Case Study 1 — October 1, 1983: The Day Medicare Stopped Buying Days

A real, public case — the adoption of the Inpatient Prospective Payment System — told from the documented record: statute, rulemaking, and the published research literature. Where the literature's findings are summarized, they are summarized qualitatively; this case study asserts no precise statistic, and any figure you need should come from the primary sources in the further reading.

Background

When Medicare was created in 1965, it inherited the payment method the insurance industry then used for hospitals: reasonable cost reimbursement. A hospital cared for a beneficiary, accounted for what the care had cost, and was paid accordingly, subject to rules about which costs were allowable.

The design was not foolish — in 1965 the urgent problem was persuading hospitals to participate at all — but its incentive structure is visible from one sentence: every additional day, test, and service generated additional reimbursement, and no actor in the system was rewarded for spending less. Through the 1970s, hospital spending grew faster than the economy, faster than the program's funding, and faster than Congress could tolerate. Cost-control attempts of the era — planning requirements, voluntary restraint, the cost limits of the Tax Equity and Fiscal Responsibility Act of 1982 (TEFRA) — adjusted the old machine rather than replacing it.

The replacement had been growing in an unexpected place: a research program at Yale associated with Robert Fetter and John Thompson, which set out in the late 1960s and 1970s to answer an industrial engineer's question about hospitals — what, exactly, is the product? Their answer was the diagnosis-related group: hospital stays can be classified into groups that are clinically coherent and statistically similar in resource use, and once you can classify the product, you can price it. New Jersey put the idea to work in a state rate-setting experiment beginning in the late 1970s, paying hospitals by DRG, and the federal government watched.

The issue

In the Social Security Amendments of 1983 — moving with a speed almost nothing in Medicare's history has matched, attached to must-pass Social Security legislation — Congress replaced cost reimbursement for acute care hospital operating payments with the Inpatient Prospective Payment System, effective for cost reporting periods beginning October 1, 1983, phased in over several years from hospital-specific rates toward national ones.

The design choice at the center is the one Chapter 33 teaches: the unit of purchase became the classified stay. A fixed payment per DRG, set in advance, regardless of the days and services actually consumed. A hospital that treated the patient for less than the payment kept the difference; a hospital that spent more absorbed the loss. For the first time, the payment system itself pushed back.

What happened

The documented effects arrived quickly, and they are exactly what the incentive analysis predicts — which is the case study's first lesson in itself.

Length of stay fell, substantially and immediately. The direction is undisputed across the literature; average Medicare stays shortened markedly in the first years of prospective payment, after years of much slower decline. Days had stopped being revenue and started being cost.

Admissions did not surge. A widely held fear — that hospitals paid per case would manufacture cases — did not materialize in the aggregate; Medicare admissions actually declined in the early PPS years, a movement usually attributed in part to the simultaneous growth of outpatient alternatives and utilization review.

Care moved to the settings the system did not yet reach. Outpatient departments, skilled nursing facilities, home health, rehabilitation — the sectors outside the fixed payment grew rapidly, and patients were discharged to them earlier. The phrase of the era, from the congressional hearings and the press, was "quicker and sicker."

Quality was studied, and the findings were mixed in an instructive way. The major evaluations of the era — including a well-known body of RAND Corporation research — found, broadly, that measured quality of in-hospital care did not deteriorate the way critics feared, while also documenting an increase in patients discharged in unstable condition. Both findings can be true at once, and the honest summary is the one this book keeps giving about payment policy: the system changed behavior exactly where it created incentives, in both directions.

And the classification itself became an object of attention — because when a classification prices the product, describing the product becomes a financial act. That thread is Case Study 2's.

The outcome

IPPS survived, and more than survived: it became the template. Prospective, classification-based payment spread to hospital outpatient care (the APCs of Chapter 34), skilled nursing, home health, inpatient rehabilitation and psychiatry — and, by license and imitation, into commercial and Medicaid inpatient contracts. The MS-DRG refinement of 2008 rebuilt the severity tiers this chapter teaches. Forty years on, the 1983 architecture — a coded record, a grouper, a weight, a rate — is simply how American hospital payment works.

What it shows

Payment design is behavioral design. Every effect above was produced without a single regulation telling a hospital how to practice medicine. The payment system changed what a day, a test, and a discharge were worth, and behavior followed. A reader of this book should recognize the pattern at every scale — it is Chapter 22's coverage rules and Chapter 29's denial economics, nationalized.

Classification is where the money enters. The 1983 bargain made the coded record the invoice. Everything Chapter 33 teaches — principal diagnosis discipline, CC/MCC capture, POA integrity — exists because of this design, and so does the CDI profession, the DRG audit industry, and §33.10's \$1,867.44.

And a payment system's boundary is a pressure gradient. Care flowed to the settings the fixed payment did not reach, and policy has chased that movement ever since — the transfer rule of §33.8 and the three-day payment window of Chapter 34 are both boundary patches on the 1983 design. When you meet a strange-looking payment rule, ask what boundary it is patching; the answer usually dates to this case.

Discussion questions

1. Cost reimbursement rewarded spending; prospective payment rewards economy. Neither incentive is neutral. For each system, name the patient-facing risk it creates and the control that exists to check it.

2. "Quicker and sicker" was a real finding and so was "measured quality broadly held." Reconcile the two, and explain what each was measuring. What does the pair teach about evaluating any payment change from a single metric?

3. The DRG made the coded record the invoice. Trace one specific way that fact changed the coder's professional exposure between 1982 and 1984 — and connect it to the compliance framing Chapter 5 built.

4. The transfer rule (§33.8) exists because a fixed payment per stay creates an incentive at the stay's boundary. State the incentive precisely, and explain why the rule's per-diem design — rather than a simple prohibition — is the answer a payment engineer would choose.

5. New Jersey ran the experiment before the nation adopted it. What did the pilot's existence contribute to the 1983 decision, and what is the general lesson for anyone proposing a payment or process change inside their own organization — including the reader of Chapter 29 §29.9?