There are tens of thousands of policies in Medicare’s policy manual, which makes for stiff competition for the “Most Maddening” award. But my vote goes to the policy around “observation status,” which is crazy-making for patients, administrators, and physicians.

“Obs status” began life as Medicare’s way of characterizing those patients who needed a little more time after their ED stay to sort out whether they truly needed admission. In many hospitals, “obs units” sprung up to care for such patients – a few beds in a room adjacent to the ED where the patients could get another nebulizer treatment or bag of saline to see if they might be able to go home.

Giving the hospital a full DRG payment for an inpatient admission seemed wrong, and yet these patients really weren’t outpatients either. The Center for Medicare & Medicaid Services’ (CMS’s) original definition of obs status spoke to the specific needs of these just-a-few-more-hours patients: a “well-defined set of specific, clinically appropriate services,” usually lasting less than 24 hours. Only in “rare and exceptional cases,” they continued, should it last more than 48 hours.

A recent article in JAMA Internal Medicine, written by a team from the University of Wisconsin, vividly illustrates how far the policy has veered from its sensible origins. Chronicling all admissions over an 18-month period, Ann Sheehy and colleagues found that observation status was anything but rare, well defined, or brief. Fully one in ten hospital stays were characterized as observation. The mean length of these stays was 33 hours; 17 percent of them were for more than 48 hours. And “well defined?” Not with 1,141 distinct observation codes.

To underscore just how arbitrary the rules regarding observation are, an investigation by the Inspector General of the U.S. Department of Health and Human Services released today found that “obs patients” and “inpatients” were clinically indistinguishable. Their major difference: which hospital they happened to be admitted to.

The potion that turned this particular policy into a monster was the Recovery Audit Contractor (“RAC”) audits, whose existence was authorized by the 2003 Medicare Prescription Drug Act. RAC auditors can target a hospital, pull a hundred or so charts, and, if they find improper billing, collect a bounty for every dollar they save CMS. With the determination of obs status so amorphous, hospital administrators have adopted a “better safe than sorry” stance, generally deciding that cases that are anywhere near a close call should be called obs. (Just this week, Beth Israel Deaconess Medical Center in Boston forked over $5.3 million to Medicare to settle charges related to admissions that auditors believed were really obs.) The result of all this angsty wheel-spinning: the number of obs cases in the U.S. went up by 50 percent between 2006 and 2011, with a more-than-400 percent (!) increase in Medicare patients staying more than 48 hours under observation.

If this mess were only about the question of money for Medicare, hospitals, and auditors, it would be plenty maddening but not miasmal. Unfortunately, patients and their families are unwitting victims, collateral damage. Picture this: your mother is sitting in a hospital bed, with a band on her wrist, an IV in her arm, nasal prongs in her nose, and EKG squiggles skipping across a telemetry monitor. Luckily, she does reasonably well and is discharged to a skilled nursing facility after a three-day stay. OK, your family thinks, at least we know that Medicare will pay for the SNF since she’s crossed CMS’s magic three-day threshold to trigger SNF coverage.

Only later do you learn that her hospital stay doesn’t count, because she was on obs the whole time. Or you get a co-pay bill for several thousand dollars because, while inpatient medications are covered under Medicare, “outpatient” medications are not. While she sat in her hospital bed, you see, she was really an outpatient.

In my editorial accompanying the Wisconsin paper, I cite the case of a 78-year-old woman who received a $16,000 bill for an uncovered nursing home stay following a four-day observation stay in the hospital. “I thought it was surely a mistake,” she said. “Nobody ever said I wasn’t admitted.” In a brochure that could have been written by Franz Kafka, Medicare tries to explain the unexplainable. “REMEMBER,” it says, the capital letters designed to make you fully alert for the nonsense that follows, “Even if you stay overnight in a regular hospital bed, you might be an outpatient.” Huh?

I hope you’ll take a look at both the Wisconsin paper and my editorial. In this blog, I’d like to extend the discussion to the fix, which didn’t seem like a bad idea to me when I first penned the editorial, but which I’m coming to learn might – hard as it is to believe – make things worse.

On April 16, 2013, CMS proposed a new rule that would, using a time-based criterion, clarify which patients should be on observation. Patients who stay for less than two midnights (“one Medicare day”) will be assumed to be obs; those staying longer than two midnights would be inpatients. Medicare apparently believes that turning all these two-day or more stays from obs into inpatient admissions will cost the agency money, since they have proposed a 0.2% cut in hospital reimbursement to compensate for these projected increased costs.

Yet CMS has not released any financial models that might help predict what the impact of this change is likely to be. And, whatever Medicare’s projections may be, many hospital administrators believe that the new policy will cost them huge amounts of money. How can this be? While hospitals will now receive a full DRG payment for some longer-stay patients who previously would have been observation (a win), these administrators believe that their losses – particularly on short-stay surgical patients who will now be observation, despite high resource use – will far outstrip the gains. I’ve seen some preliminary data that supports their fear.

Moreover, there is a widespread panic that, rather than soothing the RAC auditors, the new policy will be fresh meat for them. After all, with the two-midnight rule, it’s likely that RAC auditors will be suspicious (potentially with some justification) that hospitals will keep some patients an extra midnight. You might call it an inverse Cinderella effect, as patients are not sent home before midnight in order to capture an inpatient reimbursement and spare the patient the costs associated with an observation stay. So the unproductive and maddening cat-and-mouse game will continue, albeit within slightly less arbitrary boundaries.

A watchdog organization, the Center for Medicare Advocacy, recently highlighted this concern. They noted that the new policy calls for an inpatient admission when …

… the patient is expected to be in the hospital for at least two midnights. The physician “expectation [of a two-midnight stay]. . . should be based on such complex medical factors as patient history and comorbidities, the severity of signs and symptoms, current medical needs, and the risk of an adverse event.” These factors “must be documented in the medical record in order to be granted consideration.” In other words, the physician certification that he or she expects the patient will need to stay for more than two midnights is not enough: [RAC auditors] can scour the medical record and if they don’t see evidence, they can deem it an improper hospitalization.

Richard Rohr, a seasoned hospitalist leader who is now a consultant, echoed this fear, adding in a note to me:

The basic problem with observation stays is the disconnect between functional status and medical necessity. Many elderly patients who are not able to care for themselves and need more help than a family can perform come to the hospital because it is the social service agency that is open at night and on weekends and does not turn anyone away. These patients often do not have medical needs as defined [by Medicare]. Having stepped away from clinical hospital medicine to focus on medical necessity work, I talk regularly with hospitalists and other physicians, who struggle with the distinction between functional needs and medical needs in caring for patients.

A rather dense (and, at times, impenetrable) editorial the New England Journal of Medicine also questions whether the proposed policy is an advance. The authors don’t think so. They recommend that patient co-pays be capped, that the costs of medications the patient is already on at home be covered, and that obs days count toward the 3-day requirement for SNF eligibility.

While these are reasonable recommendations, they don’t go far enough to stem the madness. Rather, the line of the old song comes to mind: “Let’s call the whole thing off.” It’s time to restore obs to its original meaning. Medicare should develop a new payment code for those patients who need several, perhaps up to 24, hours of very specific therapy, in a physical observation unit, to determine whether they need admission: nebulizers, fluids, maybe a unit or two of blood. For everyone else admitted to a regular bed on a hospital ward, they are (and it seems silly to have to say this) admitted to the hospital, and the reimbursement system should reflect this. Utilization Review can look to see if there was medical or social justification for admission – if not, the day or days can be denied. Isn’t that simpler?

While the policy around obs is important and frustrating, there are larger issues at play. As I wrote in my editorial, Medicare is in the process of transforming itself from a “dumb payer” into an active shaper of the medical marketplace, through policies such as public reporting, “no pay for errors,” and value-based purchasing. On balance, that’s a good thing. But when Battleship CMS turns, it sends out giant wakes, some of which are unanticipated, even by the organization itself. Policies like the one related to obs status risk capsizing a lot of boats.

Whether the problems with the new observation policy will represent unintended consequences, a lackluster analysis, or a purposeful cost-cutting strategy will doubtless be in the eyes of the beholder. But whatever the motivation, it will be crucial for interested parties (AHA, SHM, and others) to thoroughly vet the proposed changes and push back hard on CMS if they truly are wanting. It’s a sad reality that CMS may well be in the process of turning a vague but maddening policy into one that is less vague but equally problematic.

Bob Wachter is professor of medicine, University of California, San Francisco. He coined the term “hospitalist” and is one of the nation’s leading experts in health care quality and patient safety. He is author of Understanding Patient Safety, Second Edition, and blogs at Wachter’s World, where this post originally appeared.