Three hundred and thirty-three community dwelling older people (157 men, 176 women), who were enrolled in a larger longitudinal study of cognitive function and ageing (Sydney Memory and Aging Study), agreed to participate in this prospective cohort study with a 1-year follow-up for falls. Participants were community-dwelling men and women living in Eastern Sydney, recruited into the Sydney Memory and Ageing study via the electoral role (8). Participants were excluded if they scored less than 24 in the Mini-Mental State Examination, had insufficient knowledge of English language, had a previous diagnosis of dementia or developmental disability, psychotic symptoms, Parkinson’s disease, multiple sclerosis, motor neuron disease, central nervous system inflammation or if they had medical or psychological conditions that may have prevented them from completing assessments. The Human Studies Ethics Committee at the University of New South Wales granted approval for the study, and informed consent was obtained from individuals prior to participation. Participants were classified into two groups based on yes or no answers to the following question asked at the baseline questionnaire: “Do you currently suffer from any following conditions/diseases? Pain – Knee / Leg?”
Participants completed a set of questionnaires and underwent assessments of health and physical performance. The questionnaires provided information on demographics, cognitive status and medical history, health and physical activity, concern about falls and fall history.
Demographic information and medical history
Participants were asked to list major medical conditions that they had such as heart conditions, high blood pressure, high blood cholesterol levels, diabetes, stroke, respiratory conditions (chronic lung disease and asthma) and others (e.g. cancer). Participants also listed other musculoskeletal conditions such as joint replacements, osteoarthritis, rheumatoid arthritis and gout. The type and number of prescribed medications, including nonsteroidal anti-inflammatory drugs (NSAIDs) and non-narcotic analgesics were also recorded.
Health and physical activity questionnaires
The presence of depressive symptoms was assessed using the Patient Health Questionnaire – 9 (9). Anxiety was assessed with the Generalised Anxiety Disorder Scale 7 (10). The Incidental And Planned Exercise Questionnaire (IPEQ) for older people was used to assess participants’ physical activity the during the last three months (11). Total time spent in planned (planned exercise and walks) and incidental activity (casual day-to-day activities) was expressed as hours per week.
Physical performance tests
Physical Performance was assessed using a battery of sensorimotor and balance tests that have been shown to discriminate significantly between fallers and non-fallers (12). Visual contrast sensitivity was assessed using the Melbourne Edge Test (12). Proprioception was measured as the average alignment error (°) from 5 trials of a task requiring aligning the great toe either side of a vertical protractor while seated and with eyes closed (12). Lower limb muscular strength was measured as the maximal (from three trials) isometric knee extension force (kg) with participants seated, knee flexed to 90° and a custom-built strain gauge attached to the lower leg (12), multiplied by the length of the force application from the knee joint to represent knee extension torque and presented as a percentage of body weight*height Reaction time involved a random-delay light stimulus and finger-press response with the average time (ms) of 10 trials recorded (12). Postural sway was measured in 4 conditions using a sway-meter that recorded displacements of the body (mm) at the level of the waist while participants stood on the floor or a foam rubber mat with eyes open or closed for 30 s (12). Maximal balance range was measured using a sway-meter that recorded anterior and posterior displacements of the body while participants leaned as far as they could forwards and then backwards from the ankles while keeping the rest of their body rigid (13). Co-ordinated stability was measured by attaching the sway-meter to the participant at waist level with a pen attached to the rod extending anteriorly. Participants viewed the pen and were asked to navigate it through a convoluted track without moving their feet, with the better of two trials being taken (13).
The participant’s ability to perform transitional movements was measured using the Five-Times Sit to Stand Test. Participants were instructed to stand up and sit down five times as quickly as possible with arms crossed over the chest. Timing began when the examiner said ‘go’ and stopped when the participant’s buttocks touched the chair on the fifth repetition (14). Mobility was measured using the Timed Up and Go Test. The participant was asked to rise from a chair, walk 3 metres, turn, return and sit back in the chair as quickly as possible (15). Average gait velocity was calculated using a 5.7m GAITRite Platinum Portable Walkway System (GAITrite, CIR Systems Inc. Franklin NJ, USA). Participants performed three walks across the mat at their usual, comfortable walking speed starting from a point 2m before the mat and stopping at point 2m after the mat (16). Quick, accurate stepping was assessed with the Choice Stepping Reaction Time (CSRT) test. Four randomly presented visual stimuli (arrows: front left, front right, left, right) were displayed on a screen ahead and participants were required to step onto the corresponding panel of the step mat as quickly as possible (17). with Step time was recorded and averaged for 20 trials.
Falls efficacy and falls surveillance
The Falls Efficacy Scale - International (FES-I) was used to assess confidence in performing activities without falling (18). The FES-I consists of 16 items rated using a Likert scale. The total score ranges from 16 (no concern) to 64 (severe concern).
Falls were monitored over a 12-month period using monthly falls calendars. Follow-up calls were made to participants if calendars were not returned. A fall was defined as “an unexpected event in which the person comes to rest on the ground, floor, or lower level” (19).
To permit parametric analyses, data with right-skewed distributions were square root or inverse transformed. Participants who were unable to safely complete a physical assessment due to physical incapacity were given a score of three standard deviations above or below the group mean to reflect their poor performance on that assessment. Participants who scored worse than this had their score censored at this level. An expectation maximisation analysis was completed to fill gaps in missing data (< 5% in each case).
Independent T-tests were conducted to determine differences in continuous variables between the pain and no pain group. Chi-square tests were used to compare the prevalence of fallers in the pain and no pain group. The Relative Risk (RR) of having multiple falls in the knee pain group, relative to no pain, was determined using modified Poisson Regression (20). Posited explanatory variables that were associated with both fall risk and knee pain measures in independent samples t-tests (P<0.1) were examined separately to determine how much each reduced the RR between knee pain and fall status. According to the criteria outlined by Barron and Kenny (1986) mediation was verified when the factor of choice was significantly associated with the independent variable (pain) and the dependent variable (faller status) and both the dependent variable and independent variables were significantly associated in univariate analyses (21). Identified mediating covariates were then combined, stepwise in order of RR reduction magnitude, into a final modified Poisson regression model and the percentage reduction in RR was computed. All statistical analyses were performed using SPSS (version 25 for Windows; SPSS Science, Chicago, IL), and P < .05 was considered significant.