genuinely effective for high-volume, low-complexity encounters such as radiology and routine outpatient visits. Where it's less reliable is the complex, high-dollar inpatient encounters, where a single missed diagnosis can swing reimbursement by tens of thousands of dollars. Automation tends to absorb the more straightforward share of the workload well. The harder cases, where much of the revenue risk lies, still call for the kind of experienced judgment that underpins strong coding audit and education programs.
That's part of the riddle and worth sitting with: software alone doesn't fully resolve a middle rev cycle problem. It often just makes the underlying gap easier to see because, once the routine volume is automated, what's left is a clearer picture of how much depth there is on the expert side.
The more durable fix is structural
Nearly half of large health systems are already using AI in their revenue cycle, according to a 2026 PayZen/HFMA survey, and that share will continue to grow. Within a few years, some form of automation will be table stakes rather than a differentiator. What will matter more is what surrounds it: who's reviewing the AI's judgment calls, who's catching ambiguous documentation before it turns into a denial, and who's accountable when DNFB days start to creep up.
That's less a technology question than an operating model question, and it's one most health systems haven't fully worked through yet. Middle rev cycle has become a strategic function rather than a purely administrative one, even though most organizational structures haven't kept pace with that shift.
We think this is where a meaningful amount of competitive advantage in healthcare finance will be won or lost over the next few years: less in front-end registration efficiency or back-end collections tactics, and more in whether an organization treats coding, documentation integrity, and revenue integrity as one connected discipline with a clear, accountable owner, rather than three separate functions each hoping the others catch what falls through the cracks.
Where this series is going
Over the next two posts, we'll dig into this in more detail:
- Part 2 looks at coding specifically: not as an "AI vs. humans" debate, but as an allocation question: where automation should carry volume, and where expert judgment should be focused on the dollars that matter most.
- Part 3 looks at auditing and makes the case that, as automation scales, a growing share of the audit opportunity is no longer about catching human error alone. It's about auditing the machine, too.
If your organization is still treating middle rev cycle as three separate problems rather than one connected discipline, that's often an early sign there's recoverable revenue sitting in the gaps between departments. We'd welcome the chance to help you find it.
