The RCM Workforce Trap: Why Cutting Jobs and Outsourcing Without Oversight Is a Slow-Moving Disaster

Three health systems made revenue cycle workforce announcements this week. UnityPoint eliminated 207 IT and RCM positions and shifted those functions to outside vendors. Northwell cut less than 4 percent of its IT workforce. Rochester Regional eliminated 22 coding positions at two hospitals, drawing a community petition with more than 1,000 signatures urging leadership to reconsider.

None of these decisions happened in a vacuum. Health systems are operating in one of the tightest margin environments in decades. Federal Medicaid policy is shifting in ways that are compressing revenue. Labor is expensive. Automation is accelerating. The pressure to find savings anywhere available is real, and I am not going to pretend otherwise.

But I have watched this pattern play out enough times to know that not all of these decisions will age well. Some organizations will look back at what they cut this year and spend the next two years paying to rebuild it. The ones that get it right are not the ones that move fastest. They are the ones that are honest about what they are actually giving up before they give it up.

Here is what I mean.

Outsourcing a function is not the same as solving the problem that function was struggling with

The most common mistake I see in RCM outsourcing decisions is treating the vendor as a solution to an operational problem rather than a delivery mechanism for a function you have already figured out. If your coding accuracy is poor, outsourcing coding does not fix the underlying workflow. It exports the problem to someone who has less context about your patient population, your payer mix, and your documentation culture than your internal team did. And now you are paying them to produce the same results your internal team was producing, plus a margin.

The organizations that get the most out of outsourcing are the ones that have already done the hard work of defining what good looks like. They know their clean claim rate, their denial rate by payer and code type, their cost to collect per claim. They have written those benchmarks into the vendor contract as enforceable KPIs with defined remedies if performance falls short. They have kept internal staff whose job is to monitor vendor performance and escalate problems before they become cash flow events.

The organizations that regret outsourcing are the ones that handed the function over because it was struggling, assumed the vendor would fix it, and found out 18 months later that it was still struggling, just at a higher total cost and with less internal visibility into why.

When you lose the people, you lose the knowledge that does not live anywhere else

This is the part that is hardest to quantify and easiest to underestimate. Revenue cycle operations run on institutional knowledge that is not documented anywhere. The coder who knows that a particular payer started rejecting a specific modifier combination six months ago and quietly adjusted the workflow. The billing specialist who knows which payer reps to call to get a complex denial resolved in two days instead of two weeks. The analyst who built the denial dashboard and knows exactly which filters to apply to find the accounts that are about to age out of appealable status.

When those people leave, that knowledge leaves with them. It does not transfer automatically to the vendor. It does not live in the EHR. It does not show up in the training materials. It just disappears, and the organization does not always notice it is gone until something breaks.

I wrote in the RCM Workforce Modernization Guide that the revenue cycle of 2030 will not be organized by department as much as by capability. The skills that resist automation are not the ones tied to specific job titles. They are the ones tied to judgment: knowing when the algorithm is wrong, understanding why a payer is behaving differently than it did last quarter, being able to explain a complex billing situation to a patient in a way that does not destroy trust. Those capabilities take years to develop. They cannot be replaced on a 30-day notice period.

The vendors do not have the same incentives you do

This is the uncomfortable truth that does not come up enough in outsourcing conversations. Your revenue cycle vendor is optimizing for the metrics in the contract and for their own margin. Those are not always the same thing as your financial outcomes.

A vendor that is paid per transaction has an incentive to process volume. A vendor that is paid on collections has an incentive to prioritize easy accounts. Neither of those incentive structures automatically produces the outcome you actually want, which is the highest possible net revenue on the broadest possible claim volume with the lowest possible administrative cost and the best possible patient experience.

The gap between vendor incentives and organizational outcomes is manageable. But it requires active management. It requires internal staff who understand the work well enough to know when the vendor is performing below standard, leadership that treats vendor oversight as a real operational function rather than a quarterly check-in, and contracts that are specific enough to be enforceable rather than aspirational.

Most outsourcing relationships do not have all three of those things. Which is why most outsourcing relationships eventually disappoint.

Automation is changing the equation, but not in the way most people think

There is a reasonable argument that some of the job cuts happening right now are appropriate responses to genuine automation. If AI is handling routine prior authorization submissions, you do not need as many staff doing manual prior auth. If ambient documentation is reducing coding rework, you may not need the same size coding team you had three years ago. That math is real.

But the math only works if the automation is actually performing at the level you think it is. And as I covered in this week's newsletter, most organizations are not measuring their AI performance in a way that captures what payers are doing in response. You may think your AI prior auth tool is saving 40 percent of staff time while your payer is simultaneously increasing denial rates on AI-generated submissions. The labor savings are real. The net revenue impact may be negative.

The workforce decisions being made right now need to account for that uncertainty. If you are reducing headcount based on AI performance projections rather than demonstrated, measured results, you are making a staffing decision on an assumption. That assumption may turn out to be right. It may also turn out that you need those people back when the payer response to your automation becomes visible in your denial rate, and they have already moved on to other jobs.

What the organizations getting this right are doing differently

The health systems navigating this well are not refusing to outsource or refusing to reduce headcount. They are doing both of those things more deliberately than their peers.

They are defining what they are keeping internally before they decide what to outsource. The functions closest to payer intelligence, the people who understand denial patterns, payer behavior, and contract performance, tend to stay internal because that knowledge is a competitive asset. The functions that are genuinely commoditized, high-volume, rules-based work that a vendor with more scale can do more efficiently, are the better outsourcing candidates.

They are writing contracts that reflect what they actually need. Outcome-based KPIs tied to clean claim rates, denial rates, net collection rates, and cost to collect. Audit rights that let them verify vendor performance at the claim level, not just in aggregate reports. Escalation mechanisms that give them recourse when performance slips before it becomes a financial problem.

They are investing in the people who stay. When you reduce a team and ask the remaining members to take on more responsibility, including oversight of the vendor doing what their former colleagues used to do, you need to invest in those people proportionally. Training, tools, and protected time to actually do the oversight work rather than just adding it to an already full plate.

And they are treating the workforce transition as a multi-year project rather than a budget line item. The organizations that make a round of cuts, declare victory, and move on are the ones that end up making a second round of cuts a few years later to pay for the cost of rebuilding what they eliminated the first time.

The bottom line

Reducing RCM headcount and outsourcing functions are legitimate strategies in the right circumstances. The margin environment health systems are operating in right now makes them sometimes unavoidable. I am not arguing against them categorically.

What I am arguing against is doing them without a clear-eyed accounting of what you are giving up. The institutional knowledge that walks out the door with your experienced coders and billing specialists. The oversight capacity that disappears when you eliminate the internal staff who understood the work well enough to know when it was going wrong. The negotiating leverage that evaporates when your vendor knows you no longer have the internal capability to do the work yourself.

Those costs are real. They are just delayed. And the organizations that are honest about them now are the ones that will not be explaining an unexpected revenue shortfall to their board in 2027.

If you want a practical framework for thinking through which RCM workforce decisions to make and how to structure the oversight infrastructure that makes outsourcing actually work, the RCM Workforce Modernization Guide covers it in depth. It is built specifically for CFOs and HR leaders navigating exactly this transition.

And if you are not already subscribed to RCM 2030 Weekly, I cover the news that matters for revenue cycle leaders every Sunday. Subscribe here: https://www.linkedin.com/build-relation/newsletter-follow?entityUrn=7441481079762518016