Study Design
This descriptive and cross-sectional study was conducted in a number of districts in the state of Selangor, Malaysia, where the number of older adults are most dense, namely Tanjung Sepat, Cheras, Kajang, Dato Keramat, Petaling Jaya, Klang, Sepang, Rasa, Kuala Selangor and Sekinchan between September and November 2018. The study comprised of community-dwelling older adults aged 60 and above.
Recruitment of participants
Participants from only one state (Selangor) representing the central region of Peninsular Malaysia, that were previously involved in longitudinal study on neuroprotective model for healthy longevity (LRGS-TUA) (20) participated in the present study. Participants in LRGS-TUA longitudinal study were recruited using multistage random sampling and the details has been reported previously (20,21).
Inclusion and exclusion criteria
Older adults who attended the session were screened to meet the inclusion criteria and screened for cognitive impairment using Mini Mental State Examination (MMSE) and depression using Geriatric Depression Scale (GDS-15). Inclusion criteria included older adults who had no permanent disability or impairments, were able to give consent, able to ambulate independently with or without assistive devices for at least 6 meters, with scores of MMSE ≥25 and GDS-15 ≤5. The exclusion criteria were those who scored MMSE scores ≤24, GDS-15 ≥6, having recent lower limb fractures and acute illnesses.
Data collection
Verbal and written information regarding the procedure of study were provided to all participants. Subsequently, participants signed informed consent forms prior to data collection.
Demographic data
A face-to-face interview was conducted to obtain the sociodemographic data and medical history of participants and included questions focusing on age, race, gender, employment status, education levels, living status, number of falls within the past twelve months, self-reported medical history (hypertension, diabetes, heart disease, joint pain, incontinence, vision impairments) and medications taken.
Outcome Measures
Falls awareness was assessed using Fall Awareness Behaviour Questionnaire (FaB) (22). The FaB is a self- rating scale used to assess the actions and behaviours that older adults usually practice to prevent falls. The scale is made of thirty items and ten subscales. The subscales are: (1) protective mobility (5 items); (2) cognitive adaptations (6 items); (3) awareness (4 items); (4) avoidance (5 items); (5) pace (2 items); (6) practical strategies (3 items); (7) being observant (1 item); (8) displacing activities (1 item); (9) changes in level (2 items) and; (10) getting to phone (1 item). For each scale, the participants were required to give a score based on four categories, never, sometimes, often and always. The data was analysed based on its total score and scores could range from 30 (risky fall behaviour) to 120 (safest falls prevention behaviour). The lower the score, the more likely a person would engage in risky behaviours. Higher scores indicated a person who was more likely to be aware of falls prevention. FaB was found to have a high validity with internal consistency of 0.84 and content validity index of 0.93 (22). The scale has been adapted for Malaysian culture in a previous local study. For the adapted scale, the Cronbach’s Alpha coefficients was 0.723 with the removal of two items, indicating an acceptable internal consistency.
Fall Risk Assesment Questionnaire (FRAQ) was used to assess falls prevention knowledge among older adults (23). It consists of twenty-two questions and these questions evaluate falls prevention knowledge based on the apsects of: (1) behaviours (eight items); (2) environmental (five items); (3) medical condition (six items) and; (4) medication (three items). The total full score for this questionnaire is 21. Higher score suggests higher knowledge of falls among older adults (24). FRAQ was found to have a strong agreement with clinical evaluation (kappa= 0.875, p<0.001) and good validity. The kappa value for individual items is ranged from 0.305 to 0.832 (24). The Malay version of FRAQ used in this study was one that had been translated for a previous local sudy. Its internal consistency had a Cronbach’s alpha value of 0.748 after removing one item.
Participants performed timed-up and go (TUG) test to assess balance and mobility status (25). TUG test was found to have high sensitivty and specifity value of 87% (26). The time taken from standing up from an armless chair with 46cm height, walking 3m towards a cone at their usual and comfortable pace, turning, walking back towards the chair and sitting down was recorded in seconds. The participants were instructed to wear their shoes and were allowed to walk with or without assistive devices. Time taken to complete the test was recorded in seconds.
Six metre gait speed was performed by participants as a measure of functional mobility. Time taken to complete walking at their own pace for a distance of six meters was taken in seconds.
30 seconds chair stand test was used to measure lower limb muscle strength among the older adults. A 17 inches (43.2cm) plastic chair without armrests was placed against the wall to prevent the chair from slipping during the test. Participants were required to sit with their backs against the backrest, arms crossed against their chest, feet with shoulder distance apart and placed firmly on the floor. Participants were instructed to stand and sit as many times as possible within 30 seconds. One repetition was considered as a complete sit and stand. Chair stand test has been shown to have a high test-retest correlation for both men (0.84) and women (0.92) (27). This test also has a good criterion related validity, r =0.78 in men and r =0.71 (27).
Dominant hand grip strength testing has been shown to have good validity with high correlation (r=0.99, p<0.001) (28). The participants were asked to sit upright with elbow and shoulders positioned at a 90-degree angle, and forearms placed in neutral position. The participants were required to squeeze the dynamometer handle as hard as possible using their dominant hand on the command ‘start’ and sustain it for three seconds. This test was conducted twice and the maximum reading was taken as the result.
Data Analysis
Data was analysed using IBM Statistical Package for Social Sciences (SPSS) software version 23.0, IBM Corporation, United States. Statistical analysis of Alpha level 0.05 was used in all statistical tests. Distribution of all data were analysed using Kolmogorov-Smirnov, Shapiro-Wilk and other non-parametric tests. Means and standard deviations were calculated for the following variables: (1) age; (2) MMSE scores; (3) GDS-15 scores; (4) BMI; (5) FRAQ scores; (6) FaB scores; (7) TUG; (8) gait speed; (9) hand grip strength; (10) chair stand tests. While, percentage were described for the following variables: (1) gender; (2) race; (3) education status; (4) marital status; (5) living status; (6) working status; (7) number of comorbidities; (8) number of medication; (9) fall history.
Stepwise linear regression analysis with level at Alpha level 0.05 was used to analyse the association between the practice of falls prevention behaviour (dependent variables) and sociodemographic factors (age, gender, race, number of comorbidity, marital status, employment status, living status and education level), physical factors (mobility status, lower limb and upper limb muscle strength) and clinical factors (knowledge of falls, comorbidity, history of falls and number of falls).
Ethics Approval and Consent to Participate
This study was approved by the Medical Research and Ethics Committee of Universiti Kebangsaan Malaysia (UKM PPI/111/8/JEP-2018-559).