Identification of fall risk requires examination of the factors that increase the likelihood of falling. Factors such as mobility and balance, which can be described by several clinical assessment tools such as the Berg Balance Scale (BBS), the Balance Evaluation Systems Test (BESTest), and the Timed Up and Go (TUG), have been used to characterize fall risk in older adults.1–4 Other factors associated with fall risk, such as age and obesity (as defined by body mass index [BMI]), may also be measured.5–8 Increased age is associated with lower scores on the BBS and longer times for the TUG.4 Similarly, higher BMI is associated with lower BBS scores, suggesting an increased fall risk for those individuals.9 Adverse drug effects are a unique and separate potential contributor to fall risk, yet no comprehensive metric exists to account for their influence.
Many pharmacodynamic mechanisms contribute to drug-related fall risk. Orthostatic hypotension, for example, is an adverse effect of many drugs used in the treatment of cardiovascular diseases and has been associated with falls.10–13 Drugs used in the treatment of central nervous system (CNS) disorders have also been shown to increase fall risk through psychomotor delay and reduced cognitive function.10–14 In addition, drugs that have CNS anticholinergic effects also cause confusion, sedation, and delirium, resulting in increased fall risk.10 , 12 , 13 Other pharmacodynamic mechanisms, which may contribute to increased fall risk, include electrolyte alterations, generalized asthenia, and visual dysfunction.12 , 13
Drug-related fall risk in older adults is a concern in all health care settings and at home. The incidence of mortality after an injurious fall rises sharply after the age 65 years.15 In the United States, a review of emergency hospitalization resulting from adverse drug effects in adults 65 years and older found that falls were most commonly associated with cardiovascular and CNS drug classes.16 Combined, these 2 classes resulted in a total of 2404 emergency department cases annually, with subsequent hospital admissions of 40% and 36%, respectively. Surprisingly, the majority of these hospitalizations were not associated with inappropriate prescribing. Thus, even drugs that are appropriately prescribed contribute to elevated fall risk for older adults. The risk of falling increases as the number of drugs an individual takes increases, but only when one or more of the drugs is independently associated with falls.17 This challenges the concept that accurate information related to fall risk can be determined simply by counting the number of drugs taken by an individual (ie, drug counts).
There are several broad categories of tools designed to capture drug-associated fall risk in older adults. The first category consists of prescription guidelines which compare the clinical benefit versus risk and include Beers Criteria, Screening Tool of Alert Doctors to the Right Treatment (START) and Screening Tool of Older Person's potentially inappropriate Prescriptions (STOPP).11 , 18 These tools allow prescribers to make informed decisions based on known adverse effects. The second category consists of indexes, which quantify the adverse effects of drugs, and include the Drug Burden Index (DBI), the Anticholinergic Risk Scale (ARS), and the Anticholinergic Cognitive Burden (ACB).19–21 These quantitative assessments focus primarily on adverse anticholinergic effects, although the DBI also includes adverse sedative effects. Although both sedation and anticholinergic effects are associated with falls, there are additional previously described pharmacologic mechanisms that may also precipitate falls and are not captured by these indices.10–14 Thus, these quantitative assessments may miss drugs that contribute to fall risk.
For the purposes of this investigation, the variables considered that contribute to fall risk were age, BMI, adverse drug effects, fall history, and drug counts. This investigation examined whether these variables were explanators of mobility and balance testing scores. Second, the investigation examined whether drug adverse effects were independently associated with mobility and balance test scores. A potential clinical benefit of this approach would be to incorporate these variables into electronic medical records. Thus, individuals who have not fallen yet but who may be at risk of falling could receive more comprehensive balance and mobility testing by physical therapists.
This study was approved by the institutional review boards at the University of Maryland, Temple University, and East Tennessee State University. Informed consent was obtained from all individual participants included in this study, and individual rights of all participants were protected. Participants were recruited for a balance training clinical trial (ClinicalTrials.gov #366151-1) via advertisements in a newspaper for older adults, and via flyers placed at local retirement communities from 2012 to 2015. Results reported here are a secondary analysis of these data and do not reflect results from the randomized clinical trial. Each participant completed a series of forms providing medications and whether or not they experienced symptoms of pain or dizziness or had a history of falling. A total of 61 participants were evaluated at the initial time point for this part of the investigation. All participants were, by self-report, free from uncontrolled or unstable cardiovascular disease. All participants reported problems with their balance. Eligibility criteria included passing the Mini-Mental State Examination (MMSE) with a score of 24 or higher,22 and the ability to walk independently on a treadmill at a self-selected speed for 2 minutes. These inclusion criteria were designed for the larger clinical trial, but were also applied to this secondary analysis. Four participants were excluded (2 were unable to walk independently on the treadmill at a self-selected pace, 1 with an MMSE score of less than 24, and the drug list for 1 participant was not recorded). Data from 57 subjects participating in the original clinical trial were analyzed for this investigation.
Setting and Outcome Measures
All participants underwent the following series of tests. The MMSE was conducted to determine eligibility. Balance ability and mobility were characterized with the following standard clinical assessments: BESTest,1 BBS,2 TUG, and cognitive dual task Timed Up and Go (TUGc).3 Each of the balance and mobility outcome measures was included as part of the larger clinical trial, and is included here to demonstrate consistency across multiple different measures of balance. In addition to the battery of balance and walking tests, all participants reported the number of falls they experienced during the previous 12 months. Participant characteristics of age, weight, height, and gender were recorded, and BMI was calculated. Finally, the Activities-Specific Balance Confidence (ABC) scale was completed to characterize individuals' confidence in their ability to perform daily activities without losing their balance. The BMI was not calculated for 1 participant as the weight was not taken and 1 participant refused to complete the TUGc.
The mobility and balance tests were selected to capture a range of standing and walking balance ability. The BESTest is a 27-item physical performance test, and each item is scored on a 4-point ordinal scale; the final score is reported as a percentage with a maximum of 100.1 The BESTest has excellent test reliability for individuals with a wide variety of balance difficulties and scores less than 69 discriminate fallers from nonfallers among individuals with Parkinson's disease.1 , 23 The BBS is a 14-item scale (each item scored on a 5-point ordinal scale) designed to identify balance deficits in older adults, with a maximum score of 56 and excellent test-retest reliability.2 Scores on the BBS scale below 46 are associated with an increased fall risk.24 The TUG and TUGc are single- and dual-task versions of a screening test designed to identify fall risk with high test-retest reliability.3 , 4 Individuals who require more than 13.5 seconds to complete the TUG or 14.5 seconds to complete the TUGc are considered to have increased risk for falling.3 The ABC is a self-report questionnaire that measures an individual's perceived balance ability and has been related to functional balance in older adults.25 The ABC has both high test-retest reliability and association with several other balance and mobility tests,25 , 26 and scores below 67% are associated with an increased fall risk.24
Development and Scoring of the Quantitative Drug Index
Drug lists for participants were obtained by recording the drugs that were brought in at the initial evaluation. Alternatively, some participants provided a written list of the drugs they were currently taking. All of the following were recorded on the drug list spreadsheet: prescribed and over-the-counter drugs, vitamins, and supplements. From each participant's drug list, the adverse effects for each drug that could impact fall risk were identified based on pharmacodynamic effects. Drug adverse effects were compiled from entries in LexiComp (Indianapolis, Indiana), and then the impact of each adverse effect on balance or falls was determined by the authors. A list of the adverse effects for this investigation are included in Table 1. Both prescription and over-the-counter drugs were evaluated for fall risk. Herbal and vitamin supplements were also evaluated if information was found in LexiComp. The quantitative drug index (QDI) was calculated as follows: (1) A drug was first identified as having adverse effect(s) related to fall risk. (2) All of the fall-associated adverse effects for that drug were summed to provide a score for that drug. (3) Then the total score for each participant was calculated by summing the individual scores for each of the drugs the participant was currently taking. For example, if drug A is associated with sedation and hypotension and drug B is associated with diplopia, ataxia, and myalgia then drug A would have a score of 2 and drug B would have a score of 3. If an individual is taking both drugs A and B, then the individual's QDI would be 5. A drug with no adverse effects related to falls would have a score of 0. Drug dosages were not included as part of the QDI calculation. Similarly, the potential for drug-drug interactions to alter, decrease or increase, concentrations of other drugs was not incorporated into the scale. Drug counts were calculated by counting the number of drugs an individual was taking, regardless of whether the drug was considered to have an adverse effect related to falls.
Data Collection and Analysis
Data analyses were completed using IBM SPSS Statistics 22 (IBM Inc, Armonk, New York).27 , 28 A multiple linear regression analysis was conducted with BESTest total score as the dependent variable and the age, BMI, fall history, QDI and drug counts as independent explanatory variables. A stepwise forward regression method was used in the model with the entry and exit criteria for the independent variables in the model set at 0.05 and 0.10, respectively.27 Age, BMI, and QDI were selected by the step forward regression model, and were subsequently used in a forced entry model to repeat the analysis, with test scores for the ABC scale, BBS, TUG, and TUGc as the dependent variables. For each regression analysis, actual residuals were plotted against predicted residuals to determine normality of distribution about the regression function, using a scatter plot and frequency histogram. Durbin-Watson test scores for autocorrelation of residuals were reported for all regression analyses as were the variance inflation factors for age, BMI, and QDI. Multiple regression functions were assessed 2-tailed post hoc for power when sample size, number of predictor variables (k = 3), coefficient of determination (R 2), and error probability (α= 0.05) were known, using the software program G*Power (Version 126.96.36.199.2, 2014).29 , 30 Subsequently, scores for the ABC scale, mobility and balance tests, MMSE, drug counts, age, and BMI were divided into 2 groups based on the QDI scores. The low-impact drug group (LIDG) was defined as a QDI score equal to 0, and the high-impact drug group (HIDG) was defined as a QDI score greater than 0. Several variables were not normally distributed (the Appendix, Item 1); therefore, nonparametric Mann-Whitney U tests for independent samples were performed to determine group differences between the LIDG and HIDG groups. Significance was set at P ≤ .05 for all analyses.
The average age of the participants was 79 (range 66-92) years and 72% (41) were female. A multiple regression analysis was conducted with age, BMI, drug counts, fall history, and QDI as independent predictor variables and BESTest total score as the outcome variable. The final model included age, BMI, and QDI as the predictors for the BESTest total score (Table 2). Subsequently, these same 3 independent variables were examined to determine their explanatory potential when ABC scale, BBS, TUG, and TUGc were the dependent outcome variables. In all analyses, the regression models for explaining these mobility and balance test scores were significant (P < .001) (Table 2). Individually within the multiple regression models, each of the 3 variables significantly (P < .05) contributed to the prediction of the outcome scores for each mobility and balance test. Age, BMI, and QDI were all negatively associated with the BESTest and BBS. An increase in any of the explanatory variables would result in lower scores for these tests, suggesting higher fall risk. Age, BMI, and QDI were all positively associated with TUG and TUGc scores. TUG and TUGc times increased as age, BMI, and QDI scores increased, again suggesting elevated fall risk. The relative independent contribution of age, BMI, and QDI toward explanatory potential for the mobility and balance test scores is shown by the standardized beta values in each multiple regression analysis (Table 2). In general, age demonstrated the greatest explanatory contribution for the mobility and balance test scores. BMI had a greater effect than the QDI for BESTest, BBS, and TUGc. In contrast, the QDI had a greater effect than BMI for the TUG.
The multiple regression function for the ABC scale resulted in several differences compared with the mobility and balance tests examined earlier. The overall multiple regression model with ABC scale as the dependent variable was significant when age, BMI, and QDI were the independent explanatory variables (Table 2). However, only BMI and QDI individually were significant within the model and BMI was the strongest explanatory of balance confidence. BMI and QDI were negatively associated with balance confidence, suggesting an increase in either of these variables would lead to a decrease the ABC scale score and increased fall risk.
The coefficients of determination from these multiple regression analyses for the 4 mobility and balance tests varied between 0.31 for the BESTest and 0.37 for the BBS. The ABC scale coefficient of determination was 0.29. Post hoc power analysis on the coefficients of determination demonstrated power for the balance and mobility tests as BESTest (0.99), BBS (0.99), TUG (0.99), and TUGc (0.99). Power analysis for the ABC scale was 0.99. All of these values represent a large effect size for the power analysis. The Durbin-Watson H statistic for BESTest and BBS demonstrated significantly more positive autocorrelation of the residuals with scores less than 1, compared with the ABC scale, TUG, and TUGc values, which were between 1.57 and 1.69. The variance inflation factors for age, BMI, and QDI in all regression analyses were approximately 1, suggesting minimal multicollinearity of these predictors. Residuals from the regression analyses were normally distributed (the Appendix, Item 2). Forty-four percent of the participants reported no falls within the previous 12 months, and a total of 74% reported 1 or less falls. This frequency would make fall history a poor outcome variable in a linear regression analysis (the Appendix, Item 3).
The participants in the HIDG demonstrated significantly lower scores on the BESTest (P = .02) and BBS (P = .02), compared with participants in the LIDG (Table 3). Individuals in the HIDG were also significantly slower on the TUG (P = .01) and TUGc (P = .04) compared with individuals in the LIDG. As expected, the HIDG had significantly higher drug counts compared with the drug counts of the LIDG (P < .001). In addition, the HIDG scored marginally, but not statistically worse (P = .11) on the ABC scale compared with LIDG. No significant differences were found between the HIDG and LIDG for reported fall history, age, BMI, or MMSE scores.
This investigation demonstrates that outcome scores of 4 commonly used mobility and balance tests may all be partially explained by age, BMI, and adverse drug effects (QDIs). Both age and BMI are readily available (or easily calculated) from electronic records. The information for calculating the QDI is also electronically available, and thus may be incorporated into electronic medical records. The purpose of this study was to determine whether these 3 independent variables could be used to assist in preliminary screening of fall risk, a determination that would then require additional mobility and balance testing. Each of the mobility or balance tests reported in this study is used to screen or more deeply probe balance ability and fall risk. There should be no surprise that as scores become worse on one test other tests may also be similarly impacted. That age, BMI, and QDI were found to be significant explanators for all 4 mobility and balance tests are due to the strong associations between these mobility and balance tests.9 , 23 , 26 These observations are even more surprising considering the low percentage of reported recurrent fallers, defined as 2 or more falls in the past year, among the participants in this sample. Only 26% of the 57 subjects in this investigation self-reported more than 1 fall in the previous year. Post hoc power analysis also demonstrated large effects for these coefficients of variation and sample sizes.
In addition, division of the participants into LIDG and HIDGs, based on the QDI scores, resulted in significant differences in the mobility and balance scores between these 2 groups. Even more interesting was that the median scores on the mobility and balance tests for the HIDG were within the normal ranges for these 4 tests.4 , 23 These results suggest that adverse drug effects may influence mobility and balance ability, even when the scores for these tests are within ranges that otherwise were not associated with elevated fall risk.
Both greater age and higher BMI contribute to increased fall risk. Age is associated with lower mobility and balance outcome scores. Both the BBS and the TUG demonstrated age-related declines in test scores for participants between the ages of 61 and 89 years.4 Age was also used as a covariate adjustment in predicting differences in fall risk for TUG under traditional, cognitive, and manual conditions.3 The current results also support the negative influence of increasing age on BBS, TUG, and TUGc scores, and are the first to identify a similar negative influence on overall balance as measured by the BESTest. In contrast to the negative age-related influence on mobility and balance scores, the influence of BMI on balance and mobility is less well defined. This lack of association is surprising as several factors that are positively associated with the severity of BMI have also been linked to fall risk, and may act as mediators between the 2.6 These BMI-related variables mediating the elevated fall risk include chronic moderate to extreme pain, sedentary lifestyle, cardiac disease, diabetes mellitus, sedative and hypnotic drugs, and anxiety or depression and the drugs to treat these latter 2 disorders. BMI was previously reported to have a negative association with scores on the BBS, but not the TUG.9 The influence of BMI on BESTest scores has not been previously reported. The current investigation identified a negative influence of BMI on scores for all 4 mobility and balance tests such that higher BMI was associated with worse performance. The differing results for the influence of BMI on TUG outcomes, between the current and previous investigation, will require further evaluation.
This study represents the initial presentation of the QDI. There may be some overlap in identified drug-related adverse effects between the current QDI and previously published drug indices. This is not surprising, as a subset of the adverse effects identified in the development of the QDI are also represented in other drug indices. However, there are several important differences between the QDI used in the current investigation and the ACB, ARS, and DBI. The ACB identified a list of drugs with documented anticholinergic effects. These drugs were then scored on a scale of 1 to 3, with 3 being the most severe, based on their reported clinical or laboratory documentation.20 Similarly the ARS examined 500 of the most prescribed drugs for anticholinergic activity, and scored these drugs on a 0 to 3 scale, with 3 being the most severe. Again, the scoring for the drugs was based on their observed laboratory effects or clinical anticholinergic activities.21 A comparison of these 2 indices demonstrates that both the ACB and ARS capture decreased cognitive and subjective activities of daily living scores.31 The DBI examines the potential fall risk for drugs that have either anticholinergic activity or sedating potential, with higher scores on the DBI related to lower functional capabilities.19 Thus, the DBI expands the list of drugs that may contribute to falls risk compared with the ACB and ARS, but is still limited to only 2 pharmacologic mechanisms of action. In contrast, the QDI examines all potential adverse drug effects that may increase the risk of fall, regardless of the drug's principal mechanism(s) of action. As such, the QDI encompasses a larger number of drugs in more classes than previous indices. Another advantage the QDI has over the DBI, ARS, and ACB is that there is no upper limit on the number of adverse effects related to fall risk for any drug on the QDI. The scores for each individual's drugs are summed, and this sum represents the QDI score for that individual at that time. Capturing a larger number of the pharmacologic adverse effects, which contribute to falls, and creating an index score with no upper limit would theoretically increase the sensitivity of the QDI compared with the previously discussed quantitative indices.
A separate but related issue to determining which drug-associated adverse effects may be influencing fall risk is how to weight these fall risk-associated adverse effects. To date, all previously published investigations, which used drug-associated adverse effects to determine fall risk, have weighted these adverse effects equally.19–21 , 31 , 32 In the present investigation, the QDI continued the convention of equal weighting for drug-associated adverse effects. Part of the limitation to weighting adverse effects may be the mechanism(s) of the weighting. That is, should the weighting be based on the severity of the adverse effect? Such a weighting would require an agreement among researchers in the field as to which adverse effects are associated with falls and a rank ordering of the severity for these adverse effects. Alternatively, the weighting for adverse effects could be based on the frequency of the occurrence for these in the general population medicated with these drugs. There may be additional alternative weighting schemes for these adverse effects not recognized in this discussion as well. Overall, weighting of adverse effects may or may not increase the sensitivity of any drug-associated fall risk score. Future studies are needed to investigate whether unequal weighting of the drug adverse effects improves the sensitivity of the drug-associated fall risk score.
In addition, both the drug dosage and the potential for drug-drug interactions may influence the drug-associated fall risk. To date, there is no consistency about incorporating dosage into the calculations of published drug-associated fall risk scoring. Both the DBI and the ARS include dosage in their calculations for drug-associated fall risk.19 , 32 In contrast, neither the QDI in the present investigation nor the ACB scale incorporated drug dosage into the drug-associated fall risk scoring.20 The fact that drug dosage was not incorporated into either the QDI, in the present investigation, or the ACB scale, in the previous investigation, suggests that addition of the drug dosage may increase the accuracy of calculating drug-associated fall risk, but may not be essential. The presence or absence of drug dosage in calculating fall risk is extremely important as not all records will list drug dosages. Having a drug-associated fall risk scale, such as the QDI, which reflects fall risk without inclusion of dosage, may lend itself to wider clinical use. The potential for drugs to alter the pharmacokinetics of other drugs, increasing or decreasing their concentrations in the body, is a recognized clinical problem influencing both clinical efficacy and adverse effects. To date, neither the QDI nor any of the previously published investigations have included drug-drug interactions in their calculation for drug-associated fall risk. The absence of drug-drug interactions in these calculations may be related to the complexity of developing a scale that includes these interactions, yet such interactions should be explored in future research.
Several previous investigations have examined the adverse effects of drugs on functional activity. Gnjidic and colleagues33 examined adverse drug effects on 5 functional tests in older community-dwelling Australian men using the DBI. The investigation utilized time to complete 5 chair stands, gait speed over a 6-m distance, narrow gait speed, balance, grip force, and a subjective activities of daily living scale. High DBI scores were associated with slower gait speeds in normal and narrow walking conditions, balance difficulties, decreased grip force, and lower scores in the activities of daily living scale. A second investigation also found that participants with higher DBI scores took longer to complete the TUG and 5 chair stands, had decreased grip force, slower gait speed, and lower activity of daily living scores.34 In the present investigation, the participant cohort was divided into LIDG and HIDG based on QDI scores. The results of the present investigation are consistent with the previous investigation, in that the HIDG scored worse on all 4 mobility and balance tests. Furthermore, this statistical difference occurred even when the median scores for the HIDG were within the low fall risk range of scores for these mobility and balance tests.3 , 23 , 24 These results suggest that drug effects on fall risk may begin when global fall risk as identified by those same tests may be subclinical. A separate investigation used the fall risk-increasing drugs (FRIDs) list to examine fall risk in older participants.35 The FRIDs is a list of prescription drugs without quantitation and similar to Beers List. The FRIDs lists a wider category of drugs associated with fall risk and includes anticholinergics, those with sedating properties, and drugs that decrease blood pressure. The spectrum of drugs assumed to be associated with falls in the FRIDs list investigation more closely parallels the drugs of the QDI in the present investigation. The TUG times were recorded at baseline and approximately 6 months following removal or reduced dosing of drugs from the FRIDs list. After modifying the dosage of these drugs, there was a significant reduction in TUG times (better performance) compared with the participants' baseline attempts. One limitation of the investigation with the FRIDs list was a lack of quantitation or rating of these fall risk-associated drugs. Thus, current and previous investigations demonstrate a direct relationship between drug profiles and mobility and balance test scores, ultimately influencing fall risk. In this study, we classified a larger group of drugs listed as being associated with fall risk, and compared the influence of the drug-derived QDI on multiple mobility and balance test scores within a single cohort of participants. This provides support that drugs with specific adverse effects have an independent influence on overall balance and fall risk.
Balance confidence as measured by the ABC scale was not associated with age, but was found to be associated with BMI and QDI. These results are surprising given that the ABC scale scores have strong associations with the other mobility and balance test scores examined.1 , 24 , 26 Future research is needed to determine the mechanisms through which balance confidence is influenced by the adverse effects of drugs. Potential mechanisms may include, but not be limited to, altered alertness or decreased mental cognition. The marginal difference for the ABC scale between the LDIG and HDIG may be biased because 74% of this cohort did not report falls in the past year. Despite enrolling only individuals with a history of falling or a self-identified balance problem, many individuals did not score at elevated fall risk on the BESTest, BBS, TUG, or TUGc. This likely reflects a high functioning cohort concerned about their balance, but who largely are still very confident in their abilities. Future research on populations with greater balance impairments is needed to clarify the discordant results between balance confidence and balance ability reported here.
Drug counts would be expected to demonstrate similar discrimination to the QDI in the multiple regression functions; however, this similarity may be accounted by the fact that the adverse effects of drugs are what account for the fall risk, not just the number of drugs.17 This demonstrates that the QDI is at least as powerful in the prediction of drug-related fall risk as drug counts, the current gold standard. Future investigations need to specifically address the interactions between drug dosage, drug-drug interactions, and weighting of adverse effects.
An alternative interpretation of the drug-associated fall risk scales/indexes is that they represent the severity of the physical disabilities in individuals with balance difficulties. The treatment of dysfunction often results in the prescription of drugs, which are both clinically efficacious and minimize adverse effects compared with other drug alternatives. When these first-line drugs are no longer efficacious, less optimal drugs are often prescribed. These latter drugs are equally efficacious, or sometimes have greater efficacy, but almost always have a higher severity or frequency of adverse effects (or entirely different and more severe adverse effects). Thus, the QDI may simply be a surrogate for measuring the severity of the physical dysfunction for these individuals. To date, there has been no discussion in the literature of this alternative explanation for these scale/indexes, and the present investigation was not structured to answer this question. Future research is needed to address which if any physical dysfunctions are associated with these fall risk scales/indexes.
The major limitation of the present investigation is that the underlying mechanisms by which age, BMI, and adverse drug effects influence balance, mobility, and potential fall risk were not delineated. Such an analysis would require a larger sample of participants. In addition, there may be independent explanatory variables beyond the 3 identified in the multiple regression modeling for this investigation. Only 31% to 37% of the variations in the 4 mobility and balance test scores, and 29% for the ABC scale, were accounted for in these models. Additional independent variables may increase the explanatory potential for the multiple regression analyses. However, age, BMI, and QDI were chosen as they would all be potentially available in electronic medical records and could provide a preliminary screen for fall risk.
The present investigation demonstrated that age, BMI, and QDI independently explained the outcome scores for BESTest, BBS, TUG, and TUGc. Whereas age is not a modifiable fall risk, both body morphology and drug adverse effects are, allowing health care professionals an avenue to reduce fall risk. In addition, these variables are either readily available in electronic medical records or could be easily calculated. Thus, these variables may provide an initial screen, which could identify individuals who are at fall risk, but not yet fallen, and trigger a referral for more comprehensive balance evaluation.
1. Horak FB, Wrisley DM, Frank J. The balance
evaluation systems test (BESTest) to differentiate balance
deficits. Phys Ther. 2009;89(5):484–498.
2. Conradsson M, Lundin-Olsson L, Lindelof N, et al Berg balance
scale: intrarater test-retest reliability among older people dependent in activities of daily living and living in residential care facilities. Phys Ther. 2007;87(9):1155–1163.
3. Shumway-Cook A, Brauer S, Woollacott M. Predicting the probability for falls in community-dwelling older adults using the Timed Up & Go Test. Phys Ther. 2000;80(9):896–903.
4. Steffen TM, Hacker TA, Mollinger L. Age- and gender-related test performance in community-dwelling elderly people: six-minute walk test, Berg balance
scale, timed up & go test, and gait speeds. Phys Ther. 2002;82(2):128–137.
5. Iwasaki S, Yamasoba T. Dizziness and imbalance in the elderly: age-related decline in the vestibular system. Aging Dis. 2015;6(1):38–47.
6. Mitchell RJ, Lord SR, Harvey LA, Close JC. Obesity and falls in older people: mediating effects of disease, sedentary behavior, mood, pain and medication use. Arch Gerontol Geriatr. 2015;60(1):52–58.
7. Barin K, Dodson EE. Dizziness in the elderly. Otolaryngol Clin North Am. 2011;44(2):437–454, x.
8. Kelsey JL, Berry SD, Procter-Gray E, et al Indoor and outdoor falls in older adults are different: the maintenance of balance
, independent living, intellect, and zest in the Elderly of Boston Study. J Am Geriatr Soc. 2010;58(11):2135–2141.
9. Benavent-Caballer V, Sendin-Magdalena A, Lison JF, et al Physical factors underlying the timed “Up and Go” test in older adults. Geriatr Nurs. 2016;37(2):122–127.
10. Hayes BD, Klein-Schwartz W, Barrueto F Jr. Polypharmacy and the geriatric
patient. Clin Geriatr Med. 2007;23(2):371–390, vii.
11. Society AG. American Geriatrics Society updated Beers Criteria for potentially inappropriate medication use in older adults. J Am Geriatr Soc. 2012;60(4):616–631.
12. Katzung B, Masters S, Trevor A, eds. Basic and Clinical Pharmacology. 12th ed. New York, NY: McGraw-Hill Co; 2012.
13. Panus P, Katzung B, Jobst EE, Tinsley S, Masters S, Trevor A, eds. Pharmacology for the Physical Therapist. 1st ed. New York, NY: McGraw-Hill Co; 2009.
14. Chang CM, Chen MJ, Tsai CY, et al Medical conditions and medications as risk factors of falls in the inpatient older people: a case-control study. Int J Geriatr Psychiatry. 2011;26(6):602–607.
15. Alamgir H, Muazzam S, Nasrullah M. Unintentional falls mortality among elderly in the United States: time for action. Injury. 2012;43(12):2065–2071.
16. Budnitz DS, Lovegrove MC, Shehab N, Richards CL. Emergency hospitalizations for adverse drug events in older Americans. N Engl J Med. 2011;365(21):2002–2012.
17. Ziere G, Dieleman JP, Hofman A, Pols HA, van der Cammen TJ, Stricker BH. Polypharmacy and falls in the middle age and elderly population. Br J Clin Pharmacol. 2006;61(2):218–223.
18. Hill-Taylor B, Sketris I, Hayden J, Byrne S, O'Sullivan D, Christie R. Application of the STOPP/START criteria: a systematic review of the prevalence of potentially inappropriate prescribing in older adults, and evidence of clinical, humanistic and economic impact. J Clin Pharm Ther. 2013;38(5):360–372.
19. Hilmer SN, Mager DE, Simonsick EM, et al A drug burden index to define the functional burden of medications in older people. Arch Intern Med. 2007;167(8):781–787.
20. Boustani M, Campbell N, Munger S, Maidment I, Fox C. Impact of anticholinergics on the aging brain: a review an practical application. Aging Health. 2008;4(3):311–320.
21. Rudolph JL, Salow MJ, Angelini MC, McGlinchey RE. The anticholinergic risk scale and anticholinergic adverse effects
in older persons. Arch Intern Med. 2008;168(5):508–513.
22. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state.” A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12(3):189–198.
23. Leddy AL, Crowner BE, Earhart GM. Functional gait assessment and balance
evaluation system test: reliability, validity, sensitivity, and specificity for identifying individuals with Parkinson disease who fall. Phys Ther. 2011;91(1):102–113.
24. Lajoie Y, Gallagher SP. Predicting falls within the elderly community: comparison of postural sway, reaction time, the Berg balance
scale and the activities-specific balance
confidence (ABC) scale for comparing fallers and non-fallers. Arch Gerontol Geriatr. 2004;38(1):11–26.
25. Powell LE, Myers AM. The Activities-Specific Balance
Confidence (ABC) Scale. J Gerontol A Biol Sci Med Sci. 1995;50A(1):M28–34.
26. O'Hoski S, Sibley KM, Brooks D, Beauchamp MK. Construct validity of the BESTest, mini-BESTest and brief BESTest in adults aged 50 years and older. Gait Posture. 2015;42(3):301–305.
27. Field A. Discovering Statistics Using SPSS. 3rd ed. London, England: Sage; 2009.
28. Portney LG, Watkins MP. Foundations of Clinical Research: Applications to Practice. 3rd ed. Norwalk, CT: Appleton & Lange; 2009.
29. Cohen J. A power primer. Psychol Bull. 1992;112(1):155–159.
30. Faul F, Erdfelder E, Buchner A, Lang AG. Statistical power analyses using G*Power 3.1: tests for correlation and regression analyses. Behav Res Methods. 2009;41(4):1149–1160.
31. Pasina L, Djade CD, Lucca U, et al Association of anticholinergic burden with cognitive and functional status in a cohort of hospitalized elderly: comparison of the anticholinergic cognitive burden scale and anticholinergic risk scale: results from the REPOSI study. Drugs
32. Lowry E, Woodman RJ, Soiza RL, Mangoni AA. Associations between the anticholinergic risk scale score and physical function: potential implications for adverse outcomes in older hospitalized patients. J Am Med Dir Assoc. 2011;12(8):565–572.
33. Gnjidic D, Cumming RG, Le Couteur DG, et al Drug burden index and physical function in older Australian men. Br J Clin Pharmacol. 2009;68(1):97–105.
34. Gnjidic D, Bell JS, Hilmer SN, Lonnroos E, Sulkava R, Hartikainen S. Drug burden index associated with function in community-dwelling older people in Finland: a cross-sectional study. Ann Med. 2012;44(5):458–467.
35. van der Velde N, Stricker BH, Pols HA, van der Cammen TJ. Withdrawal of fall-risk-increasing drugs
in older persons: effect on mobility
test outcomes. Drugs
1. Shown are the frequency histograms for the dependent variables: drug counts, age, BMI, falls, MMSE, BESTest total score, Berg Balance Scale, ABC score, 6-minute walk, TUG and TUGc. Based on these histograms, the authors conclude the majority of gait and balance tests, age, drug counts, falls, and ABC score for this sample population were nonparametric, thus a nonparametric Mann-Whitney 2-group comparison was conducted.
2. Comparison of actual and predicted variables for all balance and mobility tests. The authors examined the residuals for normality of the regression function. The scatterplots and frequency histograms for the predicted compared with the actual residuals are included in the figures for the reviewers as frequency distributions and scatterplots.
Although the scatterplots and histograms have some variation from absolute normal distribution, the authors believe the results are acceptable in part due to the small sample size of 57.
3. Forty-four percent of the participants had no falls and a total of 74% had 1 or less falls. The histogram distribution of fall history is shown here. The histogram shows that falls in the last 12 months do not have a normal distribution and would make a poor dependent variable in a multiple regression function.