The Metrics of Operationally Efficient Emergency Rooms

The Metrics of Operationally Efficient EDs

Operationally efficient EDs, by virtue of their focus on efficiency have different metrics and benchmarks than other emergency departments. In essence, many classic “Best-Practices” benchmarks do not even apply to these EDs. For example, the benchmark of less than 2% of patients leaving without being seen is meaningless as these departments usually keep those numbers below 0.5%. The truth of the matter is that these entities are in a league of their own and as such can only be benchmarked against each other. Nevertheless, the metrics pertaining to these departments have not been widely published. Through this post we draw on our personal experience implementing such departments in an attempt to define a standard. Furthermore, operationally efficient EDs can be set-up in different ways and in our experience; success is not so much dependent on the particular methodology used but rather on the metrics that must be attained to reach this capability.

In the discussion that follows we don’t intend to differentiate between departments that use Provider Intake vs. Nurse Rapid Screening or Immediate Bedding methodologies. Our purpose is to illustrate metrics shared by all “30-Minute” EDs rather than suggest a particular strategy. Discussion of the tactics used to obtain the ancillary turnaround times presented is also beyond the scope of this article.

1. Intake (Door to Provider)

The Intake category includes all the processes that happen from the time a patient enters through the door to the time the patient is actually seen by a provider of care. The consensus metric for an operationally efficient ED is an average Intake time of 22 minutes or less. This is the easiest metric to attain in the short-term as most intake processes are under direct influence and control of the ED. By the same token, it is also the most brittle metric as it fully depends on efficient Non-Admit Throughput times.

The rationale for the 22-minute cut-off is that when the ER is able to maintain this Intake time the vast majority of outliers tend to fall within the next 8 minutes. This is critical for departments that advertise a 30-Minute guarantee. By the same token, if many outliers fall out of the time to service goal, trying to decrease Intake below 22 minutes to compensate is futile. In fact, the only way to get the outliers in line with the performance goal is to achieve better Non-Admit Throughput times (Door to Release).

2. Non-Admit Throughput (Door to Release)

Throughput of non-admitted patients is by far the most important and dominant metric. In fact, without efficient non-admit throughput times it does not matter how efficient an emergency department is in Intake and Admit times, it will never be a “30-Minute” ED. This metric goes to the core of operational efficiency, as it is a measure of both, efficient bed turnover and staff productivity.

This is also the most difficult to influence, as some of the processes related to this metric are in hands of stakeholders that might not have a clear-cut incentive to attain the efficiency required to sustain the strategy. This metric actually represents a combination of factors that in average must provide for an overall non-admit throughput time of less than 180 minutes. Although a “bulk” of the improvement in this metric is achieved by the ED itself through a concept known as patient flow redirection (explained in a later post), we will now concentrate in the ancillary contributing factors:

a. Radiology turnaround times (Plain films)
b. Lab-work turnaround times (CBC, Chemistries, CK, MB, Troponin)
c. Transport turnaround times

As part of the implementation strategy the ability to obtain plain radiology turnaround times in the 25-minute range from order to completion is imperative. That means that your process reengineering must allow an X-ray response lag time of only 5 to 10 minutes to initiate the process. If the response time takes longer radiology backlogs will start to appear and the cut-off will not be met.

Some might be wondering why we do not include CT, Ultrasound, and MRI turnaround times and rather concentrate only on plain films. The answer is simple. These specialized tests have too much variability in the way they are performed, read, and prepped for. For example, a non-contrast Head CT must take no longer that 45 minutes from order to reported result but an abdominal CT with oral contrast might take 2 hours just to be performed. In the end, the actual amount of these tests performed in any given day pales in comparison to the amount of plain films an ED generates. Therefore, the ability to standardize plain film turnaround is more critical to overall efficiency.

Similarly, the ability to obtain a full cardiac work-up turnaround time in the 35-minute range from order to completion is desirable. That means that your process must allow a response lag time of only 4 minutes to initiate the blood draw process (order-to-vein) for the lab to be able to meet the mark. The rationale of using Cardiac work-up as the defining metric is that it represents a worse-case scenario. In fact, very seldom do we order just a CBC or just Chemistries or just Enzymes. Nevertheless, individual metrics for CBC, Lytes, Etc. can also be utilized although they might not be inclusive markers of efficiency. Individual times must be around 18 minutes for a CBC and around 28 minutes for simple chemistries. Better times can be obtained through Point-of Care testing but if these metrics are met by your current or future processes they are enough to sustain the strategy.

Some might also wonder why we do not include Urinalysis, pregnancy tests, and other commonly performed laboratory tests turnaround times and rather concentrate only on CBC, Lytes and Cardiac Enzymes. The answer is also simple. These tests have also too much variability in the way they are performed and obtained. For example, obtaining the urine for a Urinalysis or a pregnancy test can be a quick process (if the patient can urinate “on the spot”) or a prolonged process that might include hydration or bladder catheterization.

Furthermore, Point of Care testing for pregnancy tests can be used to improve the efficiency of resulting a particular sample, but it still does not obviate the need to obtain the urine first. Therefore Urinalysis and urine pregnancy tests do not lend themselves to benchmarking. Nevertheless, by concentrating on commonly performed tests like CBC and Lytes you do not only obtain a measurable metric but also provide “by-proxy” a rough guideline for staff to follow. In other words, if all lab tests are back in 35 minutes the staff tends to adapt itself so urine samples are obtained in a way that they can be resulted within that time.

Finally, institutions utilize transport in several different ways and capacities. If they are involved only in transporting patients to the floors then their response time does not affect Non-Admit Throughput times as much as Admit Throughput. Nevertheless, if they are involved in transporting patients to radiology and other specialized testing areas their response lag time cannot be more than 10 minutes from the time they are called to the time they are physically moving the patients.

3. Admit Throughputs

There has been emphasis given by JCAHO and EM Specialty Societies to prevent the practice of boarding admitted patients in the ED. They actually view this as the major determinant factor in ED overcrowding.

We do not contend that bed boarding is a negative and undesirable situation in any emergency department. Furthermore, we agree that, for traditional emergency departments, the statements made by those societies are right. In essence, the ability to see new patients and provide a disposition is severely curtailed whenever admit turnaround times lengthen and admitted patients are boarded. This situation can stop patient flow and produce incremental backlogs that can take hours to process. When sustained, the end-result of this breakdown is what is referred to as ED overcrowding. This further translates into a general sense of defeat and lack of control that paradoxically causes a decrease in staff productivity and efficiency at the time that is needed the most. Patients are also negatively influenced by this situation, as the “chaotic” appearance of an overcrowded emergency department undermines their confidence in the department’s ability to provide them proper care.

Having said that, operationally efficient departments behave and respond in ways that cannot be compared to traditional emergency departments. Because of the emphasis on efficient Intake and Non-Admit Throughput Times these departments have a significant “cushion” against this kind of operational breakdown. The most direct evidence is the fact that many operationally efficient EDs advertise their service. Logically, this could not happen if these departments where not able to handle previous patient loads in a more efficient way after conversion. Furthermore, after the advent of marketing, most “30-Minute” EDs have reported dramatic increases in patient volumes (25%-40%), without eroding the ability to meet the metrics above and maintain their 30-minute service advantage.

The most basic explanation for this is that only about 20% of ED patients are admitted while 80% are treated and released every day. This means that ED internal processes (along with ancillary department interfaces), are responsible for the turnover of the vast majority of ED patients. Operationally efficient EDs are able to process that 80% promptly and efficiently even while the Hospital is having problems moving the ED admits that constitute the other 20%.

By the same token, no system is perfect. Although a properly implemented “30-Minute” ED can handle increasing volumes and avoid overcrowding under most circumstances, you can only pack so much efficiency within a constrained number of treatment spaces and staff. Expectedly, a breaking point can eventually be reached if volumes continue to surge while the hospital’s ability to admit patients in an acceptable time frame lingers or deteriorates due to bed capacity or staffing issues. This needs to be anticipated and preempted to prevent a decline in service performance. For those reasons we suggest following the “Best-Practices” benchmark of 3.5 hours from door to admission.

Note: Although at the beginning of the article we explained that the classic “Best-Practices” benchmarks do not usually apply to “30-Minute” EDs, and we used the example of Left Without Being Seen, this does not imply that they must not be followed. As long as these metrics are being compared to those of other operationally efficient EDs they can still give you significant information about your intrinsic efficiency. For example, although operationally efficient EDs keep patient Leaving Without Being Seen under 0.5%, if such a department suddenly goes above that number this should raise a “red-flag”. In addition, in all reportable quality measures (i.e. “Core-Measures”), “30-Minute” EDs fare better than most other emergency departments due to timely identification of requisite pathology and a more controlled work environment that prevents errors of omission.

In conclusion, the importance of strict adherence to the stated metrics above cannot be understated. A focus on operational efficiency demands that these metrics be attained and, in particular, that Non-Admit Throughput times be kept below 180 minutes. To illustrate, imagine an ED that sees 72,000 patients a year and has Non-Admit Throughput Times of 230 minutes. If it admits 20 % of their patients that means this department discharges 160 patients a day out of an average daily census of 200. Decreasing Non-Admit Throughput Times to 180 minutes represents an average of 133 hours a day or 4,000 hours a month reclaimed from what was previously non-productive or idle time. These 133 hours is essentially the “penalty” this department pays for its operational inefficiency every day.

Obviously, Emergency Departments utilize many metrics and benchmarks to track performance. The purpose of this post is not to list them all but emphasize those essential to operationally efficient EDs while providing a standard whereby future implementation strategies can be analyzed and compared.