Building a Rational Clinical Information System for Older Adults in Acute Care: The Role of the interRAI Acute Care Suite

Abstract

Prior research has identified gaps in the ability of hospital systems to efficiently and meaningfully characterize older adults with complex health needs. We recruited community-dwelling older adults presenting to 10 Emergency Departments (EDs) across Ontario, Quebec, and Newfoundland, Canada, from April 2017 to July 2018. We deployed a staged assessment strategy based on the interRAI Acute Care Suite to identify and characterize older adults at high risk of Alternate Level of Care designation. More than 5,700 patients underwent the ED-Screener, 53.3% of whom were not self-reliant. Subsequent focused screening and assessment identified 457 patients, 93.3% of whom were not self-reliant, and who had significant impairments in function, mobility, and cognition, as well as social vulnerability. A staged assessment approach based upon the interRAI Acute Care Suite can efficiently identify older adults with risk factors for Alternative Level of Care designation.

Introduction

With increasing age, many adults develop significant multimorbidity, disability, and frailty.^1,2 Our healthcare system is poorly configured to support people with complex and inter-related health and social needs who also represent its greatest users, leading to unstable health, accelerated functional decline, and hospitalization.^1,2 Hospital stays are often complicated by iatrogenic delirium and deconditioning, leading to delayed discharges, high readmission rates, and institutionalization.^3,4

In Canada, patients who no longer require hospital services but who cannot be discharged home receive an Alternate Level of Care (ALC) designation. ⁵ In 2023-2024, there were over 3 million reported ALC days across Canada, representing 17% of all hospital days, predominantly for patients 65 years and over. ⁶ High ALC rates lead to hospital and Emergency Department (ED) overcrowding and increased wait times, delays in ambulance off-loading, and ambulance diversion. ⁷ Solutions intended to improve “patient flow” through hospitals include building more Long-Term Care (LTC) homes, though wait lists for these remain lengthy because of insufficient beds and post-pandemic human resource shortages.^8,9 Some provinces have created Transitional Care Units to provide short-term restorative care to people pending discharge home or to an LTC bed. ¹⁰

A recent qualitative study of 300 hospital managers in Western Canada suggests that patient flow interventions for older patients with complex needs have been ineffective, in large part due to a mismatch between available resources and the characteristics and needs of ALC patients.^11,12 In contrast to relatively homogenous, less complex, and more easily described patient populations (e.g., joint replacement), the breadth of factors underlying the complexity of older ALC patients is perceived to be more difficult to capture, and this lack of pertinent data precludes the development of appropriate care pathways. ¹¹

The patient characteristics associated with ALC designation among older adults fall into three main categories: functional, cognitive, and social vulnerability.^13-15 Acute care clinical information systems are generally poorly configured to capture such information, but expanding their capacity to do so must be done carefully to minimize the associated documentation burden.^16,17 A graduated strategy starting with brief case-finding followed by focused geriatric assessment has been proposed to efficiently identify at-risk older adults upon presentation to EDs and provide them timely senior-friendly care measures to avert further functional and cognitive decline and promote a safer discharge back home.^18,19

The objective of this manuscript is to determine whether such a staged strategy based upon the interRAI Acute Care Suite of instruments can identify older adults with functional, cognitive, and social vulnerability risk factors for ALC designation.

Methods

Study Design

We conducted a prospective cohort study evaluating the feasibility of a staged assessment strategy for older adults presenting to acute care hospital EDs across Canada. The strategy deploys three standardized interRAI instruments designed for acute care: (1) the ED-Screener ²⁰ ; (2) the ED Contact Assessment (EDCA) ²⁰ ; and (3) the Acute Care Comprehensive Geriatric Assessment (AC-CGA). ²¹ We describe the study using the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist for cross-sectional studies and the Reporting of studies Conducted using Observational Routinely collected Data (RECORD) Extension. ²² Ethics approval was obtained by relevant university and hospital research ethics boards.

Setting and Participants

We used convenience sampling to recruit older patients presenting to 10 EDs in Ontario (n = 5), Quebec (n = 4), and Newfoundland (n = 1) between April 2017 and July 2018. Participating EDs were in high volume urban centres except for one located in a small rural town. Trained research assistants at each ED, either seconded from the hospitals or hired for the project, recruited consenting community-dwelling individuals aged 75 years and older (or younger based on clinical judgment). Data collection began formally after a 1 to 2-month start-up period, with patients recruited over 4 to 6 months on weekdays between 8 AM and 7 PM. Participants were screened with the ED-Screener App implemented on Blackberry tablets provided by software vendor HealthIM. Those identified as high risk with the ED-Screener underwent further assessment with the more comprehensive EDCA. Patients expected to die in the following 24 hours or who did not speak English or French were excluded from the study. Patients confirmed as high risk with the EDCA and who were admitted to hospital underwent the AC-CGA. Clinicians at study sites were not provided with any assessment results.

Assessments were completed by research assistants. Where possible, research assistants also ascertained patient disposition, though not all sites were staffed to do so. Not all eligible patients initially assessed with the ED-Screener went on to further assessment according to the stage strategy because of variable availability of research assistants across sites, patient disposition occurring at all hours of the day, or because of illness severity. Figure 1 demonstrates the flow of patients through the study.

Figure 1.

Staged Assessment Strategy for the Study

Measurements and Variables

The ED-Screener, EDCA, and AC-CGA are interoperable instruments developed by interRAI, an international non-profit scientific organization specialized in developing standardized assessment systems for vulnerable populations (https://www.interrai.org/).

ED-Screener

The ED-Screener is based on the Assessment Urgency Algorithm (AUA), a 1-2 minute decision tree with 11 items assessing key health indicators, including cognition, mood, function, and caregiver burden. ²⁰ An AUA score of 5 or 6 out of 6 predicts the need for comprehensive geriatric assessment with a sensitivity of 93%, positive predictive value of 82%, and false discovery rate of 18%.^23,24 The ED-Screener meets several core requirements for ED screening, having been shown to be easy to use, convenient, and feasible to implement. ²⁵ Outcome variables of interest pertaining to the ED-Screener are the AUA score and Self-Reliance Index (SRI), a global measure of function defined as the presence of impairments in completing basic activities of daily living or in cognitive skills for daily decision-making. ²⁶

EDCA

The EDCA is a 15-20 minute 39-item clinical assessment instrument to support care and disposition planning for older adults presenting to the ED, including those with complex needs who would benefit from targeted referrals to community or specialized geriatric services. ²⁰ Completed by a trained assessor using clinical information from multiple available sources, the EDCA provides risk indicators and health status measures for cognition, delirium, mood, substance abuse, function, falls, pain, frailty, and institutional risk. The EDCA has high inter-rater reliability and good to excellent content and predictive validity. Outcome variables of interest pertaining to the EDCA include the AUA, functional indicators (SRI, capacity to perform activities of daily living with or without set-up assistance), mobility indicators (needing supervision or assistance for stairs, any fall in last 90 days), cognitive indicators (supervision for medication management, not independent with daily decision-making, any behaviour symptoms, and acute change in mental status), and social vulnerability indicators (caregiver overwhelmed, patient does NOT have a support person that is positive towards discharge).

AC-CGA

The AC-CGA is a 60-75 minute 110-item instrument for the comprehensive geriatric assessment of hospitalized older adults with complex needs. ²¹ It provides risk indicators and health status measures for cognition, delirium, mood, instrumental and basic activities of daily living, frailty, pain, wounds, malnutrition, falls, and institutional care. ²¹ The AC-CGA has undergone extensive reliability and validity testing.^27-31 Outcome variables of interest include functional indicators (SRI, Activities of Daily Living Hierarchy,^32,33 Instrumental Activities of Daily Living Capacity Hierarchy, ³³ mobility indicators (needing supervision or assistance for stairs, any fall in last 90 days), cognitive indicators (Cognitive Performance Scale), ³⁴ behaviour symptoms, and acute change in mental status) and social vulnerability indicators (patient does NOT have a support person that is positive towards discharge).

Data Sources

Participating facilities recorded data in their own electronic systems and shared de-identified data with the research team. Research assistants used a data collection system for the interRAI ED Contact Assessment provided by HealthIM to enter de-identified assessment information. These data were shared with the research team and stored on the interRAI Canada server at the University of Waterloo.

Bias

We obtained written informed consent from all participants or a verified substitute decision maker. This allowed us to include individuals living with cognitive impairment, who are often excluded from ED research. Research assistants were instructed to follow informed consent procedures for their specific hospital, including plain-language discussions about confidentiality, voluntary participation, and the right to withdraw. We aimed to recruit outside major urban areas, including a hospital in a smaller rural town. We confirm that the results align with our expectations and previous literature.

Study Size

Duration of recruitment and numbers of participants were pre-specified by each site based on their unique resource limitations, and thus there were significant differences in sample size between sites. Because of low recruitment at some sites, we present data for the overall sample only.

Quantitative Variables

Before analyzing the data, we collapsed several variables to ensure privacy, prevent small cell counts that could cause analytical problems, and ensure that differences between levels would allow for clinically meaningful comparisons. Grouped variables included organization, age, sex, and outcomes.

Statistical Methods

Descriptive statistics were used to depict the samples according to each specific assessment instrument. The research team used SAS Enterprise Guide Version 9.4 to conduct all statistical analyses (SAS Institute, Inc., Cary, NC).

Data Access and Cleaning Methods

We had full access to the database used to create the study population. Since there were minimal missing data (<5%), we removed observations with missing data from the analysis.

Results

A total of 5,707, 1,075, and 457 ED-Screeners, EDCAs, and AC-CGAs were completed, respectively, of which 678 (11.9%) were under 75 years of age. Table 1 presents the functional, cognitive, and social vulnerability characteristics of patients in the study. Figure 2 describes the flow of patients in the study.

Table 1.

Characteristics of Patients Assessed With the ED-Screener, EDCA, and AC-CGA

ED-Screener	All		Women		Men
	n	Age %85+	n	Age %85+	n	Age %85+
All	5,707	39.5%	3,247	43.2%	2,460	34.6%
AUA 1-3 (self-reliant)	2,663	82.4 ± 5.3	1,423	82.6 ± 5.5	1,240	81.7 ± 5.1
AUA 4-6 (not self-reliant)	3,044	85.4 ± 6.0	1,824	86.1 ± 6.2	1,220	84.4 ± 5.5

Legend: AUA = Assessment Urgency Algorithm; EDCA = Emergency Department Contact Assessment; AC-CGA = Acute Care Comprehensive Geriatric Assessment; ADLH = Activities of Daily Living Hierarchy (scored from 0 to 6, 0 being independent and 6 completely dependent); IADLCH = Instrumental Activities of Daily Living Hierarchy (scored from 0 to 6, 0 being independent and 6 completely dependent); CPS = Cognitive Performance Scale (scored from 0 to 6, 0 being intact and 6 comatose)

^a4 patients had “other” gender

^b6 patients missing data

^c7 patients missing data

^d6 patients missing data

Figure 2.

Patient Flow Through the Study

ED-Screener and EDCA

Slightly over 53% of patients assessed with the ED-Screener were not self-reliant based on AUA scores of 4 to 6; 45.3% had an AUA of 5 or 6 and were eligible for EDCA assessment. Correspondence between AUA scores from the ED-Screener and EDCA was very good: 166 ED-Screener patients with an initial AUA of 5-6 were found to have a lower AUA with the EDCA, though 70% of these were an AUA of 4.

Of the 1,075 patients undergoing an EDCA, almost 84% were not self-reliant. High proportions of these patients faced challenges in one or more of the three risk factor categories for ALC designation if hospitalized. Pre-morbid functional impairment was common, with 91.8% needing more than set-up assistance for at least one activity of daily living. Mobility problems affected 74.9%, with 48.8% having had falls in the last 90 days and 54.6% requiring supervision or assistance with stairs.

Challenges with cognitive function were common, with 80.7% requiring at least supervision for medication management. Pre-morbid daily decision-making was impaired in 29.7%, rising to 41.7% on presentation to the ED. An acute change in mental status, indicative of possible delirium, was present in 23.5% of patients assessed with the EDCA.

Finally, 55.3% of the patients had indicators of social vulnerability, with 47.0% of their informal caregivers endorsing feeling overwhelmed, and 17.8% not having a support person positive about a discharge home.

AC-CGA

Risk factors for ALC designation were even more common among the 457 patients assessed with the AC-CGA, 93.3% of whom were not self-reliant. Of these, 84.3% were dependent for at least one basic activity of daily living and 98.4% required assistance for at least one instrumental activity of daily living. Challenges with mobility were present in 93.3% of patients, with 90.6% requiring at least supervision with stairs and 52.5% having had a fall in the last 90 days.

Only 31.8% of the patients assessed with the AC-CGA had intact cognition as measured by a Cognitive Performance Scale of 0. An acute change in mental status, suggestive of delirium, was present in 25.7%. The proportion of those without a support person positive about discharge home was 21.3%.

Discussion

This manuscript demonstrates how staged assessment may offer an efficient approach to assess older adults presenting to EDs with a high burden of functional, cognitive, and social vulnerability, and thus identify those at risk of ALC designation. This approach minimizes documentation burden by using brief screeners to triage high-risk geriatric patients who require a lengthier comprehensive assessment. Initial screening identified slightly over half of patients as not self-reliant, a figure rising to over 93% of those receiving the comprehensive assessment. Virtually all those assessed with the AC-CGA had some degree of difficulty completing instrumental activities of daily living and mobility problems, almost 70% had cognitive impairment, and over 20% did not have a support person feeling positive about discharge to the community.

These results build upon those from other studies demonstrating the advantages of staged assessment of older adults presenting to EDs, including one showing that patients at high risk based on the ED-Screener had an over 50% decreased odds of being discharged home. ³⁵ In addition to minimizing assessment burden, a staged approach enables clinicians and decision-makers to concentrate efforts and limited human resources on those most at risk. ¹⁹

These results demonstrate how staged assessment based on risk factors for ALC designation resolves the challenge of efficiently characterizing hospitalized older adults with complex health needs, as issue previously identified in the Western Canada hospital manager study. ¹¹ The predictive capacity of the interRAI Acute Care Suite can inform care pathways to target high-risk patients with senior friendly interventions. In turn, hospitals can be enabled to prevent hospital-acquired functional decline and delirium, initiate proactive referrals to specialized geriatric services and inpatient rehabilitation, promote earlier discharge planning with home and community service to support informal caregivers and address social vulnerabilities, enhance rates of discharge home, and reduce readmissions. ¹⁹

Implementing such a strategy requires that hospitals review existing documentation practices to identify redundancies and fill gaps. A survey of 11 hospitals in the state of Victoria, Australia found 152 standard assessment forms containing over 3,700 items, 17% of which were duplicated by multiple forms, with domains such as cognition inconsistently assessed. ¹⁶ A similar survey in Nordic countries examined documentation practices pertinent to older adults among acute care physicians and nurses and found both significant redundancies and gaps in the documentation of key geriatric domains such as function and cognition. ¹⁷ To capitalize on their full functionality, the interRAI instruments should be embedded into electronic medical records, and professional practice leads engaged to develop integrated interprofessional care pathways based upon their outputs. The investments required to update clinical information systems in hospitals should more than offset the maintenance of inefficient legacy systems that are ultimately more costly to patients and the healthcare system. ¹⁶

An important limitation of this study is that not all patients could be fully assessed as per protocol for reasons noted above, and the assessments were conducted outside of routine clinical care. However, our results are consistent with those of other studies. Moreover, it is likely that a large proportion of patients who did not undergo the full AC-CGA assessment may have been too sick, thus underestimating the true efficiency of staged assessment.

Additional benefits of staged assessment based on acute care interRAI instruments is their high degree of compatibility with community-based interRAI instruments, such as the interRAI HC (home care), widely deployed across most of Canada. ³⁶ Interoperability is required to create an integrated health information system, bringing hospital systems in line with those in home and community care, LTC, mental health, and other sectors, and allowing for a longitudinal understanding of a patient’s health trajectory. It facilitates information sharing and streamlines acute care documentation if community-based assessments are available. ³⁷ Similarly, the use of interRAI instruments in acute care can facilitate referrals to home and community care services. ³⁷ Home care and LTC assessments, even up to 6 months old, have been shown to still have utility in predicting hospital outcomes and ALC designation.^15,38

Conclusion

We have shown that staged assessment starting in hospital EDs and based on interRAI acute care instruments can efficiently identify older adults with risk factors for ALC designation.

Ethical Approval

Institutional review board approval was not required.