Most healthcare leaders I speak with already know overtime is too high. What's less recognized is that most organizations don't believe it's actually within their control.

Overtime doesn't announce itself as a structural problem. It shows up as a shift that needs filling, a manager making the fastest call available, and a cycle that repeats often enough that it stops looking like a problem at all and starts looking like the plan.

Before getting into why that happens, here is what it actually takes to move the number. Overtime rises and stays high when the rules governing it exist on paper but not in the system. Organizations that bring it down do not do so by approving less of it. They do so by encoding the decision logic so that overtime becomes the last option considered, not the first one reached for.

Understanding why that cycle persists and what it actually costs requires looking at three things: why healthcare systems consistently over-award overtime, why it costs far more than most organizations track, and what regaining control actually looks like in practice.

Why Overtime Keeps Rising: When the Pressure Valve Becomes the Plan

Overtime is not chosen carelessly. It is chosen because in the moment, it is the only tool that works fast enough.

When a shift needs filling in two hours, overtime is reliable in a way that recruiting, rebalancing, or restructuring simply is not. Used occasionally, for genuine emergencies, it is a reasonable pressure valve. The problem is that in most hospitals, the emergencies never stop.

And a pressure valve that runs continuously is no longer a safety mechanism. It is the system.

The average hospital carries 43 unfilled nursing positions at any given time, and one in three reports a vacancy rate above 10%, according to a 2026 NSI Report. When that many shifts are chronically uncovered, overtime stops being a reactive decision and becomes a scheduling assumption. Managers build it into their coverage plans not because they want to, but because there is no realistic alternative visible from where they are standing.

What makes this worse is that overtime in healthcare is not supposed to be freely awarded. Collective agreements, seniority rules, and overtime thresholds exist precisely to govern when and how it gets approved. But when those rules live in binders, spreadsheets, or the memory of one experienced scheduler, they do not function as real constraints.

The failure is not the rules themselves. It is that the rules have no way of enforcing themselves at the moment a decision is made. Overtime gets awarded to the wrong person, at the wrong rate, past the threshold that should have triggered a different solution. And by the time HR has a chance to validate the decision, the shift has already been worked.

That is where the real cost starts to surface and where most organizations underestimate the impact. It is also harder to see than most are willing to admit.

The Real Cost of Overtime

The cost of chronic overtime shows up in three distinct ways and most organizations only track one of them.

Wages. Overtime pay typically runs at a 50% premium above the regular hourly rate, though thresholds vary by state and employment classification. Most organizations accept that cost when it is occasional. What becomes harder to absorb is what happens when it stops being occasional and starts being structural. Hospital labor costs rose more than 20% between 2019 and 2022. Leadership teams expected those costs to normalize once pandemic pressures eased. They did not. By 2025, total hospital labor spending had crossed $1 trillion, with compensation now accounting for 56% of the average hospital's total expenses according to the AHA. A 50% hourly premium lands very differently when the baseline underneath it has been climbing for six consecutive years with no correction in sight.

Grievances. When a nurse files a grievance claiming they were bypassed in the seniority order for an overtime offer, the hospital's first task is reconstruction. Who made the call. When. Who was available. Who had already been offered overtime that week. In a manual environment, that information is scattered across spreadsheets, phone logs, and the memory of whoever was on shift that day. Pulling it together can take hours or days, and what it produces is rarely airtight. The arbitration fees themselves are relatively modest. The staff time, legal counsel, and reputational exposure with the union are not. And unlike the overtime premium that started the chain, none of those costs appear on a labor report.

Burnout and turnover. This is where the cost stops being a line item and becomes a cycle. Chronic overtime compounds into poor work-life balance, and poor work-life balance compounds into burnout. When a nurse eventually leaves, which the NSI data tells us is increasingly likely given that scheduling conflict now ranks in the top five reasons nurses resign, replacing them costs an average of $60,090. Their colleagues then absorb the coverage gap through more overtime for an average of 78 days until the position is filled. The NSI Report puts the average annual cost of RN turnover at $5.19 million per hospital. Overtime did not cause all of that. But it is reliably upstream of it.

Premium labor. When overtime alone stops being enough, organizations move to the next most expensive tool. Travel nurses currently cost hospitals an average of $91 per hour compared to $59 for a staff nurse, a 53% hourly premium layered on top of the overtime premium already being paid. This is not a separate problem. It is the same problem one level further along the same escalation. At that point, the organization is no longer managing a scheduling challenge. It is funding a dependency that compounds every quarter that it’s being unaddressed.

To put the full picture in one place: a hospital managing chronic overtime is not paying a wage premium. It is paying a wage premium, absorbing grievance exposure, funding a turnover cycle, and in many cases subsidizing a travel nurse dependency that grows harder to unwind the longer it runs.

How to Reduce Overtime: What Regaining Control Actually Requires

The organizations that have meaningfully reduced overtime did not do it by telling managers to approve less of it. They did it by changing what the decision looks like before it gets made. What I notice in organizations that have made this shift is that the change rarely starts with a technology decision. It starts with someone asking a different question: what would it take for overtime to be the last option instead of the first?”

The answer, consistently, comes down to three shifts. I think of them as the encoding shift, the audit shift, and the signal shift.

The encoding shift: building the rules into the process. The most effective shift I have seen is when collective agreement rules, seniority logic, and overtime thresholds stop living in binders and start living in the system. When a gap opens, the platform surfaces who is eligible, in what order, at what rate, before any offer is made. What changes is not who makes the decisions. It is what information is available at the moment of them making the decision. And critically, an activated system does not default to overtime. It first checks whether a shift swap or a move-up is possible without crossing a threshold. Overtime becomes the last option, not the first.

One health system we work with described the change this way: before, the manager on shift carried the rules in their head. Now, the system carries the rules, and the manager makes the call. Same person, same authority, different outcome.

The audit shift: making every decision traceable. One of the most undervalued benefits of automated overtime management is not speed or accuracy. It is traceability. Every offer made, every declination recorded, every threshold crossed exists in a log that is independent of the person who made the decision. Critically, that log does not require a separate system or a manual audit process. It is a byproduct of how the decision was made in the first place. When a grievance is filed, the reconstruction that currently takes days takes minutes. More importantly, the organization can demonstrate compliance rather than assert it. That distinction matters more than most teams realize until they are sitting across from a union representative.

The signal shift: treating overtime as a signal, not a solution. The goal is not zero overtime. That is neither realistic nor desirable. The goal is overtime that gets awarded for the right reasons, to the right people, at the right frequency. Organizations that reach this point share one common characteristic: they stopped treating overtime as a scheduling tool and started treating it as a diagnostic. A unit running consistent overtime is telling you something about its coverage model, its vacancy rate, or its scheduling assumptions. Reading that signal early enough to act on it is only possible when the data exists in a form that supports pattern recognition, not just shift-by-shift reaction.

What Comes Next

Getting overtime under control is not the end of the story. It is a precondition for a more stable workforce, one where staff experience more consistency, managers spend less time firefighting, and the organization stops compounding its own pressure.

A stable workforce requires more than containing one cost. It requires the entire scheduling operation to run reliably: decisions made consistently, exceptions handled without heroics, coverage built on process rather than memory. Overtime is the most visible place where that breaks down. It is not the only one. 

The harder question is not how much overtime costs. Most organizations already know. The question worth sitting with is why the system underneath it keeps producing it, and whether that system is actually built to handle what healthcare asks of it every day. If you have not yet read the first article in the series, it explores why the structural conditions that produce this pressure exist in the first place, and why they are not going away on their own.