ED Observation Units: The Physician Guide to Reducing Admits and Readmissions

Your emergency department just admitted a 45-year-old woman with atypical chest pain, a normal ECG, and a first troponin of less than the 99th percentile. She is going to occupy an inpatient bed for 18 to 24 hours while cardiology completes a rule-out. The hospitalist is irritated because she does not meet inpatient criteria. Case management is already flagging it for potential denial. And somewhere in your department, two ambulances are on divert because you do not have a bed to put them in. This is not a clinical problem. It is a systems problem. And the ED observation unit is the most underutilized solution in emergency medicine today. I have built observation programs in both academic and community settings, and I can tell you the three things most departments get wrong: patient selection, protocol standardization, and the financial model that makes the whole thing sustainable.

Why Most EDs Are Hemorrhaging Revenue Without an Obs Unit

The average emergency department admits between 15% and 25% of its patients. Within that admitted population, studies consistently show that 20% to 30% of those admissions could have been safely managed in an observation medicine protocol — a structured, time-limited clinical pathway that avoids the inpatient designation entirely.

Let me put that in financial terms. A typical inpatient admission for chest pain observation generates revenue that is frequently clawed back by payers who determine — often retrospectively — that the admission did not meet medical necessity criteria. The denial rate for short-stay admissions has climbed steadily, and CMS continues to tighten the two-midnight rule enforcement. Every denied claim is not just lost revenue; it is administrative cost spent on appeals that succeed less than half the time.

An ED observation unit flips this equation. Observation status is billed under outpatient revenue codes. The reimbursement per hour is lower than an inpatient day, but the collection rate approaches 95% because payer denials virtually disappear. When you multiply that across the 400 to 800 patients per year that a moderately-sized ED can divert from inpatient admission to observation, the financial impact is substantial — typically $1.5 to $3 million in net revenue recovery annually.

But this is not just about money. Emergency department efficiency improves dramatically when observation patients are not boarding in your main ED. In my experience, a well-run 8-bed observation unit reduces main ED boarding by 15% to 20%. That translates directly to decreased door-to-provider times, reduced left-without-being-seen rates, and improved patient satisfaction scores. Every metric your hospital administrator cares about moves in the right direction.

The institutions that fail to build observation capacity are not just leaving revenue on the table. They are actively degrading the operational performance of their emergency departments while simultaneously exposing themselves to increasing payer scrutiny on short-stay admissions.

Building an Obs Unit Protocol That Survives the Real World

The number one reason observation units fail is not lack of funding or administrative support. It is poorly designed protocols. I have seen units launch with beautiful written policies that collapse the first week they encounter actual patients. Here is the approach that works.

Start with your top five diagnoses. For most emergency departments, these will be chest pain, syncope, TIA, asthma or COPD exacerbation, and abdominal pain. Do not try to build fifteen protocols at launch. Pick five. Get them right. Expand later.

Each protocol needs four components: explicit inclusion criteria, explicit exclusion criteria, a time-stamped order set, and a mandatory reassessment and disposition decision point. The inclusion and exclusion criteria are where most programs fail. They are either too broad — and the unit fills with patients who should have been admitted — or too narrow, and physicians stop using it because nothing qualifies.

For chest pain as an example, your inclusion criteria should specify: low to intermediate HEART score (4-6), negative initial troponin, no acute ECG changes, hemodynamically stable, and able to ambulate for stress testing. Your exclusion criteria should include: known coronary artery disease with prior intervention, active heart failure, hemodynamic instability, or any finding that mandates inpatient workup. This is not complicated, but it has to be written down, agreed upon by your cardiology colleagues, and enforced consistently.

The time-stamped order set is critical for reducing emergency department admissions to observation rather than inpatient status. Every protocol should have a maximum observation window — typically 24 hours — with mandatory reassessment at 8 and 16 hours. If the patient is not on a clear trajectory toward discharge at the 16-hour mark, they convert to inpatient. No exceptions. This prevents observation units from becoming holding pens.

I detail the complete protocol development process, including templates for all five core diagnoses, in my ebook Emergency Department Efficiency. It is the operational playbook I wish someone had given me when I built my first observation unit fifteen years ago.

The Staffing Model Everyone Gets Wrong

Here is the insight that separates observation units that thrive from those that quietly shut down within two years: the staffing model must be emergency medicine driven. Not hospitalist driven. Not a shared-coverage afterthought. Emergency medicine.

The logic is straightforward but counterintuitive to many hospital administrators. Observation medicine is a time-pressured, protocol-driven, disposition-focused discipline. That is exactly what emergency physicians are trained to do. We make rapid assessments, execute structured workups, and make disposition decisions under time constraints every single shift. Hospitalists are outstanding at managing complexity over days and weeks. Observation medicine operates on a 6-to-24-hour window. The cognitive framework is fundamentally different.

In the units I have directed, we use a dedicated EM physician or an advanced practice provider with EM training who covers the observation unit as their primary assignment — not as an add-on to a main ED shift. This is essential. When observation coverage is tacked onto an already-busy ED physician's assignment, the obs patients get seen last, reassessments get delayed, and the 24-hour window blows out. The unit becomes an expensive boarding area instead of a high-efficiency clinical operation.

The nursing model matters equally. Observation nurses should carry ratios of 4:1 to 5:1 — lower than med-surg but higher than the main ED. They need to be protocol-trained, meaning they can advance patients through order sets with minimal physician re-intervention for predictable decision points. If the 8-hour troponin is negative and the patient is asymptomatic, the nurse should be activating the stress test order — not paging the physician for permission.

This model costs money upfront. There is no way around that. But the return on investment is measurable within the first fiscal quarter if your patient selection and protocols are sound. The departments that try to run observation units on the cheap — borrowing staff from the main ED, splitting physician coverage, using med-surg nurses without protocol training — are the ones that generate the failure stories that scare other institutions away from the concept entirely.

The Bottom Line

ED observation units represent one of the highest-impact, lowest-risk operational improvements available to emergency departments today. The clinical evidence supports them. The financial model is proven. And the operational benefits to your main ED are immediate and measurable.

Three things to do this month. First, pull your admission data for the last 12 months and identify the percentage of short-stay admissions — less than 48 hours — for your top five diagnoses. That number is your observation unit business case. Second, schedule a meeting with your CMO and CFO to present the revenue recovery model. Frame it as payer denial mitigation, not new program cost — that language resonates with administrators. Third, identify your physician champion. This needs to be an EM doc who is operationally minded, politically savvy, and willing to own the unit through the first year of growing pains.

--> For the complete operational blueprint — including financial modeling templates, protocol development guides, and the staffing models that actually work — pick up my ebook ED Observation Units. It is the playbook I built from fifteen years of medically directing and practicing Observation Medicine in the real world.

--> If you’re an emergency physician (or any clinician treating patients daily) trying to understand how AI will actually impact your clinical practice—not just the hype—I put together a free practical AI in EM Survival guide. You can download it here.

👉 Link → Free AI in EM Survival Guide