Why can’t long-staying hospital patients leave?

by Mike Miesen

Some patients stay at the hospital far longer than their condition would predict. Why?

At first glance, the answer’s easy: because they’re sicker. Duh.

But spend some time in the bowels of a hospital – the administrative and operational departments most patients rarely see – and you’ll come away with a different answer: it’s almost never because the patient needs to be in the bed any longer.

A new study, published in JAMA Surgery, backs this up and shows what every administrator already knows: the system is sicker than the patients are.

The study looked at trauma patients with a hospital length of stay (LOS) – a time from admission to discharge – two standard deviations above their diagnosis related group (DRG) average.

Or, in English: they looked at the 2.5% of trauma patients that stayed in the hospital the longest. This is a common insurance metric, which the authors term “Excessively prolonged hospitalization,” or ExProH. Even for academics that’s a pretty clunky acronym, so let’s just go with “LOS Long-Tailers.”

Of those patients, only one in five’s excess stay was due to a clinical issue. Almost half were delayed due to difficulty in rehab placement, and one-fourth were due to in-hospital operational delays.

mm827-1Trauma patients often need time post-hospitalization to recover at a facility; they’re healthy enough that a hospital bed is overkill, but not so healthy that discharging home is a tenable option. Depending on their condition and needs, they may spend time at a long-term acute care facility (LTAC), a skilled nursing facility (SNF), a nursing home, or a rehabilitation center.

Usually, a hospital “case manager” (an experienced nurse, often with special certification) is responsible for arranging this stay, a process which almost always requires a good deal of faxing, phoning, and negotiating. If a patient needs a specific rehab facility – maybe they’re restricted by condition, geography, or payer – they have to wait in the hospital bed until the facility has space for them.

The study showed that only independent predictors of whether a trauma patient will be a LOS Long-Tailer are if he/she discharges to a rehab facility or other post-acute facility, and if he/she has Medicare/Medicaid or is self-pay. (The data here doesn’t split up Medicare and Medicaid, which is odd. Since they’re older, often sicker, patients, you might expect a Medicare beneficiary to be a more likely LOS Long-Tailer than Medicaid patients.)

Conceptually, this makes sense. Just like primary care physicians, rehab facilities are probably less likely to accept Medicaid recipients or the uninsured [1].

LOS Long-Tailers are much more likely to have Medicare/Medicaid or no insurance (p-values .002 and .02, respectively)

LOS Long-Tailers are much more likely to have Medicare/Medicaid (p≤0.002) or no insurance (p≤0.02)

This situation is basically sub-optimal for everyone involved. Because they stay longer than necessary in a hospital bed, patients are at higher risk of an adverse incident, and will leave with a much heftier hospital bill.

Hospitals often lose money on these patients, as they require fewer tests and procedures on the final waiting days. If a hospital could otherwise fill the bed with a new patient (“backfill demand”), they also lose out on the revenue from incoming high-dollar patients.

Of course, “backfill demand” is a euphemism for uncomfortable patients boarding in an Emergency Department bed, a situation which is associated with  a longer length of stay in the hospital when they do get admitted.

Insurers – including Medicare and Medicaid – pay more for each day of unnecessary hospitalization, and costs trickle down to the rest of us via higher premiums or taxes [2].

As the researchers note, trauma LOS Long-Tailers are expensive; at 5% of total admissions, they “account for approximately 70% of the unnecessary hospital cost.” Though much too optimistic, the authors note that if all non-clinical trauma delays were eliminated, Massachusetts would save about $130 million per year.

The obvious solution? More rehab beds to meet demand. But hospitals are loathe to turn high-dollar inpatient units into medium-dollar rehab units, and private companies may not want to make the long-term investment in a new rehab building.

And anyway, it’s too simple to blame a lack of rehab beds for all of the placement issues. In addition to the 26% designated as “in-hospital operational delays,” I’d bet that a good portion of the rehab delay could be ameliorated by proactivity and better communication amongst the care team.

Setting up a post-acute placement takes time, so case managers need to be proactive and get started with referrals days before a patient is expected to discharge – easier said than done for over-cased workers who spend most of the day putting out fires. Still, workflow modifications and system tweaks can help case managers be more proactive, which cuts down on rehab delays.

To even know when a patient is expected to discharge requires consistent care team communication that is often lacking; just getting all of the care team in a room at the same time is a gargantuan task [3]. When communication is poor, things get dropped; improving it can help case managers plan ahead and get patients out faster.

There are some obvious limitations to this study: service specificity, geography, and data bias. Trauma patients have a high incidence of post-acute referral, so it makes sense that LOS Long-Tailers are disproportionately delayed because of rehab delays. This study doesn’t help us understand what the specific delays are for LOS Long-Tailers under other services.


Geography also plays a significant role with post-acute referrals. The Boston area may have more, or fewer, rehab beds than other markets, so this study doesn’t help us understand if rehab placement would be such a problem in, say, Minneapolis.

Finally, check out the delay list at right; you’ll notice that “Case Manager Error” isn’t one of the codes used. Only case managers enter a delay code, and they’re customizable (read: political), so there’s some inherent bias in the coding.

But, on the whole, it’s a good proof of concept study that confirms something all administrators and clinicians know: the system causes all sorts of delays, some of which can be reduced or removed, others that must just be dealt with.


1.Anecdotally, I saw this constantly when I worked in hospitals as an operations consultant.

2. You could make a case that physicians and nurses are marginally better off with LOS Long-Tailers staying, for a number of reasons –- but because I don’t want hate mail and because it’s beside the point, I’ll leave that argument for someone else to make.

3. Speaking from experience.


Mike is a healthcare consultant turned aid worker turned traveler (currently: East Africa) and freelance journalist. Follow him on Twitter @MikeMiesen or subscribe to the blog.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: