Farrell, Mary K. PT, DSc; Rutt, Richard A. PT, PhD; Lusardi, Michelle M. PT, DPT, PhD; Williams, Ann K. PT, PhD
Individuals with dementia or Alzheimer disease (AD) exhibit both a progressive decline in cognitive function and an accumulating difficulty with physical function.1 As cognitive impairment begins, individuals will initially demonstrate difficulties with short-term memory and with instrumental activities of daily living (ADL). As the dementia progresses, individuals will increasingly demonstrate difficulties with long-term memory, communication, and basic ADL.1
In addition to the progressive decline in function, decreases in postural control are evident in early stages of AD. In a comparison of community-living subjects with and without mild AD who were independent in gait, Franssen et al2 and Petterson et al3 found that those with AD demonstrated significantly lower scores on balance measures. While the balance scores for those with mild AD did not indicate a risk for falls, they did suggest the beginnings of a decline in postural control. Chong et al4 found that individuals with AD (mean Mini-Mental State Examination [MMSE] score of 19 ± 6) had more difficulty than those without cognitive impairment in maintaining balance under conditions that required suppression of incongruent or distracting visual and somatosenory information and attention to vestibular input. Similarly, Manckoundia et al5 found that the addition of a cognitive task and answering questions about a recently viewed videotape significantly altered quiet standing center of pressure sway in those with AD. Compared with age-matched healthy older adults, those with AD (mean MMSE of 21 ± 2) had significantly greater increases in the total area of sway and the path of the sway as a result of the cognitive task. These increases in sway are indicative of less-effective postural control when attention must be divided.
Decreases in postural control place people with dementia at an increased risk for falling. The rate of falls for those with dementia is approximately twice the rate of falls for those without dementia.6 Tinetti et al7 followed 336 community-dwelling adults older than 75 years for 1 year and found that 50% of those with cognitive impairment experienced a fall, while only 30% of those without cognitive impairment experienced a fall. In individuals living in nursing homes, van Doorn et al8 found that people with dementia were twice as likely to fall as people without dementia. Following newly admitted residents over 2 years, they established a rate of falls of 4.05 per person year in those with dementia compared with a rate of falls of 2.33 per person year in those without dementia. Risk of injury has also been found to be higher in individuals with dementia compared with those without dementia.8,9
The presence of cognitive decline and dementia have, in fact, been identified as independent risk factors for falls and for injurious falls in older adults living in the community.7,10 However, within a population of people with dementia, the relationship between degree of cognitive decline and falls is less well understood. In 124 outpatients with probable AD and mean MMSE score of 16.2 ± 5.9 followed over 1 year for fall occurrence, Horikawa et al11 found no significant difference in MMSE scores between fallers and nonfallers. Similarly, Bassiony et al12 found no association between fall status and MMSE scores in community-dwelling older adults with a mean age of 77 ± 8 years and a mean MMSE score of 15 ± 7. In people with dementia living in geriatric settings in Sweden, Kallin et al13 did find a relationship between the level of cognitive function and falls; however, it was not a linear relationship. They found that those with the highest and lowest levels of cognitive function were at lower risk for falling and those with intermediate levels of cognitive function were at the greatest risk for falling. At this time, knowing the degree of cognitive impairment in someone with dementia is not sufficient to understand risk of falling.
Risk of falls in older adults is multifactorial. Known risk factors include older age, history of a fall, and decreased function and postural control. Fall incidence increases with increasing age in the general population.14 Older age has also been found to be independently associated with fall occurrence in people with AD.12 Previous occurrence of a fall is a strong predictor of a subsequent fall in older adults without dementia and has been found to be associated with falls in those with dementia.13,15
Assessment of physical performance is a common way of determining postural control and is used to determine risk of falls in older adults. While there are a number of performance-based measures available to assess postural control and risk of falls in older individuals without cognitive impairment, few measures have been evaluated for use in individuals with dementia. Given the decline in function and increased risk of falls and fall-related injury associated with dementia, a need exists for standardized tools that assess these domains to be validated in people with dementia. To confidently identify individuals with dementia who are at the risk of falling, an assessment tool needs to be evaluated for its diagnostic accuracy within this specific population.
One barrier to using assessment tools in people with dementia is the concern that these individuals will not be able to follow the standardized instructions in a manner that will allow for accurate scoring of the assessment tool. In the few studies that have examined functional performance in subjects with dementia, it appears that the extent of cognitive involvement, complexity of testing items, and novelty of testing items influence the ability of subjects to complete the tests and the level of reliability that is established.16–19
The Physical Performance Test (PPT) is a performance-based measure that has good potential for use in people with dementia. The PPT incorporates multiple physical domains into the assessment, including ADL, gross motor activities, fine motor control, balance, and walking.20 Because items on the PPT include commonly performed functional skills, individuals with dementia may be more likely to successfully follow directions and complete the tasks. While other assessment tools (eg, Berg Balance Scale and Dynamic Gait Index) present greater challenges to postural control and are more commonly used clinically, they also include items that are novel and more complex.21,22 The concern with using one of these tools for this study population was the potential decreased likelihood that subjects with dementia would be able to successfully complete all testing items.
The validity of the PPT to assess functional performance, measure change, and identify fall risk among older adults has been established.23–27 The ability of the PPT to assess ADL function and fall risk gives it the potential to be a very useful tool in the assessment of people with dementia. In the 7-item version of the PPT, the clinician records the time it takes for the patient to write a sentence, spoon beans from a bowl into a coffee can (simulated eating), lift a book onto a shelf above shoulder height, put on and remove a jacket, pick up a penny from the floor, and walk 50 feet. One item, turning 360°, is not timed. It is rated either 0 or 2 for continuity of steps and for steadiness.20 Scores for the 7-item test range from 0 to 28, with higher scores indicating better performance.
Reliability of the PPT has been established in subjects without known cognitive deficits.20,28,29 The PPT allows for the assessment of older adults with a broad range of functional abilities.30 Subjects less frequently achieve ceiling scores on the PPT when compared to self-report measures.30 Scores on the PPT are associated with significantly different frequencies of medical diagnoses, somatic symptoms, medications taken, and number of comorbidities.23,25,26,30
The PPT has also been found to have predictive validity. In community-dwelling older adults, lower scores on the PPT were a significant predictor of death or institutionalization 18–24 months later.31 The PPT has also been identified as an independent predictor of recurrent falls in older adults without specified dementia. In a study of 84 frail community-dwelling veterans, VanSwearingen et al24 assessed multiple tests for their ability to identify recurrent fallers. Recurrent faller status was determined through a structured interview of subjects. Those who reported 2 or more falls in the previous year were identified as recurrent fallers. Of the 84 subjects, 53 were found to be recurrent fallers. The 7-item PPT was identified as an independent factor in predicting fall risk. A cutoff score of 15 of a possible 28 resulted in a sensitivity of 79.3% and a specificity of 71% in identifying those who fell repeatedly. Delbaere et al32 studied multiple intrinsic risk factors in an attempt to construct a risk model to identify frequent fallers. A total of 263 older adults with a mean age of 72 years and no diagnosis of dementia were followed for 1 year, and 33.5% of the subjects fell at least once. A cutoff score less than 19 for the 7-item PPT or less than 25 for the 9-item PPT was found to significantly increase the odds of the person being a frequent faller compared with being a nonfaller by 4 times.
With the differences in postural control and the rate of falls in people with dementia compared with older adults without dementia, it is unclear whether the PPT will be a good predictor of falls in this population. Given the potential benefits of the PPT for individuals with dementia, it is necessary to establish the validity of the PPT in this population. The purpose of the study was to determine the accuracy of the PPT in identifying people with dementia who will experience a fall.
Individuals with a diagnosis of dementia were recruited from the greater Portland, Oregon, area. Subjects were included in the study if they had mild to moderate dementia, as identified by scores between 10 and 24 on the MMSE and if they had a care provider available for testing and signing the informed consent form.33,34 Subjects were excluded from the study if (1) they had a diagnosis of stroke or Parkinson disease or other medical condition that would interfere with testing, (2) they required physical assistance for walking, or (3) they did not speak English.
Subjects gave verbal assent, and a legal guardian signed an informed consent form approved by the institutional review boards of Pacific University and Rocky Mountain University of Health Professions. A single tester, a physical therapist with previous experience administering the PPT, tested each subject at his or her residence. A questionnaire regarding medical history and information pertinent to inclusion and exclusion criteria was administered to both the subject and the caregiver. As a part of this questionnaire, subject and caregiver pairs were asked to report the number of falls that the subject had experienced in the previous 6 months. The MMSE was then administered to the subject followed by the 7-item PPT. While the tester was blind to the occurrence of future falls, she was not blind to the subject's history of a fall.
After testing, each subject and caregiver pair was instructed in the follow-up procedures. They were given a calendar to post on the refrigerator and instructions to mark on the calendar any falls that occurred during the follow-up period. The definition of a fall was written on the calendar and verbally explained to the subject and caregiver. A fall was defined as “unintentionally coming to rest on the ground, floor or other lower level” and excludes “coming to rest against furniture, wall, or other structure.”35(p301) To collect data on fall occurrences and to act as reminder to mark falls on the calendar, a phone call to each residence was made by the primary tester every 2 weeks. Data on fall occurrences were collected over 4 months. This time frame was chosen in an attempt to be long enough to capture fall occurrences but not so long that the progression of the disease process could be a confounding factor. Progression of cognitive decline has been estimated at 0.24 MMSE points per month, suggesting that after 4 months, subjects in this study would, on an average, have an MMSE score that was 1 point lower.36
During the administration of the PPT, before performance, subjects received verbal instructions as well as demonstration of each item. Anticipating that subjects with dementia might have difficulty with short-term memory and with following instructions, testing protocol was modified to include verbal cueing during testing. If a subject appeared to lose focus on a task while performing it, verbal cues were given for redirection. If the subject exceeded the allotted time to complete a task, and it was determined by the examiner to be the result of a distraction, the subject was asked to redo that item. Subjects also redid any item if they had difficulty following instructions the first time. Determination of when to allow a second attempt was left to the primary tester's clinical judgment. For example, subjects who stopped performance to begin a conversation or to ask a question were allowed to redo the item. No item was completed more than 2 times. Use of assistive device was permitted for all standing activities. In this study, reliability of the PPT and tester is established and results are presented in Table 1. Description of the procedure to establish reliability is presented elsewhere.37
A priori power analysis for regression (α = .05, 3 variables, and assumed variance of 0.30) determined minimal sample size to be 31 subjects.38 There was no power analysis for logistic regression, therefore this formula was used as a close approximation to help to estimate sample size. The 3 variables that entered into the equation were age, history of a fall, and the PPT score. The assumed variance of 0.30 was estimated since falls were multifactorial and it seemed unlikely for any 1 measure to account for a majority of the shared variance. All statistics were determined by using SPSS statistical analysis software version 15 (Copyright 2006, SPSS Inc, Chicago, Illinois).
Descriptive statistics (mean and standard deviation) were calculated for the variables of age, the MMSE score, the time since diagnosis of dementia, previous falls (history of no fall or 1 or more falls in the previous 6 months), and the number of falls experienced in the follow-up period. A stepwise logistic regression was used to predict the dependent variable, fall status during the 4-month follow-up period. Fall status was dichotomized as either no fall or at least 1 fall. Variables that entered into the equation included age, history of a fall, and the PPT score. Age and history of a fall were included as they have been previously identified as risk factors for falls in people with dementia.12–15 Criteria for entry into and removal from the regression equation were α of .05 to enter and .10 to remove. Sensitivity and specificity for the predictor variables were then calculated. To determine the cutoff score, a receiver operating characteristic curve was plotted to identify a score that represented the best combination of sensitivity and specificity for those variables. Likelihood ratios were then calculated.
Thirty-four subjects with dementia, 19 women and 15 men, participated in the study. Information on their age, the MMSE scores, and time since diagnosis are presented in Table 2. Twelve subjects (35%) reported a history of a fall in the previous 6 months. Twenty-one lived at home, 6 lived in senior housing, 6 lived in either assisted living or adult foster care facilities, and 1 lived in a specialized Alzheimer unit. Nine subjects (27%) used an assistive device for ambulation: 6 using a cane occasionally when outdoors or when feeling unsteady and 3 using a wheeled walker for all ambulation.
All subjects completed the 4-month follow-up and 12 (35%) reported at least 1 fall during that period. Of these 12, 4 reported more than 1 fall. From the 3 variables entered into the stepwise logistic regression (history of a fall, the PPT score, and age), only history of a fall in the previous 6 months was a significant predictor of a fall in the 4-month follow-up period (P = .044). A previous fall increased the odds of having a subsequent fall by almost 5 times (odds ratio [OR] = 4.76, with 95% confidence interval 95% CI, 1.04–21.77). The addition of neither age nor the PPT score significantly changed the predictive ability. Table 3 shows the 2 × 2 table and calculated sensitivity, specificity, and likelihood ratios for history of a fall.
If the PPT score was forced into the logistic regression equation as a lone variable, it was not a significant predictor (P = .081). A receiver operating characteristic curve for the PPT score was created, and the area under curve was 0.674, with a 95% CI, 0.478–0.870. A cutoff score of 19 yielded the best combination of sensitivity and specificity. Table 4 shows the 2 × 2 table for the PPT score by using 19 as the cutoff score. Calculated sensitivity and specificity were 83% and 41%, respectively, and positive and negative likelihood ratios were 1.41 and 0.41, respectively.
Of the 13 subjects who were identified as a faller by their PPT score but did not experience a fall in the 4-month follow-up period (false positives), 4 had experienced a fall in the previous 6 months and 3 others had reported a fall before the 6-month period, with 2 of them experiencing a hip fracture and 1 having a significant change in condition during the follow-up, resulting in 24-hour hospice care.
The purpose of this study was to determine the accuracy of the PPT in identifying fallers within a population of people with dementia. As a part of this study, we also included other potential risk factors for fall: age and history of a fall. The results suggested that the report of a fall in the previous 6 months was the strongest predictor of a fall in the subsequent 4 months. A history of a fall in the previous 6 months increased the odds of a fall in the subsequent 4 months by 5 times. This outcome is supported by other research. Gerdhem et al15 had similar results in a population of older woman where, compared to performance-based measures, occurrence of a fall in the previous year was a stronger independent predictor of a fall in the subsequent year, with an OR of 2.91 and 95% CI, 2.09–4.06). Kallin et al13 also found that, in 2008 older adults with cognitive impairment living in geriatric settings in Sweden, those with history of a fall during their stay at the facility had greater odds of having a fall over a 1-week period, with an OR of 2.78 and 95% CI, 2.09–3.69).
While history of a fall was the strongest predictor of a subsequent fall, there were 5 subjects with no history of a fall in the previous 6 months who did experience a fall in the 4-month follow-up period. Of these fallers, 2 had had a fall before the 6-month period and 1 had reported experiencing occasional stumbles without a fall. Their PPT scores ranged from 9 to 23, the MMSE scores ranged from 15 to 21, and age ranged from 72 to 87 years. None of them used an assistive device. Two of the subjects experienced falls that were not witnessed, and therefore, the circumstances were not clear. One subject fell outside in the rain, tripping over a parking curb; 1 fell off the end of the bed while sitting, and 1 fell 2 times, once from bumping into the edge of a table and once while carrying a laundry basket. On the basis of visual inspection of these data, it is possible that the use of a fall history greater than 6 months, such as the 1-year history used by Gerdhem et al,15 may improve the predictive ability of history of a fall.
Over the 4-month follow-up, 35% of the subjects in this study experienced at least 1 fall. Within this population, it took only 4 months to approach the incidence of falls that occurred over the course of 1 year in similarly aged community-dwelling older adults without cognitive impairment.7,39 The apparent higher rate of falls in our group is consistent with the higher rate of falls in people with dementia identified in other studies. The adjusted OR for a fall in those with cognitive impairment, calculated by Tinetti et al,7 was 5 (95% CI, 1.8–13.7). Similarly, when Asada et al9 followed community-dwelling subjects over the course of 6 months, they found the rate of falls to be 3 times greater in those with dementia.
Our results regarding the predictive ability of the PPT differed from those of VanSwearingen et al24 and Delbaere et al.32 In frail older men with unknown cognitive status, VanSwearingen et al24 found the PPT to be an independent factor for predicting falls. By using a cutoff score of 15 or less, they found sensitivity of 79%, specificity of 71%, a positive likelihood ratio of 2.73, and a negative likelihood ratio of 0.29. Delbaere et al32 found that, in a group of 263 older adults with a mean age of 72 years and no diagnosis of dementia, a cutoff score of less than 19 for the 7-item PPT or less than 25 for the 9-item PPT increased the odds of the person being a frequent faller compared with being a nonfaller by 4 times.
Differences in study design may account for some of the differences in outcomes. In the current study, subjects were followed prospectively for 4 months. VanSwearingen et al24 collected fall data retrospectively over 6 months and Delbaere et al32 collected fall data prospectively over 1 year. It is possible that the shorter time frame used to collect fall occurrences in the current study was not sufficient to capture all potential fall occurrences. Another key difference in the current study was the definition used to identify a faller. The other 2 studies included only those with 2 or more falls in the faller group, while we chose to include those with 1 or more falls in the faller group and did not have a multiple-faller category. Lastly, difference in cognitive function between subjects in our study and subjects in the studies by VanSwearingen et al24 and Delbaere et al32 might also have been a contributing factor to the differences in outcomes.
One other potential threat to the validity of the outcomes was the use of self-report to identify fall occurrences. While attempts were made to ensure the accuracy of self-reported fall occurrences, it is possible that the data collected were not entirely accurate. Self-report and/or caregiver report of a fall in the previous 6 months may be inaccurate as a result of either not recalling a fall occurrence or not considering a fall to be an event worth reporting. Strategies to ameliorate this included discussing the definition of a fall; emphasizing that even if no injury occurred, it would still be considered a fall; and telling the subject and caregiver pair which month it was 6 months ago and listing the 6 months under consideration. For collection of follow-up fall occurrence data, subjects were given a verbal and written definition of a fall and were contacted every 2 weeks to gather fall data. The phone call provided a shorter recall time frame and acted as a reminder to keep track of falls. While there is no way to know for sure whether the fall occurrences reported were accurate, we believe that the strategies employed resulted in data that were as precise as possible.
On the basis of the results of this study, the PPT scores by themselves have limited usefulness in predicting falls over the course of 4 months in individuals with dementia. Contributing factors for falls in people with dementia are numerous and not fully understood.6 While components of physical function, such as gait and postural control, are known factors, many nonphysical function contributors, such as medication use and the presence of orthostatic hypotension, also contribute to fall risk.6 The fact that the PPT does not incorporate these components may help to explain why it did not have high diagnostic accuracy in predicting falls. Instead, the results of the PPT might best be used as a piece of the decision-making process. The use of likelihood ratios can assist in this process. A clinician can generate a probability for falls on the basis of factors like medication use and history of a fall and then use the results of the PPT to generate a posttest probability. For example, if an individual with dementia had a pretest probability of falling in the next 4 months of 35% (generated from our incidence data) and scored more than 19 (a negative test score) on the PPT, the probability of them falling in the next 4 months would drop to about 18%. If the individual had a positive test result (<19), then the posttest probability of falling in the next 4 months would increase to about 42%. As noted in this example, a negative test result may be more clinically useful in altering our perception of a person's likelihood of falling in the next 4 months than a positive test score does. However, caution must be taken in interpreting the likelihood ratios. As shown in Table 4, for both the positive and negative likelihood ratios, the 95% CI crossed 1. Since the 95% CI indicates the range in which the real likelihood ratio exists, there is a possibility that the likelihood ratios are not clinically useful.
Individual items within the PPT may also have contributed to the limitations in its diagnostic accuracy. In administering the PPT within subjects' residence, 1 item, placing a book on a shelf above shoulder height, was difficult to standardize. The instructions provided by Reuben and Siu20 specify only that the shelf be above shoulder height. Within each residence, it was possible to find a shelf that met this criteria; however, the distance above shoulder height differed. Subjects with higher shelves may require more time to complete the task, which potentially can have artificially lowered scores for that item. Standardizing a height for the shelf, relative to the subject, could improve this problem; however, it would also make it more difficult to find a shelf within an individual's residence that met the criteria.
One other potential factor limiting the diagnostic accuracy of the PPT is its inclusion of 2 items that do not have an apparent postural control demand to them. The first 2 items, writing a sentence and simulated eating, are upper-extremity tasks completed in a seated position (all other items were performed in standing). It is therefore possible that some subjects scored lower on the PPT because of poor performance on 1 of the first 2 items. The lower score could have inaccurately identified them as a faller. Writing a sentence did, in fact, prove to be a challenge for all 34 subjects. None of the subjects completed the item fast enough to score a 4; 10 subjects scored a 3, 5 scored a 2, 13 scored a 1, and 6 scored a 0. Their slower times indicate difficulty with the task of writing but not necessarily difficulty with postural control. It is possible that the PPT would have greater diagnostic accuracy for falls by eliminating the first 2 items or by replacing them with items with greater postural control demands. Brown et al26,27 has presented a modified version of the PPT, where those 2 items are replaced by items related to postural control: repeated chair stands (time to rise 5 times from a 16″ chair) and progressive Rhomberg (ability to maintain balance with feet together, in semitandem, and in full tandem).26,27
In our testing protocol, subjects were allowed to use their assistive device for the items done in standing (items 3–7). In the original study by Reuben and Siu,20 assistive device use was allowed for conditions 6 and 7 (turn 360° and 50′ walk). It is possible that allowing subjects to use assistive devices for additional items 3–5 (placing book on a shelf, putting on and removing jacket, and picking a penny up from floor) may have altered the time to complete the tasks or the ability of the subjects to successfully complete the tasks. While this could have inflated the scores of the 3 subjects who used a walker during testing, it most likely affected only 1 of the subject's scores, as the other 2 scored 0s on these items.
One final potential factor that may have influenced the diagnostic accuracy of the PPT was the smaller sample size. Before data collection, we calculated a power analysis for regression and estimated that a sample size of 31 subjects would be required. As a part of this equation, we made an estimation of a shared variance of 0.30. This was an estimation based on the assumption that, because falls had multifactorial causes, any 1 factor would account for only a small amount of the variance associated with falls. The purpose of this calculation was to estimate sample size. It is possible that the estimation of variance was too high, and as a result, sample size was underestimated. If this was the case, then any potential association between the PPT scores and falls would be missed.38 VanSwearingen et al24 and Delbaere et al,32 who found the PPT score to be a factor associated with falls, had larger sample sizes with 84 and 263 subjects, respectively.
On the basis of the results of this study, we cannot currently recommend the use of the PPT for identifying fallers; however, further research is warranted. It is possible that the PPT is a better indicator of postural control than of fall prediction over a 4-month period. Of the 13 false positives (Table 4), 7 had other strong risk factors for falling and 1 had a significant change in condition. It is possible that the PPT was accurately assessing postural control despite the fact that those individuals did not fall.
This study was the first to look at the diagnostic accuracy of a performance-based measure of function within a population of people with dementia. While subjects were able to complete the PPT, the outcomes were insufficient to draw a conclusion on its diagnostic accuracy for fall risk. On the basis of the results of this study, report of a fall in the previous 6 months is the strongest predictor of a fall in the subsequent 4 months in people with dementia. Neither age nor the PPT scores were independent factors for predicting falls. Further research is indicated. Considerations for future research on the predictive validity of the PPT in people with dementia should include several factors. For identifying fallers, a larger sample size and longer follow-up time should be implemented. In addition, it may be appropriate to use the modified version suggested by Brown et al.26,27 In all instances, previous history of a fall should be a part of the data collected for subjects with dementia.
The authors thank the Oregon Alzheimer's Disease Center NIH AG08017 for its contribution. This research was supported by a Pacific University Faculty Development grant. The research was presented as a platform presentation at the American Physical Therapy Association's Combined Sections Meeting, Las Vegas, February 2009.
1. Reisberg B, Ferris SH, de Leon MJ, Crook T The global deterioration scale for assessment of primary degenerative dementia. Am J Psychiatry. 1982;139:1136–1139.
2. Franssen EH, Souren LEM, Torossian CL Equilibrium and limb coordination in mild cognitive impairment and mild Alzheimer's disease. J Am Geriatr Soc. 1999;47(4):463–469.
3. Petterson AF, Engart M, Wahlund LO Activity level and balance in subjects with mild Alzheimer's disease. Dement Geriatr Cogn Disord. 2002;13:213–216.
4. Chong RKY, Horak FB, Frank J, Kaye J Sensory organization for balance: specific deficits in Alzheimer's but not Parkinson's disease. J Gerontol A Biol Sci Med Sci. 1999;54(3):M122–M128.
5. Manckoundia P, Pfitzenmeyer P, d'Athis P, Dubost V, Mourey F Impact of cognitive task on the posture of elderly subjects with Alzheimer's disease compared to healthy elderly subjects. Mov Disord. 2006;21(2):236–273.
6. Shaw FE Falls in cognitive impairment and dementia. Clin Geriatr Med. 2002;18:159–173.
7. Tinetti ME, Speechley M, Ginter SF Risk factors for falls among elderly persons living in the community. N Engl J Med. 1988;319(26):1701–1707.
8. van Doorn C, Gruber-Baldini AL, Zimmerman S, et al. Dementia is a risk factor for falls and fall injuries among nursing home residents. J Am Geriatr Soc. 2003;51:1213–1218.
9. Asada T, Kariya T, Kinoshita T, et al. Predictors of fall-related injuries among community-dwelling elderly people with dementia. Age Ageing. 1996;25:22–28.
10. King MB, Tinetti ME Falls in community-dwelling older persons. J Am Geriatr Soc. 1995;43:1146–1154.
11. Horikawa E, Matsui T, Arai H Risk of falls in Alzheimer's disease: a prospective study. Intern Med. 2005;44(7):717–721.
12. Bassiony MM, Rosenblatt A, Baker A, et al. Falls and age in patients with Alzheimer's disease. J Nerv Mental Dis. 2004;192(8):570–572.
13. Kallin K, Gustafson Y, Sandman P-O, Karlsson S Factors associated with falls among older, cognitively impaired people in geriatric care settings: a population-based study. Am J Geriatr Psychiatry. 2005;13(6):501–509.
14. Stevens JA, Mack KA, Paulozzi LJ, Ballesteros MF Self-reported falls and fall-related injuries among persons >65 years—United States, 2006. Morb Mortal Wkly Rep. 2008;57(9):225–229.
15. Gerdhem P, Ringsberg KAM, Akesson K, Obrant KJ Clinical history and biologic age predicted falls better than objective functional tests. J Clin Epidemiol. 2005;58:226–232.
16. Alexander NB, Mollo JM, Giordani B, et al. Maintenance of balance, gait patterns, and obstacle clearance in Alzheimer's disease. Neurology. 1995;45(5):908–914.
17. Logsdon RG, Gibbons LE, McCurry SM, Teri L Assessing quality of life in older adults with cognitive impairment. Psychosom Med. 2002;64(3):510–519.
18. Tappen RM, Roach KE, Buchner D, Barry C, Edelstein J. Reliability of physical performance measures in nursing home residents with Alzheimer's disease. J Gerontol A biol Sci Med Sci. 1997;52(1):M52–M55.
19. Thomas VS, Hageman PA A preliminary study on the reliability of physical performance measures in older day-care clients with dementia. Int Psychogeriatr. 2002;14(1):17–23.
20. Reuben DB, Siu AL An objective measure of physical function of elderly outpatients: the physical performance test. J Am Geriatr Soc. 1990;38:1105–1112.
21. Berg KO, Wood-Dauphinee S, Williams JI, Maki B Measuring balance in the elderly: validation of an instrument. Can J Public Health. 1992;83:S7–S11.
22. Whitney SL, Hudak MK, Marchetti GF The dynamic gait index relates to self-reported fall history in individuals with vestibular dysfunction. J Vestib Res. 2000;10:99–105.
23. Rozzini R, Frisoni GB, Bianchetti A, Zanetti O, Trabucchi M Physical performance test and activities of daily living scales in the assessment of health status in elderly people. J Am Geriatr Soc. 1993;41(10):1109–1113.
24. VanSwearingen JM, Paschal KA, Bonino P, Chen TW Assessing recurrent fall risk of community-dwelling, frail older veterans using specific tests of mobility and the physical performance test. J Gerontol A biol Sci Med Sci. 1998;53(6):M457–M464.
25. Brach JS, VanSwearingen JM Physical impairment and disability: relationship to performance of activities of daily living in community-dwelling older men. Phys Ther. 2002;82(8):752–761.
26. Brown M, Sinacore DR, Binder EF, Kohrt WM Physical and performance measures for the identification of mild to moderate frailty. J Gerontol A biol Sci Med Sci. 2000;55(6):M350–M355.
27. Brown M, Sinacore DR, Ehsani AA, Binder EF, Holloszy JO, Kohrt WM Low-intensity exercise as a modifier of physical frailty in older adults. Arch Phys Med Rehabil. 2000;81:960–965.
28. King MB, Judge JO, Whipple R, Wolfson L Reliability and responsiveness of two physical performance measures examined in the context of a functional training intervention. Phys Ther. 2000;80(1):8–16.
29. Paschal KA, Oswald AR, Siegmund RW, Siegmund SE, Threlkeld AJ Test-retest reliability of the physical performance test for persons with Parkinson's disease. J Geriatr Phys Ther. 2006;29(3):82–86.
30. Brach JS, VanSwearingen JM, Newman AB, Kriska AM Identifying early decline of physical function in community-dwelling older women: performance-based and self-report measures. Phys Ther. 2002;82(4):320–328.
31. Reuben DB, Siu AL, Kimpau S The predictive validity of self-report and performance-based measures of function and health. J Gerontol A biol Sci Med Sci. 1992;47(4):M106–M110.
32. Delbaere K, Van den Noortgate N, Bourgois J, Vanderstraeten G, Tine W, Cambier D The physical performance test as a predictor of frequent fallers: a prospective community-based cohort study. Clin Rehabil. 2006;20:83–90.
33. Folstein MF, Folstein SE, McHugh PF Mini mental state. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12:189–198.
34. Larson EB, Shadlen MF, McCormick WC, et al. Survival after initial diagnosis of Alzheimer disease Ann Intern Med. 2004;140(7):501–509.
35. Buchner DM, Hornbrook MC, Kutner NG, et al. Development of the common data base for the FICSIT trials. J Am Geriatr Soc. 1993;41:297–308.
36. Roselli F, Tartaglione B, Federico F, Lepore V, Defazio G, Livrea P Rate of MMSE score change in Alzheimer's disease: influence of education and vascular risk factors. Clin Neurol Neurosurg. 2009;111(4):327–330.
37. Farrell MK, Rutt RA, Lusardi MM, Williams AK Reliability of the physical performance test in people with dementia. Phys Occup Ther Geriatr. 2010;28(10):144–152.
38. Porteny LG, Watkins MP Foundations of Clinical Research: Applications to Practice. 2nd ed. Upper Saddle River, NJ: Prentice-Hall, Inc; 2000.
39. Boulgarides LK, McGinty SM, Willett JA, Barnes CW Use of clinical and impairment-based tests to predict falls by community-dwelling older adults. Phys Ther. 2003;83(4):328–339.
accidental falls; dementia; physical performance test