Type 1 observation unit

As characterized by Ross and colleagues, there are 4 main types of observation care, depending on the use of a dedicated space for observation and protocols to guide clinical care. Type 1 settings are dedicated OUs that are managed through protocol-driven care. Type 2 settings are dedicated OUs but clinical management is done at the discretion of each individual clinician (without the use of protocols). Type 3 settings involve observation care that occurs in any hospital bed but care is driven by protocols. Lastly, type 4 settings consist of discretionary, nonprotocolized observational care that happens in any hospital bed. The primary focus of this article is on type 1 OUs. Research has demonstrated this structure as superior to other types of OUs. Prospective randomized studies of patients with chest pain, asthma, transient ischemic attack, syncope, and atrial fibrillation managed in an emergency department OU (EDOU) have shown improved patient satisfaction, shorter lengths of stay (LOSs), lower costs, and comparable or improved clinical outcomes relative to similar patients admitted to an inpatient floor.

Open unit versus closed unit

An important element of OU structure is whether to establish an open unit versus closed unit. In closed units, the management of patient care is under the direction of a single physician group or specialty, such as emergency medicine. On the other hand, open OUs grant more than 1 group of physicians the opportunity to place a patient in the unit; as such, patient care is generally driven by the discretion of individual clinicians. As alluded to previously, evidence demonstrates the superiority of a closed, protocol-driven unit, which is the most common design. With more unified leadership in closed units, condition-specific protocols are used that incorporate inclusion/exclusion criteria, typical interventions performed, and specific criteria for discharge or inpatient admission. There is also greater consistency in ensuring appropriate patients are placed in the unit and better adherence to the specific endpoints of time-conscious protocols. In addition, as a result of a smaller and more consistent clinician group, individual clinicians themselves acquire more expertise and efficiency in managing patients in the OU setting. These closed unit settings make it easier to provide targeted feedback and education as well as enforce accountability and ensure quality control.

Hybrid unit

For smaller hospitals and those just starting an OU, 1 of the initial challenges becomes justifying the ongoing costs of maintaining nursing and ancillary support staff salaries while the fledgling unit is not consistently operating at full capacity. In these instances, creating a hybrid OU may seem desirable. Hybrid units essentially allow the dedicated space of an OU to be used by both observation patients and other patient populations, such as recovering elective procedure patients. A study by Ross and colleagues revealed that such a design demonstrated a complementary diurnal occupancy pattern, which improved both hourly census and nurse resource use. It also had the added benefit of more efficient management of scheduled procedure patients—without resulting in any adverse effects on the LOS or discharge rate of OU patients.

The appeal of hybrid functioning OUs to maximize operational capacity also pertains to hold or boarder patients who are awaiting transfer to an inpatient bed or another facility. Although including these patients in a hybrid design seems enticing in the afternoon hours (when OU occupancy reliably drops due to the efficiency of morning discharges and an expected arrival peak in the late afternoon and evening), it must be avoided—except perhaps in cases of extreme hospital census when ED waiting room conditions become unsafe due to a lack of acute care bed capacity. Avoidance of this practice is due to the reality that the hold or boarding time frames are unknown and often longer than expected. Thus, when patients in an ED need OU beds in the evening hours, capacity is lacking. A vicious cycle may ensue wherein patients eligible for shorter observation stays must be managed by less efficient inpatient teams due to a lack of OU capacity, further exacerbating the lack of inpatient capacity. The default policy should be no boarding within an OU; the rare practice of OU boarding can be justified at times, although strong leadership is required to ensure it does not become the norm.

Location

From both patient care and logistics perspectives, location is a critical design component of OU development. The highest potential for patient comfort, hospital bed use efficiency, and cost savings stems from the creation of a dedicated physical space for observation patients. Locating this dedicated space within or immediately adjacent to the emergency department (ED) aids in the success of such units, because processes are minimally disrupted and close proximity to clinicians, supplies, and equipment enables safe and efficient care. The placement of the unit at a significant distance from the ED might not change the internal dynamic but results in a loss of efficiency in clinical re-evaluation, transport, and communication. As distance from the ED increases, patient expectations of a brief hospitalization and manifestation of the ED culture of rapid throughput and continuous rounding may become more difficult to maintain. At the same time, for those hospitals with crowded EDs and limited department space, it likely does not make sense to undersize the ED to generate a close-proximity OU. In these cases, it may be more amenable to have a remote OU—remembering that the OU is only 1 resource among competing interests.

Volume capacity

Another important step in the establishment of an OU is understanding the projected daily volume capacity and thus the ideal number of beds contained within the unit. Studies have demonstrated that approximately 5% to 10% of all ED patients may be good candidates for observation care. Facilities opening a new unit can expect to start closer to 5% and increase over time as new protocols are implemented and staff comfort rises. Another metric used is to estimate OU volume as approximately 2 beds per every 10,000 ED visits. At the same time, the minimum recommended size for a dedicated unit is 5 beds to 8 beds, which allows for an efficient nurse-to-patient staffing ratio of 1:5 or 1:4. Based on an average LOS of 15 hours and a size of 5 OU beds, the expected average daily volume is 8 patients (assuming full occupancy and perfect matching of arrivals and departures). Therefore, the minimum ED volume to support 8 OU patients/d in a 5-bed unit is 80 ED visits/d (10% of ED volume) or approximately 29,000 ED visits per year. As a result, most hospitals do not consider opening an OU until ED annual visit volume rises well above 30,000.

Physical space layout

Careful thought is required in designing the physical space of a new OU. Ideally, the OU is set up in such a way to optimize clinical care—with primary considerations centering around space usage, the patient-clinician relationship, and allocated funds. Generating a design that maximizes the use of space with an eye toward balancing patient comfort and clinician workflow is critical. It is necessary to be attentive upfront to the planned nurse-to-patient ratio (1:4–1:6), because it has an impact on the number of beds included in the OU. Whether the location is built within new or renovated space also has an impact on decision making regarding OU design. Local factors, such as state and city departments of public health regulations, may also heavily influence OU design.

In addressing the patient-clinician relationship, a key design element that requires attention pertains to providing structure to the patient space within an OU. Variations in such structure exist within current OUs, with some possessing curtained cubicles and others using closed-walled rooms—both have advantages and disadvantages. Structuring the physical space in such a way that it promotes the patient-clinician relationship is equally important. One option includes having nursing stations that face all patient rooms/cubicles. At one of the authors’ institutions (Brigham and Women’s Hospital), this is done with a semicircular nursing station that enables direct visualization of all patients, allowing patients to feel better cared for during their stay. Other design considerations include creating a patient nourishment station, medication preparation area, workstations for consultants and other staff (pharmacy, physical therapy, and care coordination), and space and accommodation for visitors.

Time cutoff

Establishing an LOS time cutoff is another relevant consideration for determining how to structure an OU. Traditionally, most EDOUs used a 24-hour maximum cutoff; however, on October 1, 2013, the Centers for Medicare & Medicaid Services (CMS) introduced the Two-Midnight Rule, wherein all Medicare patients with an expected hospital LOS of less than 2 midnights should be classified as observation patients. Although the application of this rule has undergone several revisions, it continues as standard CMS policy. Depending on the hour of patient arrival for initial ED evaluation, observation patients can spend well over 24 hours in the hospital. These recent changes, however, do not remove the need for prospective, explicit endpoints to determine either a safe discharge to home or further inpatient hospitalization.

Amenable conditions

In the past few decades, several conditions have been studied and found amenable for OU management. Chest pain was the initial focus of such efforts. Further research, however, has shown the efficacy of treating several other conditions, including asthma, atrial fibrillation, transient ischemic attack, syncope, acute decompensated heart failure, and pulmonary embolism. Other conditions have been managed in an OU as well (eg, abdominal pain, dehydration, renal colic, cellulitis, back pain, headache, pyelonephritis, chronic obstructive pulmonary disease exacerbation, community-acquired pneumonia, metabolic derangement, deep vein thrombosis, and psychiatric emergency) but limited to no formal assessment of their efficacy has been published to date.

Once a survey of these potential conditions of interest is complete, the determination of optimal conditions for a new OU can be performed through data analysis and discussions with key leadership. Important areas of analysis include internal hospital data (ie, most frequent diagnoses discharged from inpatient teams in observation status or those with ED LOS greater than 8 hours) as well as using external resources, such as Medicare reports (eg, Program for Evaluating Payment Patterns Electronic Report [PEPPER] report, https://www.pepperresources.org/ ). This work to identify those conditions with the highest volume opportunity to manage in an OU, in combination with leadership approval, guides decision making. Often, however, when starting an OU, it is wise to start with more widely established pathways, such as chest pain. Over time, successful implementation of this protocol builds consensus and enables development of the additional amenable conditions discovered through analysis.

Inclusion of patients with mental illness

An additional consideration in the establishment of an OU is whether to include the population of behavioral health patients. At the core of this debate is whether care of this population fits with the definition of the purpose of observation care, which is to determine if inpatient admission is necessary. Some OU leaders argue that a behavioral health patient, such as one with suicidal ideation who is held for placement in an inpatient psychiatric facility, is inappropriate for OU care, because the disposition is not in doubt and the delay in care is related to a bed search alone. The counterargument is that some of these patients improve with new or resumed medical therapy while in observation status and, on re-evaluation, can be discharged back into the community with a safe outpatient follow-up plan; accordingly, there is legitimate uncertainty around the need for inpatient admission in many cases and observation care is justified. Nevertheless, each institution should examine its own patient populations and resources; discuss options with psychiatry, security, and compliance colleagues; and determine the practice that is most appropriate. For those who proceed with the care of this special population, the OU must be adequately equipped with resources to meet the needs of this patient population. For example, the rooms need to be made safe for patients at risk for self-harm, security should be near the unit, and the unit’s atmosphere should be calm and supportive. Patients requiring physical or chemical restraints, or those with evidence of acute intoxication, should not be managed in an OU.

Diagnostic versus therapeutic endpoints

Once the best candidate conditions have been chosen, the next branch point in this process is the selection of diagnostic versus therapeutic endpoints for these conditions of interest. Diagnostic protocols, as their name suggests, focus on using time in the OU to obtain a necessary diagnosis (or more likely, sufficiently exclude life-threatening diagnoses, such as acute coronary syndrome in patients with chest pain). Therapeutic protocols, on the other hand, concentrate on providing appropriate time-sensitive therapies (medications, fluids, and so forth) where a diagnosis has already been obtained or will be sought after in the outpatient setting (ie, a patient with asthma who needs more time for steroids and albuterol to control symptoms).

Protocol development

Once conditions of interest have been chosen and diagnostic versus therapeutic endpoints selected, the next phase is protocol development. As referenced previously, type 1 OUs use protocol-driven care to assist in achieving superior clinical outcomes, patient satisfaction, and efficiency. Generating protocols has a unique ability to enable clinicians choosing patients for OU management, as well as OU staff, to have a clear sense of patient selection and planned management. The process of protocol development commences with a thorough literature review. Then, because many institutions have a long history of OU practice and freely share protocols, gathering existing protocols may be a useful early step in the process. Discussion with local and national content experts may also inform best practices and evidence-based care. Once this is complete, enlisting all relevant stakeholders (OU leadership and appropriate consultants) allows for the joint creation of an OU pathway. A multidisciplinary approach is often best, including not only physician staff but also nurses, advanced practice providers (APPs), and all other staff who may be impacted by the protocol.

The key areas of discussion and subsequent protocol development are (1) inclusion/exclusion criteria, (2) expected OU interventions, and (3) disposition criteria. Additional considerations for the protocol include expected LOS and considerations for the medical record. Inclusion/exclusion criteria establish clinical variables for each condition of interest that enables work-up that is suitable for an OU while maintaining a reasonable discharge-to-home rate. Avoidance of patients who clearly require inpatient management or would be better managed by discharge home with clinic follow-up should be emphasized in patient selection. Expected OU interventions outline diagnostic and/or therapeutic activities that are typically done within the unit and provide a template to guide clinicians rather than a prescriptive to-do list, given the variability between patients. Disposition criteria facilitate appropriate patient flow and efficiency within the OU as well as assist in the sometimes difficult decision to admit a patient to the hospital. LOS often varies for each condition; generating an overall sense of expected LOS for each respective condition assists clinicians to make the optimal use of OU resources. Lastly, outlining specific considerations for the medical record (ie, contact with primary care provider or outpatient follow-up plan) can aid clinicians in navigating the nuances of OU medical documentation.

Protocol verification, implementation, and maintenance

With the key components identified, the next phase involves protocol verification and launch. During this time, the protocol is vetted by the relevant stakeholder groups and feedback is obtained. Once sufficient review and editing have been performed, the protocol for the condition of interest can become operational in the OU. The implementation process requires intense surveillance and solicitation of feedback in the critical period after launch. Typically, a new protocol may take several months from inception to implementation.

The final phase in this process involves protocol maintenance. Protocol development is not a static but a dynamic process. Once a set of protocols is developed, updating and/or retiring the algorithms is necessary. In the authors’ practice, this is done every year at a minimum to ensure optimal use of the OU management pathway.

Type 1 observation unit

As characterized by Ross and colleagues, there are 4 main types of observation care, depending on the use of a dedicated space for observation and protocols to guide clinical care. Type 1 settings are dedicated OUs that are managed through protocol-driven care. Type 2 settings are dedicated OUs but clinical management is done at the discretion of each individual clinician (without the use of protocols). Type 3 settings involve observation care that occurs in any hospital bed but care is driven by protocols. Lastly, type 4 settings consist of discretionary, nonprotocolized observational care that happens in any hospital bed. The primary focus of this article is on type 1 OUs. Research has demonstrated this structure as superior to other types of OUs. Prospective randomized studies of patients with chest pain, asthma, transient ischemic attack, syncope, and atrial fibrillation managed in an emergency department OU (EDOU) have shown improved patient satisfaction, shorter lengths of stay (LOSs), lower costs, and comparable or improved clinical outcomes relative to similar patients admitted to an inpatient floor.

Open unit versus closed unit

An important element of OU structure is whether to establish an open unit versus closed unit. In closed units, the management of patient care is under the direction of a single physician group or specialty, such as emergency medicine. On the other hand, open OUs grant more than 1 group of physicians the opportunity to place a patient in the unit; as such, patient care is generally driven by the discretion of individual clinicians. As alluded to previously, evidence demonstrates the superiority of a closed, protocol-driven unit, which is the most common design. With more unified leadership in closed units, condition-specific protocols are used that incorporate inclusion/exclusion criteria, typical interventions performed, and specific criteria for discharge or inpatient admission. There is also greater consistency in ensuring appropriate patients are placed in the unit and better adherence to the specific endpoints of time-conscious protocols. In addition, as a result of a smaller and more consistent clinician group, individual clinicians themselves acquire more expertise and efficiency in managing patients in the OU setting. These closed unit settings make it easier to provide targeted feedback and education as well as enforce accountability and ensure quality control.

Hybrid unit

For smaller hospitals and those just starting an OU, 1 of the initial challenges becomes justifying the ongoing costs of maintaining nursing and ancillary support staff salaries while the fledgling unit is not consistently operating at full capacity. In these instances, creating a hybrid OU may seem desirable. Hybrid units essentially allow the dedicated space of an OU to be used by both observation patients and other patient populations, such as recovering elective procedure patients. A study by Ross and colleagues revealed that such a design demonstrated a complementary diurnal occupancy pattern, which improved both hourly census and nurse resource use. It also had the added benefit of more efficient management of scheduled procedure patients—without resulting in any adverse effects on the LOS or discharge rate of OU patients.

The appeal of hybrid functioning OUs to maximize operational capacity also pertains to hold or boarder patients who are awaiting transfer to an inpatient bed or another facility. Although including these patients in a hybrid design seems enticing in the afternoon hours (when OU occupancy reliably drops due to the efficiency of morning discharges and an expected arrival peak in the late afternoon and evening), it must be avoided—except perhaps in cases of extreme hospital census when ED waiting room conditions become unsafe due to a lack of acute care bed capacity. Avoidance of this practice is due to the reality that the hold or boarding time frames are unknown and often longer than expected. Thus, when patients in an ED need OU beds in the evening hours, capacity is lacking. A vicious cycle may ensue wherein patients eligible for shorter observation stays must be managed by less efficient inpatient teams due to a lack of OU capacity, further exacerbating the lack of inpatient capacity. The default policy should be no boarding within an OU; the rare practice of OU boarding can be justified at times, although strong leadership is required to ensure it does not become the norm.

Location

From both patient care and logistics perspectives, location is a critical design component of OU development. The highest potential for patient comfort, hospital bed use efficiency, and cost savings stems from the creation of a dedicated physical space for observation patients. Locating this dedicated space within or immediately adjacent to the emergency department (ED) aids in the success of such units, because processes are minimally disrupted and close proximity to clinicians, supplies, and equipment enables safe and efficient care. The placement of the unit at a significant distance from the ED might not change the internal dynamic but results in a loss of efficiency in clinical re-evaluation, transport, and communication. As distance from the ED increases, patient expectations of a brief hospitalization and manifestation of the ED culture of rapid throughput and continuous rounding may become more difficult to maintain. At the same time, for those hospitals with crowded EDs and limited department space, it likely does not make sense to undersize the ED to generate a close-proximity OU. In these cases, it may be more amenable to have a remote OU—remembering that the OU is only 1 resource among competing interests.

Volume capacity

Another important step in the establishment of an OU is understanding the projected daily volume capacity and thus the ideal number of beds contained within the unit. Studies have demonstrated that approximately 5% to 10% of all ED patients may be good candidates for observation care. Facilities opening a new unit can expect to start closer to 5% and increase over time as new protocols are implemented and staff comfort rises. Another metric used is to estimate OU volume as approximately 2 beds per every 10,000 ED visits. At the same time, the minimum recommended size for a dedicated unit is 5 beds to 8 beds, which allows for an efficient nurse-to-patient staffing ratio of 1:5 or 1:4. Based on an average LOS of 15 hours and a size of 5 OU beds, the expected average daily volume is 8 patients (assuming full occupancy and perfect matching of arrivals and departures). Therefore, the minimum ED volume to support 8 OU patients/d in a 5-bed unit is 80 ED visits/d (10% of ED volume) or approximately 29,000 ED visits per year. As a result, most hospitals do not consider opening an OU until ED annual visit volume rises well above 30,000.

Physical space layout

Careful thought is required in designing the physical space of a new OU. Ideally, the OU is set up in such a way to optimize clinical care—with primary considerations centering around space usage, the patient-clinician relationship, and allocated funds. Generating a design that maximizes the use of space with an eye toward balancing patient comfort and clinician workflow is critical. It is necessary to be attentive upfront to the planned nurse-to-patient ratio (1:4–1:6), because it has an impact on the number of beds included in the OU. Whether the location is built within new or renovated space also has an impact on decision making regarding OU design. Local factors, such as state and city departments of public health regulations, may also heavily influence OU design.

In addressing the patient-clinician relationship, a key design element that requires attention pertains to providing structure to the patient space within an OU. Variations in such structure exist within current OUs, with some possessing curtained cubicles and others using closed-walled rooms—both have advantages and disadvantages. Structuring the physical space in such a way that it promotes the patient-clinician relationship is equally important. One option includes having nursing stations that face all patient rooms/cubicles. At one of the authors’ institutions (Brigham and Women’s Hospital), this is done with a semicircular nursing station that enables direct visualization of all patients, allowing patients to feel better cared for during their stay. Other design considerations include creating a patient nourishment station, medication preparation area, workstations for consultants and other staff (pharmacy, physical therapy, and care coordination), and space and accommodation for visitors.

Time cutoff

Establishing an LOS time cutoff is another relevant consideration for determining how to structure an OU. Traditionally, most EDOUs used a 24-hour maximum cutoff; however, on October 1, 2013, the Centers for Medicare & Medicaid Services (CMS) introduced the Two-Midnight Rule, wherein all Medicare patients with an expected hospital LOS of less than 2 midnights should be classified as observation patients. Although the application of this rule has undergone several revisions, it continues as standard CMS policy. Depending on the hour of patient arrival for initial ED evaluation, observation patients can spend well over 24 hours in the hospital. These recent changes, however, do not remove the need for prospective, explicit endpoints to determine either a safe discharge to home or further inpatient hospitalization.