Readmission penalties dodged by placing patients on observation status

Quality Improvement: ‘Become Good At Cheating And You Never Need To Become Good At Anything Else’

By David Himmelstein and Steffie Woolhandler
Health Affairs Blog, August 27, 2015

The Centers for Medicare and Medicaid Services (CMS) has trumpeted the recent drop in hospital readmissions among Medicare patients as a major advance for patient safety. But lost amidst the celebration is the fact that hospitals are increasingly “observing” patients (or treating returning patients in the emergency department) rather than “readmitting” them. But while re-labeling helps hospitals meet CMS’ quality standards (and avoid costly fines), it probably signals little real quality gain and often leaves patients worse off financially.

Observation Status

In most cases, observation patients receive care in a regular inpatient unit, and get treated just like other inpatients. And in many cases, observation stays stretch out to several days: in 2012, 26 percent lasted two nights and 11 percent at least three. But from Medicare’s point of view, this is outpatient care, which leaves patients responsible for more of the bill, and ineligible for Medicare-paid rehab or skilled nursing care.

Hospitals started designating more stays as “observation” after Medicare’s auditors began disallowing the entire payment for some brief hospital “admissions.” Even though “observation stays” pay less than inpatient admissions, hospitals took a better safe than sorry approach, classifying many brief stays as “observation.” Between 2006 and 2013, observation stays increased by 96 percent, accounting for more than half of the apparent decline in total Medicare admissions during that seven-year period.

Observation Classification

Medicare’s recent adoption of penalties for readmissions offered hospitals a new incentive to shift some patients returning within 30 days of their discharge to observation status. A patient stay labeled “observation” doesn’t count as a readmission, allowing hospitals that might otherwise be fined for having too many readmissions to skirt the penalty.

About 10 percent of all hospital stays occurring within 30 days of discharge are now classified as “observation;” a quarter of hospitals classified 14.3 percent or more of all repeat stays as “observation.” Moreover, analysis of time trends in observation stays makes it clear that they account for a significant chunk of the reduction in readmissions. Between 2010 and 2013, 36 percent of the claimed decrease in readmissions was actually just a shift to observation stays.

Emergency Department Use

And it’s not just observation stays that have been on the increase. More of the recently discharged patients are being treated in emergency departments (EDs) — without being admitted — as well.

Factoring in the 0.4 percent increase in ED visits within 30 days of discharge, the fall in the percent of discharged patients returning to hospitals for urgent problems is only 0.3 percent over the past three years — less than one-third of the improvement that CMS claims. And even this 0.3 percent overall fall may be partly an artifact of hospitals’ “upcoding” (exaggerating the severity of patients’ illnesses), which boosts diagnosis-related group (DRG) payments, and could also corrupt the formula used to risk-adjust expected readmission rates.

Medicare’s readmission penalties are among the growing number of pay-for-performance (P4P) and value-based purchasing initiatives that offer bonuses to high performers and/or penalize the laggards. We previously pointed out that the evidence for this carrot and stick approach is unconvincing. More recently, a long-term follow-up of the English hospital P4P program found that P4P generated no improvement in patient outcomes, damping the enthusiasm generated by the rosy short-term findings, and reinforcing the need for skepticism.

Adopting unproven everywhere P4P strategies that have been proven nowhere risks quality failure on a monumental scale. It pressures hospitals to cheat, saps doctors’ and nurses’ intrinsic motivation to do good work even when no one is looking, and corrupts the data vital for quality improvement.

As the graffiti artist Banksy once said: “Become good at cheating and you never need to become good at anything else.”



By Don McCanne, MD

In lieu of adopting comprehensive health care financing that actually would improve value - a single payer system - our national leaders have elected to continue with our current dysfunctional system and try to make it work. One measure that has been introduced is the assessment of penalties for patients who are readmitted within 30 days of being discharged from a hospital - on the theory that the patients should be fully stabilized at discharge with followup arranged that would prevent the need for readmission.

Physician and hospital administrators do not want patients readmitted soon after discharge. Their efforts are directed at providing the best care appropriate to improve patient outcomes. But there are some clinical conditions that are inherently unstable, or that can have unavoidable complications, that can result in the need for readmission. Also, in spite of outreach efforts, socioeconomic variables can result in destabilization of the patient’s condition. Some readmissions are absolutely inevitable.

This report is important because it shows that the improvement in lowering readmission rates that occurred after the introduction of penalties is not so much due to improved inpatient and post-discharge management, but rather is due to gaming rules that are characteristic of pay-for-performance (P4P) schemes.

Specifically, there has been a dramatic increase in placing patients on observation status instead of formally admitting them to the hospital. The care may be identical - provided in the inpatient service departments - but by not certifying the patient as being a formal readmission within 30 days of the last admission, the penalty is avoided. Medicare comes out ahead since outpatient services are priced lower than inpatient services, even if they are identical services - thus the trumpeting by CMS of another success in their reform efforts. But the patient loses since the cost sharing requirements for outpatient services are much higher than they are for inpatient. But then who ever said that health care reform was to benefit patients?

A variation of this gaming is to manage the followup hospitalization completely within the emergency department, perhaps even holding the patient overnight. The result is essentially the same as holding them on observation status.

This P4P-type gamesmanship is really a form of cheating. But it works. Instead of shifting to a much more efficient financing infrastructure - a single payer system - we can continue to follow the Banksy dictum - “Become good at cheating and you never need to become good at anything else.”

Maybe we should start to think about becoming good at financing health care instead.