Study design

This is a planned secondary analysis of the Canadian Emergency and Trauma Initiative (CETI),^{Reference Sirois, Griffith and Perry5} a multicentre prospective cohort study (N = 3,350) conducted between 2011 and 2016 in six university-affiliated EDs (Quebec City, Montreal, Ottawa, Toronto, Hamilton, and Calgary). Full protocol details were previously reported.^{3,Reference Brousseau, Dent and Hubbard13}

Study setting and population

Community-dwelling older patients were included if they were ages ≥ 65, independent in all activities of daily living (ADLs) (see measures below), and presented to a participating ED with chief complaints of minor injury (i.e., not requiring admission/surgery) sustained in falls. Hospitalized patients and those unable to give consent or to speak French or English were excluded.

Eligible patients were identified by emergency physicians or research staff 24 hours/day, 7 days/week. In-person (for patients presenting to the ED during research staff office hours) or telephone interviews were conducted by trained research assistants within 72 hours of the ED visit. An in-person assessment of the participants’ physical condition (functional, mobility, and frailty status) was conducted during the in-person interview. In order to minimize loss to follow-up, the study's steering committee determined that a hybrid combination of in-person and phone interviews at 3 and 6 months would be conducted.

Study measures

Fear of falling was assessed using the Short Falls Efficacy Scale-International (SFES-I).^{Reference Kempen, Yardley and van Haastregt17} This validated tool assesses how concerned older adults are about falling when performing seven daily activities (getting dressed/undressed, taking a bath/shower, getting in/out of a chair, going up/down stairs, reaching for something above the head or on the ground, walking up/down a slope, going out to social events). Each component is rated from 1 (not concerned) to 4 (severe concern). Total scores range from 7 (not concerned) to 28 (most severe concern) and are interpreted by stratifying the scores into three categories: mild (7–8), moderate (9–13), and severe fear of falling (14–28).^{Reference Kempen, Yardley and van Haastregt17} Participants were questioned regarding any subsequent falls that might have occurred following their initial ED visit at the 3- and 6-month follow-ups. Participants were asked if they “had fallen hard enough to feel pain afterwards” (yes/no)¹⁸ and the number of falls they had sustained if they answered yes. Returns to the ED were also self-reported during both follow-up interviews, in which participants were asked if they had consulted to the ED (yes/no) in the past 3 months, and the number of times they did. Medical files of patients returning to the ED at Hôpital de l'Enfant-Jésus were analysed to assess the causes of return to the ED (medical or traumatic).

We collected demographic data, including age and sex. Self-reported comorbidities were recorded (yes/no) using a list of 18 physical^{Reference Daveluy, Pica, Audet, Courtemanche and Lapointe19} (e.g., cardiac, neurological diseases) or psychological conditions (i.e., anxiety, depression, irritability) that may have an impact on the participants’ function.^{Reference Daveluy, Pica, Audet, Courtemanche and Lapointe19} The Identification of Seniors at Risk (ISAR) ED screening tool was used at the index ED-visit. ISAR ≥ 2/6 reflects needs for further geriatric evaluation.^{Reference McCusker, Bellavance and Cardin20} The 2005–Abbreviated Injury Scale²¹ codes were used to compute the Injury Severity Score (ISS), which ranges from 1 (minor) to 75 (unsurvivable).²¹ ISS values ≤ 8; 9–15; and ≥ 16 reflect minor, moderate, and severe trauma, respectively.^{21,Reference Copes, Champion and Sacco22}

Patients’ functional status was measured using the validated Older Americans Resources Services (OARS) scale, which aims to assess a patient's ability to perform seven basic ADLs (eating, grooming, dressing, transferring, walking, bathing, continence) and seven instrumental ADLs (meal preparation, homemaking, shopping, using transportation, using the phone, managing medication, and money).^{Reference Bissett, Cusick and Lannin23,Reference McCusker, Bellavance, Cardin and Belzile24} The total OARS score ranges from 0 (dependent) to 28 (independent). Basic mobility was assessed by the Timed “Up-and-Go” (TUG), during which patients are asked to stand up, walk 3 meters away from their chairs, turn around, walk back to their chairs and sit down. The amount of time that each participant takes to perform these tasks was recorded by a research assistant. Participants’ mobility status is then considered as free (TUG < 10 seconds), mostly independent (< 20 seconds), variable (20–29 seconds), and impaired (> 30 seconds). The TUG was shown to predict functional decline in older patients with minor injuries.^{Reference Eagles, Perry and Sirois25} The Study of Osteoporotic Fractures (SOF)^{Reference Kiely, Cupples and Lipsitz26} index was used to determine frailty status of the patients. Higher SOF scores are associated with functional decline in older ED patients presenting with minor injuries.^{Reference Sirois, Griffith and Perry5} Cognitive impairment was evaluated in-person with the Montreal Cognitive Assessment (MoCA)^{Reference Luis, Keegan and Mullan27} or by the Telephone Interview for Cognitive Status-modified (TICS-m)^{Reference Knopman, Roberts and Geda28} with respective cut-offs of < 23/30 and ≤ 32/50 for cognitive impairments.^{Reference Kiely, Cupples and Lipsitz26}

Data analysis

Descriptive statistics using the Student's t-test and the chi-square test (means, proportions) were used to describe patients’ characteristics according to fear of falling levels. Logistic regressions were used to examine whether fear of falling alone predicted the outcomes (yes/no) at 3 and 6 months. The risk of return to the ED and falls according to fear of falling levels were estimated using odds ratios (OR, with 95% confidence intervals [95% CI]), using mild fear of falling as the reference. Predictive statistics with 95% CI were computed (sensitivity, specificity, predictive positive value [PPV_ and negative predictive value (NPV]). Sensitivity analyses were performed to determine the best (no return to the ED/fall) and worst (positive return to the ED/fall) case scenarios for the missing data. Statistical significance was defined as p-values ≤ 0.05. Data were analysed using SAS software (SAS Institute, Inc., Cary, NC), version 9.4.