Enhancing emergency department triage for older patients: a prospective study on the integration of the identification of seniors at risk

Abstract

Background

Older adults are a growing demographic in emergency departments (EDs) worldwide, yet traditional triage systems often fail to account for their unique risks, leading to under-triage and adverse outcomes. The Identification of Seniors at Risk (ISAR) tool offers a pragmatic approach to enhance risk stratification, but its integration into ED triage systems remains underexplored.

Methods

This prospective single-center observational cohort study assessed older patients (≥ 65 years) using both the standard ED triage system and the ISAR scale. After a 30-day follow-up, triage levels were retrospectively adjusted upward by one level for patients with ISAR scores ≥ 2. The predictive accuracy of the revised triage system was compared to the original system using logistic regression and receiver operating characteristic (ROC) curve analysis.

Results

Among 973 patients completing follow-up, 38.1% had an ISAR score ≥ 2. Older patients (≥ 75 years) were more likely to be classified as high risk and had significantly higher rates of adverse outcomes, including ICU admission and 30-day mortality. The revised triage system slightly improved discriminative ability in patients aged ≥ 65 years (AUC 0.697 to 0.714), with stable performance maintained in those aged ≥ 75 years (AUC 0.703). Sensitivity declined slightly, while specificity improved.

Conclusion

Integrating ISAR into ED triage modestly enhanced the identification of older patients at risk for short-term adverse outcomes, particularly among those aged ≥ 65 years. These findings support the value of incorporating geriatric screening into routine triage to enable more tailored risk stratification. Further studies are needed to evaluate implementation feasibility across different healthcare settings and to inform integration into routine practice.

Clinical trial number

Not applicable.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12873-025-01253-5.

Background

The global aging population has profoundly strained healthcare systems, particularly in emergency departments (EDs) where older adults now represent a rapidly growing patient cohort [1]. Compared to younger populations, this demographic is more likely to experience complex medical conditions, such as multiple comorbidities, cognitive decline, and functional impairments [2]. These age-related vulnerabilities predispose them to adverse outcomes such as prolonged hospitalization, unplanned ICU admissions and even death, necessitating precision in ED triage to mitigate risks of clinical deterioration [3]. Despite this pressing need, widely used triage tools such as the Emergency Severity Index (ESI) and the Rapid Emergency Triage and Treatment System (RETTS) were primarily developed for the general adult population and rely heavily on vital signs and chief complaints [4]. However, such tools may fail to detect geriatric-specific risks—such as frailty, cognitive impairment, functional decline, and polypharmacy—that are not readily apparent during initial assessment but are strongly associated with poor outcomes [5]. This leaves older patients vulnerable to under-triage, in which high-risk individuals are misclassified as lower acuity due to atypical or subtle presentations [6].

Internationally, various efforts have been made to adapt triage systems to the needs of older adults. For example, the Canadian Triage and Acuity Scale (CTAS) has introduced changes to better account for frailty in the triage process, but challenges remain in standardizing frailty assessments across diverse clinical settings [7]. Additionally, some countries have implemented multidisciplinary approaches by integrating geriatric specialists into the triage workflow [8, 9]. While these approaches provide comprehensive assessments, they often require significant time and resources, making them difficult to sustain in fast-paced ED environments. In China, there is growing recognition of the need to improve triage for older adults in emergency departments [10]. However, this awareness is still in the early stages, and practical implementation remains limited. Many hospitals have yet to develop standardized protocols that adequately address the unique risks associated with aging, leaving room for improvement in the triage process [11].

The Identification of Seniors at Risk (ISAR) scale provides a practical, streamlined approach for assessing older patients in ED [12]. Developed as a simple six-item tool, ISAR evaluates key risk factors such as functional dependence, recent hospitalizations, cognitive impairment, vision problems, and polypharmacy—factors that are critical for identifying older patients at risk of adverse outcomes [13]. Given its simplicity and ease of use, ISAR offers a feasible solution for triage systems in resource-limited EDs.

This study aims to more accurately identify high-risk older patients, referring to those with an increased likelihood of experiencing adverse outcomes such as emergency department revisit within 72 h, hospital admission or death within 30 days, by integrating the ISAR scale into the existing triage system. Ultimately, this integration seeks to enhance emergency geriatric care by providing a more tailored, efficient approach to managing the complex needs of older patients in the ED.

Methods

Study design

This was a prospective, single-center, observational cohort study conducted in the emergency departments (EDs) of a large tertiary hospital in central China. Although the hospital operates three geographically distinct EDs across different campuses, all departments share the same institutional governance, ethical oversight, and triage protocols. The study was therefore designed and implemented as a single-center investigation.

Participants were recruited using a convenience sampling approach between March 7 and March 26, 2024. Screening was conducted by 15 trained triage nurses during their clinical shifts in the three EDs. All 15 triage nurses involved in the study received standardized training on the use of the ISAR scale prior to data collection to ensure consistent application. The training included a detailed review of the ISAR scale criteria, scoring guidelines, and case-based simulations. Due to variable workloads, shift-based staffing, and the limited number of ISAR-trained nurses, not all eligible patients were assessed. Screening was performed only when trained personnel were on duty and clinical conditions permitted, in order to minimize disruption to routine emergency care.

The ISAR assessment was conducted by trained nurses in real time at the triage station, immediately upon patient arrival. The tool was embedded in the department’s intelligent triage system and administered face-to-face with the patient or their accompanying family member. Responses were entered directly into the digital platform as part of routine triage. Importantly, the ISAR score was not displayed to clinicians and had no influence on the patient’s initial triage level or clinical management.

Following initial triage and ISAR assessment, patients were prospectively followed for 30 days. Structured telephone follow-up was used to identify adverse outcomes, and hospital electronic medical records were reviewed to supplement outcome data. After follow-up data collection was complete, a retrospective triage reclassification was performed: patients with an ISAR score ≥ 2 had their original triage level upgraded by one category. The predictive validity of these revised triage levels was then compared with the original triage assignments.

The decision to apply ISAR-based triage reclassification retrospectively was made to ensure patient safety and preserve the integrity of the triage process during the evaluation phase. This approach allowed us to simulate the potential impact of ISAR without influencing real-time clinical decision-making. It also avoided introducing unvalidated modifications to the triage system that could have affected patient flow or outcomes. The study was approved by the institutional ethics committee, and informed consent was obtained from all participants or their legal proxies.

Study setting and population

The study was conducted at three urban EDs across three campuses of a large tertiary hospital in central China, which has over 3,700 beds and an annual ED volume exceeding 200,000 patients. Approximately 20% of ED visits are by patients aged 65 years or older, reflecting a growing demand for geriatric emergency care in this setting. Participants were selected via convenience sampling in March 2024. Inclusion criteria: (1) age ≥ 65 years; (2) informed consent from the patient or their proxy. Exclusion criterion: severe cognitive impairment without a formal proxy.

Instrument

The ISAR tool used in this study is the authorized Chinese version of The Revised Identification of Seniors at Risk (see appendix for details). This six-item screening tool assesses the following questions commonly associated with adverse outcomes in older adults: need for regular help before the current illness/injury, increased need for help in the past 24 h, recent hospitalization (past 6 months), vision impairment, memory problems, polypharmacy (≥ 6 medications). Each item is scored 1 if the patient or caregiver affirms the presence of the issue, and 0 otherwise, yielding a total score ranging from 0 to 6. A score of ≥ 2 indicates a higher risk of adverse outcomes. The tool is simple, non-invasive, and can be administered within minutes, making ISAR a practical and time-efficient tool for use in busy emergency settings.

The triage system used in this study was based on the latest National Expert Consensus on Emergency Triage in China [14]. Compared to international triage systems such as CTAS, ESI, and MTS, the Chinese system emphasizes rapid assessment through objective vital parameters and vital complaints, which enhances reliability and reduces triage time in high-volume emergency settings [15]. Triage was conducted by experienced emergency nurses using a digital triage platform that integrates national triage criteria with automated prompts for symptom classification and vital sign thresholds. The system categorizes patients into four levels: Level I (critical), Level II (severe), Level III (urgent), and Level IV (non-urgent), aiming to standardize decision-making and support efficient patient flow.

Data collection and outcomes

At the first visit, general demographic information and clinical information was collected, including age, gender, and whether the patient arrived by ambulance. Additional variables were extracted automatically through the electronic triage system. “Time of arrival” referred to the patient’s registration time in the emergency department. “Triage department” indicated the clinical unit (e.g., emergency internal medicine, emergency surgery) where the patient was directed after triage. Pain intensity was assessed using the Numeric Rating Scale (NRS).

Vital signs, including respiratory rate, heart rate, systolic and diastolic blood pressure, oxygen saturation, and body temperature, were measured and recorded by triage nurses for all patients at presentation. In rare cases where measurements could not be obtained (e.g., due to patient refusal), this was documented in the triage system.

Fall risk was assessed at triage using the Fall Risk Assessment Standard for Outpatient and Emergency Patients, an institution-wide tool embedded in the triage system. Patients were classified as high risk if they met predefined criteria, such as unstable gait, visual impairment, use of assistive devices, history of falls and recent sedation.

Adverse outcomes were defined as a composite endpoint comprising: (1) ED revisit within 72 h of the index visit (to any facility), (2) hospital admission within 30 days (regardless of admission pathway), or (3) all-cause death within 30 days. Each component was assessed through structured follow-up interviews with patients or family members. Importantly, as some individuals experienced more than one event (e.g., ED revisit followed by hospitalization), the individual event rates are not mutually exclusive and should not be summed. The primary analysis was based on the occurrence of any one or more of these events, with a composite adverse outcome rate calculated to avoid double counting.

Statistical analysis

Categorical variables are presented as numbers and percentages and were compared using the chi-squared test. Continuous variables are presented as means and standard deviations or as medians and interquartile ranges, depending on the distribution of the data. Student’s t-test was used for normally distributed continuous variables, and the Mann–Whitney U test was applied to non-normally distributed continuous variables. Univariable and multivariable logistic regression models were used to identify variables associated with primary outcomes. Odds ratios and 95% confidence intervals are reported for logistic regression analysis. To assess the predictive capacity of the triage system before and after the retrospective triage upgrade, the area under the receiver operating characteristic curve (AUC) was calculated. Missing values were handled using multiple imputation methods to ensure analytical robustness.

For subgroup analysis, patients were categorized into two age groups: 65–74 and ≥ 75 years. This stratification was based on prior literature indicating increased heterogeneity in health risks and care needs among the oldest old, and aimed to explore whether the predictive performance of ISAR varies across older age strata [16]. For the purpose of outcome comparison and regression modeling, the original four-level triage categories were consolidated into two groups: Level I–II (critical care group) and Level III–IV (non-critical care group). This grouping reflects actual clinical practice in the study setting, where Level I and II patients are treated in the resuscitation or intensive monitoring areas, while Level III and IV patients are managed in general care areas.

All statistical analyses were performed using SPSS Statistics 26.0. A two-tailed p-value < 0.05 was considered statistically significant.

Results

Demographic and clinical characteristics of patients

A total of 1324 eligible patients were screened, of whom 973 (73.5%) completed 30-day follow-up and were included in the final analysis (Fig. 1). Among the 351 patients lost to follow-up, 48 could not be reached due to invalid or missing contact information, 132 did not respond to repeated phone calls, and 171 declined to answer follow-up questions despite being successfully contacted. The median age was 73.00 years (IQR: 69.00–80.00), and 45.7% were male. Compared to the 65–74 age group, a higher proportion of patients aged ≥ 75 years were female (58.2% vs. 51.3%, p = 0.033), arrived during the daytime shift (p = 0.044), and were classified as high risk for falls (50.7% vs. 27.9%, p < 0.0001). A total of 38.1% of patients had an ISAR score ≥ 2, with a significantly higher proportion in the older age group (53.6% vs. 26.7%, p < 0.0001). The proportion of ambulance transport was also higher among older patients (18.6% vs. 11.4%, p = 0.002). Triage level distribution differed by age group, with a greater proportion of older patients categorized as higher acuity level (34.0% vs. 20.8%, p < 0.0001). Regarding vital signs, older patients had significantly higher systolic blood pressure, lower diastolic blood pressure, and lower oxygen saturation (Table 1). The median length of hospitalization was also longer in the older age group. Regarding ED disposition, 77.0% of patients were discharged, 13.6% were admitted to a ward, and 9.5% were transferred to the ICU. ICU admission was significantly higher among older patients (13.0% vs. 6.8%, p < 0.0001). At 30-day follow-up, 48.8% of patients experienced at least one adverse outcome, including 45.1% who required hospitalization and 2.6% who died. The incidence of adverse outcomes was significantly higher in the ≥ 75-year-old group (58.5% vs. 41.7%, p < 0.0001) (Table 1).

Fig. 1.

Flow diagram of patient enrollment

Table 1.

Patient’s demographic and clinical characteristics

	Total(N = 973)	Age group		Statistical value (X2, Z)	P-value
		65–74(N = 559)	≥ 75(N = 414)
Age, years, median (IQR)	73.00(69.00,80.00)	69.00(67.00,71.00)	81.00(77.00,86.00)	-26.731	< 0.0001
Gender, n (%)				4.524	0.033
Male	445(45.7)	272(48.7)	173(41.8)
Female	528(54.3)	287(51.3)	241(58.2)
Time of arrival, n (%)				6.231	0.044
8:01–16:00	486(49.9)	260(46.5)	226(54.6)
16:01–24:00	380(39.1)	234(41.9)	146(35.3)
0:01–8:00	107(11.0)	65(11.6)	42(10.1)
ISAR score, n (%)				73.327	< 0.0001
<2	602(61.9)	410(73.3)	192(46.4)
≥ 2	371(38.1)	149(26.7)	222(53.6)
Triage department, n (%)					0.516
Internal Medicine	723(74.3)	408(73.0)	315(76.1)
Surgery	143(14.7)	85(15.2)	58(14.0)
Others	107(11.0)	66(11.8)	41(9.9)
High risk of falling, n (%)				52.771	< 0.0001
No	607(62.4)	403(72.1)	204(49.3)
Yes	366(37.6)	156(27.9)	210(50.7)
Admission by ambulance, n (%)				9.813	0.002
No	832(85.5)	495(88.6)	337(81.4)
Yes	141(14.5)	64(11.4)	77(18.6)
Triage level, n (%)				29.138	< 0.0001
1	10(1.0)	6(1.1)	4(1.0)
2	257(26.4)	116(20.8)	141(34.0)
3	344(35.4)	195(34.9)	149(36.0)
4	362(37.2)	242(43.3)	120(29.0)
ED disposition, n (%)				21.448	< 0.0001
Discharged	749(77.0)	460(82.3)	289(69.8)
Admission to ward	132(13.6)	61(10.9)	71(17.1)
Admission to ICU	92(9.5)	38(6.8)	54(13.0)
Living alone, n (%)				0.912	0.339
No	918(94.3)	524(93.7)	394(95.2)
Yes	55(5.7)	35(6.3)	20(4.8)
Stable informal support, n (%)				2.465	0.116
No	79(8.1)	52(9.3)	27(6.5)
Yes	894(91.9)	507(90.7)	387(93.5)
Pain scale, median (IQR)	0.00(0.00,2.00)	0.00(0.00,2.00)	0.00(0.00,2.00)	-1.378	0.168
HR, median (IQR)	87.00(75.00,100.00)	87.00(75.00,100.00)	87.00(74.00,100.00)	-0.252	0.801
B, median (IQR)	18.00(17.00,20.00)	18.00(17.00,20.00)	18.00(17.00,20.00)	-0.699	0.485
SBP, median (IQR)	143.00(127.00,159.00)	142.00(126.00,155.00)	144.00(128.00,162.00)	-2.575	0.01
DBP, median (IQR)	78.00(68.00,86.00)	78.00(70.00,87.00)	76.00(68.00,84.25)	-3.802	< 0.0001
SPO, median (IQR)	97.00(96.00,98.00)	98.00((96.00,98.00)	97.00(95.00,98.00)	-3.832	< 0.0001
T, median (IQR)	36.20(36.00,36.60)	36.20(36.00,36.60)	36.20(36.00,36.60)	-0.013	0.989
Length of hospitalization, median (IQR)	7.00(5.00,10.00)	7.00(5.00,9.00)	7.00(5.00,11.00)	-2.259	0.024
Combined Adverse outcomes, n (%)				26.779	< 0.0001
No	498(51.2)	326(58.3)	172(41.5)
Yes	475(48.8)	233(41.7)	242(58.5)
Hospitalization (within 30 days), n (%)				33.188	< 0.0001
No	534(54.9)	351(62.8)	183(44.2)
Yes	439(45.1)	208(37.2)	231(55.8)
Emergency Department Revisit (within 72 h), n (%)
No	923(94.9)	528(94.5)	395(95.4)	0.446	0.504
Yes	50(5.1)	31(5.5)	19(4.6)
Death (within 30 days), n (%)				4.830	0.028
No	948(97.4)	550(98.4)	398(96.1)
Yes	25(2.6)	9(1.6)	16(3.9)

Univariable and multivariable logistic regression analysis for adverse outcomes

In the univariable regression analysis, older age, male, ISAR score (≥ 2), high risk of falling, admission by ambulance, lower SPO₂, level I or level II of both triage criteria were significantly associated with the adverse outcomes. In the multivariable regression analysis, age (OR:1.036, 95% CI: 1.015 ~ 1.057), gender (female vs. male) (OR:0.667, 95% CI: 0.504 ~ 0.883), admission by ambulance (OR: 2.332, 95% CI: 1.403 ~ 3.879), and the new triage level (OR: 2.595, 95% CI: 1.535 ~ 4.385) were still significantly associated with the adverse outcomes (Table 2).

Table 2.

Univariable and multivariable regression models on the adverse outcomes^a

Variables	Odds ratio	95%CI	p-value
Univariable regression
Age, year	1.061	1.042 ~ 1.080	< 0.0001
Gender (Female vs. Male)	0.663	0.514 ~ 0.854	0.001
ISAR score (≥ 2vs < 2)	2.705	2.070 ~ 3.535	< 0.0001
Primary triage level (I, II vs. III, IV)	4.802	3.499 ~ 6.590	< 0.0001
New triage levelb (I, II vs. III, IV)	5.008	3.784 ~ 6.630	< 0.0001
High risk of falling (Yes vs. No)	2.321	1.779 ~ 3.027	< 0.0001
Admission by ambulance (Yes vs. No)	5.509	3.543 ~ 8.565	< 0.0001
SPO2, %	0.906	0.870 ~ 0.943	< 0.0001
Multivariable regression
Age, year	1.036	1.015 ~ 1.057	0.001
Gender (Female vs. Male)	0.667	0.504 ~ 0.883	0.005
New triage level (I, II vs. III, IV)	2.595	1.535 ~ 4.385	< 0.0001
Admission by ambulance (Yes vs. No)	2.332	1.403 ~ 3.879	0.001

^a: A composite outcome defined as the occurrence of any of the following events: ED revisit within 72 h, hospitalization within 30 days, or death within 30 days

^b: Patients with a ISAR score < 2 remained in the primary triage level, while patients with a ISAR score ≥ 2 are upgraded one triage level

Comparison of the predictive power of primary vs. ISAR-enhanced triage criteria for adverse outcomes

As shown in Fig. 2; Table 3, the ISAR-enhanced triage criteria modestly improved predictive performance in patients aged ≥ 65 years, with the AUC increasing from 0.697 (95% CI: 0.664–0.730) under the primary criteria to 0.714 (95% CI: 0.682–0.746). In patients aged ≥ 75 years, the AUC remained unchanged at 0.703 (95% CI: 0.652–0.753). The Youden index was also higher with the ISAR-enhanced criteria in both age groups. Compared with the primary criteria, the ISAR-enhanced triage showed lower sensitivity (86.7% vs. 78.3% for ≥ 65 years; 83.7% vs. 69.2% for ≥ 75 years) but higher specificity (42.3% vs. 58.1% for ≥ 65 years; 48.3% vs. 66.1% for ≥ 75 years).

Fig. 2.

Comparison of Area under the receiver operating characteristics curves (AUCs) for predicting adverse outcomes using two triage criteria (Left panel: Patients aged ≥ 65 years; Right panel: Patients aged ≥ 75 years)

Table 3.

Predictive performance of two triage criteria for adverse outcomes

Triage criteria	AUC	SE	95%CI	Youden index	Sensitivity (%)	Specificity (%)
Patients aged ≥ 65 years
Primary Triage Criteria	0.697	0.017	0.664 ~ 0.730	0.291	86.7	42.3
New Triage Criteria	0.714	0.017	0.682 ~ 0.746	0.364	78.3	58.1
Patients aged ≥ 75 years
Primary Triage Criteria	0.703	0.026	0.652 ~ 0.753	0.321	83.7	48.3
New Triage Criteria	0.703	0.026	0.652 ~ 0.753	0.353	69.2	66.1

Discussion

To the best of our knowledge, this is the first study in China to evaluate the integration of the Identification of Seniors at Risk (ISAR) tool into an ED triage system. Our findings suggest that incorporating ISAR modestly improves the predictive performance for adverse outcomes, primarily by enhancing specificity. Triage level distribution differed by age, with patients aged ≥ 75 years more likely to be categorized as high acuity. This likely reflects recognition of more overt risk factors such as ambulance transport and abnormal vital signs. However, the higher prevalence of underlying vulnerabilities, such as frailty and functional decline, may still go underrecognized in standard triage. These findings underscore the added value of geriatric-specific tools like ISAR, which can complement existing triage systems by capturing risk dimensions not fully addressed by traditional measures. Multivariable analysis further confirmed that the revised triage classification was independently associated with adverse outcomes. Although the overall gain in predictive accuracy was limited, these refinements may nonetheless support more nuanced triage decisions and promote earlier identification of high-risk older patients.

Toward efficient geriatric triage: interpreting modest gains in a real-world context

The updated analysis revealed that the ISAR-enhanced triage system improved discriminative ability in patients aged ≥ 65 years, with the AUC increasing from 0.697 to 0.714. For those aged ≥ 75 years, performance remained stable at 0.703. While these gains may appear modest, surpassing the 0.7 threshold marks a meaningful step toward clinically relevant risk stratification. In the fast-paced emergency department (ED) setting, even slight improvements can facilitate more precise identification of at-risk older adults without adding procedural burden [17]. This refinement may contribute to better identification of high-risk older adults, which in turn could help inform interventions aimed at reducing adverse outcomes such as prolonged hospitalizations or ICU admissions, although further research is needed to confirm such effects. A key feature of this enhancement is the sensitivity-specificity trade-off. Sensitivity decreased, while specificity rose markedly (42.3–58.1% for ≥ 65 years; 48.3–66.1% for ≥ 75 years). By reducing over-triage, this shift ensures low-risk patients avoid unnecessary interventions, a critical advantage in resource-limited EDs where efficiency is essential [18].

In the ≥ 65-year group, the ISAR-enhanced triage system showed a more substantial improvement in predictive performance, reflected by a higher AUC and a notable gain in specificity. In comparison, the ≥ 75-year subgroup exhibited stable AUC values, but with an improved specificity and a higher Youden index, suggesting enhanced risk discrimination for this more vulnerable population. The ability to reduce over-triage while still identifying high-risk individuals is particularly important for older adults who are more susceptible to adverse outcomes [19]. These findings indicate that while the revised triage model offers clear benefits for the broader elderly population, its effects may differ across age subgroups and should be considered accordingly in clinical practice.

Importantly, our findings also highlight the limitations of relying on a single screening tool such as ISAR. Although selected for its simplicity and feasibility, ISAR alone lacks the granularity to differentiate frailty severity. Other instruments, such as the Clinical Frailty Scale (CFS), have demonstrated superior performance in predicting outcomes like 30-day mortality and could be explored as alternative or complementary components in ED triage [20]. Future research should aim to develop a more comprehensive, frailty-informed triage framework that integrates tools like ISAR and CFS with functional assessments and clinical judgment, in order to better identify high-risk older adults in the ED.

Pragmatic use of ISAR in Chinese ED settings: value, limitations, and contextual considerations

The integration of ISAR into emergency department (ED) triage was intended as a feasible and scalable method for identifying high-risk older adults, particularly in busy clinical environments where time and resources for comprehensive geriatric assessment are limited [21]. Its brevity, clarity, and focus on functional vulnerability make it especially appealing for rapid screening at the point of entry [22].

Nevertheless, our study found that the discriminative ability of ISAR-enhanced triage was moderate. While this may appear limited in isolation, it must be interpreted in light of the structural realities of Chinese EDs. Unlike systems with well-developed primary care, China’s EDs often serve as the initial contact point for a wide range of health concerns, many of which are non-emergent [23]. This results in a diagnostically heterogeneous patient population and contributes to persistent ED overcrowding, both of which pose challenges for any triage tool to achieve high discriminative performance across the acuity spectrum [3]. In such settings, the value of ISAR lies not in pinpoint accuracy but in pragmatic utility: it provides a rapid, structured means to flag vulnerable older adults early in the triage process—without overburdening staff or delaying care. This function is particularly relevant where triage nurses must make timely decisions amidst high volumes and resource constraints.

Importantly, the moderate predictive performance observed should not be taken as a failure of ISAR, but rather as a reflection of systemic complexity. In this light, ISAR should be seen as a first-layer screening tool that represents a pragmatic compromise between ideal comprehensiveness and operational feasibility. It helps EDs move beyond a purely acuity-based model and towards an approach that at least partially accounts for age-related vulnerability [24]. This does not preclude the use of more robust frailty measures. The challenge lies not in choosing one tool over another, but in designing a triage framework that is responsive to both patient complexity and system constraints.

In sum, ISAR’s incorporation into ED triage represents a feasible and meaningful step toward geriatric-sensitive emergency care in China. Its moderate discriminative power underscores the need for continued innovation, including tiered triage strategies, digital support tools, and staff training in geriatric principles, to achieve more precise and equitable risk stratification in this vulnerable population.

From tool to pathway: toward structured models of geriatric emergency care

While ISAR was initially introduced as a practical screening tool to identify high-risk older adults during emergency department triage, its potential extends beyond risk stratification. When embedded into broader care frameworks, ISAR can serve as the entry point to more structured geriatric care models. One notable example is the Geriatric Emergency Department Innovations (GEDI) model, which incorporates ISAR scores as a threshold to trigger early multidisciplinary interventions [25]. In this model, patients identified as high risk are promptly evaluated by specialized geriatric teams who initiate targeted actions such as medication reconciliation, mobility assessment, and discharge planning. These interventions are designed to occur during the ED visit and aim to prevent avoidable admissions and improve post-discharge outcomes.

From a policy perspective, ISAR’s integration aligns with broader efforts to enhance geriatric care in China’s healthcare system [26]. As the country faces rapid population aging, EDs are increasingly pressured to address the complex needs of older patients. Incorporating ISAR into national triage guidelines could standardize geriatric risk assessment across diverse settings, promoting more equitable care. However, the implementation must account for regional disparities [27]. While urban hospitals with advanced electronic health records (EHRs) might integrate ISAR seamlessly, rural or resource-limited EDs may require simpler adaptations, such as paper-based scoring systems. Tailoring implementation strategies to local contexts will be essential for widespread adoption.

Limitations

This study has several limitations. First, it was conducted in a single urban emergency department in China, which may limit the generalizability of the findings to other settings with different patient populations, healthcare systems, or triage protocols. Second, the exclusion of patients with severe cognitive impairment without a formal proxy may introduce selection bias, although the rarity of such cases in our cohort likely limits its impact. Third, although the ISAR-based triage adjustment demonstrated improved predictive performance, it was applied retrospectively and did not influence real-time clinical decision-making. As such, our findings reflect a research scenario rather than a direct clinical intervention. Finally, the loss to follow-up represents a potential source of attrition bias, particularly given that such missingness may be non-random—for example, involving patients with poorer outcomes or unstable contact information. As recent studies in emergency care research have emphasized, systematic missingness can distort outcome estimates, especially in high-risk populations such as older adults. Future work should explore the mechanisms of follow-up loss and adopt strategies to mitigate its impact on data validity and generalizability [28].

Conclusion

In summary, the integration of ISAR into ED triage systems represents a promising step toward enhancing geriatric care through simple, evidence-based risk stratification. Its success will depend on thoughtful implementation strategies, supportive policies, and a commitment to addressing the unique challenges of diverse healthcare settings. By embedding geriatric risk assessment into routine practice, we can better address the needs of older adults and lay the foundation for more responsive, equitable, and age-friendly emergency care systems.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

AUC

The area under the receiver operating characteristic curve

CTAS

The Canadian Triage and Acuity Scale

ESI

Emergency Severity Index

MTS

Manchester Triage System

NRS

Numeric Rating Scale

EDs

Emergency departments

GEDI

Geriatric Emergency Department Innovations

ISAR

The Identification of Seniors at Risk

ROC

Operating characteristic

Author contributions

L.O. contributed to manuscript writing and statistical analysis. S.Y. was responsible for manuscript drafting. Z.H. played a key role in implementing the new triage method in clinical practice. Y.L. facilitated the progression of the study protocol. D.L. oversaw project coordination and provided critical revisions. All authors reviewed and approved the final manuscript.

Funding

None.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The present study was approved by the Ethics Committee of The Central Hospital of Wuhan (Ethics approval number: WHZXKYL2024-033). All methods were carried out in accordance with relevant guidelines and regulations, including the Declaration of Helsinki. Written informed consent was obtained from all participants or their legal guardians prior to their inclusion in the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.