For Whom the Noon Bell Tolls

The clock struck noon, and my patient was still in her room. My team sighed and my case manager rolled her eyes. Even if we had provided flawless clinical care throughout her hospital stay, she was now going to “count against us” in our quest to achieve our Discharge By Noon (DBN) goals – a metric thrust upon us with some explanation about throughput and how it will help patients stuck boarding in the emergency department, or something.

I must admit that I am emotionally torn about this goal. I do remember being in the hospital with my wife and newborn daughter, frustrated that we were ready to go home and we were left sitting and waiting… and waiting… and waiting… presumably for an intern to input an order and then a nurse to print a form. I would have loved if that medical team was doing all they could and holding their breath hoping that we would walk out the door before lunch was served.

The logic of DBN seems to be straightforward: overcrowding and boarding in the emergency department is an important patient safety and hospital efficiency issue, so achieving earlier discharges should free up those needed beds. However, there seems to be only one study published (in the October 2015 issue of Journal of Hospital Medicine) so far looking at the effect on emergency department throughput following a DBN program. After achieving a sustained DBN rate of 35%, this hospital found a statistically significant change in median arrival time of emergency department admissions to the floor from 5pm to 4pm. There is very weak evidence on other outcomes of DBN programs, such as how it affects patient and staff experience. Dr. Dan Shine, who turns out to be the retired chief of medicine from the hospital where the above study about DBN was published, called DBN an “urban legend,” in an editorial last year for the American Journal of Medicine, and he pointed out that in the absence of compelling evidence, the University HealthSystem Consortium (UHC) inexplicably describes 50% discharges before 11AM as “best practice.”

I have seen some really great changes in response to this metric, such as the institution of “tee time,” where the case managers meet with one of the senior members of the medical team (usually the senior resident or the attending) at around 3pm each day to discuss potential next-day discharges and any aspects of care coordination to still be sorted out. Arranging transportation to nursing facilities seems to have been expedited recently, which definitely helps since these discharges can oftentimes be days in the making so there is no reason these patients can’t leave early in the day.

I even am a cheerleader for hospital efficiency and a vocal proponent of high-value care. How could I possibly be against a program that advocates for discharging patients early in the day? Well, I am not. Not exactly. The problems with this metric for me are the unattended effects.

Every time I am on clinical service, I strive to see my hospital medicine team outperforming the others on our numerous “dashboard” metrics, including the one that seems to be getting the most attention lately from our administrators and chiefs: Discharge By Noon. In doing so, I have caught myself asking patients whether their husband could just get up early tomorrow and leave before dawn to beat the traffic over the Golden Gate Bridge. Oh, you were hoping to eat lunch before the drive home, let me see if we can get that for you to go.

As my boss, Dr. Bob Wachter, recently wrote in a fabulous article in the New York Times about how measurement is failing doctors, “…the focus on numbers has gone too far. We are hitting the targets, but missing the point.”

With DBN, my interns are more likely to break-off during rounds to finish up discharge orders, rather than listening and engaging in the care plans of other patients on the team. Basically, our working mornings – already busting at the seams with patient care and educational responsibilities – now oftentimes feels even more compressed, lest all our patients turn into pumpkins when the clock strikes noon.

There is one other unintended effect that may be even more insidious.

A number of my hospitalist colleagues at UCSF recently analyzed nearly 40,000 hospitalizations over the course of three years at our hospital. The findings, published ahead-of-print in December in the Journal of Hospital Medicine, found that DBN seemed to be associated with a longer length of stay. After adjusting for a number of factors, Rajkomar and colleagues showed that patients who were admitted emergently to the hospital had approximately a 12-hour increase in length of stay if they ended up being discharged before noon.

Could it be that we were now actually keeping people overnight to discharge them in the morning? Well, even if not done maliciously, measurement sure does have a way of tipping fence-straddling decisions in one consistent direction. Of course, this retrospective study could only determine association, which as we all know cannot imply causation. But, the explanation sure seems plausible.

So what should we do?

The authors conclude that their study provides “a rationale for alternate approaches to measuring an early discharge program’s effectiveness,” such as using “multiple evaluation metrics including the effect on emergency department wait times, ICU or PACU transitions, and on patient reported experiences of care transitions.”

And we certainly need more studies examining the outcomes and unintended effects of instituting comprehensive DBN programs.

Perhaps, in closing, we can quote our father of modern medicine, Sir William Osler: “Medical care must be provided with the utmost efficiency. To do less is a disservice to those we treat, and an injustice to those we might have treated.”

Maybe this is just as true at 7 pm as it is at 7 am as it is at 12:01 pm.