Introduction

The emergency department (ED) plays a critical role in caring for older adults with over 19 million ED visits a year, yet emergency care in the USA is not attuned to their needs [1]. During an ED visit, 76% of delirium is missed [2], 12–16% of older adults receive potentially harmful medications [3, 4], and 80% of patients presenting for a fall-related injury do not receive fall prevention counseling [5, 6]. These missed opportunities contribute to the poor outcomes seen in the 6 months after an ED visit for a fall or minor injury: 42% return to the ED, 25–35% suffer significant functional decline, and the mortality rate is 10 times higher than older adults without an ED visit [7,8,9,10]. The national multispecialty Geriatric ED Guidelines recognize this problem and endorse Multidisciplinary Geriatric Assessment for all high-risk ED patients [11]. Multidisciplinary assessment by geriatricians, case managers, pharmacists, and physical therapists (PTs) in the ED can identify geriatric syndromes and address needs, which leads to decreased unnecessary hospitalizations [12,13,14,15,16,17,18]. Studies of multidisciplinary assessment in the ED show measurable benefits with decreases in hospitalizations, intensive care unit admissions, ED revisits, and functional decline at 6 months [12, 15, 17, 19,20,21,22,23,24].

Despite these proven benefits, implementation of multidisciplinary care in EDs has been limited. Barriers to implementation include the 24/7 ED care model, cost of multidisciplinary staff [25], difficulties risk stratifying or choosing which older adults should receive the intervention [26], incentives for short length of stay in the ED and expedited care [27], and a lack of data on implementation in this dynamic and complex healthcare setting [28, 29]. The data supporting multidisciplinary geriatric assessment has come from EDs with external funding from research or philanthropic organizations. To be able to provide the benefits of multidisciplinary geriatric assessment to all patients, more information on implementation and operational models that do not rely on external funding is needed.

One operational model of multidisciplinary assessment in the ED uses an observation unit (Obs Unit) [30]. Obs Units are a promising solution to these barriers. Over 36% of EDs have Obs Units, which provide a setting for typically 8–24 more hours of further monitoring and testing [31,32,33]. Providing multidisciplinary assessments in an Obs Unit addresses the barriers of long stays in the ED, personnel costs, and consultant availability. For example, instead of needing to have a PT available at 8 pm when the patient is getting their ED evaluation, the patient can be kept in observation to see the PT at 8 am. This process has been shown to be feasible and effective in Obs Units around the world, but is not standard care in the USA [18, 34,35,36,37].

Currently, over 100 EDs have been accredited as geriatric EDs, and more are applying each year. However, only 10 have reached level 1 status, which requires the availability of multidisciplinary geriatric assessment in the ED setting. This study will develop a protocol incorporating risk stratification and multidisciplinary geriatric assessment. By providing clear information on ways to implement this type of program in a sustainable manner, this study has the potential to increase access to this valuable service in EDs across the nation.

This study will investigate the implementation of a two-step geriatric Obs Unit protocol which begins with (1) ED nurses using validated tools to assess the patients for fall risk, delirium, polypharmacy, and frailty. If needs are identified, and the patient does not meet admission criteria for their other medical issues, the patient is (2) placed in the Obs Unit to be evaluated by geriatricians, PTs, pharmacists, and/or case managers. A pilot of step 2 of the protocol resulted in new interventions for 76% of patients who were assessed by the multidisciplinary geriatric consultants [37]. However, we found a significant performance gap: the screening is being done in less than 2% of older patients. Nurses and physicians have received training on how to use the geriatric screening tools, and the consultant teams for multidisciplinary geriatric assessment are available, but they are rarely consulted. Monthly EHR reports show that geriatricians are consulted only five times a month, despite 1600 older adults being cared for in the Obs Unit per year. Currently, we do not know why the ED staff are choosing to assess some patients and not others, or whether the intervention would have benefitted those who did not receive it. Therefore, this study has two aims: evaluation of the implementation process in the ED and Obs Unit setting and evaluation of protocol effectiveness. Studying the implementation of this process is essential and will aid in potential dissemination of successful protocols for maximal national impact.

We will conduct a hybrid II effectiveness-implementation study with before and after cohort analysis, adapting the Consolidated Framework for Implementation Research: Care Transitions Framework (CFIR) [38] to the ED setting. We will track implementation processes and measures, and Lean Six Sigma rapid cycle process improvement will be used to overcome barriers to protocol use and fidelity. Lean Six Sigma is frequently used in our and other medical centers for process improvement and therefore will allow us to speak a common language with other centers for future dissemination. We will also report on measures needed for sustainability by continuing to monitor screening rates and consultant use while phasing out implementation support. Finally, to determine the effectiveness for patients, both operational metrics and patient-centered outcomes will be assessed using a pre-/post-cohort evaluation.

Methods

This study is approved by its institutional review board and registered on clinicaltrials.gov (NCT04068311). The aims are to develop, implement, and sustain a two-step intervention providing ED geriatric assessments by combining (1) ED nurse-based screening for geriatric syndromes of all older ED patients with (2) multidisciplinary geriatric assessment in an Obs Unit. Secondly, it will describe the effect of this protocol on reducing functional decline after an ED visit.

Setting

The study will be carried out in an academic, tertiary care referral ED with over 80,000 ED visits annually. Of those visits, 25.2% are made by adults ≥ 65 years old. The ED has an embedded 20 Obs Unit that is staffed 24/7 by advanced practice providers and 8 h a day by an emergency medicine physician. The ED and Obs Unit have access to 24-h social work and case management services, 18 h a day of pharmacist coverage, and 12 h a day of physical therapy coverage Monday through Saturday. The inpatient geriatric consult team prioritizes Obs Unit consult requests and is available during business hours.

External setting context includes the Joint Hospital Accreditation Council mandates on fall risk screening, several national reported quality measures (falls and ED recidivism), and the recent accreditation of geriatric ED programs through the American College of Emergency Physicians (www.acep.org/geda). The ED for this study has Level 1 Geriatric ED accreditation in part for the ability to provide and track geriatric screening, but the screening is being done inconsistently and therefore to maintain accreditation improvement in this area is required.

Intervention

The intervention is a two-step protocol beginning with screening for geriatric syndromes for patients aged ≥ 65 years in the ED (Table 1). The screening is done by the bedside nurse and entered directly into the electronic health record (EHR). It takes 3 min, including the time needed to get the patient out of bed and to document results in the EHR [42]. Positive results are relayed to the provider team verbally or via EHR chat function. The second step is multidisciplinary assessment in the Obs Unit for those who have needs identified by the screening tools but do not meet medical necessity for this evaluation as an inpatient (Fig. 1). Patients who have acute needs and require admission or admission to a skilled nursing facility are hospitalized. For those placed in the Obs Unit, the physician team chooses which elements of multidisciplinary professionals need to be consulted based on their clinical evaluation and the geriatric screening results (Table 2).

Table 1 Study intervention: nurses will perform three geriatric screening assessments that direct the need for geriatrician, pharmacist, PT, and case manager evaluations
Full size table
Fig. 1
figure 1

Patient flow through the ED visit and integration of the geriatric interventions per protocol

Full size image
Table 2 Patient-centered outcomes chosen to evaluate the effectiveness of the protocol for multidisciplinary geriatric assessment in the emergency department
Full size table

Study design—implementation

This is a T3 translation study, where a treatment effect is studied in real clinical practice [43]. For the first aim, implementation, the proportion of older adult patients receiving geriatric screening will be evaluated pre- and post-intervention. The Consolidated Framework for Implementation Research: Care Transitions Framework (CFIR) was chosen to derive barriers and facilitators to implementation and map them to the external context, setting structure (organizational characteristics), provider roles and characteristics, and patient characteristics and factors (Table 3) [44]. CFIR has been used successfully in the ED setting to guide rapid cycle process improvement [45, 46].

Table 3 Study intervention characteristics as mapped to the Consolidated Framework for Implementation Research Transitions of Care Framework, adapted from Rojas Smith et al. [38]
Full size table

The implementation method will be Lean Six Sigma, which is a commonly used tool for rapid cycle process improvement in healthcare settings. Lean Six Sigma is a QI approach that is validated in healthcare settings and used by our hospital [47]. Lean Six Sigma functions well for quality improvement in EDs [48,49,50,51,52], but there is minimal data on implementation strategies/methods in Obs Units [53, 54]. The Lean Six Sigma team will consist of frontline staff, nursing leaders, physician leaders, and members from PT, geriatrics, pharmacy, and case management teams.

The design and reporting of this study adhere to the Standards for Reporting Implementation Science (Table 4) [55]. The core component of the intervention is multidisciplinary geriatric assessment in the ED setting. The adaptable elements are the screening tools used, where and who does the screening, and where the geriatric assessments occur. For example, one study found that PT evaluations in the ED setting can be done without prolonging ED length of stay, but this service was only offered Monday through Friday, 7 am–4 pm [56]. The implementation team may discover that during regular business hours multidisciplinary assessments can be completed in a time frame that does not require observation placement. We would not consider this a “protocol violation” but an adaptation taking advantage of times when resources are high. Similarly, it may be discovered that there is a significant impediment to workflow with one of the chosen geriatric screening tools or that new data shows that a different tool has improved specificity. Any protocol adaptions and the reasons will be reported. Changes in knowledge and awareness will be assessed with before/after surveys.

Table 4 Standards for Reporting Implementation Studies study checklist and rationale for choosing measures to report, adapted from Pinnock et al. [55]
Full size table

As part of the evaluation, we will analyze sustainability of the program. There are no sustainability instruments validated in the ED; we adapted the Measurement Instrument for Sustainability of Changed Work Practices (Sustainability survey) to evaluate culture and routine changes in this new setting [57]. This survey will be given to the 150 ED nurses at 12 and 24 months after full implementation, with attention to the sections on routinization and institutionalization.

Measures and data analysis—implementation

Monthly geriatric patient visits as well as geriatric screening and consultations completed will be recorded and summarized descriptively (means, standard deviations, 95% confidence intervals; medians, interquartile ranges; proportions with 95% confidence intervals). For fidelity, we will look at the proportion that screened positive for consultation and received the appropriate consultation for their positive screening test. We will also perform fidelity checks during the post-cohort recruitment. We will note any external or internal events that may affect the process, such as new governmental or institutional mandates that may arise.

As a secondary analysis, we will examine the change in the proportion of geriatric screening using an interrupted time series analysis. We do not expect to see seasonal variation, but may see effects from confounders such as changes in total ED volume, nursing staff turnover (as evidenced by hours of float nurse pool coverage), patient volumes, and ED boarding rates [58].

Power—implementation

This phase of the study is based on all geriatric patients seen in the ED. On average, 6660 patients are seen per month in the OSUMC ED with approximately 25% or 1800–2000 age ≥ 65 years. In October 2019, we had 7252 total patients, 1439 (19.8%) aged ≥ 65 years, of which only 20 were screened (1.3%). Assuming 1300 monthly geriatric visits, the two-sided 95% CI width for the monthly proportion for screening rates will range from 0.016 if the actual proportion is 1.5% to a maximum of 0.055 if the proportion is 50%. We will have balanced time periods of 24 months before and 24 afterwards. The study timeline is 5 years, which includes a year for implementation.

Study design—effectiveness

The effectiveness study is a pre-/post-cohort analysis. We will evaluate patient-centered outcomes—functional status, health-related quality of life (HRQoL), and patient satisfaction (Table 2). Study personnel will monitor the Obs Unit electronic tracking board (7 am–11 pm M–F and select weekends) to identify and recruit 380 patients. A survey and chart review will be completed during the Obs Unit stay. Initial data elements collected include demographics, insurance status, zip code for socioeconomic status estimate, and the Charlson Comorbidity Index components and overall score [59]. Follow-up phone interviews will occur at days 30 ± 3 and 90 ± 5 post index visit. In simulations, survey completion required 8–10 min. Surveys include the Older Americans’ Resources and Services Activities of Daily Living questionnaire (OARS) and Patient-Reported Outcomes Measurement Information System (PROMIS) HRQoL, ensuring comparability to existing studies and ease of data dissemination [41, 60, 61]. Outcome scales will be measured during the ED stay, at 30 and 90 days. Primary outcome is the change in functional status based on OARS from 0 to 90 days. A planned subset analysis comparing the patients in the post-cohort who had the full screening and protocol to those in the pre-cohort who did not will be done.

To assess patient satisfaction with the protocol, a trained researcher will conduct in-person, semi-structured interviews with post-cohort patients. In addition to Likert-type questions, we will solicit brief descriptions of clinical exemplars of this intervention. Maximum variation/heterogeneous purposive sampling will provide a mix of genders, ages, and observation dispositions for the structured interviews [62]. Initial sampling will be sequential. After enrolling patient 30 into the post-cohort, the research team will begin qualitative semi-structured patient interviews. We will interview every 5th patient for 10 interviews. The team will then summarize the interviewee demographics and identify specific types of patients for the next 10 interviews to ensure that there is representative viewpoints of the oldest old (age 85+), the younger old (age 65–70), all genders, and admitted and discharged patients.

Data analysis—effectiveness

The primary effectiveness outcome is the proportion of patients in the pre- and post-cohort with a significant (≥ 3 point) decline in functional status (OARS) from day 0 to day 90. This corresponds to a complete loss of one activity of daily living or a decrease in several. We will compare proportions with decline in the pre- and post-group using a chi-square test (primary analysis). Additionally, logistic regression modeling will be used to compare the functional decline between the groups univariately and while controlling for initial ED HRQoL, demographic factors (age, race, average socioeconomic status from zip code census tract), Charlson Comorbidity Index score [43], home health services prior to the ED visit, and any other significant factors varying between the two cohorts. The total number of covariates will not exceed 12, given we estimate 122 subjects with functional decline. A similar secondary analysis will be done using the 30-day timepoint data.

For secondary outcomes, proportions of dichotomous variables will be compared between pre- and post-intervention cohorts using chi-square tests. Regression models will be used for exploring additional secondary outcomes between the two groups. The model used will be dictated by the outcome of interest such as logistic regression for dichotomous outcomes and Poisson or negative binomial regression for count data.

Qualitative analyses will be conducted from the patient interviews. Patient interviews will be transcribed verbatim and entered into Atlas.ti (Scientific Software Development GmbH). We will conduct manifest content analysis of the descriptions using phrases and sentences as our unit of analysis [63, 64]. We will categorize the exemplars and analyze for appropriateness and perceived outcomes based on the level of detail provided. Analysis will include open and axial coding procedures using techniques of constant comparison and questioning within and across cases [65, 66]. The coding schema will be created with consensus on coding definitions and grouping codes into code families/categories. Dual coding with negotiated consensus will be performed on 20% of the data to add rigor to the analysis. Themes will be reported per the consolidated criteria for reporting qualitative research guidelines [55, 67].

Power—effectiveness

We conservatively estimate that 40% of the pre-cohort and 30% of the post-cohort will experience functional decline [19, 68,69,70]. We will over-sample the post-implementation cohort due to the expected heterogeneity (screen negative/positive, with/without full intervention). With 137 subjects in the pre-intervention cohort and 206 in the post-intervention cohort, we will have 80% power to detect a difference of this magnitude, based on a chi-square test and an alpha of 0.05. To account for an estimated 10% loss to follow-up, 380 subjects (150 in the pre-cohort and 230 in the post) will be recruited.

Discussion and dissemination

This single-center, hybrid implementation/effectiveness study is the first study of multidisciplinary geriatric assessment in the ED setting which is designed to be reproducible and sustainable without external funding or increasing ED length of stay. By incorporating consultants to assist with multidisciplinary assessment, the program also does not require new staff. These were all considerations by the interdisciplinary geriatric ED team which developed this protocol. The protocol has been piloted and now requires a full implementation and evaluation. By assessing implementation factors and evaluating sustainability, the study will provide information on the implementation of protocols and screening tools in an acute care setting. This information will also be used to plan a subsequent dissemination study of the protocol if it proves to be effective for reducing functional decline and poor outcomes after an ED visit.

Publishing the full study protocol is meant to encourage discussion about implementation in acute care and short stay settings. Lean Six Sigma was chosen as the implementation strategy because it provides a framework for rapid cycle process improvement, as well as specific strategies for handling complicated systems with a variety of inputs. Lean Six Sigma is also oriented towards improving flow, which is critical to the time-sensitive ED setting. It has a focus on including frontline staff in all aspects of quality improvement, standardizing and simplifying processes and protocols, and optimizing the value to customers. In this case, both values to the hospital and to the patient will be assessed by including both hospital operational metrics as well as patient-oriented outcomes. CFIR was chosen because the transition of care framework considers all the moving parts and different people involved in emergency care, including population and community needs, patient-centeredness, different levels in the health system, and appropriate outcome measures for this type of study.

We specifically chose not to include an economic analysis as part of the study outcome (Table 4). The goal of this program is to improve the transition of care from the ED to home, allowing older adults to live safely in the community. Therefore, this program will initially increase healthcare costs. An economic analysis of healthcare cost savings should be done over the course of ensuing years, not 90 days. In essence, one could consider this study the pilot to determine if the ED effectively identifies needs and connects the older adult patient to resources, and further studies can assess for downstream healthcare cost savings if the program is successful.

This study hopes to develop a method of incorporating holistic, geriatric care into the ED setting in a sustainable fashion. The authors invite any input on the design of this and future studies that could arise from this data.