Falls risk assessment score (FRAS): Time to rethink†

  • Abstract
  • Full Text
  • Images
  • References

Abstract

Background

Early in the development of geriatric medicine, falls were identified as a “geriatric giant”, a nonspecific indicator of functional decompensation. This led to the notion of “falls prevention services”, and the concept that identification of those patients at high risk of falls is essential to approach this group of elderly people.

Objective

This work was carried out aiming to develop a model that predicts falls risk for both in- as well as outpatients using clinical variables that are easily assessed in clinical practice.

Study Design

A case-control study to determine the risk factors and the prediction rule of falls risk among older people.

Methods

Three hundred and seventy-three outpatients and 186 inpatients, with a minimum age of 65 years, were assessed for falls risk factors. The clinical characteristics with independent predictive value for the development of falls were selected using logistic regression analysis. The diagnostic performance of the prediction rule was evaluated using the area under the curve. Cross-validation controlled for over fitting of the data (internal validation) was also carried out.

Results

The prediction rule consisted of five clinical variables: history of falls in the last 12 months, slowing of the walking speed/change in gait, history of loss of balance in the last 12 months, and impaired sight and weak hand grip. The prediction score ranged from 0 to 6.5, and corresponded to the percent chance of sustaining a fall. For several cutoff values, the positive and negative predictive values were determined. The area under the curve values for the prediction rule was 0.89.

Conclusion

In elderly people, the risk of sustaining a fall can be predicted, thereby allowing individualized decisions regarding the patient’s management. Falls risk assessment score is a new self-reported tool that can be used in standard clinical practice by all health care professionals both in the outpatient and the acute hospital inpatient settings. Assessing for the falls risk would help to minimize the negative impact of falling on the patient’s physical, psychological, and social functional abilities.

Keywords:

Falls risk assessment score, FRAS, Falls prediction, Fracture

Article Outline

  1. Introduction
  2. Methods
  3. Results
  4. Discussion
  5. Appendix 1. Falls risk assessment questionnaire
  6. References

Abstract

Background

Early in the development of geriatric medicine, falls were identified as a “geriatric giant”, a nonspecific indicator of functional decompensation. This led to the notion of “falls prevention services”, and the concept that identification of those patients at high risk of falls is essential to approach this group of elderly people.

Objective

This work was carried out aiming to develop a model that predicts falls risk for both in- as well as outpatients using clinical variables that are easily assessed in clinical practice.

Study Design

A case-control study to determine the risk factors and the prediction rule of falls risk among older people.

Methods

Three hundred and seventy-three outpatients and 186 inpatients, with a minimum age of 65 years, were assessed for falls risk factors. The clinical characteristics with independent predictive value for the development of falls were selected using logistic regression analysis. The diagnostic performance of the prediction rule was evaluated using the area under the curve. Cross-validation controlled for over fitting of the data (internal validation) was also carried out.

Results

The prediction rule consisted of five clinical variables: history of falls in the last 12 months, slowing of the walking speed/change in gait, history of loss of balance in the last 12 months, and impaired sight and weak hand grip. The prediction score ranged from 0 to 6.5, and corresponded to the percent chance of sustaining a fall. For several cutoff values, the positive and negative predictive values were determined. The area under the curve values for the prediction rule was 0.89.

Conclusion

In elderly people, the risk of sustaining a fall can be predicted, thereby allowing individualized decisions regarding the patient’s management. Falls risk assessment score is a new self-reported tool that can be used in standard clinical practice by all health care professionals both in the outpatient and the acute hospital inpatient settings. Assessing for the falls risk would help to minimize the negative impact of falling on the patient’s physical, psychological, and social functional abilities.

Keywords:

Falls risk assessment score, FRAS, Falls prediction, Fracture

1. Introduction

The idea of using a falls prediction tool to target patients for fall prevention strategies is an attractive one to health care organizations and clinicians. Tools have frequently been used both in research and real life intervention programs. Despite the lack of evidence for falls risk assessment tools, many hospitals continue to use them.1,2,3 Although the use of such tools might be an attractive option, their use might be falsely reassuring that “something has been done” to target high-risk patients, whereas in fact it is an opportunity to focus on more effective interventions that has been missed.4,5

Falls have many different causes and older people may have several predisposing risk factors. Assessing falls risk can help caregivers and older people predict and even prevent falls. The relative contribution of each fall risk factor differs depending on comorbidities, level of functional independence, and environmental circumstances (e.g., the presence of hazardous conditions, such as poor lighting, slippery floor surfaces, cluttered pathways, and bathrooms without hand rail support). The more fall risk factors there are, the higher the risk of falling. Identifying older people at risk of falling would have a potentially positive effect on the person’s mobility, reducing the fear of falling, and help maintain the person’s autonomy. Potential interventions are based on identifying the fall risk factors and include medical, rehabilitative, environmental, and behavioral approaches.6

The health-related quality of life (HRQOL) assessment is a useful tool in clinical trials; but in most fall prevention trials, it remains unpopular and less commonly used. Even in the few cases of fall prevention studies where HRQOL has been used, it was usually a secondary outcome and incomplete HRQOLs measures were assessed.7 In later life, prevention of falls is a key public health priority. Physical frailty and fall-related injuries are two of the biggest threats to QOL or HRQOL. Furthermore, psychological consequences of falls, such as fear of falling, can have a detrimental effect on the perceived QOL and thereby cause further impairment of psychological and social functional abilities.8

The current limitations associated with falls risk assessment tools further the need to rethink how they may contribute to managing falls risk.4 Available tools are devoted either to community living people or, on the other hand, elderly people living in nursing homes or admitted to the hospital. In general, it has been suggested that the best approach to tackle falls should include a risk factor analysis, and if this analysis reveals a high risk, this should be taken into consideration when formulating a care plan for the patient.9 The American Geriatric Society (AGS), the British Geriatric Societies (BGS), and the American Academy of Orthopedics published their guidelines for prevention of falls in older people,10 including the results of univariate analysis of most common risk factors for falls identified in 16 studies; however, decisions to adopt any particular recommendation were left for the practitioner to be made in light of available evidence and resources.10 This work was carried out aiming to develop a scoring model that predicts falls risk among the elderly people, whether they are inpatients or assessed in the outpatient vicinity, using clinical variables that are easily assessed in standard clinical practice. The derived prediction rule was then internally validated, controlling for over fitting of the data.

2. Methods

A case-control study was set up to determine the risk factors and the prediction rule of falls risk among patients attending the osteoporosis and falls integrated service.

2.1. Patients

2.1.1. Outpatients

This inception cohort comprises more than 1100 patients living within the hospital catchment area, either at home, in a retirement complex, or in a residential or nursing home, referred for osteoporosis and falls assessment. Referral pathway was set up in coordination with the primary care practitioners where patients who sustained a low impact fracture following a fall or patients identified being at risk of falls (having history of fall[s] in the last 12 months in addition to any/or more than one of the following: muscle weakness, gait deficit, balance deficit, use assistive device, visual deficit, arthritis, impaired activities of daily living, cognitive disorders, depression, receiving more than four medications, or having Parkinson’s disease or stroke10) were referred to osteoporosis and falls integrated service. Patients with a minimum age of 65 years were included in this work. All patients were newly diagnosed referred from primary care within 1–5 days from diagnosis.

2.1.2. Inpatients

This cohort included 186 patients (aged older than 65 years) admitted under the care of the gerontorheumatology team. A detailed history from each patient regarding their falls history, comorbid medical conditions, medication use, and risk factors for falls was taken. All patients were subjected to the same assessment carried out for the outpatients group.

2.1.3. Control group

All patients of matched sex and age referred for osteoporosis and falls assessment who did not have a history of falls or low trauma fracture were included as a control group.

At the first visit, each patient completed a questionnaire to describe the incident of fall and to identify risk factors of osteoporosis as well as falls. Fall was defined as “an event that results in a person coming to rest inadvertently on the ground or floor or other lower level with or without loss of consciousness.”11 Information regarding current medical conditions as well as past medical and surgical history, medication use with particular attention to tranquillizers, sedatives, diuretics, antihypertensives, antiparkinsonian drugs, and antidepressants was taken. Past history of previous falls was also recorded. Each patient had a thorough physical examination, including cardiovascular (CV) and neurology examination. Attention was paid to the presence of any or more than one of the following: impaired vision, hearing loss, arthritis, lower limb abnormalities, gait disturbance, and confusion.10 Patients were deemed to have impaired vision if they were registered blind or partially sighted or were unable to see better than 6/60 on a Snellen chart using glasses if appropriate. Hearing impairment was defined as the inability to follow a conversation with or without using a hearing aid. A limb was considered abnormal if there was any evidence of weakness (Medical Research Council criteria Grade 4/5 or less); neuropathy; amputation; joint abnormality excluding minor osteoarthritic changes; or any condition judged to interfere with normal gait, such as cellulitis or a deep vein thrombosis. A patient’s gait was assessed by performing the Get Up and Go Test.12 Patients were considered to be confused if they scored less than 7 of 10 on the Abbreviated Mental Test score.13 Activities of daily living, transfer, and mobility were assessed using Barthel index.14

The identified risk factors were presented to the patient in a self-reported questionnaire format, and the patient was asked to tick only the box that applies to him/her.

2.2. Comprehensibility

The time to complete the questionnaire was recorded and a comprehensibility score was also assessed. After completing the questionnaire for the first time, every patient was asked to rate the questionnaire out of 10 to assess for the comprehensibility of its items.

2.3. Statistical analysis

Data were entered and analyzed using SPSS version 15.0 (SPSS Inc., Chicago, IL, USA). A Fisher’s exact (two-sided) test or Student’s t test was used to examine group differences in categorical and continuous data, respectively. Predictors of falling were identified using logistic regression analysis, with occurrence of falls as the dependent variable. The patient was used as the unit of analysis, irrespective of the number of falls. Using a backward selection procedure, the most significant independent variables were identified, using a p value greater than 0.10 as the removal criterion.

In the logistic regression model, the predicted probability of falls was related to its covariates via the following prognostic index: B1×X1+B2×X2+B3×X3…Bk×Xk. The regression coefficient (B) of the covariate indicates an estimate of the relative magnitude of the prognostic power of a specific variable. Using the prognostic index (regression coefficient), we calculated the predicted probability of falls developing for every patient. Categories were created using clinically applicable cutoff levels and percentiles.

To obtain a simplified prediction rule, the regression coefficients of the predictive variables were rounded to the nearest number ending in 0.5 or 0.0, resulting in a weighted score; subsequently, the values for the independent predictive variables were summed. The calculated prediction scores were compared with the observed percentage of patients who experienced falls. The positive and negative predictive values were determined for several cutoff values of the prediction scores. Sensitivity was defined as the total number of fallers correctly identified as high risk. Specificity was defined as total number of nonfallers correctly defined as low risk. The total predictive accuracy was the total number of patients correctly identified expressed as a percentage. The positive predictive value was defined as the number of high-risk patients who went on to fall. The negative predictive value was the number of low-risk patients who did not fall. To evaluate the diagnostic performance of the rule and to control for overfitting,15 a receiver-operating characteristic (ROC) curve was constructed. The area under the ROC curve (AUC) values provided a measure of the overall discriminative ability of a model.

Local protocols for approval of the study were followed. A signed informed consent was obtained from all participants.

3. Results

By the end of the study, 559 fallen patients were assessed as the patients group for falls risk factors. These were 373 outpatients and 186 inpatients. The control group included 426 patients. Figure 1 is a flow chart showing the distribution of the patients and control groups included in this work. Females represented 45.7% (270/599) of the faller group and 49.3% (210/426) of the control group and the gender distribution in the two groups was not statistically different. The mean age was 73.2±14.7 and 74.5±13.5 years in faller and the control groups respectively, and no statistically significant difference was observed between the two groups. Among the outpatients group, 55.8% (208/373) lived at home, whereas 165 (44.2%) were assisted living or living in nursing homes. Twenty-three patients (11.1%) of the outpatient group were living on their own. Seven percent (26/373) of the outpatient group were living on their own.

Fig. 1

Patient recruitment flow chart.

3.1. Univariate analyses

The characteristics of patients who sustained falls and those who did not sustain any fall are compared in Table 1. In the univariate analyses, history of more than one fall in the last 12 months, slowing of the walking speed or change in the patient’s gait, history of loss of balance, poor sight, and weak hand grip were significantly more reported by the patients’ group. There was an agreement between the patient-reported limitations and outcomes of clinical assessment (p<0.001). Table 2 shows a comparison between the subgroups of faller patients (inpatients vs. outpatients).

Table 1Characteristics of the patients who sustained a fall and those who did not
Predictors of falling Patients who sustained a fall (falls group) (N=559) Patients who did not fall (control group) (N=426) p
Male gender, N (%) 279 (49.9) 216 (50.7) 0.18
Age (mean±SD) 73.2±14.7 74.5±13.5 0.16
History of previous fall(s), N (%) 559 (100) 0 (0) <0.0001
History of more than 1 fall in the last 12 months, N (%) 15 (2.7) 0 (0) <0.001
Slowing of walking speed/gait change, N (%) 458 (81.9) 68 (16) <0.001
Loss of balance, N (%) 364 (65.1) 52 (12.2) <0.001
Poor sight, N (%) 208 (55.1) 39 (9.2) <0.001
Weak hand grip, N (%) 336 (60.1) 30 (7.5) <0.001
Patient agitated/confused, N (%) 54 (9.6) 31 (7.3) 0.16
Requiring frequent toileting, N (%) 68 (12.2) 54 (12.7) 0.15
Urinary incontinence, N (%) 51 (9.1) 36 (4.5) 0.16
Hearing loss, N (%) 44 (7.9) 36 (8.5) 0.17
Polypharmacy, N (%) 362 (64.8) 293 (68.8) 0.16

SD=standard deviation.

Table 2Characteristics of outpatient and inpatients who experienced falls
Predictors of falling Inpatients (N=186) Outpatients (N=373) p
Male gender, N (%) 96 (51.6) 193 (51.7) 0.16
Age (mean±SD) 74.1±13.6 72.3±15.1 0.17
History of previous fall(s), N (%) 186 (100) 373 (100) NS
History of more than 1 fall in the last 12 months, N (%) 5 (2.6) 10 (2.7) NS
Slowing of walking speed/gait change, N (%) 157 (84.4) 300 (80.4) 0.14
Loss of balance, N (%) 123 (66.1) 240 (64.3) 0.16
Poor sight, N (%) 104 (55.9) 202 (54.2) 0.18
Weak hand grip, N (%) 113 (60.8) 221 (59.2) 0.17
Confusion, N (%) 28 (15.1) 83 (2.2) <0.01
Hearing loss, N (%) 14 (7.5) 27 (7.2) 0.18
Polypharmacy, N (%) 122 (65.6) 256 (68.9) 0.17

SD=standard deviation; NS=not significant.

3.2. Multivariate analyses and derivation of the prediction rule

In the logistic regression analysis, the independent predictive variables for falls and their predictive score are shown in Table 3, which shows also the coefficients for the simplified prediction score. A prediction score was calculated for every patient with falls (Table 4). Fall prediction scores were ranging between 0 and 6.5, with a higher score indicating a greater risk of sustaining a fall.

Table 3Independent predictive variables of falls based on results of logistic regression analysis
Variable B OR 95% CI p Points
Age 0.02 1.02 1.01–1.04 0.011 0.02/yr
History of any fall 0.4 1.5 0.8–2.7 0.18 0.5
History of more than 1 fall 2.2 9.3 3.0–28.7 0.003 2
Slowing of walking speed/change in gait 1.6 5.0 2.0–12.0 0.01 1.5
Loss of balance 1.2 3.5 1.5–7.5 0.02 1
Weak hand grip 1.2 2.1 1.1–4.4 0.04 1
Poor sight 1.1 2.8 1.1–7.6 0.03 1

95% CI=95% confidence interval; B values=regression coefficients; OR=odds ratio.

For the simplified prediction rule derived from the regression coefficient.
Table 4Sensitivity and specificity of different cut off score in predicting falls risk
Cutoff Sensitivity Specificity PPV (%)
1.5 0.818 0.806 81.4
2 0.828 0.817 82.3
2.5 0.928 0.839 82.0
3 0.938 0.849 83.9
3.5 0.962 0.860 85.9
4.0 0.962 0.860 85.7
5.5 0.960 0.862 85.6
6.0 0.960 0.860 85.4
6.5 0.960 0.862 85.2

PPV=positive predictive value.

Assessment of the observed percentage of patients who experienced falls in relation to the calculated prediction score revealed that 19 patients who had history of falls had a prediction score less than two (these 19 patients had a fall > 12 months before assessment), whereas 309/499 had scores of 3.5 or more; Table 5 shows the sensitivity and specificity of different cutoff values of the score in predicting falls risk. A set of sensitivity and specificity values was derived from a range of experimental cutoff points. As a screening tool, this score would be valuable if it has high sensitivity rather specificity. The cut off value “3.5” can be selected to be the threshold of high falls’ risk. This point has a sensitivity of 96.2% and a specificity of 86.0%. This is the point at which the proposed model can miss the least number of cases together with a low false-positive.

Table 5Form used to calculate a patient’s prediction score
Risk factor Points Total Score
1 yr increase from 60 yr old 0.02
>1 fall in the last 12 months 2
Slow walking speed/change in gait 1.5
Loss of balance 1
Poor sight 1
Weak hand grip 1

The range of possible scores is 0–6.5, with higher scores indicating a greater risk of sustaining a fall.

3.3. Discriminative ability

The discriminative ability of the logistic regression model and the prediction rule were evaluated with a ROC curve (Fig. 2). Both the logistic regression model and the prediction rule had a mean±standard deviation AUC value of 0.89±0.014. The finding that the AUC values for the logistic regression model and the prediction rule were equal indicates that the derivation of the prediction rule from the logistic regression model had not introduced a loss of discriminative ability.

Fig. 2

Receiver operating characteristic curve for the prediction model. The area under the curve value for the prediction rule model was 0.89.

3.4. Internal validation

Cross-validation was used to control for over fitting. This procedure yielded a value for the predicted probability of falls for every patient, based on results of model fitting on the other patients.10 The AUC value of the cross-validated predictions nearly equaled the mean±standard deviation AUC value of the prediction score (0.89±0.015), indicating that overfitting was not a major problem.

3.5. Applicability and comprehensibility

The questionnaire format of the prediction tool (Appendix 1) was easy to answer by all the patients. Time to complete was 1.45±0.54 minutes. A mean comprehensibility score of 9.3 was reported by the patients.

4. Discussion

The long-waited, updated guidelines published jointly by the AGS, the BGS, and the American Academy of Orthopedic Surgeon in May 201016 recommended opportunistic screening of all older persons (>65 years) by asking them about a history of falls and difficulties in gait and balance. To make this applicable in standard clinical practice, a rethink to change in our tactics was necessary aiming at developing a new self-reported approach, to screen for falls risk that would be comprehensible, easily scored, and applicable for both inpatients and outpatients. The newly developed questionnaire, FRAS, provides the best starting points for all health care professionals. Results of this study revealed that the FRAS questionnaire was a comprehensive, easy to calculate, and quick tool that can be used to inform the treating doctor, in a variety of clinical settings, about the patient’s risk of falls. Once identified, all those older persons who report high risk of falls should be referred for a formal falls risk assessment.16

Earlier studies reported that to be useful, a prediction tool should have ease and speed of completion, a small number of items (not requiring specialist assessment technology or skills), transparent, simple, and evidence-based scoring.17,18,19 Results of this study revealed that the developed FRAS meets these recommendations. The developed tool predicted the development of falls in both out- as well as inpatients using five variables that are commonly assessed during the patient’s visit. Furthermore, results of the univariate analysis presented in this work agree with those published by the AGS and BGS guideline for falls management.16 The suggested cutoff point of the FRAS scoring system showed high sensitivity of 96.2% as well as specificity of 86.0%. Guidelines on the use of diagnostic tools20,21 highlighted the importance of this element. Both negative and positive predictive values depend on the prevalence of the reference event (falls). High negative predictive value and specificity mean that the reassurance can be given about low-risk patients. This should be paralleled by high total predictive accuracy and sensitivity, rendering the use of the risk-prediction tool of value and a proper use of the staff time. In another study,22 it was reported that to be operationally useful, a falls screening tool would require a predictive accuracy of more than 80%. The positive and negative predictive values of the prediction score as well as the discriminative ability were reliable, with an AUC value of 0.89 and a value of 0.87 after internal validation correcting for overfitting. Such degree of accuracy and easy applicability of the tool make the FRAS model a step forward in achieving individualized assessment of falls risk in standard clinical practice.

The FRAT22 was developed to identify people who would benefit from further assessment of their falls risk. In contrast to the results of this study as well as most of the earlier published findings about falls risk factors, the FRAT study did not find strong evidence that visual impairment was independently associated with falling. A recent review23 that assessed 20 studies found that older people with decreased visual acuity and other visual deficits are more likely to have recurrent and injurious falls compared with fully sighted populations. The authors of the FRAT study noted limitations of the cohort studies used to identify the predictive factors in the FRAT tool and that none of the studies reported multivariable analyses using all the factors finally chosen for the FRAT tool; hence, the authors were not certain whether the FRAT tool contains the most efficient combination of factors.22

The problem with frail or acutely ill older patients admitted to hospital is that their falls risk will often vary over time and this is related to the underlying medical/surgical condition. Earlier data revealed that the risk of falls is proportional to the number of risk factors.24,25 On this premise, a considerable number of risk-assessment tools have been developed for inpatients.9,17,18,26,27,28 Despite claims of high accuracy, those tools were often found to be disappointing and mostly limited to admitted patients.16,26,27,28 No wonder then that in 2008, David Oliver29 recanted the STRATIFY score and suggested “to put them to bed.” Results of this study revealed that confusion/agitation was only prevalent among the admitted inpatients and was related to the intercurrent illness. Also, most inpatients assessed in this work had urinary catheters fitted, making the frequent toileting and nocturnal micturition remote possibilities. In contrast to earlier falls risk assessment tools, agitation, frequent toileting, and psychological status were not included in the FRAS scoring model. There was no significant difference on comparing the inpatients to the outpatient group regarding the polypharmacy factor and type of medications taken. Patients who were taking sedatives were under closer observation during their period of stay in the hospital.

Slowing of the walking speed was the second most important predisposing factor for falls in this study. This finding agrees with the latest data published by Kenny.30 Using modern information technology together with semiautomated observation, a lot of new information and insight into falls mechanisms is currently being generated. The latest publication revealed slowing of the walking speed was found to be significantly correlated to occurrence of falls.30 Another study31 supported this finding as it found that older people were 50% more likely to be hospitalized with falls or fractures if they scored poorly when tested for walking pace. Walking speed has also been correlated to CV deaths, with those in the slowest tertile three times more likely to suffer CV death over 5 years than those who walked faster.32 Walking speed can be both a subjective as well as an objective measure of physical fitness. Most of the patients are aware of changes in their walking speed. In the mean time, assessment of walking speed is relatively simple and can be performed easily in the standard clinical setting.

Although developing an instrument applicable to both inpatient and outpatient populations could be a limiting factor or at least a challenge, hospitalized individuals might have unique characteristics and patients at high risk of falls might present to different health care professionals. Also, falling by itself is a multifactorial issue, which may represent different pathological etiologies. The univariate as well as multivariate analysis identified the most significant risk factors and results of the work revealed that the questionnaire is applicable to both groups. The next step is to assess the external validity of the developed FRAS in an independent cohort of patients. Head-to-head comparison of the FRAS to other instruments is another way to assess the performance of the developed tool.

Although basic professional competence in falls assessment and prevention is required from all health care professionals dealing with patients known to be at risk of falling, it is important to highlight that falls screening tools are just a beginning and is not the end of the fall prevention process. In agreement with the recent AGS/BGS guidelines, the FRAS questionnaire provides a starting point for all health care professionals irrespective of the variables given, which can serve as an introduction or a checklist to identify those patients who are in need of further assessment and management. National Institute for Health and Clinical Excellence guidelines11 for falls assessment and management highlighted that falls management is multifactorial process. Screening tools, no matter how well they are developed, will be effective only if they are integrated into a comprehensive fall prevention program, of which they are only one aspect.

In conclusion, the FRAS is a new self-reported tool that allowed screening of older adults for falls risk. FRAS was a sensitive and specific predictor of future falls and can be recommended for standard clinical practice both in the outpatient and the acute hospital inpatient settings. Predicting the falls risk would help to minimize the negative impact of falling on the patient’s physical, psychological, and social functional abilities.

Appendix 1. Falls risk assessment questionnaire

 
Falls Risk Assessment
Over the past year (please tick whatever applied to you)
I had more than one fall
My walking speed has got slower/my gait has changed
I have lost my balance
I have problems with my sight
My grip strength got weaker

References

  1. National Patient Safety Agency. Slips trips and falls in hospitals. NPSA, London, UK; 2007 (Available at)www.npsa.nhs.uk. (Date accessed: November 27, 2010)
  2. Australian Council on Safety. Falls prevention in hospitals and residential aged care facilities. Best Practice Guidelines. Australian Council on Safety, Brisbane, Ausralia; 2005
  3. Preventing falls in institutions. in: S. Lord, C. Sherrington, H. Mens, J. Close (Eds.) Falls in Older People. 2nd ed. Cambridge University Press, Cambridge, UK; 2007 (Chapter 18)
  4. Oliver, D., Britton, M., and Seed, P. Development and evaluation of evidence based risk assessment tool (STRATIFY) to predict which elderly inpatients will fall: case-control and cohort studies. Br Med J. 1997; 315: 1049–1053
  5. Mayer, D. Risk assessment. Essential Evidence-based Medicine. Cambridge University Press, Cambridge, UK; 2004 (Chapter 13)
  6. Eiksaar, R. Assessing fall risk and preventing falls in community-living elderly. (Available at)http://www.seekwellness.com/fallprevention/assessing-fall-risk.htm. (Date accessed: November 29, 2009)
  7. Vaapio, S., Salminen, M., Vahlberg, T., Sjösten, N., Isoaho, R., Aarnio, P. et al. Effects of risk-based multifactorial fall prevention on health-related quality of life among the community-dwelling aged: a randomized controlled trial. Health Qual Life Outcomes. 2007; 5: 20–26
  8. Ory, M.G., Schechtman, K.B., Miller, P., Hadley, E.C., Fiatarone, M.A., Province, M.A…., and The FICSIT Group. Frailty and injuries in later life: the FICSIT trials. J Am Geriatr Soc. 1993; 41: 283–296
  9. Oliver, D., Daly, F., and Martin, F.C. Risk factors and risk assessment tools for falls in hospital in-patients: a systemic review. Age Ageing. 2004; 33: 122–130
  10. Guideline for the prevention of falls in older persons. American Geriatrics Society, British Geriatrics Society and American Academy of Orthopaedic Surgeons Panel on Falls Prevention. J Am Geriatr Soc. 2001; 49: 664–672
  11. Guideline development group. The assessment and prevention of falls. NICE Clinical Guidelines. 2004; 21 (Available at)http://www.nice.org.uk/nicemedia/live/10956/29583/29583.pdf. (Date accessed: April 3, 2010)
  12. Mathias, S., Nayak, U.S.L., and Isaccs, B. Balance in elderly patients. The “get up and go” test. Arch Phys Med Rehabil. 1986; 67: 387–389
  13. Qureshi, K.N. and Hodkinson, H.M. Evaluation of a ten question mental test in the institutionalized elderly. Age Ageing. 1974; 3: 152–157
  14. Collin, C., Wade, D.T., Davies, S., and Horne, V. The Barthel ADL index: a reliability study. Int Disabil Stud. 1988; 10: 61–69
  15. Harrell, F.E., Lee, K.L., and Mark, D.B. Multivariate prognostic models, evaluating assumptions and accuracy, and measuring and reducing errors. Stat Med. 1996; 15: 361–387
  16. Summary of the updated American Geriatrics Society/British Geriatrics Society clinical practice guideline for prevention of falls in older persons. J Am Geriatr Soc. 2011; 59: 148–157
  17. Cannard, G. Falling trend. Nurs Times. 1996; 92: 36–37
  18. Scott, V., Votova, K., and Scanlan, A. Multifactorial and functional mobility assessment tools for fall risk among older adults in community, home support, long-term and acute care settings. Age Ageing. 2007; 36: 130–139
  19. Wyatt, J. and Altman, D. Prognostic models: clinically useful or quickly forgotten?. BMJ. 1995; 311: 539–541
  20. Meyer, J.G. Guidelines for reporting information in studies of diagnostic test accuracy. The STARD initiative. J Pers Assess. 2003; 81: 191–193
  21. Webster, J., Courtney, M., O’Rourke, P., Marsh, N., Gale, C., Abbott, B. et al. Should elderly patients be screen for their falls risk? Validity of the STRATIFY falls screening tool and predictors of falls in a large acute hospital. Age Ageing. 2008; 37: 702–705
  22. Nandy, S., Parsons, S., Cryer, C., Underwood, M., Rashbrook, E., Carter, Y. et al. Development and preliminary examination of the predictive validity of the falls risk assessment tool (FRAT) for use in primary care. J Pub Health. 2004; 26: 138–143
  23. Legood, R., Scuffham, P., and Cryer, C. Are we blind to injuries in the visually impaired? A review of the literature. Inj Prev. 2002; 8: 155–160
  24. Tinetti, M.E., Williams, T.F., and Mayewski, R. Fall risk index for elderly patients based on number of chronic disabilities. Am J Med. 1986; 90: 429–434
  25. Morse, J.M., Morse, R.M., and Tylko, S.J. Development of a scale to indentify the fall-prone patient. Can J Ageing. 1989; 8: 366–377
  26. Winjnia, J.W., Ooms, M.E., and Van Balen, R. Validity of the STRATIFY risk score of falls in nursing homes. Prev Med. 2006; 42: 154–157
  27. Nyberg, L. and Gustafson, Y. Using the Downton index to predict those prone to falls in stroke rehabilitation. Stroke. 1996; 27: 1821–1824
  28. Smith, J., Forster, A., and Young, J. Use of the STRATIFY falls risk assessment in patients recovering from acute stroke. Age Ageing. 2006; 35: 138–143
  29. Oliver, D. Risk assessment tools for falls in hospital inpatients. Time to put them to bed?. Age Ageing. 2008; 37: 248–250
  30. Kenny, R.A. Technology research for independent living (TRIL). (Available at)http://www.trilcentre.org/falls_prevention/falls_prevention.474.html. (Date accessed: December 6, 2009)
  31. Cawthon, P., Fox, K., Gandra, S., Delmonico, M., Chiou, C., Anthony, M. et al. Do muscle mass, muscle density, strength, and physical function similarly influence risk of hospitalization in older adults?. J Am Geriatr Soc. 2009; 57: 1411–1419
  32. Dumurgier, J., Elbaz, A., Ducimetiere, P., Tavernier, B., Alpérovitch, A., and Tzourio, C. Slow walking speed and cardiovascular death in well functioning older adults: prospective cohort study. BMJ. 2009; 339: b4460

 

Fig. 1

Patient recruitment flow chart.

Fig. 2

Receiver operating characteristic curve for the prediction model. The area under the curve value for the prediction rule model was 0.89.

References

  1. National Patient Safety Agency. Slips trips and falls in hospitals. NPSA, London, UK; 2007 (Available at)www.npsa.nhs.uk. (Date accessed: November 27, 2010)

  2. Australian Council on Safety. Falls prevention in hospitals and residential aged care facilities. Best Practice Guidelines. Australian Council on Safety, Brisbane, Ausralia; 2005


  3. Preventing falls in institutions. in: S. Lord, C. Sherrington, H. Mens, J. Close (Eds.) Falls in Older People. 2nd ed. Cambridge University Press, Cambridge, UK; 2007 (Chapter 18)


  4. Oliver, D., Britton, M., and Seed, P. Development and evaluation of evidence based risk assessment tool (STRATIFY) to predict which elderly inpatients will fall: case-control and cohort studies. Br Med J. 1997; 315: 1049–1053


  5. Mayer, D. Risk assessment. Essential Evidence-based Medicine. Cambridge University Press, Cambridge, UK; 2004 (Chapter 13)


  6. Eiksaar, R. Assessing fall risk and preventing falls in community-living elderly. (Available at)http://www.seekwellness.com/fallprevention/assessing-fall-risk.htm. (Date accessed: November 29, 2009)


  7. Vaapio, S., Salminen, M., Vahlberg, T., Sjösten, N., Isoaho, R., Aarnio, P. et al. Effects of risk-based multifactorial fall prevention on health-related quality of life among the community-dwelling aged: a randomized controlled trial. Health Qual Life Outcomes. 2007; 5: 20–26


  8. Ory, M.G., Schechtman, K.B., Miller, P., Hadley, E.C., Fiatarone, M.A., Province, M.A…., and The FICSIT Group. Frailty and injuries in later life: the FICSIT trials. J Am Geriatr Soc. 1993; 41: 283–296


  9. Oliver, D., Daly, F., and Martin, F.C. Risk factors and risk assessment tools for falls in hospital in-patients: a systemic review. Age Ageing. 2004; 33: 122–130


  10. Guideline for the prevention of falls in older persons. American Geriatrics Society, British Geriatrics Society and American Academy of Orthopaedic Surgeons Panel on Falls Prevention. J Am Geriatr Soc. 2001; 49: 664–672


  11. Guideline development group. The assessment and prevention of falls. NICE Clinical Guidelines. 2004; 21 (Available at)http://www.nice.org.uk/nicemedia/live/10956/29583/29583.pdf. (Date accessed: April 3, 2010)


  12. Mathias, S., Nayak, U.S.L., and Isaccs, B. Balance in elderly patients. The “get up and go” test. Arch Phys Med Rehabil. 1986; 67: 387–389


  13. Qureshi, K.N. and Hodkinson, H.M. Evaluation of a ten question mental test in the institutionalized elderly. Age Ageing. 1974; 3: 152–157


  14. Collin, C., Wade, D.T., Davies, S., and Horne, V. The Barthel ADL index: a reliability study. Int Disabil Stud. 1988; 10: 61–69


  15. Harrell, F.E., Lee, K.L., and Mark, D.B. Multivariate prognostic models, evaluating assumptions and accuracy, and measuring and reducing errors. Stat Med. 1996; 15: 361–387


  16. Summary of the updated American Geriatrics Society/British Geriatrics Society clinical practice guideline for prevention of falls in older persons. J Am Geriatr Soc. 2011; 59: 148–157


  17. Cannard, G. Falling trend. Nurs Times. 1996; 92: 36–37


  18. Scott, V., Votova, K., and Scanlan, A. Multifactorial and functional mobility assessment tools for fall risk among older adults in community, home support, long-term and acute care settings. Age Ageing. 2007; 36: 130–139


  19. Wyatt, J. and Altman, D. Prognostic models: clinically useful or quickly forgotten?. BMJ. 1995; 311: 539–541


  20. Meyer, J.G. Guidelines for reporting information in studies of diagnostic test accuracy. The STARD initiative. J Pers Assess. 2003; 81: 191–193


  21. Webster, J., Courtney, M., O’Rourke, P., Marsh, N., Gale, C., Abbott, B. et al. Should elderly patients be screen for their falls risk? Validity of the STRATIFY falls screening tool and predictors of falls in a large acute hospital. Age Ageing. 2008; 37: 702–705


  22. Nandy, S., Parsons, S., Cryer, C., Underwood, M., Rashbrook, E., Carter, Y. et al. Development and preliminary examination of the predictive validity of the falls risk assessment tool (FRAT) for use in primary care. J Pub Health. 2004; 26: 138–143


  23. Legood, R., Scuffham, P., and Cryer, C. Are we blind to injuries in the visually impaired? A review of the literature. Inj Prev. 2002; 8: 155–160


  24. Tinetti, M.E., Williams, T.F., and Mayewski, R. Fall risk index for elderly patients based on number of chronic disabilities. Am J Med. 1986; 90: 429–434


  25. Morse, J.M., Morse, R.M., and Tylko, S.J. Development of a scale to indentify the fall-prone patient. Can J Ageing. 1989; 8: 366–377


  26. Winjnia, J.W., Ooms, M.E., and Van Balen, R. Validity of the STRATIFY risk score of falls in nursing homes. Prev Med. 2006; 42: 154–157


  27. Nyberg, L. and Gustafson, Y. Using the Downton index to predict those prone to falls in stroke rehabilitation. Stroke. 1996; 27: 1821–1824


  28. Smith, J., Forster, A., and Young, J. Use of the STRATIFY falls risk assessment in patients recovering from acute stroke. Age Ageing. 2006; 35: 138–143


  29. Oliver, D. Risk assessment tools for falls in hospital inpatients. Time to put them to bed?. Age Ageing. 2008; 37: 248–250


  30. Kenny, R.A. Technology research for independent living (TRIL). (Available at)http://www.trilcentre.org/falls_prevention/falls_prevention.474.html. (Date accessed: December 6, 2009)


  31. Cawthon, P., Fox, K., Gandra, S., Delmonico, M., Chiou, C., Anthony, M. et al. Do muscle mass, muscle density, strength, and physical function similarly influence risk of hospitalization in older adults?. J Am Geriatr Soc. 2009; 57: 1411–1419


  32. Dumurgier, J., Elbaz, A., Ducimetiere, P., Tavernier, B., Alpérovitch, A., and Tzourio, C. Slow walking speed and cardiovascular death in well functioning older adults: prospective cohort study. BMJ. 2009; 339: b4460