Healthcare Administrative Costs: Read the Estimate Before Building the Business Case
National administrative-cost estimates answer different questions. Compare years, populations and definitions before deciding how they can inform a local workflow business case.
The most useful next step for an executive is a local baseline tied to a defined workflow. Count the work, observe the staff effort, identify the costs that would actually change and make the assumptions visible. That creates a decision a finance team can test.
Ask what the estimate is measuring
An administrative activity may be necessary even when its execution can improve. Paying staff to resolve a record discrepancy is a cost; identifying how much of that work could be avoided requires another analysis. A spending total is not automatically an addressable-savings total.
Use four distinct labels in the evidence record:
| Label | Question it answers | Mistake to avoid |
|---|---|---|
| Administrative spending | What cost falls within the author's definition? | Treating every dollar as waste |
| Estimated waste | What spending does the analysis classify as unnecessary or inefficient? | Treating the estimate as a current audited budget |
| Cost avoidance | What cost is estimated not to have occurred under a comparison scenario? | Calling it new cash available to one organization |
| Savings opportunity | What further reduction does a model suggest is possible? | Reporting it as an achieved result |
When a presentation switches between these labels, ask for the source definition before comparing the numbers. A familiar headline is not sufficient provenance.
Read the 2019 waste estimate in its original context
The often-cited JAMA analysis by Shrank, Rogstad and Parekh was published in 2019. It synthesized estimates across six categories of healthcare waste, using annual national estimates in 2019 dollars where conversion was necessary. Administrative complexity was estimated at $265.6 billion.
That figure is a historical estimate of waste within the authors' defined category, not a measured 2026 total for all healthcare administration. The review also reported that it did not identify intervention studies targeting administrative complexity. Its estimated savings from interventions therefore did not supply a demonstrated recovery rate for that category.
Use the study to explain why administrative complexity deserves attention. Do not divide its estimate by the number of US practices, apply a vendor automation percentage, or present the result as a forecast for a clinic. Those steps introduce assumptions the study did not establish.
The practical source ledger should retain the study year even when a slide deck uses a rounded figure. Rounding cannot turn a historical estimate into a current measured cost.
Distinguish transaction estimates from broader waste estimates
The 2025 CAQH Index announcement, now hosted by DataSpring, was published in February 2026 and describes 2024 activity. It reports an estimated $258 billion in administrative costs avoided through electronic transactions and improved data exchange, alongside a remaining $21 billion savings opportunity from further automation of manual and partially manual transactions.
These are different measures within a transaction-focused analysis. They are not alternative current measurements of the JAMA administrative-complexity estimate, and they should not be added to it.
DataSpring's Index methodology FAQ explains that participating plans report transaction counts and costs, while participating providers report transaction cost and time. The report covers defined administrative transactions. A transaction estimate should not be silently relabelled as the cost of an entire patient episode or every activity in a referral team.
The current Index report page identifies the 2025 edition and its methodology resources. For a transaction-level comparison, read the relevant edition's definitions before using a per-transaction figure. A public headline does not supply the complete method for every local calculation.
| Source | Time and scope | Appropriate use | Unsupported extrapolation |
|---|---|---|---|
| Shrank and colleagues, JAMA | 2019 synthesis of national waste estimates | Historical context for administrative complexity | A practice-level savings forecast |
| 2025 CAQH Index announcement | 2024 activity in the Index's administrative transaction scope | Context for cost avoidance and remaining modeled opportunity | Total administrative waste recoverable by one platform |
| Local workflow record | Your defined period, team and included work | A baseline for a bounded operating decision | Generalizing to workflows or sites not measured |
If your business case currently starts with a national total, book a demo to map the local coordination workflow underneath it. Bring the available volume and time records so assumptions can be separated from observations.
Build a local baseline without double counting
Choose one unit of work. It might be a referral episode, an authorization request or an appointment event. Keep sub-tasks linked to that unit so a status call and a document correction are not mistaken for two complete episodes.
Observe the work across a defined period and record the measurement method. Separate active staff effort from elapsed waiting time. A request can wait several days without consuming several days of staff labor. Conversely, a short episode can involve repeated touches by several people.
Use this worksheet with locally verified inputs. It intentionally contains no assumed staffing, salary or conversion figures.
| Input | Evidence to record | Double-counting check |
|---|---|---|
| Included work units | Source report, period and exclusions | One episode counted once |
| Staff effort | Observed time or documented sampling method, by role | Shared tasks allocated once |
| Labor cost basis | Finance-approved rate and included components | Benefits or overhead not added twice |
| Current external fees | Invoices and allocation method | Fees not already inside another cost line |
| Proposed operating cost | Written scope and recurring charges | Usage and support assumptions explicit |
| Implementation effort | Named internal tasks and approved estimates | One-time costs separated from recurring costs |
| Exceptions after change | Defined human work expected to remain | Residual effort not assumed to disappear |
Use the manual prior authorization cost guide for that narrower workflow. Referral-specific scenarios belong in the referral automation ROI calculator, with its assumptions inspected against your own inputs.
Separate capacity, cash and service outcomes
Time released from repetitive work creates capacity. Whether it reduces cash expenditure depends on what happens next. If staff use the time for unresolved cases, the organization may gain useful capacity without a payroll reduction. State that benefit in its own terms.
The same discipline applies to additional appointment activity. More scheduled visits are not automatically additional collected revenue. Any financial scenario needs finance-approved assumptions about attendance, reimbursement and incremental costs. Leave those assumptions blank when they are not known.
Build separate lines for observed baseline, proposed change, residual work, recurring cost and one-time effort. Show which values are measured and which are estimates. If a range is useful, derive its endpoints from defensible assumptions rather than choosing a range that makes the investment attractive.
The related guide to ROI in AI referral automation addresses that evaluation. National research can explain the wider problem while the local evidence determines the business case.
Where Linear Health fits
Linear Health provides operational AI for healthcare. A useful evaluation starts with the coordination task, its current evidence and the human exceptions that will remain. Ask for a workflow demonstration against that scope before translating a product claim into a financial model.
Book a demo to examine one administrative workflow and its baseline. The outcome should be a clearer scope and a list of inputs your operations and finance teams can verify.
Frequently asked questions
What counts as healthcare administrative cost?
How does waste differ from all administration spending?
Can a national estimate predict one practice's savings?
Why do administrative-cost estimates disagree?
What local evidence supports an automation case?
Sources
Sources accessed October 14, 2026. The JAMA abstract and the DataSpring pages describe their own methods and figures.