Fall Predictor Research Articles

Comparison of different maximal isometric strength of lower limb muscle groups in predicting fall-risk among older persons.

Many independent studies have investigated the role of normalized maximal voluntary isometric strength (MVIS) of lower limb muscle groups (MVISLLMG) by body weight and summed knee and ankle muscle strength in predicting the risk of falling among older persons. However, it is unknown which MVISLLMG is better at predicting the fall risk. This study aimed to compare different MVISLLMG in predicting fall-risk among older persons against the reference standard (history of falls). This study had a cross-sectional retrospective diagnostic research design. 47 fallers and 93 non-fallers were recruited from Anand district, Gujarat, India, using sequential sampling. The MVISLLMG was measured with a microFET®2 hand-held dynamometer. Following feature selection, four machine learning (ML) models (Random Forest (RF), k-Nearest Neighbors (KNN), Navie Bayes (NB), and Kernel Support Vector Machines (SVM)), were utilized to assess the diagnostic characteristics of every measured MVISLLMG in comparison to the reference standard. The best ML model was chosen based on the highest diagnostic performance in predicting fall-risk. Among the ML models, the NB revealed that the non-normalized summed MVIS of knee and ankle muscle (Sensitivity (Se)= 87%, Specificity (Sp)= 91%, Accuracy (Ac)= 90%, Precision (Pr)= 84%) has the best diagnostic characteristics in fall-risk prediction against the fall history, followed by non-normalized MVIS of hip abductor, knee extensor, plantar flexor, and dorsiflexor, normalized summed MVIS of hip sagittal and knee muscle, and normalized MVIS of hip sagittal and frontal, knee, and plantar flexor. These results suggest that non-normalized summed MVIS of knee and ankle muscles is the better fall predictor in older persons compared to other index measures. This finding may assist clinicians in playing a better role in selecting suitable MVISLLMG data for fall risk assessment and predicting falls.

Elasticity of leg muscles and incidence of falls in older adults: a prospective cohort analysis.

Aging impacts muscle strength and elasticity, which in turn influence dynamic balance, walking speed, and physical performance. We aimed to evaluate the relationship between the elasticity of leg muscles and incidence of falls in older adults. We conducted a prospective cohort analysis with outpatients from a geriatric clinic. Any history of falls in the past year was recorded. Timed up and go test, muscle thickness, and handgrip strength tests were performed. Elasticities of the gastrocnemius medialis (GM) and rectus femoris (RF) muscles were evaluated using shear wave elastography. Patients self-recorded their falls, and additional phone calls were made to them each month for 6months. The median age of the patients (n = 55) was 72years (66-86); and 72% were women. The GM showed significantly lower elasticity in patients with history of falls in the past year than in those without it (8.08 kPa [3.90-16.17] vs. 9.70 kPa [4.99-20.95]; p = 0.028). A similar negative correlation between GM and fall incidence was noted among those with additional falls during the follow-up period (6.96 kPa [3.90-12.41] vs. 9.13 kPa [4.99-20.95]; p = 0.019). GM elasticity was significantly correlated with the timed up and go test score (r = -0.612, p < 0.001), handgrip strength (r = 0.384, p = 0.015), and muscle thickness (r = 0.232, p = 0.049). No such associations were observed for the RF muscles. GM muscle elasticity is associated with alterations in muscle structure that may lead to falls in older adults. Therefore, muscle elasticity may be a fall predictor in older adults.

Self-perceived balance confidence is independently associated with any subsequent falls and injurious falls among community-dwelling older fallers: A prospective cohort study

IntroductionBalance confidence contributes to activity restriction, decline in physical function, and falls, potentially. This study aimed at examining whether balance confidence is a significant fall predictor independent of physical function measures in community-dwelling older adults. MethodsThis was an observational cohort study. Self-perceived balance confidence was measured by the Activities-specific Balance Confidence scale (ABC) and the short form of it, ABC-6. Other potential predictors included gait speed, Five Times Sit-to-Stand Test, Physiological Profile Assessment (PPA), age, sex, cognition, depression, arthritis, body mass index, recurrent fall history, number of comorbidities, number of prescribed medications, and the use of walking aids. Fall occurrence and subsequent injuries were followed up trimonthly for 12 months. The association of the potential predictors with falls was examined with multivariable logistic regression analyses. Sensitivity analysis for their association with injurious falls in a year was performed. Results461 community-dwelling older adults with a fall history in the past year completed the 12-month follow-up (mean age 70.6±7.1 years, range 60–92 years; 81% females). Participants with high balance confidence (ABC score ≥76 and ABC-6 score ≥64) were less likely to fall in the subsequent year, independently of the physical function measures and other covariates (adjusted OR 0.51-0.57, 95%CI 0.30-0.97, p<0.05). ConclusionBalance confidence is a significant fall predictor independent of physical function measures and other covariates in community-dwelling older adults with a history of falls. Future studies might investigate the causal effects of balance confidence on the risk of falling with randomized controlled trials.

Fall predictors in hospitalized patients living with cancer: a case-control study.

To identify fall predictors and develop an assessment tool to be used for screening hospitalized cancer patients at risk for fall. A retrospective case-control study was conducted in 2018 at a cancer center in Northern Italy. The study participants were 448 adult cancer patients admitted to the oncology ward from 2009 to 2013. The case group consisted of 112 patients presenting at least one fall, while controls were randomly chosen by matching each case for age, sex, and admission period with three patients who did not fall. Data for the fall predictors were extracted from the electronic medical records. Conditional logistic regression was used to evaluate the association between patient's characteristics and fall risk. The overall prevalence of patients having at least one candidate fall predictor was high (98%). Seven of the studied variables showed an independent association with fall risk at multivariate analysis. These were tumor site, the presence of neurologic diseases, gait imbalance disorders, fatigue, and the assumption of certain medications such as diuretics, hypnotics, and opioids (odds ratios and 95% confidence intervals in brackets were 3.78 (1.78-8.13), 2.26 (1.08-4.77), 4.22 (1.87-9.52), 2.76 (1.45-5.26), 2.66 (1.52-4.66), 2.41 (1.20-4.85), and 3.03 (1.68-5.45), respectively). In this study, we identified falling risk factors in an Italian population of hospitalized cancer patients and developed a new risk assessment tool. An external validation is necessary before implementing our screening tool in clinical practice.

Falls in Rheumatoid Patients: Does Ankle and Foot Ultrasonography have a Predictive Role? A Single-blind Study

ABSTRACT Background Rheumatoid arthritis (RA) patients have a higher prevalence of falls compared with the healthy population. Several risk factors of falls in RA have been postulated, including high disease activity, low balance, muscle weakness and non-treatment with biologics. Aim of the work We investigated our hypothesis that the sonographically detected ankle and foot changes in RA patients can predict falls in this population. To our knowledge, no previous study had investigated this before. Methods In a total of 101 RA patients, we performed assessments of disease activity, disability level, gait speed, balance status, clinical examination of ankle and foot and an MSUS assessment of the ankle and foot joints and tendons as possible risk factors of falls. Results The Berg balance test had the highest fall-predicting power (71.3%), followed by a gait speed test and restricted range of motion (ROM) of the Rt. subtalar joint, each with a predictive power of 70.3%. Of the sonographic findings, erosion of the first metatarsophalangeal (MTP) joint was the most accurate fall predictor, followed by erosion of the talonavicular joint and tenosynovitis of the ankle dorsal flexors with an area under the curve of 0.656, 0.642 and 0.614, respectively. Conclusion The use of the MSUS as an adjunct objective method for predicting falls in RA patients has not been studied before. It was found that clinical foot and balance testing was a superior and easier way of predicting falls in RA patients than using ultrasonography.

Cog-y Balance Test: Validity and Reliability for Fall Prediction in Community Dwelling Elders

Background: Years lived with disability (YLDs) by the end result of falls were 631 per 1 lakh population inIndia in 2010 and it is raising till date. The avoidance of fall among older is perhaps one of the prime publichealth question.Objective: Aim of our study is to assess the concurrent validity, inter-rater and intra-rater reliability ofCognitive-Y Balance Test (Cog-YBT).Methods: In this cross sectional study, 29 male and 27 female participants (64.77± 4.66 years) were recruitedfrom the community area. Data collection for the Berg Balance scale (BBS), Cognitive Time Up and Go(Cog-TUG) and Cog-YBT were done for all. Cog-YBT data were taken by two different Data collector(Rater 1 and 2). Rater 1 have taken data for 2 time with adequate rest between two sessions.Results: Statistically significant (p &lt; 0.05), though weak to moderate correlations were found for Cog-YBTto BBS and Cog-YBT to Cog-TUG comparison (spearman’s correlation coefficient = 0.35 to.0404). ICCscore for intra-rater and inter-rater reliability was excellent for Cog-YBT (ICC &gt;0.9).Conclusion: Though the reliability of cognitive Y Balance test is high in our context, the concurrent validityof this test is not adequate to use it as fall predictor measure in community dwelling elder yet.

Posturographic measures did not improve the predictive power to identify recurrent falls in community-dwelling elderly fallers

OBJECTIVE:This study aimed to evaluate if posturography can be considered a recurrent fall predictor in elderly individuals.METHODS:This was a cross-sectional study. A total of 124 subjects aged 60 to 88 years were evaluated and divided into two groups—the recurrent fallers (89) and single fallers (35) groups. Patients' sociodemographic characteristics were assessed, and clinical testing was performed. The functional test assessment instruments used were timed up and go test (TUGT), Berg Balance Scale (BBS), five times sit-to-stand test, and Falls Efficacy Scale (to measure fear of falling). Static posturography was performed in a force platform in the following three different situations—eyes open (EO), eyes closed (EC), and EO dual task.RESULTS:There were significant differences between the single and recurrent fallers groups regarding the fear of falling, the Geriatric Depression Scale score, the mean speed calculated from the total displacement of the center point of pressure (COP) in all directions with EO, and the root mean square of the displacement from the COP in the mediolateral axis with EC. Based on the hierarchical logistic regression model, none of the studied posturographic variables was capable of significantly increasing the power of differentiation between the recurrent and single fallers groups. Only TUGT with a cognitive distractor (p<0.05) and the BBS (p<0.01) presented with significant independent predictive power.CONCLUSION:TUGT with a cognitive distractor and the BBS were considered recurrent fall predictors in elderly fallers.

Impaired local dynamic stability during treadmill walking predicts future falls in patients with multiple sclerosis: A prospective cohort study

BackgroundFalling is a significant problem in patients with multiple sclerosis (MS) and the majority of falls occur during dynamic activities. Recently, there have been evidences focusing on falls and local stability of walking based on dynamic system theory in the elderly as well as patients with cerebral concussion. However, in patient with MS, this relationship has not been fully investigated. The aim of this study was to investigate local stability of walking as a risk factor for falling in patients with MS. MethodsSeventy patients were assessed while walking at their preferred speed on a treadmill under single and dual task conditions. A cognitive task (backward counting) was used to assess the importance of dual tasking to fall risk. Trunk kinematics were collected using a cluster marker over the level of T7 and a 7-camera motion capture system. To quantify local stability of walking, maximal finite-time Lyapunov exponent was calculated from a 12-dimensional state space reconstruction based on 3-dimensional trunk linear and angular velocity time series. Participants were classified as fallers (≥1) and non-fallers based on their prospective fall occurrence. Findings30 (43%) participants recorded ≥1 falls and were classified as fallers. The results of multiple logistic regression analysis revealed that short-term local dynamic stability in the single task condition (P<0.05, odds ratio=2.214 (1.037–4.726)) was the significant fall predictor. InterpretationThe results may indicate that the assessment of local stability of walking can identify patients who would benefit from gait retraining and fall prevention programs.

Risk factors for falls among older Aboriginal and Torres Strait Islander people in urban and regional communities.

To examine associations between fall risk factors identified previously in other populations and falls among Aboriginal people aged 60years and older, living in New South Wales, Australia. Interviews were conducted with older Aboriginal people in five urban and regional communities. Associations between past falls and 22 fall predictor variables were examined using linear and multiple regression analyses. Of the 336 participants, 80 people (24%) reported at least one fall in the past year, and 34 (10%) reported two or more falls. Participants had anincreased fall risk if they were female; used three or more medications; had arthritis, macular degeneration, depression, history of stroke; were unable to do their own housework; or were unable to do their own shopping. Falls were experienced by one-quarter of study participants. Fall risk factors identified for older Aboriginal people appear to be similar to those identifiedin the general population. Understanding of fall risk factors may assist with the development of appropriate and effective community-led fall prevention programs.

Relationship between Center of Pressure and Local Stability of the Lower Joints during Walking in the Elderly Women

Objective: The purpose of this study was to determine the relationship between center of pressure (CoP) and local stability of the lower joints, which was calculated based on approximate entropy (ApEn) during walking in elderly women. Method: Eighteen elderly women were recruited (age: <TEX>$66.4{\pm}1.2yrs$</TEX>; mass: <TEX>$55.4{\pm}8.3kg$</TEX>; height: <TEX>$1.56{\pm}0.04m$</TEX>) for this study. Before collecting data, reflective marker triads composed of 3 non-collinear spheres were attached to the lateral surface of the thigh and shank near the mid-segment to measure motion of the thigh and shank segments. To measure foot motion, reflective markers were placed on the shoe at the heel, head of the fifth metatarsal, and lateral malleolus, and were also placed on the right anterior-superior iliac spine, left anterior-superior iliac spine, and sacrum to observe pelvic motion. During treadmill walking, kinematic data were recorded using 6 infrared cameras (Oqus 300, Qualisys, Sweden) with a 100 Hz sampling frequency and kinetic data were collected from a treadmill (Instrumented Treadmill, Bertec, USA) for 20 strides. From kinematic data, 3D angles of the lower extremity's joint were calculated using Cardan technique and then ApEn were computed for their angles to evaluate local stability. Range of CoP was determined from the kinetic data. Pearson product-moment and Spearman rank correlation coefficient were applied to find relationship between CoP and ApEn. The level of significance was determined at p<.05. Results: There was a negative linear correlation between CoP and ApEn of hip joint adduction-abduction motion (p<.05), but ApEn of other joint motion did not affect the CoP. Conclusion: It was conjectured that ApEn, local stability index, for adduction/abduction of the hip joint during walking could be useful as a fall predictor.

Motor Performance and Physical Activity as Predictors of Prospective Falls in Community-Dwelling Older Adults by Frailty Level: Application of Wearable Technology

Background: Few studies of the association between prospective falls and sensor-based measures of motor performance and physical activity (PA) have evaluated subgroups of frailty status separately. Objective: To evaluate wearable sensor-based measures of gait, balance, and PA that are predictive of future falls in community-dwelling older adults. Methods: The Arizona Frailty Cohort Study in Tucson, Arizona, followed community-dwelling adults aged 65 years and over (without baseline cognitive deficit, severe movement disorders, or recent stroke) for falls over 6 months. Baseline measures included Fried frailty criteria: in-home and sensor-based gait (normal and fast walk), balance (bipedal eyes open and eyes closed), and spontaneous daily PA over 48 h, measured using validated wearable technologies. Results: Of the 119 participants (36% non-frail, 48% pre-frail, and 16% frail), 48 reported one or more fall (47% of non-frail, 33% of pre-frail, and 47% of frail). Although balance deficit and PA were independent fall predictors in pre-frail and frail groups, they were not sensitive to predict prospective falls in the non-frail group. Even though gait performance deteriorated as frailty increased, gait was not a predictor of prospective falls when participants were stratified based on frailty status. In pre-frail and frail participants combined, center of mass sway [odds ratio (OR) = 5.9, 95% confidence interval (CI) 2.6-13.7], PA mean walking bout duration (OR = 1.1, 95% CI 1.0-1.2), PA mean standing bout duration (OR = 0.94, 95% CI 0.91-0.99), and a fall in previous 6 months (OR = 7.3, 95% CI 1.5-36.4) were independent predictors of prospective falls (area under the curve: 0.882). Conclusion: This study suggests that independent predictors of falls are dependent on frailty status. Among sensor-derived parameters, balance deficit, longer typical walking episodes, and shorter typical standing episodes were the most sensitive predictors of prospective falls in the combined pre-frail and frail sample. Gait deficit was not a sensitive fall predictor in the context of frailty status.

Hyponatremia as a fall predictor in a geriatric trauma population

IntroductionApproximately one in three older adults fall each year, resulting in a significant proportion of geriatric traumatic injuries. In a hospital with a focus on geriatric fall prevention, we sought to characterize this population to develop targeted interventions. As mild hyponatremia, defined as a serum sodium <135meq/L, has been reported to be associated with falls, unsteadiness and attention deficits, we hypothesized that hyponatremia is associated with falls in our geriatric trauma population. MethodsGender, age, pre-existing conditions (cardiac disease, diabetes, hematologic disorder, liver disease, malignancy, musculoskeletal disorder, neurological disorder, obesity, psychiatric disorder, pulmonary disease, renal disease, thyroid disease), mechanism of injury and admitting serum sodium level were queried for all geriatric trauma admissions from 2008 to 2011. Mechanism of injury was coded as falls admissions and non-falls admissions. Admitting serum sodium levels were coded as hyponatremic (<135mmol/L) and not hyponatremic (≥135mmol/L). ResultsOf the 2370 geriatric trauma admissions during the study period, there were 1841 (77.7%) falls admissions and 293 (12.4%) patients who were hyponatremic. Gender, age, neurological disorder, hematologic disorder, and hyponatremia were found to be significant predictors of falls in both univariate and multivariable analyses. ConclusionHyponatremic patients are significantly more likely to be admitted for a fall than non-hyponatremic patients, when adjusting for age, neurological disorder, and hematologic disorder. Consequently, hyponatremia identification and management should be an integral part of any geriatric trauma fall prevention programme. Additionally, if hyponatremia is found during a geriatric fall workup, it should be corrected prior to discharge and closely monitored by a primary care physician to prevent recurrent episodes of falls.

AB0350 Prospective Observational Study in Rheumatoid Arthritis Adults in Russia: Osteoporosis, Sarcopenia, Falls and Fractures

BackgroundOsteoporosis (OP) and falls lead to fractures especially in patients with rheumatoid arthritis (RA) due to pain and joint abnormalities, low muscle mass and strength, functional impairment, that have consequences...

OP0230-HPR Falls, Fear of Falling and Risk Factors in Adults with Rheumatoid Arthritis: A Prospective Study

BackgroundThe risk of falls may be higher in people with RA due to muscle weakness, impaired mobility and balance, pain and joint abnormalities however there are few studies that have...

LEARNING TO PREDICT HUMANOID FALL

Falls are undesirable in humanoid robots, but also inevitable, especially as robots get deployed in physically interactive human environments. We consider the problem of fall prediction: to predict if the balance controller of a robot can prevent a fall from the robot's current state. A trigger from the fall predictor is used to switch the robot from a balance maintenance mode to a fall control mode. It is desirable for the fall predictor to signal imminent falls with sufficient lead time before the actual fall, while minimizing false alarms. Analytical techniques and intuitive rules fail to satisfy these competing objectives on a large robot that is subjected to strong disturbances and exhibits complex dynamics. We contribute a novel approach to engineer fall data such that existing supervised learning methods can be exploited to achieve reliable prediction. Our method provides parameters to control the tradeoff between the false positive rate and the lead time. Several combinations of parameters yield solutions that improve both the false positive rate and the lead time of hand-coded solutions. Learned solutions are decision lists with typical depths of 5–10, in a 16-dimensional feature space. Experiments are carried out in simulation on an ASIMO-like robot.

Predicting Falls of a Humanoid Robot through Machine Learning

Although falls are undesirable in humanoid robots, they are also inevitable, especially as robots get deployed in physically interactive human environments. We consider the problem of fall prediction, i.e., to predict if a robot's balance controller can prevent a fall from the current state. A trigger from the fall predictor is used to switch the robot from a balance maintenance mode to a fall control mode. Hence, it is desirable for the fall predictor to signal imminent falls with sufficient lead time before the actual fall, while minimizing false alarms. Analytical techniques and intuitive rules fail to satisfy these competing objectives on a large robot that is subjected to strong disturbances and therefore exhibits complex dynamics. Today effective supervised learning tools are available for finding patterns in high-dimensional data. Our paper contributes a novel approach to engineer fall data such that a supervised learning method can be exploited to achieve reliable prediction. Specifically, we introduce parameters to control the tradeoff between the false positive rate and lead time. Several parameter combinations yield solutions that improve both the false positive rate and the lead time of hand-coded solutions. Learned predictors are decision lists with typical depths of 5-10, in a 16-dimensional feature space. Experiments are carried out in simulation on an Asimo-like robot.

Person—environment fit predicts falls in older adults better than the consideration of environmental hazards only

Objective: To test the hypotheses that the empirical consideration of objective person—environment fit in the home environment is a stronger predictor of indoor falls among older adults than the assessment of environmental barriers only, and that perceived aspects of home play a role as predictors for falls. Design: Survey study with data collection at home visits, followed up by self-reports about falls at home visits one year later. Setting: Urban districts in Sweden, Germany, Latvia. Participants: Eight hundred and thirty-four single-living, older adults (75—89 years), in ordinary housing. Measurements: An assessment of objective person—environment fit in the home environment (housing enabler), a self-rating of the perceived home environment (usability in my home) and retrospective self-reports on indoor falls. Results: The participants reporting falls tended to be frailer than the non-fallers. The number of environmental barriers in the home was similar for the fallers and non-fallers; the magnitude of person—environment fit problems was higher among the fallers. The person—environment fit problem variable was a stronger fall predictor (odds ratio (OR) = 1.025; P=0.037) than number of environmental barriers (n.s.), even after controlling for confounders. Fallers also experienced lower usability of their home. Conclusion: The results suggest that much of the inconclusiveness of the data in the relationship between environmental hazards and falls in the previous falls literature could be due to the neglect of person—environment fit assessment. The effectiveness of environmental interventions based on the notion of person—environment fit compared with traditional home hazard checklists remains to be tested.

Differentiating fall-prone and healthy adults using local dynamic stability

Variability in kinematic and spatio-temporal gait parameters has long been equated with stability and used to differentiate fallers from non-fallers. Recently, a mathematically rigorous measure of local dynamic stability has been proposed based on the non-linear dynamics theory to differentiate fallers from non-fallers. This study investigated whether the assessment of local dynamic stability can identify fall-prone elderly individuals who were unable to successfully avoid slip-induced falls. Five healthy young, four healthy elderly and four fall-prone elderly individuals participated in a walking experiment. Local dynamic stability was quantified by the maximum Lyapunov exponent. The fall-prone elderly were found to exhibit significantly lower local dynamic stability (i.e. greater sensitivity to local perturbations), as compared to their healthy counterparts. In addition to providing evidence that the increased falls of the elderly may be due to the inability to attenuate/control stride-to-stride disturbances during locomotion, the current study proposed the opportunity of using local dynamic stability as a potential indicator of risk of falling. Early identification of individuals with a higher risk of falling is important for effective fall prevention. The findings from this study suggest that local dynamic stability may be used as a potential fall predictor to differentiate fall-prone adults.

The Physical Performance Test as a predictor of frequent fallers: a prospective community-based cohort study

Objective: To construct a risk model in order to identify elderly individuals at risk of frequent falling. Design: Prospective community-based cohort study over 12 months. Setting: Baseline measures were performed at a local community centre. Subjects: Two hundred and sixty-three community-dwelling elderly people (mean age 72 years). Measurements: A variety of variables were evaluated, including medical, psychological, sensory, physical and postural control measurements. Fall incidence was monitored retrospectively and during one-year follow-up. Results: Logistic regression analysis showed that polypharmacia was the most prominent medical fall predictor with an odds ratio (OR) of 1.29 (P =0.005), poor visual acuity the best sensory predictor (OR=0.84; P =0.009) and general fear of falling the most crucial psychological predictor (OR=3.25; P B=0.001). Increased postural sway in near-tandem stance with eyes open was selected as the best balance predictor for falls (OR=5.60; P =0.010), followed by delayed anteroposterior movement velocity during rhythmic weight shifts (OR=0.42; P =0.004). The best physical predictor was a low score on the Physical Performance Test (OR=4.16; P <0.001), followed by decreased maximal handgrip strength (OR=0.87; P <0.001) and increased timed chair-stands (OR=1.13; P <0.003). Step-by-step regression analysis revealed a risk model for the prediction of future falls, as a combination of the Physical Performance Test and maximal handgrip strength. Conclusion: This study confirms the multicausality of falls, since medical, psychological, sensory, postural control as well as physical variables provides a predictive value. The composed fall risk model was mainly physically oriented.

Fall-related factors in a cohort of elderly community residents

To identify factors associated to fall and recurrent fall episodes among elderly living in a community, and to determine the relative risk of each factor as a fall predictor. A two-year follow-up study with two segments of multi-dimensional household surveys (1991-1992 and 1994-1995) was carried out in a cohort of 1,667 elderly, aged 65 or older, living in a community in the city of São Paulo, Brazil. The instrument used for data collection was the BOMFAQ, the Brazilian version of the Multidimensional Functional Assessment Questionnaire (OARS). A stepwise logistic regression analysis was conducted with p<0.05 and 95% CI. About 31% of the elderly referred a fall episode and around 11% referred two or more fall episodes in the previous year of the first survey. After follow-up, 53.4% didn't refer any fall episodes, 32.7% referred a fall episode either in the first survey or in the second one and almost 14% referred fall episodes in both surveys. The predictive model of recurrent falls was composed by the following variables: absence of a spouse (OR=1.59 95% CI 1.00-2.52), not having a reading habit (OR= 1.56 95% CI 1.03-2.37), history of fractures (OR=4.6 95% CI 2.23-9.69 difficulty to perform one to three activities of daily life (OR=2.37 95% CI 1.49-3.78), difficulty to perform four or more activities of daily life (OR=3.31 95%CI 1.58-6.93) and among those whose sight is most impaired (OR=1.53 95%CI 1.00-2.34). Population ageing and increase in life expectancy demand preventive and rehabilitation actions in order to reduce risk factors for falls, such as impaired functional capacity, impaired eye sight and lack of cognitive stimulation.