Why Clinical Variation Is Expensive — and Why Most Organizations Underestimate the Cost

A quality director at a large regional health system recently told us something worth sitting with. After a lengthy internal analysis, her team had calculated what they believed was the annual cost of unwarranted clinical variation — roughly $8 million in excess spending across five high-volume DRGs. Leadership was pleased. The number was concrete, bounded, and manageable.

It was also, almost certainly, wrong. Not because the analysis was flawed, but because it was only measuring the costs that were easy to see.

Clinical variation is one of the most studied problems in American medicine. The Dartmouth Atlas has spent three decades documenting it. The Institute of Medicine called it out. Every major health system in the country has a quality initiative touching it. And yet, despite all the attention, the problem persists — in part because organizations keep measuring its cost incorrectly, and as a result, they consistently under-invest in fixing it.

The true cost of unwarranted variation is bigger than any single dashboard captures. Understanding why matters, because getting the number right changes the economic case for the intervention.

The Costs Health Systems Do Measure

The most common approach to quantifying clinical variation is claims-based: look at utilization patterns across similar patients, identify outliers, calculate the difference in resource consumption. It's a reasonable starting point, and it's what most quality departments are trained to do.

This analysis typically surfaces real costs. Unwarranted variation in surgical technique drives excess complications. Variation in antibiotic selection — especially when clinicians work without local antibiogram guidance — fuels resistance and treatment failure. Variation in care transitions leads to preventable readmissions. These are measurable, and the dollars are significant: estimates from independent actuarial analysis place preventable clinical variation at $200–$250 billion in annual U.S. healthcare spending.

But claims data captures the output, not the process. And the most expensive part of clinical variation often happens before any claim is ever generated.

The Costs Most Organizations Miss

Consider what's not in a claims analysis.

Physician and nursing cognitive load. When clinicians lack clear, locally-approved guidance at the point of care, they improvise — often well, but at a cost. Every unnecessary decision cycle burns time and attention. At scale, across thousands of daily clinical decisions, the aggregate drain is enormous. It doesn't appear on a cost report. It shows up in burnout rates, in throughput, in the microfrictions that slow a floor or an ED without ever triggering a billing event.

The cost of managing the variation you already have. Health systems spend considerable effort on the downstream consequences of unwarranted variation — peer review, quality huddles, retrospective chart audits, denials management, payer-facing appeals. These activities don't eliminate variation; they account for it after the fact. The labor costs are real and rarely attributed to variation as their root cause.

Compliance and regulatory exposure. CMS quality measures don't exist in isolation — they carry financial stakes. Under value-based payment programs, excess variation in process adherence translates directly into payment adjustments. Joint Commission and CMS surveyors expect evidence-based protocols governing high-risk clinical situations. Organizations without a disciplined, documented approach to clinical standardization are carrying regulatory exposure that rarely appears in a variation cost model.

The opportunity cost of pathways that don't get used. This may be the most expensive line item in the variation ledger, and it's almost never included. Clinical pathways that exist on a shared drive — or in a system no one opens — cost roughly the same as no pathway at all. When utilization is low or untracked, the ROI of the clinical content investment is zero, and the variation the pathway was meant to address continues unchecked. The sunk cost of pathway development compounds annually.

Why the Underestimation Is Structural

Health systems don't systematically underestimate variation costs because they lack data. Most have significant data infrastructure. They underestimate because the measurement tools available to them — claims analysis, utilization reports, payer reconciliation — are retrospective, output-focused, and organizationally siloed.

Process costs sit in operations. Regulatory exposure sits in compliance. Physician attention is not measured at all. Pathway utilization is frequently unknown. When a finance team runs a variation analysis, they're drawing from a dataset that structurally excludes most of what variation actually costs.

The result is a number that's legible but incomplete. And incomplete numbers drive incomplete responses.

This matters practically. An organization that believes its variation problem costs $8 million will price its intervention accordingly — a targeted quality initiative, some new order sets, a committee. An organization that understands its variation problem costs $25 million — when compliance exposure, operational drag, and utilization failure are properly counted — makes a different investment decision. Often the right one.

Closing the Gap Requires More Than Pathways

The standard prescription for clinical variation is better clinical pathways: build more, keep them current, distribute them widely. That's right, as far as it goes. The evidence base is unambiguous — systematic use of locally-relevant clinical guidelines reduces length of stay, lowers complication rates, decreases inappropriate antibiotic prescribing, and improves adherence to evidence-based process measures. Hundreds of peer-reviewed studies make this case.

But pathway existence is not pathway utilization. And utilization — not the mere presence of content — is where the cost reduction actually happens.

Health systems that get this right are doing something specific: they're building governance infrastructure around their clinical content, not just the content itself. That means knowing who used which pathway, when, and with what result. It means ensuring that a new sepsis protocol is actually seen by the nurses and physicians making sepsis decisions — not filed somewhere and forgotten. It means connecting the clinical content to the moment of care, not the educational moment that happens three weeks before.

It also means that the AI now entering clinical decision support cannot be a black box. Clinicians — and the CMOs and medical directors who are accountable for their decisions — will not trust AI-generated guidance unless that guidance visibly reflects their own institution's standards: their formulary, their antibiogram, their approved protocols. The trust architecture matters as much as the technology. Generic AI recommendations, however accurate, are not the same as locally-governed recommendations that carry the authority of the institution's own leadership.

Getting the Number Right

For health systems serious about understanding what clinical variation actually costs, the exercise starts with an honest accounting of what's being left out of the current model. That means looking beyond claims data to include operational drag, compliance exposure, and pathway utilization rates alongside traditional utilization metrics.

It also means asking a harder question: not just what unwarranted variation costs today, but what the failure to address it costs over time, compounded against the regulatory and market trajectory moving toward value-based care.

Organizations that invest in real governance infrastructure for clinical standardization — one that encompasses content creation, maintenance, distribution, and utilization tracking — tend to find that the economics improve significantly when the full cost base is visible. The intervention that looked expensive against an $8 million problem looks different against a $25 million one.

The variation problem is not unsolvable. It's been systematically undersolved, in part because it's been systematically underpriced. Getting the number right is where it starts.

Curbside Health builds clinical pathway and guideline infrastructure for health systems — from AI-accelerated content creation to EHR-integrated decision support and utilization analytics.