Article, Geriatrics

Validation of the Elderly Risk Assessment Index in the Emergency Department

a b s t r a c t

Objectives: The Elderly Risk Assessment (ERA) score is a validated index for primary care patients that predict hospitalizations, mortality, and Emergency Department (ED) visits. The score incorporates age, prior hospital days, marital status, and comorbidities. Our aim was to validate the ERA score in ED patients.

Methods: Observational cohort study of patients age >= 60 presenting to an academic ED over a 1-year period. Re- gression analyses were performed for associations with outcomes (hospitalization, return visits and death). Me- dians, interquartile range (IQR), odds ratios (OR) and 95% confidence intervals (CI) were calculated.

Results: The cohort included 27,397 visits among 18,607 patients. Median age 74 years (66-82), 48% were female and 59% were married. Patients from 54% of visits were admitted to the hospital, 16% returned to the ED within 30 days, and 18% died within one year. Higher ERA scores were associated with: hospital admission (score 10 [4-16] vs 5 [1-11], p b 0.0001), return visits (11 [5-17] vs 7 [2-13], p b 0.0001); and death within one year (14 [7-20] vs 6 [2-13], p b 0.0001). Patients with ERA score >= 16 were more likely to be admitted to the hospital, OR 2.14 (2.02-2.28, p b 0.0001), return within 30 days OR 1.99 (1.85-2.14), and to die within a year, OR 2.69

(2.54-2.85).

Conclusion: The ERA score can be automatically calculated within the electronic health record and helps identify patients at increased risk of death, hospitalization and return ED visits. The ERA score can be applied to ED pa- tients, and may help prognosticate the need for advanced care planning.

Background

The United States (US) population is aging. The average life expec- tancy at birth has increased from 47 years in 1900 to 79 years in 2014. Currently, the US population over age 65 is estimated to be roughly 14% [1,2]. By 2030, the proportion of the population over 65 years of age will exceed 20%, or over 70 million people [3]. Worldwide, the num- ber of adults over 60 years of age will exceed 2 billion by 2050 [4]. Con- cordant with this increase, the number of older adults presenting to the

Abbreviations: ERA, Elderly Risk Assessment; ED, emergency medicine; IQRs, interquartile ranges; OR, odds ratios; CI, confidence intervals; US, United States; ISAR, identification of seniors at risk; TRST, triage risk screening tool; EHR, electronic health re- cord; STROBE, strengthening the reporting of observational studies in epidemiology; IRB, Institutional Review Board; CHF, congestive heart failure; MI, myocardial infarction; CAD, coronary artery disease; CVA, cerebrovascular accident or stroke; COPD, chronic ob- structive pulmonary disease.

* Corresponding author at: 200 First Street SW, Generose Building G-410, Mayo Clinic, 55905, USA.

E-mail address: [email protected] (F. Bellolio).

Emergency Department (ED) has also increased over time, accounting for 30% of ED visits, a disproportionally large number [5].

As the “baby boomer” generation continues to age, emergency med- icine must prepare to proactively care for this vulnerable population, and consider focused efforts to decrease morbidity and mortality. The ability to accurately identify which older adults are at increased risk of an adverse outcome, such as return visit or death, after an ED visit will allow us to take preventive measures [6-10]. Different Screening tools have been created in the past years, such as the Runcinam question- naire, the Rowland questionnaire, the Silver code, the Identification of Seniors at Risk (ISAR) tool, the Triage Risk Screening Tool . The ISAR tool and the TRST tool are the most studied screening tools for el- derly patients in the ED [11]. The ISAR tool [12] has been recommended to screen older adults [13], however, a recent systematic review showed that the ISAR is not sufficiently accurate to predict increased risk of ED return visits, functional decline, Hospital readmission or adverse out- comes [14]. ISAR is particularly difficult to use with older adults who may have cognitive impairment or poor recall due to acute illness or in- jury [15]. The TRST showed a limitED capacity to discriminate between

https://doi.org/10.1016/j.ajem.2019.11.048

older adults who did or did not have an adverse outcome following dis- charge from the ED [16]. Given these limitations, we need a tool to accu- rately identify Geriatric patients at higher risk to allow the medial system to proactively intervene in hopes of reducing morbidity and im- proving quality of life. As service needs increase our system will become overstretched and high quality care will still need to be provided. Fac- tors related to an ED visit impact not only the individual person, but the health care system as a whole as service needs increase [17].

In this study we evaluate the Elderly Risk Assessment (ERA) score, a tool that has been shown to predict hospitalizations and ED visits in adults older than 60 years of age in the Primary care setting [18]. The ERA score is automatically calculated within the electronic health record and has been implemented in daily practice within primary care [19]. The score has been previously validated as a predictor of mortality in community-dwelling adults who sustain a Hip fracture [19]. An ERA score of 16 or greater has been associated with higher mortality. This tool has also been used to identify subgroups of patients that may ben- efit from advanced care planning or referral to a transitional care pro- gram [20]. ERA score is currently used in our primary care practice to assist with clinical decision making and to identify patients who may benefit from referral to an outpatient palliative care program. Recogniz- ing a patient needs assistance early can minimize the risk of readmissions and adverse events post-discharge [21,22]. The ERA score can predict the need for referral to a social worker, physiothera- pist, occupational therapist, and gerontology clinical nurse specialist. Utilization of the ERA in this way for elderly patients being discharged may help facilitate safer discharge plan [23]. This should include connecting patients at risk for rapid return to the ED or hospital with community social support and medical resources [24].

The aim of this study was to validate the ERA score as a predictor of hospitalization, mortality and return visits in ED patients.

Methods

Study design and setting

This was an observational cohort study of consecutive patients age 60 years and older who presented to a quaternary care academic ED with 77,000 annual patient visits. This study was approved by the Insti- tutional Review Board. All patients included in analysis provided Min- nesota research authorization for Medical records review. We followed reporting guidelines including STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) for cohort studies [25].

Study protocol

A protocol was developed prior to initiation of data collection. This study was done in collaboration with the Department of Internal Medi- cine. We defined criteria for inclusion and exclusion and data was auto- matically harvested from the electronic medical record (EHR).

Selection of participants

Electronic data were retrieved for consecutive patients age 60 and older that presented to the ED between January and December 2017. Only patients with a primary care physician within our health care sys- tem were included to ensure accuracy of the ERA score.

Data collection

Each component of the ERA score and the calculated total score were retrieved from the EHR. All patient visits were included, and a single pa- tient could have multiple ED visits during the study period. All visits were screened for 30-day returns.

The ERA incorporates a weighted score of age, number of hospital days in the prior 2 years, marital status, medical diagnoses of congestive heart failure (CHF), myocardial infarction (MI), coronary artery disease (CAD), diabetes mellitus, cerebrovascular accident or stroke (CVA), chronic obstructive pulmonary disease (COPD), cancer, and dementia (Table 1). The minimum score on the index is -1 and the maximum score possible is 34 [26].

A cut-off score of 16 has been associated previously with higher mortality and was used for binary analyses [20].

Key outcome measures

We followed patients through health record review for 1 year from the index ED visit to determine whether there was a return visit and/ or death. Mortality at 30, 90, 180 days, and one year, disposition after the ED visit and return visits to the ED within 30 days were collected.

Data analysis

Continuous features were summarized with medians and interquar- tile ranges (IQRs), odds ratios (OR) with 95% confidence intervals (CI) were calculated to estimate the strength and direction of each associa- tion. Categorical features were summarized with frequency counts and percentages. Comparisons of interest were: outcomes of patients with an ERA score b 16 versus score of 16 and higher, as well as return visits and mortality. Chi-square and Wilcoxon test were used for statis- tical analysis depending on the type and distribution of the data. Statis- tical analyses were performed using version 9.4 of the SAS software package (SAS Institute; Cary, NC). All tests were two-sided and p- values b 0.05 were considered statistically significant.

Results

We had a total of 29,991 ED visits by 20,452 patients 60 years and older. Patients age 60 and above represent 39% of all visits to our ED. Of these, 8.6% of the visits did not consent for medical record review or did not have a primary care doctor within our Health Care System and were excluded. Patients came from home, independent living, assisted living or nursing homes. The final cohort included 27,397 visits among 18,607 patients. The median age was 74 years (IQR 66-82), 48% were female and 59% were married. A total of 54% of patients in our co- hort were admitted to the hospital and 16% of the patients returned to the ED within 30 days. There were 1189 deaths within 30 days (4%) and 4735 (18%) within one year. (Table 2)

The median ERA score was 7 (IQR 2-14). When comparing those ad- mitted to the hospital during the index ED visit, we found that those with higher ERA scores were more likely to be admitted to the hospital with a median ERA score of 10 (4-16) vs 5 (1-11), p b 0.0001. When comparing those that returned to the ED versus not, we found that those with higher ERA score were more likely to return to the ED within 30 days, with a median ERA score of 11 (5-17) for those who returned

Table 1

Elderly Risk Assessment Index (ERA) – scoring system.

Married -1

Age 70-79 1

Age 80-89 3

Age 90 or more 7

1-5 Hospital days in the previous 2 years 5

6 or more hospital days in previous 2 years 11

Comorbidities

Diabetes mellitus	2
Stroke	2
COPD	5
Cancer	1
Dementia	3
CAD/MI/CHF	3

Table 2

Demographics (N = 27,397 visits, 18,607 patients).

N = 27,397

attainable and can be automatically calculated via the electronic medi- cal record. The ERA score identifies ED patients at higher risk of death within 1 year, hospitalization and return ED visits.

Age at ED visit (median, years) 74 (IQR 66-82)

60-79 years 18,716 (68.3%)

>=80 years 8681 (31.7%)

Sex

Female 13,157 (48.0%)

Male 14,240 (52.0%)

Marital status

Married 16,247 (59.3%)

Widow 5935 (21.7%)

Divorced 2910 (10.6%)

ERA Score Index 7 (IQR 2-14)

Inpatient length of stay days 10 (6-18) Comorbidities

Coronary artery disease 12,590 (46.0%)

Myocardial infarction 4609 (16.8%)

Congestive heart failure 7659 (28.0%)

Cerebrovascular accident or stroke 5122 (18.7%)

Chronic obstructive pulmonary disease 5512 (20.1%)

Diabetes mellitus 9707 (35.4%)

Dementia 5066 (18.5%)

Cancer 7963 (29.1%)

vs 7 (2-14) for those who did not, p b 0.0001. Overall male patients were more likely to return to the ED within 30 days compared to fe- males (OR 1.16, 95% CI 1.1-1.2, p b 0.001). In addition, patients with a higher ERA score were more likely to die within 1 year, with a median score of 14 (7-20) for those that died vs 6 (2 to 13) for those alive at 1 year, p b 0.0001.

When evaluating patients with ERA >= 16, these patients were more likely to be admitted to the hospital (OR 2.14, 95% CI 2.02-2.28, p b 0.001), more likely to return to the ED within 30 days (23% vs 13%, OR 1.99, 95% CI 1.85-2.14, p b 0.001), and to die within one year, (34%

vs 13%, HR 2.69, 95% CI 2.54-2.85, p b 0.0001) compared to those with scores b16. An ERA score of >=16 was predictive of 30 day mortality and 9% of patients with ERA >=16 died within 30 days of ED visit versus 3% of those with ERA b16 (OR 2.68, 95% CI 2.38-3.02, p b 0.001). (Table 3)

Discussion

Our study of over 18,000 ED patients 60 years of age or older suc- cessfully validates the ERA score in the ED patient. This score is easily

We found that a large proportion of older adults (54%) were admit- ted to the hospital and those with higher ERA scores were more likely to be admitted. Biese et al. described a similar hospital admission distribu- tion from ED encounters ranged from 4% to 52% (mean 32%) [27]. This is not surprising as one would expect higher ERA scores to be associated with greater Need for hospitalization given the variables used to calcu- late the score.

There is strong evidence suggesting ED visits without hospitalization are also associated with serious negative consequences among community-living older adults [28]. Being evaluated in the ED without a hospitalization is associated with increased risk and worse outcomes including worse disability scores, functional decline, nursing home ad- missions and mortality in the months subsequent to discharge [28]. Nearly half of our older patients are dismissed from the ED back to the community or nursing home. These patients do not routinely receive an evaluation of their likelihood of return visit. Previous studies have re- ported 20% to 35% of older ED patients experiencing some functional de- terioration in the 6 months after an ED visit [29-31]. Thus we are missing an opportunity identify those who may decline subsequently and to prevent this deterioration.

The 30-day readmission rate has been shown to be a useful indicator of increased risk of major cardiac adverse event, All-cause death and myocardial infarction (64 vs 21%; 28 vs 6%; and 20 vs 2.7%, respectively) [32]. A study in 2018 reported percentage of ED encounters with 30-day return from the 167 EDs ranged from 18% to 39% (median = 23%) [27]. In our study, a total of 16% of patients returned to the ED within 30 days; those with ERA score >=16 had almost a 2 fold increase in 30 day ED re- turn (23% vs 13%).

We found that 18% of patients in our cohort died within one year, and these patients had a higher ERA score. Takahashi et al. reported sig- nificantly higher 2-year mortality among 12,650 community adults with a relative risk of 51.4 among those with an ERA score of 16 or greater compared with those in the lower risk group [20].

Older adults represent an ever-growing population that utilizes ED services, with 12-21% of all visits being made by older adults. ED use for medical complaints account for nearly 80% of all visits by older adults [33]. Older adults’ seek care in the ED for many reasons. Many acute is- sues arise like change in cognitive status, abdominal pain or syncope but often times older patients visit for exacerbations of chronic conditions

Table 3

Comparison of features for visits by patients with ERA Score b16 and >=16 (N = 27,397 visits).

	b16 (n = 21,629)	>=16 (n = 5768)	OR (95% CI)	p-Value
Age at ED visit (years)			2.75 (2.59-2.92)	b0.0001
60-79 years	15,840 (73.2%)	2876 (49.9%)
>=80 years	5789 (26.8%)	2892 (50.14%)
Gender			1.20 (1.14-1.28)	b0.0001
Female	10,598 (49%)	2559 (44.4%)
Male	11,031 (51%)	3209 (55.6%)
Marital status				b0.0001
Married	13,753 (63.6%)	2494 (43.2%)
Widow	4005 (18.5%)	1930 (33.5%)
Divorced	2086 (9.6%)	824 (14.3%)
Death within 1 year (N = 26,699)	2766 (13.2%)	1969 (34.3%)	HR 2.69 (2.54-2.85)	b0.0001
Death within 30 days (N = 26,699)	701 (3.4%)	488 (8.5%)	2.68 (2.38-3.02)	b0.0001
Return within 30 days (N = 26,699)	2796 (13.3%)	1347 (23.5%)	1.99 (1.85-2.14)	b0.0001
Inpatient days (length of stay) median (IQ)	0 (0-0)	10 (6-18)		b0.0001
Comorbidities
Coronary artery disease	8130 (37.6%)	4460 (77.3%)	5.66 (5.29-6.06)	b0.0001
Myocardial infarction	2456 (11.4%)	2153 (37.3%)	4.64 (4.34-4.5)	b0.0001
Congestive heart failure	3982 (18.4%)	3677 (63.8%)	7.79 (7.31-8.30)	b0.0001
Cerebrovascular accident or stroke	2728 (12.6%)	2394 (41.5%)	4.92 (4.6-5.25)	b0.0001
Chronic obstructive pulmonary disease	2591 (11.2%)	2921 (50.6%)	7.54 (7.06-8.05)	b0.0001
Diabetes mellitus	6352 (29.4%)	3355 (58.2%)	3.34 (3.15-3.55)	b0.0001
Dementia	2639 (12.2%)	2427 (42.1%)	5.23 (4.89-5.59)	b0.0001
Cancer	5671 (26.2%)	2292 (39.7%)	1.86 (1.75-1.97)	b0.0001

resulting from CAD, COPD, HF, DM, or dementia. Frail adults visit after a fall, for problems conducting activities of daily living and complex psy- chosocial needs [34-36]. Identifying elderly patients at risk for adverse outcomes using a simple and easily calculated tool like the ERA index will allow providers in the ED and out-patient to deploy successful in- terventions to reduce return ED visits. The ED may mobilize home care or begin Physical therapy to mitigate falls which ultimately impact safety and overall health of older patients. Earlier intervention for this at-risk group has the potential to lead to an improved quality of life and medical care that is concordant with patient beliefs.

As we evaluate the prior studies we can plot the course of an older patient who is seen in the ED and dismissed home without an evalua- tion of subsequent risk. We can see that dismissal home after an ED visit, by itself, predicts a functional decline [28]. The higher ERA score is related to a higher probability of patients returning to the ED, hospi- talization and death [37].

Evidence indicates that adequate post-discharge interventions fo- cused on high-risk elderly patients can decrease rate of readmission and have subsequent cost savings [38,39]. Project Boost (Better Out- comes for Older adults through Safe Transitions) is one example of a successful intervention to decrease hospital readmissions [40]. Koehler et al. developed a personalized care intervention, based on medication counseling and condition education which reduced 30 day readmission and ED visits compared to the control group (10% versus 38%) [41]. Connecting the concepts of validated screening tools for geriatric ED populations combined with proven interventions to decrease readmissions will improve the quality of life for our growing population of older patients.

Similar to previously identified uses of the ERA score, we found that using a cut off of 16 can help identify a high risk cohort; these patients may benefit from additional resources such as advanced care planning, multi-disciplinary team coordination in the ED and palliative care con- sultation or referral. Identifying this high-risk population has potential to allow for early referral, which may reduce readmission and morbid- ity. The score may also be used to target patients who would benefit from goals of care discussion while in the ED to guide subsequent treatment.

Limitations

Data quality relies on the accuracy of the medical record and missing data may occur. Additionally, the score is only available to patients with a primary care physician from the institution, which may increase cer- tainty from the score but limits its generalizability. This study was con- ducted in a single center with a predominantly white cohort, which may limit external validity. A small percentage of the patients did not allow their medical records to be used for research and were excluded; per the IRB report, there were no differences in age or gender between those who consent and declined their records for review.

Conclusion

The ERA score can be automatically calculated within the elec- tronic medical record and can help identify older ED patients at higher risk for adverse outcomes, including death, hospitalization and return visits. This score identifies patients who will benefit from additional resources such as geriatric referral, social worker evaluation in the ED, coordination of home care resources, goals of Care discussions, and palliative care consultation or referral. An ED visit is an opportunity to assess risk, the ERA score helps identify those at higher risk and will allow implementation of Preventive strategies aimed to mitigate adverse outcomes by coordinating addi- tional patient resources, reducing readmission and promoting care concordant with patient values.

Author contributions ERA

Espinoza Suarez: Data curation; Roles/Writing – original draft; Writing – review & editing.

Walker: Roles/Writing – review & editing.

Jeffery: Formal analysis; Investigation; Methodology; Supervision; Validation; Visualization; Roles/Writing – review & editing.

Stanich: Roles/Writing – review & editing. Campbell: Roles/Writing – review & editing.

Lohse: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Re- sources; Software; Supervision; Validation; Visualization; Roles/ Writing – original draft; Writing – review & editing.

Takahashi: Conceptualization; Data curation; Resources; Roles/ Writing – review & editing.

Bellolio: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Re- sources; Supervision; Validation; Visualization; Roles/Writing – orig- inal draft; Writing – review & editing.

Funding sources

This research was supported though the CCaTS Small grant program, part of Mayo Clinic CCaTS grant number UL1TR000135 from the Na- tional Center for Advancing Translational Sciences (NCATS), a compo- nent of the National Institutes of Health (NIH). Its contents are solely the responsibility of the authors and do not necessarily represent the of- ficial view of NIH.

References

AJEM