How Hospitals Should Evaluate Whether to Expand a Service Line

By Two Pillars & Co Advisory Group · Sources checked October 2026

What question comes first when considering expansion?

Expansion decisions are often framed as “should we add a cardiology program?” or “should we open more orthopedic capacity?” The better first question is: what problem does the expansion solve, and for whom? A service line expands successfully when there is unmet or misdirected demand, the hospital has a credible way to capture it, and the volume pays enough to cover its costs. If any of the three is missing, the expansion is a risk.

This matters because expansion commits money, space and people for years. A new service that underperforms is hard to reverse, and the staff and equipment may not be easily redeployed.

How do you assess demand?

Start with the market. The Dartmouth Atlas of Health Care defines hospital service areas as local health care markets for hospital care, made up of ZIP codes whose residents receive most of their hospitalizations from the hospitals in that area, and hospital referral regions as regional markets for tertiary medical care, each containing at least one hospital that performs major cardiovascular procedures and neurosurgery.^[1] Those definitions give a practical way to size demand for different service types: local geography for common services, regional geography for tertiary ones.

Then look at where residents currently go. Public discharge data, such as the Healthcare Cost and Utilization Project’s all-payer, encounter-level databases, which begin in 1988 and are compiled from state data organizations, can show how many patients from an area received a service and at which hospitals.^[4] In New York, the SPARCS public use files serve a similar purpose. The question is not just how large demand is, but how much of it already leaves the area, which is the subject of the outmigration articles in this series.

How do you test whether the hospital can capture it?

Capture depends on referral patterns, clinical reputation, access and capacity. Referral patterns determine whether physicians in the area will send patients. The Centers for Medicare & Medicaid Services publishes Physician Shared Patient Patterns files for 2009 through 2015, at 30-, 60-, 90- and 180-day intervals, which can show which physicians have patients in common, although they do not by themselves prove referral relationships.^[5] Local intelligence from physicians, payers and community leaders fills in what the data cannot.

Capacity includes beds, operating rooms and equipment, but staff is usually the binding constraint. A service line that needs specialized nurses, technologists or physicians the hospital cannot recruit will not reach its volume targets however strong demand is. Test supply realistically before committing.

Will the volume pay?

Build a contribution model with the payer mix expected from the new volume. The payer mix is critical because payment varies. MedPAC reported that fee-for-service Medicare payments were well below hospitals’ costs in aggregate in FY 2024, with a Medicare margin of −12.1 percent, while hospitals’ all-payer operating margin was 6.5 percent.^[2] A line that depends mainly on Medicare patients needs unusually low costs to be profitable, and a line with strong commercial volume may subsidize others.

Include step-fixed costs: if the expansion needs a new nursing unit or an additional operating room team, those costs arrive before the volume does. Then run a sensitivity test with lower volume, slower ramp-up and a payer mix shifted toward lower-paying sources. The decision should survive the pessimistic case, or at least the hospital should know the size of the loss it is accepting.

Costing method matters. Kaplan and Porter argue in the Harvard Business Review that health care organizations tend to measure the wrong things the wrong way, including costs at the department rather than the patient level.^[3] An expansion case built on a department’s average cost may overstate or understate the actual cost of the new patients.

What does a good expansion case look like?

A good case states the demand estimate and its source, the share the hospital expects to capture and why, the staffing plan, the contribution model with base, low and high scenarios, the capital required, the break-even volume and the date by which results will be reviewed. It also states the exit: what would trigger a pause or a reversal.

Sizing the downside is as important as sizing the upside. A line with a modest expected return but a manageable loss in the worst case can be better than one with a high return and an unbounded loss. The interactive example below lets you vary volume, payer mix and cost to see how contribution changes. It uses invented figures.

Frequently asked questions

What to do next

1. Define the problem the expansion solves and for whom.
2. Size demand and current leakage using public data.
3. Build a contribution model with step-fixed costs and a pessimistic scenario.
4. Set a review date and an exit trigger.