student

Journal of Cemnlology: MEDtCAL SCtENCES 2002, Vol. 57A. No. 8, M504-M510

Copyriglu 2002 hy The Gcromohsical Society of America

Risk Factors for Falling Among Community-Based Seniors Using Home Care Services

Paula C. Fletcheri and John P. Hirdes^

'Department of Kinesiology & Physical Education, Wilfrid Laurier University, Waterloo, Ontario, Canada. ^Department of Health Studies & Gerontology, University of Waterloo, and Canadian Collaborating Centre—interRAI,

Homewood Research Institute, Guelph, Ontario, Canada.

Background. Despite the plethora of information concerning risk factors for falls, limited research efforts have fo- cused on the issue of the differences in risk factors for falls based on fall status, or more specifically one-time versus chronic/recurrent fallers. Given that multiple falls have been found to be associated with negative outcomes, such as an increased risk of institutionalization, more research in this area is warranted.

Methods. The purpose of this investigation was to determine the risk factors for nonfallers versus fallers (1 + falls), and for nonfallers/one-time fallers versus recurrent fallers (2+ falls). All participants {N = 2304) in this study were re- ceiving home care services from 10 community-based agencies (Community Care Access Centres) in Ontario, Canada. The Minimum Data Set-Home Care (MDS-HC) is an assessment instrument that covers several key domains, such as service use, function, health, and social support. Nurses trained to administer the MDS-HC assessed each of the partici- pants within their homes.

Results. Of the 2304 participants in the study, 27% fell one or more times, and 10% experienced multiple falls (2 + falls). In the two final logistic regression models for risk of falling (0 falls vs 1 + falls) and multiple falling (0 falls/1 fall vs 2 + falls), the independent variables that remained significant included gender, gait, environmental hazards, and the Changes in Health, End Stage Disease and Signs and Symptoms of Medical Problems Scale. Also significant in the model for multiple falls was the Cognitive Performance Scale, Parkinson's disease, and perceived health status.

Conctusions. Overall, distinguishing individuals into different fall status classifications is important from a clinical perspective, as it is the recurrent faller who would benefit to the greatest extent from fall prevention efforts and from the negative outcomes associated with multiple falls (i.e., mortality). One of the most significant barriers in determining risk factors for falls is the lack of consistency in the variables/tools used in the research. As such, utilizing a standardized tool, such as the MDS-HC, would assist researchers in making comparisons between different settings.

FALLING results in substantial disability, morbidity, andmortality among seniors. For example, falling is the leading cause of injury admissions to acute care hospitals and in-hospital deaths (1,2). Each year, approximately one third of seniors experience a fall, (3-5), although the major- ity of falls do not lead to serious injury, hospitalization, or death (6,7). However, many elderly individuals experience complications, including restricted activity, soft tissue inju- ries, or fractures (8-11) as a direct result of the fall event.

The literature consistendy identifies multiple risk factors for falling among the community-based elderly population, including having a history of previous falls (4,12), being functionally impaired (3,7,11,13), being of advanced age (3,4,13), being female (4,14-16), using various medications or multiple medications (4,17-22), having specific condi- tions, diseases, or physiological limitations (3-5,23,24) and comorbidity (16), having cognitive impairments (4,16,25), having factors contributing to postural instability and gait impairments (7,16,26,27), performing activities such as bed transfers (3), climbing stairs (28,29), and night urination (24,30), and environmental influences or engaging in rou- tine activities (i.e., walking on stairs) (13,3f ,32).

Despite the volume of information concerning these risk factors, few studies have dealt with the issue of the differ- ences in risk factors for falls based on fall status, or more

specifically one-time versus chronic/recurrent fallers. Nevitt and colleagues (25) suggested that risk factors for one-time fallers appeared to be less robust than for chronic fallers. Further, single falls were generally less predictable and may have been the result of an accident (e.g., environmental haz- ard) or an overwhelming incident (e.g., myocardial infarc- tion), whereas multiple falls may have been more indicative of intrinsic factors (e.g., physiological predisposition to fall- ing, chronic disease, physiological disability) (25,33). After completing a one-year prospective study to determine the risk factors for falling using a sample of community-based seniors, Tinetti and colleagues (5) concluded that the risk factors for multiple fallers as compared to one-time fallers were the same; however, the magnitude of the associations were stronger for recurrent fallers. Use of sedatives, cogni- tive impairments, lower-extremity disability, palmomental reflex, i'oot problems, and number of balance-and-gait ab- normalities were significant predictors of falls, and the risk of falling increased linearly with the number of these risk factors present (5).

Graafmans and colleagues (34) found that mobility im- pairments and dizziness were associated with falls (1 + i'alls) and recurrent falls (2+ falls); however, history of stroke, poor mental state, and postural hypotension were as- sociated with being a recurrent faller only. Lord and col-

M504

RISK FACTORS FOR FALLING M505

leagues (27) found five factors that significantly distin- guished recurrent fallers from nonfallers/one-time fallers: sway, proprioception in the lower limbs, visual contrast sen- sitivity, quadriceps strength, and reaction time. However, this analysis was restricted to physiological factors associ- ated with falls. Other studies of different patterns of falls fo- cused only on certain types of variables (e.g., balance [35] or medications 119|) or failed to examine multivariate mod- els (36). Therefore, they failed to provide comprehensive information about risk factors for recurrent fallers from a multifactorial perspective.

Based on information to date, one-time fallers and multi- ple/recurrent fallers appear to represent two distinct groups. For example, it would appear that single falls are often chance events that may not be modifiable through interven- tion, whereas multiple falls would seem to be characteristic of a group of older and more frail seniors with a greater number of comorbid conditions or physiological impair- ments. Given that multiple falls have been found to be asso- ciated with negative outcomes, such as an increased risk of institutionalization (37), more research in this area is war- ranted. This investigation aims to determine the risk factors for nonfallers versus fallers (1-t- falls) and for nonfallers/ one-time fallers versus recurrent fallers (2+ falls) using data from a comprehensive assessment completed by home care professionals.

METHODS

Subjects and Data Collection Measure All participants (A' = 2304) in this study were receiving

home care services from 10 community-based agencies (Community Care Access Centres [CCACs]) in Ontario, Canada. Each of the CCACs utilized the home care version of the Resident Assessment Instrument on a pilot basis. The Resident Assessment Instrument-Home Care (RAI-HC) is a comprehensive and standardized assessment tool used to evaluate the needs and ability levels of older adults utilizing home care services. The RAI-HC consists of two core ele- ments: the Minimum Data Set-Home Care (MDS-HC) and the Clinical Assessment Protocols (CAPs). The MDS-HC is the screening portion of the instrument, which serves as a brief assessment instrument covering several key domains, such as service use, function, health, and social support (Ta- ble 1). In addition, the MDS-HC identifies individuals who

Table I. Domain Areas Assessed in the MDS-HC

Table 2. Clinical As.sessment Protocols (CAPs) Triggered by the MDS-HC

Demographics Rclerral Cognilion Coniiminicalion Vision Mood and behavior Social t'unclioning Inlorniai support ADLs and IADLs Continence

Disease diagnoses Health conditions Preventive health measures Nutrition/hydration Dental .status Skin condition Environmental assessment Service utilization Medications

Functional Performance ADL rehabilitation potential

Health promotion IADLs Institutional risk

Sensory Performance Communication disorders Visual function

Health Problems/Syndromes Cardiorespiratory Falls Oral health

Skin and foot conditions Dehydration Nutrition Pressure ulcers Pain

Continence

Bowel management Urinary incontinence and

indwelling catheters Service Oversight

Adherence Medication tiianagement Preventive health care measures Reduction in formal .services Brittle support system Palliative care Psychotropic drugs Environmental assessment

Note: MDS-HC = Minimum Data Set-Home Care; ADLs = activities of daily living; IADLs = instrumental activities of daily living.

Notes: MDS-HC = Minimum Data Set-Home Care; ADL = activity of daily living; IADLs = instrumental activities of daily living.

may benefit from more extensive evaluation and care plan- ning through 30 problem-oriented Clinical Assessment Pro- tocols or CAPs (Table 2) (38,39). Prior to data collection, the nurses were trained to administer the MDS-HC and sub- sequently assessed each of the participants in his or her home.

The MDS-HC was developed by investigators from Can- ada, France, Italy, Japan, Netherlands, Switzerland, the United Kingdom, and the United States. The reliability and validity of the instrument were established through a five- country study that included Canada and the United States (39). Studies of the use of the MDS-HC for preventive home screening are currently underway in Canada and in 10 European countries. In short, the MDS-HC provides a com- prehensive assessment of the full range of client needs, and it directly supports the development of person-specific care- plans. The MDS-HC takes approximately 45 to 60 minutes to complete.

The independent variables representing the risk factors for falls have been grouped into the following sections: (i) sociodemographic and social relationship variables (age, gender, marital status, education, living arrangements, change in social activities, amount of time alone during the day); (ii) measures of frailty (various chronic diseases, per- ceived health status, cognition or Cognitive Performance Scale [CPS] or Changes in Health, End Stage Disease and Signs and Symptoms of Medical Problems [CHESS] Scale); and (iii) exposure to risk variables (gait, environmental haz- ards, various medications). The Cognitive Performance Scale, a measure that describes cognitive status, is based on four items from the MDS-HC: short-term memory, cogni- tive decision making, making self understood, and depen- dent eating. The CHESS Scale, which is indicative of the degree of frailty or medical instability, utilizes a combina- tion of the following items dealing with changes in health (activities of daily living and cognitive decline), end-stage disease, and signs and symptoms of medical conditions (i.e..

M506 FLETCHER AND HIRDES

edema, shortness of breath, weight loss, dehydration, loss of appetite, diarrhea, vomiting). Scores range Ürom 0 (indicat- ing no instability) to a high of 5.

The dependent variable was fall status. Specifically, indi- viduals were asked whether they had fallen in the past 90 days. The outcome variable was dichotomized in two ways, as two analyses were completed: (i) 0 falls versus 1 or more falls and (ii) 0 falls/1 fall versus 2 or more falls.

Data Analysis Logistic regression was utilized to analyze the data with

fall status as the dependent variable and other MDS-HC variables as the independent variables. Stepwise methods were not used in the logistic regression analyses. Rather, a variety of models were examined in order to rule out order effects prior to specification of the final model. Only the in- dependent measures found to be significant at the bivariate level {p < .05) were examined in multivariate models. The final logistic regression models were used to estimate the adjusted odds ratios for the main and interactive effects for the measures investigated.

Table 3. Percentage (Frequency) Distributions of Sociodemographic Variables and Social Relationship Variables

Utilizing the MDS-HC

Variable , •

Age

65-69 years of age 70-74 years of age 75-79 years of age 80-84 years of age 85 years of age and older

Gender Women Men

Marital Status Never married Married Widowed Other

Education Elementary/no schooling Secondary/some secondary Technical/trade or some post secondary Diploma/university/graduate degree

Living Arrangements (at Referral) Lived alone Lived with spouse only Other

Change in Social Activities No decline Decline, not distre.ssed Decline, distressed

Amount of Time Alone During Day Never or hardly ever About one hour Long periods of time (i.e.. all morning) All of the time

Percentage (Frequency)

7.8(179) 13.9(321) 23.1 (531) 25.7(591) 29.6 (682)

71.8(1653) 28.3(651)

6.0(138) 33.2 (762) 55.6(1276) 5.1 (118)

33.8 (773) 43.4 (992) 16.6 (379) 6.2(141)

47.7(1048) 28.8 (634) 23.5(517)

69.7(1598) 20.8 (476) 9.6 (220)

33.3 (766) 11.3(261) 30.6 (703) 24.8(571)

Note: MDS-HC = Minimum Data Set-Home Care.

Table 4. Percentage (Frequency) Distributions of Measures of Frailty Utilizing the MDS-HC

Variable Percentage (Frequency)

Stroke No stroke 84.9(1952) Stroke 15.1(347)

Heart Disease No heart disease 82.3 ( 1893) Heart disease 17.7(406)

Hypertension No hypertension 62.9 ( 1447) Hypertension 37.1 (852)

Parkinson's Disease No Parkinson's disease 95.6 (2200) Parkinson's disease 4.4(101)

Alzheimer's Disease No Alzheimer's disease 93.9(2158) Alzheimer's disease 6.1(141)

Arthritis No arthritis 52.5(1208) Arthritis 47.5(1091)

Osteoporosis No osteoporosis 88.4 (2032) Osteoporosis 11.6 (267)

Hip Fracture No hip fracture 95.7 (2201 ) Hip fracture 4.3 (987)

Glaucoma or Cataracts No glaucoma or cataracts 78.0 ( 1798) Glaucoma or cataracts 22.0 (506)

Vision Adequate vision 71.9 ( 1656) Impaired moderately 24.4 (562) Severely impaired 3.7 (86)

Perceived Health Status Perceived good health 69.2 ( 1594) Perceived poor health 30.8(710)

Note: MDS-HC = Minimum Data Set-Home Care.

RESULTS

Univariate Results The univariate distributions have been summarized in Ta-

bles 3 through 6. Table 3 provides results for the sociode- mographic and social variables. The majority of the 2304 participants sampled were between the ages of 75 and 79 years of age (23%), 80 to 84 years of age (26%), or 80 years of age and older (30%) (Table 3). Women comprised 72% of the sample, while 28% were men. With respect to marital status, 56% and 33% were widowed and married, respec- tively. Approximately 77% of the sample had obtained a sec- ondary education or less. Forty-eight percent of the sample lived alone, while 29% lived with their spouses. Seventy percent of the participants had not had a change in their so- cial activities during the past 90 days, while 21% reported a decline in their social activities but they were not distressed by the change. An additional 10% were distressed by the changes that had occurred in their social activities (Table 3).

Univariate distributions for measures of frailty have been summarized in Table 4. The majority of the sample (69%) reported perceiving their health to be good. With respect

RISK FACTORS FOR FALLING M507

Table 5, Percentage (Frequency) Distributions of Exposure to Risk Measures tJtilizing the MDS-HC

Table 6, Percentage (Frequency) of Fall Status, Utilizing the MDS-HC

Variable Percenlage (Frequency) Variable

Gail* Nol impaired Impaired

Environmenlal Hazards*

0 environmenlal bazards 1 environmental bazard 2 or more environmental bazards

Antipsychotic/Neuroleptie Medications

Non use Use

Anxiolytic Medications

Non use Use

Antidepressant Medications

Non use Use

Hypnotic Medications Non use U.se

46.5 (1070) 53.5(1232)

88.4 (2046) 7.69(177) 3.47 (80)

95.0(2184) 5.0(114)

83.1 ( 1913) 16.9(388)

81.8(1880) 18.2(418)

96.0 (2204) 4.0 (92)

Note: MDS-HC = Minimum Data Set-Hotne Care. *ln tbe MDS-HC Section on unsteady gait, tbe following information is pro-

vided: "A gait tbat places tbe client at risk of falling. Unsteady gaits take many forms. Tbe client tiiay appear unbalanced or walk witb a sway. Otber gaits may bave tnicoordinated or jerking movements. Exatnples of unsteady gaits tTiay in- clude fast gaits witb large, careless movements; abnormally slow gaits witb small sbul'lling steps; or wide-based gaits witb balling, tentative steps."

t|n tbe section on "Environmental Hazards," tbe assessor is asked to check any of tbe following tbal make the home environment bazardous: lighting in evening (including inadequate ur no lighting in living room, sleeping rootn, kilcben, toilet, corridors); flooring and carpeting (e.g., boles in floor, electric wires where client walks, scatter rugs); batbroom and toilet room (e.g., nonoper- ating toilet, leaking pipes, no rails tbougb needed, slippery batbtub, outside toi- let); kitcben (e.g., dangerous stove, inoperative refrigerator, infestation by rats or bugs); beatitig and cooling (e.g., too bot in summer, too cold in winter, wood stove in a bome witb an astbtnatic); access to bome (e.g., difftculty entering/ leaving bome); access to rooms in house (e.g., unable to climb stairs).

to medical conditions, the most frequently reported con- sisted of arthritis (48%), hypertension (37%), glaucoma or cataracts (22%), heart di.sease (18%), and stroke (15%), Less frequently occurring conditions included osteoporo- sis (12%), Alzheimer's disease (6%), Parkinson's disease (4%), fractures of the hip (4%), and impaired (24%) or se- verely impaired (4%) vision.

Measures of exposure to risk variables are summarized in Table 5, About 54% of the sample were categorized as hav- ing a gait impairment, while 12% had one or more environ- mental hazards within their homes. With respect to medica- tion, anxiolytics (17%) and antidepressants (18%) were used more frequently than hypnotics (4%) and neuroleptics (5%).

The distribution for the dependent variables employed within the logistic regression analyses is summarized in Table 6. For the analysis of nonfallers versus fallers (1 + falls), 73% of the sample were classified as nonfallers, while 27% were deemed fallers. For the second logistic analysis, 90% of the sample experienced no falls or one fall, while 10% experienced multiple falls (2+ falls).

Frequency (Percentage)

Falls

No falls I or more falls

Falls 0 or 1 falls Multiple falls (2 or more falls)

73.0(1679) 27.0(621)

89.7 (2063) 10.3(237)

Notes: The MDS-HC defines a fall as "an unintentional ebatige in position wbere tbe elder ends up on tbe floor or ground. A fall tiiay result frotn intrinsic or extrinsic causes or botb." MDS-HC = Minimutii Data Set-Home Care.

Multivariate Results In the final logistic regression model for risk of falling (0

falls vs 1+ falls), the independent variables that remained significant included gender, gait, environmental hazards, and the CHESS Scale (Table 7). No interaction terms were significant in the final model. Specifically, being male was associated with an increased risk of falling as indicated by an odds ratios of 1.31. Individuals with impaired gait were also more likely to experience falling (odds ratio [OR] = 2.50). Seniors who had one or more environmental hazards within their homes were 1.20 times more likely to experi- ence a fall, and those with higher scores on the CHESS Scale (OR = 1,20) were at greater risk of falling, a differ- ence of 5 points on the CHESS, representing the highest and lowest possible scores, resulting in an OR of about 2,50.

The risk factors significant in the second logistic model, namely nonfallers/one-time fallers versus recurrent fallers (2+ falls), included the same four variables in the first model (Table 7), Also significant in this model were CPS,

Table 7. Multiple Logistic Regression Models for Risk of Falling

Independent Variables

Gender Female Male

Gait Not impaired Impaired

Count of Environtiiental Hazards

CHESS Scale (single-point increment)

Cognitive Perfortnance Scale (single-point incretiient)

Parkinson's Disease (PD) Do not bave PD Have PD

Healtb Status Good bealtb Poor bealtb

Model One Odds Ratios

(Confidence Intervals) for Fallers ( 1 + falls)

1.00 1.31 (1.07, 1.62)*

1.00 2.50 (2.05, 3.07)t

1.20(1.01, 1.43)*

1.20(1.11, l.3l)t

Model Two Odds Ratios

(Confidence Intervals) for Multiple Fallers

1.00 1.45(1.08, 1.95)*

1.00 2.80 (2.01, 3.89)t

1.35(1.04, 1.59)*

1.29(1.15, l.47)t

1.13(1.02, 1.25)*

1.00 2.47 (1.50, 4.07)t

1.00 1.35(1.01, 1.82)*

Note: CHESS = Changes in Health, End Stage Disease and Signs and

Sytnptoms of Medical Problems.

*/)< .05;V< .01; V<.OOI.

M508 FLETCHER AND HIRDES

Parkinson's disease, and perceived health status. Men were 1.45 times more likely to be at risk for multiple falls than women. Individuals with impaired gait were also more likely to experience 2+ fall events (OR = 2.80). Having en- vironmental hazards within the home increased risk of re- current falls by 1.35 times, whereas higher scores on the CHESS Scale increased fall risk by 1.29 per single unit in- crement. Clients with Parkinson's disease and impaired cognition scores were also more likely to fall two or more times, as indicated by odds ratios of 2.47 and 1.13, respec- tively. A six-point differential on the CPS, again represent- ing extreme scores, results in a 2.08 increase in the odds of multiple falls. Lastly, those with poor self-rated health had an increased risk of multiple falls by 1.35 times. There were no significant interaction terms in either of the final logistic regression models.

DISCUSSION

This study examined two different outcome measures for fall events among home care clients: no falls versus one or more falls, and no falls/less than one fall versus recurrent falls (2+ falls) (Table 7). Categorization and analyses of falls in this manner is consistent with the literature (34,40). Differentiating "one-time fallers" and "recurrent fallers" is important from a clinical viewpoint as the latter group is more likely to be targeted and to benefit from preventive ef- forts (34). Analyses revealed four variables that were signif- icant in both models, namely gender (males), impaired gait, presence of environmental hazards within the home, and higher scores on the CHESS Scale for medical instability. However, three other factors were significant predictors of multiple falls: impaired cognition, Parkinson's disease, and poor self-rated health.

The present findings differ from the majority of research concerning gender. Men were 1.31 times and 1.45 times more likely to be at risk for a fall or recurrent falls, respec- tively. The bulk of literature suggests that women have an increased risk of falling in comparison to men (3,4,13- 16,21,41,42); however, there have been speculations that in- creased falls among women may be the result of the reluc- tance of men to report falling, the result of variables not examined (i.e., differences in gait, knee action) (15,42), or factors associated with being female gender, like osteoporo- sis (43,44) or medication use (i.e., psychotropic medica- tions) (15,21). Other research suggests that more women may be injured by a fall, but more men die from fall-related injuries (45). This may be related to the fact that men not only take more risks, but also experience more traumatic in- juries. It is conceivable that this group of male fallers re- ceiving home care services may represent a distinct group as compared to other men in the fall literature. For example, the men within this study may be receiving home care ser- vices as a direct result of injurious falls, or this study may include men who are able to remain in the home with the support of their spouses and home care services. Similar op- portunities may not be available to women who experience injurious falls. Alternatively, unpublished results also using this data revealed that women were more fearful of falling than men (Fletcher & Hirdes, unpublished data, 2002). As such, it is possible that women feared that they may fall so

they restricted their activity and thus experienced fewer falls. Another possibility is the men in this group were more severely impaired or widowed and thus required more home care than the women within this group; however, there were no significant interaction terms between gender and the CHESS or marital status. Regardless, more research in this area is needed to ascertain whether community-based men receiving home care experience more falls than women, whether the group in question is distinct from other groups of community-based seniors, or whether men experienced more falls than women because women restrict their activity because of their fear of falling.

Individuals with impaired gait were also found to be as- sociated with an increased risk of falls and recurrent falls. For example, individuals with impaired mobility were 1.65 times more likely to experience a fall, as compared to those with no impairments in mobility (42). Wickham and col- leagues (46) determined that seniors with impaired mobility were 2.0 times more likely to experience a fall. Other sup- port for the relationship between impaired mobility and fall risk, as measured by impairments in balance and gait, were found to be associated with falls (3,4,14,18), multiple falls (25,27), serious injuries resulting from falling (16,28), and mortality associated with falls (13). Given the significant in- fluence that impaired gait, balance, and mobility have on el- derly adults, it would seem essential that preventive inter- ventions include a component that restores, improves, or maintains the balance control system.

Environmental hazards may not always be used as pre- dictors or risk factors in fall research, but rather are defined as causal agents or contributing factors in the fall event. Therefore, they may not be entered as risk factors in multi- variate models. However, the contribution of environmental hazards is an important consideration related to the need for environmental modifications in prevention efforts (47-50). Campbell and colleagues (51), Connell (52), and Hind- marsh and Estes (53) found that environmental hazards con- tribute to falls among seniors, which is consistent with the findings for falls and recurrent falls in this research.

The CHESS Scale, which measures changes in health and medical signs and symptoms, was also a significant predic- tor of falls and recurrent falls. This measure is a relatively new scale and has not been used in other falls research. However, many of the key components inherent in the scale itself have been established as risk factors for falls (e.g., in- continence [24,25,30,42], activities of daily living [54], and cognition [10]). Future work in the predictive value of the CHESS Scale is warranted given that it is a standard com- ponent of the MDS-HC, which is now being adapted by home care agencies in several U.S. states and Canadian provinces.

Impaired cognition and Parkinson's disease were two of the risk factors that distinguished failers from recurrent fall- ers. Individuals with impairments in cognitive status as measured by the CPS were more likely to experience two or more falls, but not one or more falls. Cognitive impairment has been linked to increased risk of falling (5,11,55) and re- current falling (5,34). Graafmans and colleagues (34) re- ported that individuals with compromised mental status were 2.4 times more likely to experience recurrent falls. In

RISK FACTORS FOR FALLING M509

addition, impairments in cognition may be indicative of the early stages of dementia, and, as such, multiple falls may be a sign that intervention is required. Parkinson's disease, a progressive degenerative neurologic disorder, was also pre- dictive of recurrent falls, but not one or more falls. Lipsitz and colleagues (40) showed that Parkinsonism was one of the causes of recurrent falls in frail, community-based se- niors. Ambulating safely through one's environment re- quires physical stability and sound mental capacity to react to hazardous conditions. As such, the contribution that cog- nition and Parkinson's disease have on fall status is not sur- prising.

Poor self-rated health is the remaining significant vari- able in the final model for recurrent falls. Studies by Vellas and colleagues (32) and Stalenhoef and colleagues (56) pro- vide confirmatory evidence of the predictive value of health status for falls, injurious falls, or recurrent falls. Perceived measures of health may be useful in identifying those at risk of falling, or more specifically recurrent falling, particularly when clinical assessments may be too expensive or difficult to conduct (32). More research in determining whether per- ceived poor health has predictive value in and of itself or is a surrogate measure of other conditions such as various chronic conditions, physiological impairments, or use of certain medication is warranted.

Underreporting of falls is one limitation that may have af- fected the present results. The number of falls may have been underreported because of problems with recall (18) given the retrospective design of the study. Tideiksaar (57) suggests that reliance on self-reporting may also be prob- lematic if seniors do not want to admit they have experi- enced a fall because they (i) attribute the fall to conse- quences of normal aging; (ii) deny the fall because it reminded them of increasing frailty and dependency; or (iii) fear that reporting it would lead to restriction of activities or to institutionalization. Further, the cross-sectional nature of the study may also limit the conclusions that can be drawn from the multivariate analyses, as cross-sectional designs do not allow researchers to establish a temporal order for fac- tors associated with the fall event.

The risk factors found to be significant in the final logis- tic regression models for one or more falls and recurrent falls (2-1- falls) are consistent with previous falls research with seniors, with the exception of the two scales inherent in the MDS-HC and the male gender finding. Distinguishing individuals into different fall status classifications is impor- tant from a clinical perspective, as it is the recurrent fallers that would benefit to the greatest extent from fall prevention efforts, and from the negative outcomes associated with multiple falls (i.e., mortality) (58). One of the most signifi- cant barriers in the area of determining risk factors for falls is the lack of consistency in the variables/tools used in the research. As such, utilizing a standardized tool, such as the MDS-HC, would assist researchers in making comparisons between different settings. Currently, five Canadian prov- ince.s/territories, seven states, and Veterans Affairs in the United States are either in the process of implementing or are utilizing the MDS-HC as their home care assessment in- strument. Further, an assessment instrument, like the MDS- HC, provides thorough comprehensive health information

about clients and indicates those who may benefit from more extensive evaluation and/or care planning.

ACKNOWLEDGMENTS

We acknowledge the financial contribution from the Health Transition Fund, Health Canada. The views expressed herein do not necessarily repre- sent the official policy of Health Canada. The MDS-HC is a comprehensive assessment instrument developed and owned by interRAI, a not-for-profit network of researchers and clinicians.

An earlier version of this article was presented at the 17th World Con- gress of the International Association of Gerontology Conference, Vancou- ver. British Columbia, Canada, July 2001.

Address correspondence to Paula C. Fletcher, PhD, Department of Ki- nesiology and Physical Education, Wilfrid Laurier University, Waterloo, Ontario N2L 3C5 Canada. E-mail: [email protected]

REFERENCES

1. Canadian Institute for Health Information. (2000). Retrieved Decem- ber 2001 from www.cihi.ca/medrls/jan26-99/figl.htm

2. Canadian Institute for Health Information. (2000). Retrieved Decem- ber 2001 from www.cihi.ca/medrls/rlsntr.htm

3. Campbell AJ, Reinken J, Allan BC, Martinez GS. Falls in old age: a study of frequency and related clinical factors. Age Ageing. 1981; 10: 264-270.

4. Prudham D, Grimley-Evans J. Factors associated with falls in the el- derly: a community study./4ge/4ge/)ig. 1981; 10:141-146.

5. Tinetti ME, Speechley M, Ginter SF. Risk factors for falls among el- derly persons living in the community. N Eng J Med. I988;319(26): 1701-1707.

6. Lilley M, Arie T, Chilvers CED. Accidents involving older people: a review of the literature. Age Ageing. 1995;24:346-365.

7. O'Loughlin JL, Robitaille Y, Boivin JF, Suissa S. Incidence of and risk factors for falls and injurious falls among community-dwelling el- derly. Am J Epidemiol. l993;137(3):342-354.

8. Baker SP, Harvey AH. Fall injuries in the elderly. Clin Geriatr Med. 1985;I(3):5OI-512.

9. Gryfe Cl, Amies A, Ashley MJ. A longitudinal study of falls in an el- derly population: (I) incidence and morbidity. Age Ageing. I977;6: 201-210.

10. Tinetti ME, Williams CS. The effect of falls and fall injuries on func- tioning in community-dwelling older pensons. J Gerontot Med Sei. 1998;53:MI12-M119.

11. Vellas B, Cayla F, Bocquet H, de Pemille F, Albarede JL. Prospective study of restriction of activity in old people after falls. Age Ageing. I987;16(3):189-193.

12. Teno J, Kiel DP, Mor V. Multiple stumbles: a risk factor for falls in community-dwelling elderly: a prospective study. J Am Geriatr Soe. I99O;38(I2): 1321-1325.

13. Wild D, Nayak USL, Isaacs B. How dangerous are falls in old people at home? B Am 7. 1981 ;282(24):266-268.

14. Craven R, Bruno P. Teach the elderly to prevent falls. J Gerontohgi- calNurs. l986;12(8):27-33.

15. Campbell AJ, Spears GF, Borrie MJ. Examination by logistic regres- sion modelling of the variables which increa.se the relative risk of elderly women falling compared to elderly men. J Clin Epidemiol. I99O;43(12): 1415-1420.

16. Tinetti ME, Doucette J, Claus E, Marottoli R. Risk factors for serious injury during falls by older persons in the community. J Am Geriatr Soe. I995;43(11):12I4-1221.

17. Campbell AJ. Drug treatment as a cause of falls in old age: a review of the offending agents. Drugs Aging. 1991 ; 1 (4):289-302.

18. Campbell AJ, Borrie MJ, Spears GF. Risk factors for falls in a com- munity-based prospective study of people 70 years and older. J Gerontol Med Sei. 1989;44:MI 12-Ml 17.

19. Cumming RG, Miller JP, Kelsey JL, et al. Medications and multiple falls in elderly people: the St. Louis OASIS study. Age Ageing. 1991 ; 20:455-461.

20. Liu BA, Topper AK, Reeves RA, Gryfe C, Maki BE. Falls among older people: relationship to medication use and orthostatic hypoten- sion. 7/4m Geriatr Soe. 1995;43(IO):1141-1145.

M510 FLETCHER AND HIRDES

21. Maxwell CJ, Neutel Ct, Hirdes JP. A prospective study of falls after benzodiazepine use: a comparison of new and repeat use. Pharma- coepidemiotogy and Drug Safety. 1997;6:27-35.

22. Neutel Ct, Hirdes JP, Maxwell CJ, Patten SB. New evidence on ben- zodiazepine use and falls: the time factor. Age Ageing. 1996;25:273- 278.

23. Gabell A, Simons MA, Nayak USL. Falls in the healthy elderly: pre- disposing causes. Ergonomies. 1985;28(7):965-975.

24. Yasumura S, Haga H, Nagai H, Suzuki T, Amano H, Shibata H. Rate of falls and the correlates among elderly people living in an urban community in Japan. Age Ageing. 1994;23:323-327.

25. Nevitt MC, Cummings SR. Kidd S, Black D. Risk factors for recurrent nonsyncopal falls: a prospective study. JAMA. 1989;261(18):2663- 2668.

26. Blake AJ, Morgan K, Bendall MJ, et al. Falls by elderly people at home: prevalence and associated factors. Age Ageing. 1988;17(6): 365-372.

27. Lord SR, Ward JA, Williams P, Anstey KJ. Physiological factors asso- ciated with falls in older community-based women. J Am Geriatr Soe. 1994;42(10):1110-1117.

28. Tinetti ME, Doucette J, Claus E. The contribution of predisposing and situations risk factors to serious injuries. J Am Geriatr Soe. 1995; 43(ll):1207-1213.

29. Waller JA. Falls among the elderly: human and environmental factors. AeeidAnatPrev. 1978;10(l):21-33.

30. Stewart RB. Moore MT, May FE, Marks RG, Hale WE. Nocturia: a risk factor for falls in the elderly. J Am Geriatr Soe. 1992;41(12): 1217-1220.

31. Bergland A, Pettersen AM, Laake K. Falls reported among elderly Norwegians living at home. Physiotherapy Res Int. 1998;3(3):164- 174.

32. Vellas BJ, Wayne SJ, Garry PJ, Baumgartner RN. A two-year longitu- dinal study of falls in 482 community-dwelling elderly adults. J Gerontol Med Sei. 1998;53A:M264-M274.

33. Nicken H. Intrinsic factors in falling among the elderly. Areh Intern Med. 1985; 145:1089-1093.

34. Graafmans WC, Ooms ME, Hofstee HMA, Bezemer PD, Bouter LM, Lips P. Falls in the elderly: a prospective study of risk factors and risk proñles. Am J Epidemiol. 1996; 143:1129-1136.

35. Gunter KB, White KN, Hayes WC, Snow CM. Functional mobility discriminates nonfallers from one-time and frequent fallers. J Geron- tol Med Sei. 2000;55A:M672-M676.

36. Wild D, Nayak JSL, Isaacs B. Prognosis of falls in old people at home. J Epidemiol Community Health. 1981;35:200-204.

37. Dunn JE, Fumer SE, Miles TP. Do falls predict institutionalization in older persons? An analysis of data from the Longitudinal Study of Ag- ing. J Aging Health. I993;5(2):194-2O7.

38. Morris JN, Bernabei R, Ikegami N, et al. RAI-Home Care (RAI-HC) . Assessment Manual. Washington, DC: interRAI Corporation; 1999.

39. Morris JN, Fries BE, Steel K, et al. Comprehensive clinical assessment in community setting: applicability of the MDS-HC. JAm Geriatr Soe. 1997;45:1OI7-IO24.

40. Lipsitz LA, Jonsson PV, Kelley MM, Koestner JS. Causes and corre- lates of recurrent falls in ambulatory frail elderly. J Gerontol Med Sei. 199I;46:M114-M122.

41. Sattin RW, Lambert-Huber DA, DeVito CA, et al. The incidence of fall injury events among the elderly in a defined population. Am J Epi- demiol. 1990;131:1028-1037.

42. Fletcher PC, Hirdes JP. Risk factors for serious falls among com- munity-based seniors: results from the National Population Health Survey. Can Aging. 2002;21(1): 103-116.

43. Lindsay R. Osteoporosis. C/mGeria/r A/erf. I988;4(2):411-430. 44. Melton LJ, Riggs BL. Risk factors for injury after a fall. Clin Geriatr

Med. 1985;l(3):525-540. 45. Waller PF. Preventing injury to the elderly. In: Phillips HT, ed. Aging

and Publie Health. New York, NY: Springer; 1985:103-141. 46. Wickham C, Cooper C, Margetts BM, Barker DJP. Muscle strength,

activity, housing and the risk of falls in elderly people. Age Ageing. 1989; 18:47-51.

47. Kellogg International Work Group. The prevention of falls in later life. Dan Med Bull. 1987;4:l-25.

48. Josephson KR, Fabacher DA, Rubenstein LZ. Home safety and fall prevention. Clin Geriatr Med. 1991;7(4):707-731.

49. Tideiksaar R. Environment adaptations to preserve balance and pre- vent falls. Top Geriatr Rehabit. 1990;5(2):78-84.

50. El-Faizy M, Reinsch S. Home safety intervention for the prevention of falls. Phys Occup Ther Geriatr. 1994;12(3):33-49.

51. Campbell JA, Robertson MC, Gardener MM. Elderly people who fall: identifying and managing the causes. Br J Hosp Med. 1995;54(IO): 520-523.

52. Connell BR. Role of the environment in falls prevention. Ctin Geriatr Med. 1996;12(5):859-880.

53. Hindmarsh JJ, Estes EH. Falls in older persons. Areh Intern Med. 1989;149:2217-2222.

54. Langlois JA, Smith GS, Nelson DE, Sattin RW, Stevens JA, DeVito CA. Dependence in activities of daily living as a risk factor for fall in- jury events among older people living in the community. JAm Geriatr Soe. l995;43(3):275-278.

55. Herndon JG, Helmick CG, Sattin RW, Stevens JA, DeVito C, Wingo PA. Chronic medical conditions and risk of fall injury events at home in older adults. JAm Geriatr Soe. 1997;45:739-743.

56. Stalenhoef PA, Diederiks JPM, de Witte LP, Schiricke KH, Crebolder HFJM. Impact of gait problems and falls on functioning in indepen- dent living persons of 55 years and over: a community survey. Patient Education and Counseting. 1999;36:23-31.

57. Tideiksaar R. Fatting in Old Age: Its Prevention and Treatment. New York, NY: Springer; 1989.

58. Dunn JE, Rudberg MA, Furner SE, Cassel CK. Mortality, disability and falls in older persons: the role of underlying disease and disability. Am J Publie Health. 1992;82(3):395-400.

Received December 10, 2001 Accepted March 11, 2002

Copyright of Journals of Gerontology Series A: Biological Sciences & Medical Sciences is the property of

Oxford University Press / UK and its content may not be copied or emailed to multiple sites or posted to a

listserv without the copyright holder's express written permission. However, users may print, download, or

email articles for individual use.