INTRODUCTION
Background
Currently, patients 65 years and older account for 12%-24% of all emergency department (ED) visits.Reference Samaras, Chevalley and Samaras 1 - 3 Although most older patients are discharged home following an ED visit, the return rate is double that of their younger counterparts, ranging from 17%-29%.Reference Lowenstein, Crescenzi and Kern 4 - Reference Aminzadeh and Dalziel 9 Multiple factors have been found to be associated with return ED visits in older patients, including polypharmacy, specific medical conditions such as heart failure or kidney disease, poor social supports, and poor physical functioning.Reference Friedmann, Jin and Karrison 5 , Reference Caplan, Brown and Croker 6 , Reference McCusker, Cardin and Bellavance 8
Rationale
The Society for Academic Emergency Medicine Geriatric Task Force advocates for a functional assessment, such as the Get Up and Go test, as a mandatory component of the evaluation of an older ED patient. They cited as one of their top research priorities the development of minimal functional status measures that evaluate the ability to transfer and ambulate independently.Reference Carpenter, Heard and Wilber 10 Multiple guidelines, including the recently published Geriatric Emergency Medicine Guidelines, recommend a mobility assessment prior to discharge in older patients presenting to the ED following a fall. 11 - 13 Unfortunately, most physicians do not conduct a mobility assessment within the ED.Reference Baraff, Lee and Kader 14 , Reference Tirrell, Sri-on and Lipsitz 15 Minnee et al. found that older patients were less likely to return to the ED within 7 days if there was documentation of mobility status prior to presentation and on discharge.Reference Minnee and Wilkinson 16 Thus, from a health care expenditures and a quality of care perspective, it is imperative to gain an increased understanding of what mobility assessments are performed in the ED and to evaluate their association with adverse outcomes.
Objectives
This systematic review was undertaken to identify mobility assessments that are used in ED patients of 65 years and older and determine whether mobility test measures are associated with reported outcomes of hospitalization, repeat visits to the ED, future falls, or frailty.
METHODS
Protocol and registration
This systematic review conforms to the PRISMA guidelines for the reporting of systematic reviews. A protocol was not registered.
Eligibility criteria
Randomized controlled trials and observational studies were considered for inclusion. Editorials, commentaries, letters, case reports, systematic reviews, and meta-analyses were excluded. For abstract citations that met criteria for the full-text review, primary authors were contacted, where contact information was available, for further study information. No language restrictions were placed on the electronic search strategies. Non-English papers were translated prior to the full-text review.
We sought to include studies on patients of 65 years or older who underwent a physical mobility assessment in the ED that reported association with hospitalization, repeat visits to the ED, future falls, or frailty. Examples of mobility assessments included but were not limited to a Timed Up and Go (TUG) test,Reference Podsiadlo and Richardson 17 six-metre walk test (6MWT),Reference Tiedemann, Shimada and Sherrington 18 sit-to-stand test with five repetitions (STS-5),Reference Whitney, Wrisley and Marchetti 19 sit-to-stand (STS) test,Reference Bohannon 20 , Reference Csuka and McCarty 21 half-turn test,Reference Berg, Wood-Dauphine and Williams 22 alternate step test (AST),Reference Berg, Wood-Dauphine and Williams 22 and the lie-to-sit-to-stand test.Reference Reicherz, Brach and Cerny 23 There was no restriction placed on who could perform the mobility assessment. The assessment could be performed by emergency physicians, registered nurses, physiotherapists, occupational therapists, patient care technicians, or research assistants. A broad criterion was used for types of mobility tests and the assessors so as not to miss any relevant studies. Physical assessments of balance or strength were excluded. Our outcomes included a 90-day return to the ED (with or without subsequent hospitalization), recurrent falls, increased frailty, or functional decline.
Information sources
The electronic search strategies were developed with the guidance of a professional health sciences research librarian. A comprehensive search strategy was conducted to identify all relevant articles from the following electronic databases: Ovid MEDLINE (1946 to September 7, 2016), Embase (September 7, 2016), CINAHL (September 11, 2016), Cochrane Library (September 7, 2016), PEDro (September 7, 2016 ), and OTseeker (September 7, 2016). The full literature search is available in the Supplementary Appendix S1. References of all articles included for full-text reviews and systematic reviews were reviewed for any publications that may have been missed by our electronic search strategy.
Study screening and selection
EndNote software (Thomson Reuters, NY) was used to store and deduplicate included studies for the review. Two independent reviewers (KY, DE) independently screened all of the citations from the electronic literature search. The first phase of screening involved a review of titles and abstracts only. The second phase involved a review of full-text articles that met the inclusion criteria. Three articles (two French, one Spanish language) were translated prior to the full-text review. At the conclusion of each phase, the independent reviewers resolved discrepancies and involved a third party (JJP) when required in order to achieve consensus. Kappa was calculated to determine statistical agreement between the two reviewers.
Data collection
Both reviewers (KY, DE) abstracted data from all included publications. A data extraction form (see Supplementary Appendix S2) was used to aid screening of abstracts and for the full-text review of potentially relevant articles. Key datapoints included type of publication; sample size; age criteria; mobility assessments with applicable cut-off points; and 90 day outcomes, including admission, return to the ED, future falls, and frailty with effect estimates and confidence intervals (CIs).
Risk of bias
Randomized control trials were to be assessed using the Cochrane Risk of Bias tool.Reference Higgins, Altman and Gotzsche 24 The Newcastle-Ottawa Scale (NOS) (quality assessment) for cohort and cross-sectional studies was used to assess study quality and risk of bias in the non-randomized studies. 25 This scale was specifically designed for assessing the quality of non-randomized studies for meta-analyses. Each component (selection, comparability, and outcome) has several questions for which stars can be allocated. The maximum star rating for each component is 4, 2, 3 and 5, 2, 3 for cohort and cross-sectional studies, respectively.
Analysis of results
We planned to conduct a meta-analysis using a random effects model and assess for statistical heterogeneity; however, a qualitative systematic review without a meta-analysis was performed due to study heterogeneity.
RESULTS
Study selection
The results of the screening process are summarized in a PRISMA flow diagram (Figure 1).Reference Moher, Liberati and Tetzlaff 26 The search identified a total of 1,365 unique citations after removal of duplicates. Initial screening of titles and abstracts resulted in 1,333 citations being excluded. The reviewers had a substantial level of agreement with a kappa of 0.66.Reference Landis and Koch 27 , Reference Gisev, Bell and Chen 28 Thirty-two citations were included for the full-text review.Reference Anaf and Sheppard 29 - Reference Roedersheimer, Pereira and Jones 60 Twenty-three articles were excluded after the full-text review, leaving nine articles for inclusion in the systematic review.Reference Carpenter, DesPain and Keeling 37 - Reference Davies and Kenny 39 , Reference Lee, Ross and Tracy 44 , Reference Deschodt, Devriendt and Sabbe 55 - Reference Suffoletto, Miller and Shah 58 , Reference Walker, Bailey and Bradshaw 61
Figure 1 Flow diagram.
A list of excluded studies from the second phase of screening is provided in Supplementary Appendix S3 (see Appendix Table S1). Twelve studies were excluded because they did not meet the full-inclusion criteria (two studies had a non-applicable outcome measure,Reference Stiffler, Finley and Midha 54 , Reference Roedersheimer, Pereira and Jones 60 two did not meet age criteria,Reference Anaf and Sheppard 29 , Reference Murray, Hill and Phillips 47 four did not have a formal physical mobility assessment,Reference Ballabio, Bergamaschini and Mauri 30 , Reference Basic and Conforti 31 , Reference Lightbody, Watkins and Leathley 45 , Reference Marques Vilallonga, San Jose Laporte and Klamburg Pujol 46 and six were conducted in a non-ED setting).Reference Basic and Khoo 32 , Reference Boye, Oudshoorn and van der Velde 35 , Reference Hockings, Schmidt and Cheung 43 , Reference Roussel-Laudrin, Paillaud and Alonso 49 , Reference Bodilsen, Klausen and Petersen 59 Seven studies were excluded because of insufficient data.Reference Dresden, Aldeen and Courtney 40 - Reference Hendin, Clement and Stiell 42 , Reference Roedersheimer, Pereira and Holland 48 , Reference Schoenenberger, Bieri and Ozguler 50 , Reference Wong and Foo 52 , Reference Young, Pyburn and Fox 53 Five of these were abstracts. We attempted but were unsuccessful in getting more information from the authors; therefore, the studies were excluded. One study was a duplicate, and one other was a commentary.Reference Birks 33 , Reference Boye, Mattace-Raso and Van Lieshout 34
Study quality assessment
The quality of all nine studies was assessed using the NOS (Table 1). Overall, the studies are of moderate to poor quality with a high risk of bias. Few studies provided sufficient data about the study participants or controlled for confounding.
Table 1 Quality assessment of studies using the Newcastle-Ottawa Scale
Asterisks are the grading system used for the NewCastle Ottawa Scale.
Result of individual studies
The study design, population, mobility assessment intervention, and outcomes of the nine studies included in this review are summarized in Table 2. Included studies consisted of six cohortReference Carpenter, DesPain and Keeling 37 , Reference Walker, Bailey and Bradshaw 51 and three cross-sectional studies.Reference Crehan, O’Shea and Ryan 38 , Reference Davies and Kenny 39 , Reference Lee, Ross and Tracy 44 Study populations were from six different nations, ranged in size from 26-778 enrolled patients (total n=2,513), and ranged in mean age from 75-85 years.
Table 2 Summary of main study characteristics
CI=confidence interval (95%); OR=odds ratio; PCC=Pearson correlation coefficient.
No meta-analysis was undertaken due to the heterogeneity of the studies. There were five different mobility tests. Hospitalization was the most common outcome, described in four studies. Most studies had insufficient information to characterize their patient population. Three studies enrolled patients who presented to the ED for management of a fall, whereas the remainder enrolled patients presenting to the ED for any issue. It is likely that the severity of illness or functional disability was also dissimilar between studies given the difference in rates of admissions between the cohorts, ranging from no admissions to 77% admitted patients.
The most common mobility assessment was the TUG test, which was evaluated in five studies.Reference Crehan, O’Shea and Ryan 38 , Reference Lee, Ross and Tracy 44 , Reference Walker, Bailey and Bradshaw 51 , Reference Dresden, Alkhawam and Sarwark 56 , Reference Huded, Dresden and Gravenor 57 The TUG test consists of the number of seconds that it takes to stand up from a chair, walk 3 meters with the usual gait aid and speed, turn and return to the chair, and sit down. The inability to perform the TUG was associated with increased hospitalization in one study;Reference Lee, Ross and Tracy 44 however, in another that controlled for age, race, and the Emergency Severity Index, no association was found.Reference Dresden, Alkhawam and Sarwark 56 No association was found between TUG time and ED return visit or admission at 90 days.Reference Walker, Bailey and Bradshaw 51
Other mobility assessments included the Get Up and Go test,Reference Deschodt, Devriendt and Sabbe 55 , Reference Suffoletto, Miller and Shah 58 tandem gait,Reference Carpenter, DesPain and Keeling 37 subjective “gait abnormality,”Reference Davies and Kenny 39 and the Tinetti fall risk screen.Reference Lee, Ross and Tracy 44 Deschodt et al. reported that patients assessed as dependent in the Get Up and Go test were more likely to be admitted to the hospital; however, it was not associated with a return to the ED.Reference Deschodt, Devriendt and Sabbe 55 Suffoletto found no association between the Get Up and Go test and a composite score of a return ED visit, post-discharge hospitalization, or death at 30 days.Reference Suffoletto, Miller and Shah 58 Tandem gait was not associated with future falls.Reference Carpenter, DesPain and Keeling 37
DISCUSSION
Summary
In this systematic review, we conducted a comprehensive electronic search of the medical literature using six databases to identify any studies of older adults (age ≥65 years) who presented to the ED and received a mobility assessment. Our search identified nine relevant studies, totaling 2,513 patients. Study quality of the six cohort and three cross-sectional studies was assessed to be moderate to poor. There was substantial clinical heterogeneity between the studies, such that a meta-analysis could not be done. This review found that the TUG was the most widely used mobility test performed in older ED patients (see S4 for a summary of various mobility assessments). The patient’s ability to perform the test varied widely, and no association was noted in a return ED visit or hospitalization.
There are multiple reasons that an evaluation of mobility continues to be recommended by geriatric guidelines, textbooks, and experts alike – reasons that include, at the very minimum, a need to demonstrate an ability to transfer and ambulate to be safely discharged from the ED, because self-report is not a reliable measure.Reference Roedersheimer, Pereira and Jones 60 These mobility tests do not require expensive, sophisticated equipment to perform; often only a chair is required and contributes little to the overall length of stay, taking less than 1 minute to perform, in general. Performing the mobility test does not put the patient at any greater risk of pain or harm than if the patient were to be discharged and the test performed in an unsupervised environment. Thus, the overall risk benefit balance lies in the performance of these tests in as much as an aid to making appropriate discharge decisions.
Three recent systematic reviews have evaluated use of the TUG as a tool for assessing fall risk in older adults.Reference Barry, Galvin and Keogh 62 - Reference Schoene, Wu and Mikolaizak 64 Their conclusions were similar: the TUG test has limited ability to predict future falls in older persons; the reviews contained studies that were completed in the community or in admitted patients; and it is unclear what difference on outcome, if any, this would have in an ED population. Schoene et al. found that the difference between fallers and non-fallers was 0.63 (95% CI 0.14-1.12) seconds in high-functioning subjects and 3.59 (95% CI 2.18-4.99) in low-functioning subjects.Reference Schoene, Wu and Mikolaizak 64 The meta-analysis completed by Barry et al. found that the sensitivity and specificity of the TUG with a cut-off of >13.5 seconds was 0.32 (95% CI 0.14-0.57) and 0.73 (95% CI 0.51-0.88), respectively, for all studies. The TUG was poor at predicting future falls with odds ratio (OR) 1.01 (95% CI 1.00-1.02).Reference Barry, Galvin and Keogh 62 The ED study by Walker et al. is consistent with the systematic reviews, finding no association with outcomes.Reference Walker, Bailey and Bradshaw 51 Carpenter et al. completed a systematic review of the ED literature to determine predictors of geriatric falls following an ED visit.Reference Carpenter, Avidan and Wildes 65 They reported that simple bedside functional tests, including chair stand, chair sit, raise feet while walking, turn 180 degrees, and near tandem stand were not associated with a 6-month fall risk.
Crehan et al. assessed frailty and the TUG in their evaluation of older patients presenting to the ED after falling.Reference Crehan, O’Shea and Ryan 38 Using Fried’s classification for frailty, 75% of their cohorts were classified as frail, 15% as pre-frail, and 10% as non-frail. However, 38% of patients were unable to complete the TUG, and the remainder had an abnormally slow TUG, with a mean score of 46 seconds. A 2015 meta-analysis by Clegg et al. examined the diagnostic test accuracy of various tools to predict frailty in older community dwelling persons.Reference Clegg, Rogers and Young 66 Two mobility assessments, gait speed and the TUG test, were found to be highly sensitive for predicting frailty in persons older than 65 years. A TUG time >10 seconds had a sensitivity of 0.93, a specificity of 0.62, positive predictive value of 0.17, and negative predictive value of 0.99 for frailty.Reference Clegg, Rogers and Young 66 Hastings et al. found that frail, older ED patients were at increased risk of adverse outcomes (repeat ED visit, hospital admission, nursing home admission, or death) within 30 days of discharge from the ED (HR 1.44, 95% CI 1.06-1.96).Reference Hastings, Purser and Johnson 67 The concept of frailty is widely accepted. However, there is no agreement on a standardized definition or measurement tool that contributes to a clinician’s difficulty in interpreting or managing the frail patient.
The lack of high quality studies, the heterogeneity in measures and outcomes make drawing strong conclusions or recommendations based on the result of this systematic review difficult. Foremost, it highlights a large gap in knowledge in the care of this vulnerable patient population. The challenge of disposition decisions is complex in older patients and goes beyond addressing an acute medical condition. When considering discharge back to the community, mobility assessments are performed in the elderly to determine functional ability to be safely discharged as well as attempt to identify those at high risk of returning to the ED. The lack of feasible ED tools that can predict adverse outcomes hinders the ED physician in making appropriate discharge decisions and puts the older ED patient at risk of preventable morbidity.
Implications for future research
Given the current lack of evidence, there is ample opportunity for future research. Robust research is needed with strong methodological foundations.Reference Carpenter, Heard and Wilber 10 For any meaningful clinical impact to be achieved, the following key issues must be addressed. Firstly, the target population must be specified. The ideal test would be designed to assess all older ED patients not just those who present post-fall. Secondly, the feasibility of the test for use in the ED is critical. It must be quick, not require any extra tools, and be easy for any bedside clinician (nurse, physician) to use to enhance widespread uptake in all ED settings. Thirdly, explicit outcomes must be established. The research question must address whether they are answering a diagnostic (immediate disposition decisions, such as admission to hospital v. discharge to the community) or prognostic (to establish risk of long-term outcomes, such as return to the ED, future fall, or frailty) concern. An evaluation of both harms and benefits should be included. Finally, test characteristics, including cut-off values, inter-rater reliability, and positive and negative predictive values, must be determined. Ultimately, a randomized controlled trial to investigate the effects of a mobility test would be needed. Such research would be highly informative and would provide important evidence to better inform clinical care and future geriatric ED guideline recommendations.
Strengths and limitations
This systematic review used robust methodology to complete the electronic search for relevant studies. Search strategies were developed with the input of an experienced health sciences research librarian. In total, six electronic databases were searched, and no language restrictions were used. Therefore, we feel that it is unlikely that any studies published in peer-reviewed journals would have been missed. A limited grey literature search was conducted, and we excluded some potential studies due to the inability to contact those study authors.
CONCLUSIONS
Our comprehensive systematic review identified a significant gap in knowledge surrounding physical mobility assessments for older ED patients. The TUG was the most commonly used measure. No association was found between TUG scores and a return to the ED. Clinicians should continue to evaluate mobility of older ED patients prior to discharge, although use of the TUG cannot be strongly advocated at this time. Robust research is needed to develop and evaluate standardized mobility measures to assist physicians in making appropriate disposition decisions, ultimately in an effort to improve the care and outcomes in this older, vulnerable patient population.
Acknowledgements
Lindsey Sikora, a health sciences research librarian, helped with peer review and development of the electronic search strategies. Miguel Alejandro Cortel, MD, Elizabeth Gottman, MD, and Shawn Mondoux, MD, translated the Spanish- and French-language articles. Authors KY, DE, and ME were responsible for developing the clinical research question. KY and DE developed the electronic search and served as independent reviewers for both stages of screening. JJP acted as an adjudicator in the event that any disagreement could not be resolved by consensus. KY contributed to the authorship of the initial draft of the manuscript. All authors contributed to critical revision and approval of the final version of the manuscript.
Competing interests: None declared.
Supplementary Material
To view supplementary material for this article, please visit https://doi.org/10.1017/cem.2017.46
