In older adults (≥60 y) with serious illness who present to the ED, is ED-initiated hospice and/or palliative care consultation associated with the provision of advance care planning, when compared to those who do not receive consultation? Secondary outcome of interest include:

We consulted an academic-affiliated librarian (PAB) for the systematic literature search. We will identify studies that have examined the role of HPC consults in the ED by searching the electronic databases MEDLINE (Ovid), Embase (Elsevier), Web of Science Core Collection (Clarivate), and the Cochrane Central Register of Controlled Trials (Wiley). The searches include terms for end-of-life conversations or consultations and emergency services (Multimedia Appendix 1). Controlled vocabulary terms (MeSH, Emtree) are included when available; no date or language limit is applied. We will conduct citation tracking on all eligible studies to highlight any articles potentially missed by our search strategy.

We plan to include randomized and quasi-randomized clinical trials that enrolled older ED patients (>60) with serious illness presenting for care [19]. We used the age of 60, mindful that (i) it is more inclusive that the commonly used cut-off of 65 years, and (ii) this is the definition used by the World Health Organization. We describe serious illness using the Lancet Commission framework, which defines serious illness broadly as conditions associated with serious health-related suffering that require palliative care. For ED application, serious illness includes advanced or progressive conditions with significant symptom burden, functional decline, or elevated risk of mortality that commonly prompt ED visits, such as advanced cancer, heart failure, chronic obstructive pulmonary disease, dementia, kidney disease, frailty, and severe multimorbidity [19-21]. Further mindful that this is still a relatively new body of literature, we also plan to include observational study designs, specifically purposed to evaluate HPC consultation in the ED. We plan to exclude study designs that lack a comparison group (eg, case series, case report ) Finally, we plan to exclude conference proceedings and abstracts because of limited methodological description, hindering the risk of bias (RoB) assessment. Studies will not be excluded based on language, sample size, or time of publication.

We will export titles and abstracts into Covidence software (elbourne, Australia), where duplicates will be removed. We will conduct title, abstract, and full-text screening independently and in duplicate by two trained reviewers at minimum. A standardized decision-tree was created to guide title and abstract screening (Figure 1). Any disagreement between reviewers regarding study inclusion following title, abstract, or full-text review will be resolved through a third study reviewer and/or discussion. Inter-rater agreement on title, abstract, and full-text screening will be reported using Cohen κ statistics, respectively.

Two reviewers will extract data independently and in duplicate. A standardized and pilot-tested data collection form will be created to ensure consistency of extraction. The following data from eligible studies will be extracted: author(s), year of publication, study design, single versus multisite, country of study, inclusion and exclusion criteria, recruitment time frame (mo), follow up length, total sample size, the proportion of adults ≥60 years of age, definition and timing of outcomes, number of events, baseline demographics (eg, age, sex, frailty status/score, triage acuity), unit of change for continuous outcomes, classification for categorical outcomes, and the unadjusted and adjusted point estimates of risk with confidence intervals. All extracted data values will be rounded to two decimal places, and we will import this data into Microsoft Excel.

Data will be synthesized using R software (version 4.0; R Foundation for Statistical Computing) and the ‘meta’ package [22]. We plan on generating point estimates and their respective 95% CIs using absolute risk difference as the primary effect size, and relative measures as secondary effect sizes. Results will be pooled according to the outcome of interest, assuming the definition and timing of measurement are congruent. We plan to synthesize and report both univariable and multivariable estimates. If baseline risk estimates are not reported and conversion or data retrieval from the study materials is not possible, we will attempt to contact the study authors. Our contact process involves emailing the corresponding, first, and last authors using email addresses provided in the publication or obtained through institutional websites. If no response is received within two weeks, a follow-up email will be sent. Should contact the authors fail after one month, we will proceed with subgroup analyses based on the format of the effect estimate. We will compare studies that provide information on baseline risk, where we can appropriately convert odds ratio, risk ratio, and hazard ratio, to studies where conversion is not possible or appropriate. We will conduct a sensitivity analysis to determine the influence that data imputation has on pooled estimates. A random-effects model will be used for all statistical pooling, mindful that models of care and patient populations are likely to vary between regions [23]. Adjusted effect sizes will be converted to absolute risks using methods proposed by Foroutan and colleagues [24]. Where meta-analysis is not possible, we will report outcomes narratively.

We will examine the risk of bias within each study, individually and in duplicate. We will leverage the Cochrane Risk of Bias 2.0 assessment for clinical trials. Specially, the RoB will be determined by examining six domains: the randomization process, deviations from intended interventions, missing outcome data, outcome measurement, and selection of the reported result. For observational methods, we will defer to the ROBINS-I bias assessment instrument and evaluate the following domains: confounding, participant selection, intervention classification, deviations from intended intervention, missing data, outcome measurement, and reported outcomes. We will use the individual domains, rated as low or high risk of bias, to inform each study’s overall risk of bias. A flag of high risk of bias across any single domain, regardless of study design, will be classified as high risk of bias.

Statistical heterogeneity will be assessed through the visual inspection of forest plots, examining the consistency among point estimates and overlap among the associated confidence intervals, and the χ² test for homogeneity. The inconsistency index (I²) measure will only be used if most studies include <500 patients, mindful that prior work has demonstrated a lack of variance in the I² measure among studies with large sample sizes [25]. Clinical and methodological heterogeneity will be examined to identify factors that may modify the association between HPC consultation and the outcomes of interest. Effect modifiers of interest are RoB (high vs low), study design (interventional vs observational), and geriatric population (mixed vs full).

Certainty of our estimates will be determined using the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) approach. Per GRADE recommendations, confidence will be rated as either high, moderate, low, or very low. For meta-analyses with more than ten studies, funnel plots will be produced to examine the distribution of positive and negative findings allowing for the detection of publication bias. An individual assessment of confidence will be given for each outcome. The congruency of pooled estimates between our sensitivity analysis (including imputed non-significant studies evaluating the predictor of interest) and our primary model (not including the imputed non-significant predictor) will be reported and taken into consideration when discussing our confidence in the estimates.

We are setting out to conduct a systematic review and meta-analysis, evaluating the association between ED-initiated referral for HPC and the provision of advance care planning downstream. We are also interested in evaluating the mortality, hospitalization, in-patient length of stay for those admitted, repeat health service use, costs, and satisfaction with the model of care. We hypothesized that ED-initiated referral would have a positive effect on the outcomes listed (eg, increase in provision of advance care planning, decrease in mortality, etc) We intend to disseminate our findings through presentations at academic conferences, and through the creation of a peer-reviewed publication.

Despite the expanding corpus of emergency medicine and HPC literature, a focused systematic review on the geriatric population remains conspicuously absent [10,11]. Over the past four years, numerous systematic reviews have explored various facets of EM HPC, with eight well executed systematic reviews dedicated to this domain [26-32]. However, none have specifically targeted the older adult demographic within the ED setting. This gap underscores the critical need for our systematic review protocol, which is uniquely positioned as the first to address older adult HPC consultations in the ED. The importance of this systematic review is notable, as it seeks to impact clinical outcomes, guide policy decisions, and shape future research. This ensures that emerging evidence is effectively used to improve patient care in emergency settings. By focusing on this underrepresented group, the Academy of Geriatric Emergency Medicine and the Geriatric ED Guidelines 2.0 Working Group anticipate that our findings will be instrumental in guiding the forthcoming revisions of their ED HPC guidelines, using GRADE methodology for the development of future geriatric ED clinical practice guidelines [33-35].

The strengths of our protocol are manifold. It includes a well-defined Population, Intervention, Comparison, Outcome (PICO) question, the collaboration of an interdisciplinary team comprising national and international leaders in emergency medicine and HPC, and the potential to establish new standards in emergency care for older adults. Additional strengths include a comprehensive review strategy and the use of advanced analytical techniques to ensure robust synthesis of the available data. Nonetheless, the geriatric focus of our review might limit the selection of studies, potentially affecting the breadth of evidence available. Moreover, the existing literature might not yet be sufficiently developed to answer our PICO question with the desired level of reliability. Despite these limitations, our systematic review protocol still offers substantial value, supporting needs assessments by the Geriatric Evaluation & Advanced Research (GEAR) teams and contributing to the foundational knowledge required to enhance geriatric care in emergency settings [30,34-37].

Our systematic review is novel in its evaluation of ED-initiated HPC consultation from the ED among a panel of patient-important outcomes among older ED patients. We hypothesize that ED-initiated HPC consultation would improve rates of advance care planning, decrease mortality, decrease hospital admission, decrease hospital length of stay, decrease costs, and improve satisfaction for patients and clinicians. The final systematic review will be utilized to develop the Geriatric ED Guidelines 2.0 recommendations related to HPC care in the ED for older adults.