A cross-sectional study using convenience sampling was undertaken to quantify the prevalence of falls in previous 12 months and to identify the risk factors for falls and circumstances of falls among studied population. In the present study, a fall was defined as “unintentionally coming to the ground or some lower level and other than as a consequence of sustaining a violent blow, loss of consciousness, sudden onset of paralysis as in a stroke or epileptic seizure”¹.

Participants were recruited through the distribution of flyers to all members of University of the Third Age (U3A) Selangor and Kuala Lumpur, Malaysia, based at the Institute of Gerontology, Universiti Putra Malaysia. University of the Third Age (U3A) Malaysia, is program that provide lifelong learning courses to older adult aged 50 years and above. Inclusion criteria for the study were (i) living in the community and (ii) being aged 50 years and above. The cut-off age of 50 years and above was used in this study rather than 65 years of age because of a lower life expectancy in Malaysia compared to in Western countries. The life expectancy for Malaysia was 72.2 and 77.3 years for males and female, respectively (26), and (iii) able to tolerate standing at least 6 min independently or with a single point stick support. Older adults who expressed interest in participation were contacted through an initial phone call to discuss details of what the project involved. Screening was carried out during this phone conversation to ensure all eligibility criteria for the study were met. Written consent was obtained from each participant.

Data were obtained by trained researchers based on a structured questionnaire, and a series of tests of physical performance. The structured questionnaire comprised of five main sections: (i) socio-demography information, (ii) physical and medical conditions, (iii) history of falls and falls risk assessment, (iv) psychological aspect associated with falls, and (v) self-reported level of physical activity. The physical performance measures, reported below, included both clinical and laboratory assessments of balance and mobility. The questionnaires were interspersed between the physical performance tests. The interview and assessments were conducted at the assessment laboratory and lasted ~90 to 120 min. Participants were allowed to have a rest between the tests if required.

Height (SECA Bodymeter), weight and body composition (BF-418, TANITA, Japan) were measured by trained assessor. Level of physical activity was assessed using Human Activity Profile, a questionnaire with 94 activity items listed in order of increasing energy expenditure. Each item was rated as “still doing,” “have stopped doing,” or “never did.” The Adjusted Activity Score (AAS) (highest numbered item listed as “still doing” less the number of items rated as “have stopped doing”) was used in the analysis.

Visual edge contrast sensitivity was assessed using the Melbourne Edge Test. This test presents 20 circular patches containing edges with reducing contrast (27, 28), with the highest numbered circle where the direction of the edge of contrast is correctly identified recorded and were reported in the study.

Descriptive analysis was applied to participant demographic data and all outcome measures. Performance on all outcome measures were compared between participants who reported one or more falls in the preceding 12 months and those reporting no falls, using t-test for continuous/normally distributed variables, and chi squared for categorical variables. Univariate logistic regression was performed to determine variables associated with falls and variables with a p < 0.1 were then included in a multivariate logistic regression (38). All analyses were conducted using SPSS vs. 22.0, and the critical value for all analyses was p < 0.05.

A total of 156 participants were assessed (Table 1). The age of study participants ranged from 50 to 78 years old, with mean (SD) age was 63.2 (6.2), and more than half of the participants (57%) were females. The majority of participants were Malay (83.4%) and most were still married (76.3%). In terms of education level, all participants were literate (able to write and read) with more than half of participants (56.5%) having received secondary education. Most participants (73.1%) were retired and the median income for study participants was MYR1900 per month. There was no significant difference of socio-demographic between faller and non-faller.

In terms of health and medical conditions, 56% of participants reported of having at least one to three health problems and 7.5% had four and more health problems that might affect their balance performance (Table 2). The most common health conditions reported were diabetes mellitus (19.2%), arthritis (16.7%), cardiac problems (14.7%), and back pain (14.7%). Thirty percent of participants reported taking three or more types of medication daily and antidepressant (29.5%) was the most common taken medication. Only two participants in present study, used a single point of stick during outdoor activity. The majority of participants were considered not to have depressive symptoms based on a GDS score of between 0 and 4 (39, 40), with an overall mean score for GDS-15 of 2.22 ± 2.3 SD. Study also found that the mean score of FES-I was 17.6 (6.1 SD) and 16.8 (6.4SD) among non-fallers and fallers respectively.

Study found 49 (31.4%) participants had at least one fall in the previous 12 months, and seven of these participants (14.3%) reporting experiencing two falls within the 12 months prior to the date of interview (Table 3). Out of the 49 participants who fell, 31 (63.3%) were female and 18 (36.7%) were males. The mean age of fallers and non-fallers were 63.3 years (6.5 SD) and 62.5 years (6.8 SD), respectively. There was no difference in the proportion of fallers who were male (36.7%) or female (63%), χ² (1, n = 156) = 1.36, p = 0.243 nor between the three age groups (36.7%) of fallers were aged 50–59, 32.7% aged 60–69 and 30.6% aged 70 years and above, χ² (2, n = 156) = 1.36, p < 0.05.

The self-perceived causes and consequences of fall were obtained from the participant's explanation regarding their falls. Most fallers fell at home (63.3%). The bathroom, kitchen and dining area, and bedroom were the most common areas where falls occurred inside the home, and the backyard area was the most prevalent area reported for falls occurring outside the home. Thirty seven percent of fallers reported falling in community areas, with the most frequent location for these falls being in front of shops, streets (including curb, uncovered drain or uneven walking paths), and park areas. Trips and slips were the most prevalent causes, accounting for 77% of falls, with a further 20.4% being described as due to loss of balance. Almost a third of fallers fell in the forwards direction and 18.4% fell backward. Based on self-report regarding the severity of injuries incurred during falls, nine fallers (18.4%) reported no injury, while 40 fallers (81.6%) suffered injuries from their fall. More than half of fallers (71.4%) sustained minor injuries such as grazes, bruises, sprains, and cuts, while 10.2% reported suffering a severe injury such as fracture or dislocation as a consequence of their fall. Among those five participants who experienced severe injuries, one participant had a fractured patella, one experienced an ankle fracture, one had a wrist fracture and two had ankle dislocations. Meanwhile, those fallers who had minor injuries, 22 (62.8%) did not require medical treatment and 37.2% did seek medical attention. Consequently, most of the fallers perceived these injuries has restricted their mobility for at least 3 days up to a month.

Independent group t-test indicates a number of significant differences were identified between fallers and non-fallers with fallers having higher BMI, lower MET score, and higher depression score (Table 4), as well as poor balance (Step Test), and more sway and took longer time to perform the step quick turn test, and slower walking speed (Table 5). In addition, presence of back pain was significantly associated with falls status, χ²(1) = 3.56 (p ≤ 0.05). Participants with those reporting back pain are three time odd to fall than those without back pain (O.R: 2.87).

A multiple logistic regression was performed to ascertain the effect of BMI, step test, turning performance, degree of sway and turn taken during step quick turn test, level of depression, having back pain problem and edge contrast sensitivity on the odd that participants have fall as listed in Table 6. The logistic regression model was statistically significant, χ²(3,7) = 7.885, p = 0.048. The model explained 20.7% of the variance in having a fall, and correctly classified 71.8% of fallers. Those who were identified as having imbalance during turning were 4.8 times more likely to have a fall compared to those who did not have turning impairment.