validasi metode skrining pasien risiko jatuh

8/10/2019 Validasi Metode Skrining Pasien Risiko Jatuh

1/4

MJA Volume 189 Number 4 18 August 2008 193

R E S E A R C H

The Medical Journal of Australia ISSN:

0025-729X 18 August 2008 189 4 193-

196The Medical Journal of Australia 2008www.mja.com.auResearch

alls are a significant burden on thehealth care system, with a substantialproportion of the associated health

care costs occurring in residential aged carefacilities (RACFs).1RACF residents have anelevated risk of falls due to high numbersof comorbidities; however, reduced physi-cal activity,2 safer environments andincreased supervision may actually reducefall risk in this setting.3

Identifying people at risk of falls inRACFs is complex, and risk assessmenttools developed for community and acutehospital settings cannot be translated toRACFs.4Previous studies have shown thatfalls occur more frequently in mobile5,6andphysically active nursing home residents.7

As part of the Fracture Risk Epidemiologyin the Elderly (FREE) study,3,8we reporteda non-linear association between standingbalance and falls, with low fall rates inthose with the worst balance as well asthose with the best balance. The FREEstudy also found that many fall risk factorsin those who could stand were either notevident or were actually protective in those

who could not stand, further indicatingthat fall risk identification in RACF resi-dents is not straightforward. None of thecurrently available screening tools forRACFs can be considered definitive forpredicting residents at risk of falls, as theyhave not been adequately validated in largepopulations.

The aim of this study was to use theFREE study database to develop and vali-date screening tools that take into accountthe complexities relating to fall risk in frailolder people living in RACFs. Such vali-dated screening tools would help maximise

efficiency and cost-effectiveness of inter-ventions in RACFs, where resources areoften limited.

METHODS

Participants

The study sample comprised 2005 peoplewho took part in the FREE study betweenJune 1999 and June 2003.3,8 Participantswere recruited from randomly selectedRACFs (898 participants from 80 nursing

homes and 1107 participants from 50intermediate-care hostels) in northern Syd-

ney, New South Wales. The participationrate of eligible residents (ie, those notexclusively confined to bed) was 44.8% fornursing home residents and 55.2% forhostel residents. Participants were aged65104 years (mean SD, 85.7 7.1 years),and 1532 (76.4%) were women. About halfof the population (1033; 51.5%) hadexperienced at least one fall during theprevious year.

The Northern Sydney Area Health ServiceEthics Committee approved the study, andinformed consent was obtained from theparticipants or from a person legally able to

give consent on their behalf.

Baseline risk factor assessments

Medical conditions, medication use, andcognitive and psychological statusResident assessment by a research nurse,self-report, care provider interviews, andmedical records were used to determine thepresence of medical conditions. Medicationuse, urinary incontinence, falls history, anduse of assistive devices were documentedfrom residents medical records. Cognitive

status was assessed with the Mini-MentalState Examination.9

Balance and physical functionStanding balance was assessed using a staticbalance test.10 Participants were classifiedinto five grades: 1= unable to stand on thefloor for any period without support fromanother person or a walking aid; 2 = unableto maintain balance on the floor for 30 s;3 = able to maintain balance on the floor for30 s but unable to maintain balance on amedium-density foam rubber mat (70cm 60cm 15 cm thick) for any period of time;4 = able to maintain balance on the floor butunable to maintain balance on the foam

rubber mat for 30s; and 5 = able to maintainbalance when standing on the floor and thefoam mat for 30s each. Twelve residents didnot complete the static balance test and wereexcluded from the analysis.

Participants sit-to-stand ability was meas-ured by assessing their ability to rise from astandard-height (0.43 m) chair with armrests.11 Participants were graded on a four-point scale: 1 = unable; 2= able with thehelp of another person; 3= able with the useof arm support; and 4 = able without theneed for arm support.

Development and validation of fall risk screening toolsfor use in residential aged care facilities

Kim Delbaere, Jacqueline C T Close, Hylton B Menz, Robert G Cumming, Ian D Cameron,Philip N Sambrook, Lyn M March and Stephen R Lord

ABSTRACT

Objective: To develop screening tools for predicting falls in nursing home andintermediate-care hostel residents who can and cannot stand unaided.

Design and setting: Prospective cohort study in residential aged care facilities innorthern Sydney, New South Wales, June 1999 June 2003.

Participants: 2005 people aged 65104 years (mean SD, 85.7 7.1 years).

Main outcome measures: Demographic, health, and physical function assessmentmeasures; number of falls over a 6-month period; validity of the screening models.

Results: Ability to stand unaided was identified as a significant event modifier for falls.In people who could stand unaided, having either poor balance or two of three other riskfactors (previous falls, nursing home residence, and urinary incontinence) increased therisk of falling in the next 6 months threefold (sensitivity, 73%; specificity, 55%). In people

who could not stand unaided, having any one of three risk factors (previous falls, hostelresidence, and using nine or more medications) increased the risk of falling twofold(sensitivity, 87%; specificity, 29%).

Conclusions: These two screening models are useful for identifying older people livingin residential aged care facil ities who are at increased risk of falls. The screens are easy toadminister and contain items that are routinely collected in residential aged care

MJA 2008; 189: 193196

facilities in Australia.

F


2/4

194 MJA Volume 189 Number 4 18 August 2008

R E S E A R C H

Falls

All residents were followed up for a periodof 6 months or until death, if sooner. Fallswere ascertained from incident reports and

medical records and were classified usingthe Kellogg definition.12 Forty-seven resi-dents died within 3 months without havinga fall and were excluded from the analysis.

Statistical analysis

Ability to stand unaided was identified as asignificant event modifier for falls. Accord-ingly, separate logistic regression modelswere used to calculate univariate odds ratiosfor the associations between demographic,health, and physical measures and falls inthose who could (static balance grades 25)

and could not (static balance grade 1) standunaided. In subsequent multivariate regres-sion models, the best set of independent andsignificant risk factors for falls were soughtfor the two groups. The predictive accuracyof combinations of the identified independ-ent and significant risk factors were thenexamined using MantelHaenszel statistics.Finally, the sample was divided randomlyinto two groups and the validity of themodels assessed with split-half analyses.

The data were analysed using SPSS forWindows (SPSS Inc, Chicago, Ill, USA).

RESULTS

Of the 1946 residents assessed, 813 (41.8%)fell once or more during the 6-month fol-low-up period. Of these, 410 fell twice or

more, and 460 suffered at least one fall thatresulted in an injury.2

Screening model for falls in residentswho could stand unaided

Univariate analyses showed that the risk ofexperiencing at least one fall during theprospective period significantly increasedwith age, and was higher in nursing homeresidents and in those who had fallen in thepast year. Factors associated with fallsincluded poor static balance and sit-to-standability, greater illness severity, impaired cog-

nitive status, Parkinson disease, urinaryincontinence, knee osteoarthritis, use of awalking aid, and use of many medications(Box 1).

Stepwise logistic regression analysis iden-tified four significant independent risk fac-tors for falls: nursing home residence,impaired balance, a history of falls in thepast year, and urinary incontinence (Box 2).In a second step, impaired balance wasdichotomised based on examination of areceiver operating characteristic (ROC)curve as the inability to stand on a foam mat

(static balance grades 23). The model sig-nificantly predicted falls (2 = 159.5, df = 4,P < 0.001) and accounted for 14% of thevariance in faller status, with 80% of the

non-fallers successfully predicted, 44% ofthe fallers successfully predicted, and 65%overall faller status successfully predicted.

Additional analyses examining the predic-tive accuracy of all combinations of inde-pendent and significant risk factors showedthat the risk of experiencing a fall was 3.55times greater when the person had poorbalance, and 2.73 times greater for any twoother risk factors. According to this model,the proportion of people at risk of a fall was57%, with a sensitivity of 73% and a specifi-city of 55%.

Screening model for falls in residentswho could not stand unaided

Fewer risk factors were evident for the resi-dents who could not stand unaided (Box 1).In this group, residents with a history of fallsand those who were taking more medica-tions had an increased risk of falls. Nursinghome residence, increased care levels andreduced ability to rise from a chair wereassociated with fewer falls.

The logistic regression analysis identifiedthree significant independent risk factors for

1 Investigated risk factors for falls in people who could and could not stand unaided

IQR = interquartile range. MMSE= Mini-Mental State Examination. na= not applicable. *Values are defined in parentheses in Risk factor column. Denominators forpercentages vary due to missing data for some risk factors. P< 0.05. See Methods for details of classification.

Could stand unaided (n = 1569) Could not stand unaided (n =377)

Risk factor Value* Odds ratio (95% CI) Value* Odds ratio (95% CI)

Demographics

Age (mean [SD] in years) 85.6 (6.9) 1.03 (1.011.04) 85.6 (7.6) 1.02 (0.991.05)

Sex (no. [%] of women) 1195 (76.2%) 1.00 (0.791.27) 297 (78.8%) 0.76 (0.461.26)

Nursing home residence (no. [%] of residents) 505 (32.2%) 2.75 (2.213.41) 349 (92.6%) 0.18 (0.070.42)

Previous fall in the past year (no. [%] of residents) 773 (50.6%) 2.16 (1.752.66) 231 (63.6%) 2.00 (1.263.18)

Medical conditions

Implicit Illness Severity Scale (median score [IQR]) 3 (23) 2.21 (1.832.66) 3 (33) 1.68 (0.933.01)

Standardised MMSE score (mean [SD]) 21.5 (8.1) 0.94 (0.930.96) 15.6 (10.3) 1.00 (0.981.02)

Parkinson disease (no. [%] of residents) 73 (4.7%) 1.70 (1.062.72) 39 (10.9%) 0.83 (0.411.68)

Urinary incontinence (no. [%] of residents) 863 (56.0%) 1.75 (1.422.15) 274 (75.9%) 1.05 (0.641.73)

Osteoarthritis of knee (no. [%] of residents) 532 (35.8%) 1.26 (1.011.56) 118 (37.1%) 0.81 (0.511.29)

Use of aid during walking (no. [%] of residents) 985 (61.1%) 1.39 (1.121.71) 0 na

Psychoactive medication (no. [%] of residents) 180 (12.5%) 0.86 (0.621.18) 59 (15.9%) 1.51 (0.862.64)

Number of medications used (median [IQR]) 6 (49) 1.05 (1.021.08)

7 (410) 1.06 (1.011.12)

Physical measures

Static balance (median grade [IQR]) 4 (35) 0.58 (0.520.64) 1 (11) na

Sit-to-stand ability (median grade [IQR]) 2 (12) 1.84 (1.522.23) 3 (34) 0.54 (0.370.77)


3/4

MJA Volume 189 Number 4 18 August 2008 195

R E S E A R C H

falls: hostel residence, a history offalls in the past year, and use ofmany medications (Box 3). In asecond step, the number of medi-cations was dichotomised basedon an ROC curve inspection,with the cut-off point at nine or

more medications. The modelsignificantly predicted falls (2 =31.43, df = 3, P < 0.001) andaccounted for 11% of the vari-ance in faller status, with 97% ofthe non-fallers successfully pre-dicted, 16% of the fallers success-fully predicted, and 67% overallfaller status successfully pre-dicted.

Additional analyses examiningthe predictive accuracy of eachcombination of independent andsignificant risk factors showed

that the risk of experiencing a fallwas 2.09 times greater when theperson scored positive on anyone of these risk factors. Accord-ing to the model, the proportionof people at risk was 77%, with asensitivity of 87% and a specifi-city of 29%.

Validation of the screening models

The split-half validation of the regressionmodel for residents who could standunaided revealed that the sensitivity and

specificity were 74% and 56%, respectively,in the exploratory analyses, and 73% and54% in the confirmatory analyses. For thepeople who could not stand unaided, thesensitivity and specificity were 86% and

30%, respectively, in the exploratory analy-ses, and 90% and 29% in the confirmatoryanalyses.

DISCUSSION

Two fall risk screening models emerged fromthis study: one for people who could standunaided and one for people who could not.The stronger model was achieved in peoplewho could stand unaided. Having poor bal-

ance or a positive score for two ofthree other risk factors (previousfalls, nursing home residence,and urinary incontinence) pro-duced a threefold increased riskof falling in the next 6 months.These factors have consistently

been identified as important riskfactors for falls.3,8,13-17

The model in people whocould not stand unaided wasless robust. Having had a previ-ous fall, using nine or moremedications, or residing in ahostel were associated with atwofold increased risk of fallingduring the follow-up period.Thus, standard physical riskfactors do not appear to bepresent in this subpopulation, inaccordance with previous find-ings.3,6,15Measures such as pro-viding high-level care, usingalarm devices, and regular medi-cation review may be particu-larly beneficial for this group.

Several fall risk screeningtools for residential care have

previously been published.6,18 However,their small to moderate sample sizes (rang-ing between 78 and 472 participants), aswell as the frequent use of univariateapproaches, have been insufficient to pro-duce validated tools with acceptable sensi-

tivityspecif icity ratios. One studyundertaken in Germany used multivariatemodelling, similar to our approach, to con-struct an algorithm for predicting falls innursing homes.6 It found that falls history,

2 Fall risk models for people who could stand unaided

* Using multivariate logistic regression analyses. Static balance grades 23.

Using MantelHaenszel analyses.

VariableRegressioncoefficient P

Odds ratio(95% CI)

Regression model*

Nursing home residence 0.650


4/4

196 MJA Volume 189 Number 4 18 August 2008

R E S E A R C H

vision impairment and incontinence werefall risk factors in residents able to transferindependently from bed to a standing posi-tion, compared with risk factors of fallshistory and restraint use in residents unableto transfer independently. Our study buildson this work by including hostel as well as

nursing home residents and by including asimple balance assessment, which assists inidentifying the more mobile people at risk offuture falls. Our sample size was also suffi-ciently large to allow investigation of multi-ple measures and split-half validation. Theresultant screening models are quick andeasy to administer and require only oneinexpensive and readily available piece ofequipment (a 15cm thick, medium-densityfoam rubber mat).

A simple algorithm, shown in Box 4,demonstrates how the screening models

could be used in RACFs. By using them thisway, only 56% of people who can standunaided would need to be targeted to iden-tify 75% of all fallers. Similarly, 77% ofpeople who cannot stand unaided wouldneed to be targeted to identify 87% offallers. While these models assist in identify-ing most residents at risk of falls, weacknowledge that significant proportions offallers would not be identified. Therefore, amultifaceted approach to fall preventionshould be considered as part of routine carefor all older people in RACFs.1These careplans should include evidence-based strate-

gies such as education of staff19-21and resi-dents,19 environmental modifications,19-21

regular medication reviews,20,21and exerciseto improve strength, balance, gait, safetransfers and walking aid use.19,20,22 Ourscreening model augments this care, not onlyby identifying residents most at risk, but alsoby providing information on risk factors toguide tailored intervention strategies.

As yet, no randomised controlled trialshave been undertaken for fall prevention inRACFs in Australia, and disparate findingsregarding the effectiveness of interventions

have been published in other countries.19-23This discordance likely reflects differencesin what constitutes an RACF, staffing andcasemix within RACFs, and study designand quality of interventions. The screeningtools developed in this study have someclinical utility in terms of identifying at-riskfallers, but further work is required to teaseout whether screening tools offer clinicalefficacy and cost-effectiveness. In theabsence of a consensus approach to routineassessment and intervention in RACFs, thesuggested screening models offer these facil-

ities the option of applying a validated toolto identify residents at high risk of falling.

ACKNOWLEDGEMENTS

This research was conducted as part of the FREEstudy, which has been funded by a National Healthand Medical Research Council (NHMRC) grant and

the Arthritis Foundation of Australia. The analysisfor this paper was funded by NSW Health. HyltonMenz is currently an NHMRC Clinical Career Devel-opment Research Fellow (ID 433049) and StephenLord is an NHMRC Senior Principal ResearchFellow.

COMPETING INTERESTS

None identified.

AUTHOR DETAILS

Kim Delbaere,MPT, PhD, Research Officer1,2

Jacqueline CT Close,MD, FRCP, SeniorLecturer and Geriatrician1

Hylton B Menz,BPod(Hons), PhD, AssociateProfessor and Director3

Robert G Cumming,MB BS, PhD, Professor,Centre for Education and Research on Ageing4

Ian D Cameron,MB BS, PhD, Professor,Rehabilitation Studies Unit4

Philip N Sambrook,MD, LLB, Professor,Institute of Bone and Joint Research4

Lyn M March,MB BS, PhD, Associate Professor,Institute of Bone and Joint Research4

Stephen R Lord,PhD, Professor and SeniorPrincipal Research Fellow1

1 Falls and Balance Research Group, Prince ofWales Medical Research Institute, Universityof New South Wales, Sydney, NSW.

2 Department of Experimental-Clinical andHealth Psychology and Department ofRehabilitation Sciences and Physiotherapy,Ghent University, Ghent, Belgium.

3 Musculoskeletal Research Centre, Faculty ofHealth Sciences, La Trobe University,Melbourne, VIC.

4 Faculty of Medicine, University of Sydney,Sydney, NSW.

Correspondence: [email protected]

REFERENCES

1 Safety and Quality Council. Preventing falls andharm from falls in older people. Best practice

guidelines for Australian hospitals and residen-tial aged care facilities. Canberra: AustralianCouncil for Safety and Quality in Health Care,2005.

2 Tinetti ME. Factors associated with seriousinjury during falls by ambulatory nursing homeresidents. J Am Geriatr Soc1987; 35: 644-648.

3 Lord SR, March LM, Cameron ID, et al. Differingrisk factors for falls in nursing home and inter-mediate-care residents who can and cannotstand unaided. J Am Geriatr Soc 2003; 51:1645-1650.

4 Oliver D, Masud T. Preventing falls and injuriesin care homes. Age Ageing2004; 33: 532-535.

5 Thapa PB, Brockman KG, Gideon P, et al. Injuri-ous falls in nonambulatory nursing home resi-

dents: a comparative study of circumstances,incidence, and risk factors. J Am Geriatr Soc1996; 44: 273-278.

6 Becker C, Loy S, Sander S, et al. An algorithmto screen long-term care residents at risk foraccidental falls. Aging Clin Exp Res 2005; 17:186-192.

7 Norton R, Galgali G, Campbell AJ, et al. Is physi-cal activity protective against hip fracture in frail

older people? Age Ageing2001; 30: 262-264.8 Chen JS, March LM, Schwarz J, et al. A multi-

variate regression model predicted falls in resi-dents living in intermediate hostel care. J ClinEpidemiol2005; 58: 503-508.

9 Folstein MF, Folstein SE, McHugh PR. Mini-mental state. A practical method for gradingthe cognitive state of patients for the clinician.J Psychiatr Res1975; 12: 189-198.

10 Lord SR, Clark RD, Webster IW. Physiologicalfactors associated with falls in an elderly popu-lation. J Am Geriatr Soc1991; 39: 1194-1200.

11 Seaby L, Torrance G. Reliability of a physiother-apy functional assessment used in a rehabilita-tion setting. Physiother Can1989; 41: 264-271.

12 Gibson MJ, Andres RO, Isaacs B, et al. Theprevention of falls in later life. A report of theKellogg International Work Group on the Pre-vention of Falls by the Elderly. Dan Med Bull1987; 34 Suppl 4: 1-24.

13 Lord SR, Clark RD. Simple physiological andclinical tests for the accurate prediction offalling in older people. Gerontology 1996; 42:199-203.

14 Ganz DA, Bao Y, Shekelle PG, Rubenstein LZ.Will my patient fall? JAMA2007; 297: 77-86.

15 Thapa PB, Gideon P, Brockman KG, et al. Clini-cal and biomechanical measures of balance asfall predictors in ambulatory nursing home resi-dents. J Gerontol A Biol Sci Med Sci1996; 51:M239-M246.

16 Kiely DK, Kiel DP, Burrows AB, Lipsitz LA. Iden-

tifying nursing home residents at risk for falling.J Am Geriatr Soc1998; 46: 551-555.

17 Perell KL, Nelson A, Goldman RL, et al. Fall riskassessment measures: an analytic review. J Ger-ontol A Biol Sci Med Sci2001; 56: M761-M766.

18 Scott V, Votova K, Scanlan A, Close J. Multifac-torial and functional mobility assessment toolsfor fall risk among older adults in community,home-support, long-term and acute care set-tings. Age Ageing2007; 36: 130-139.

19 Becker C, Kron M, Lindemann U, et al. Effec-tiveness of a multifaceted intervention on fallsin nursing home residents. J Am Geriatr Soc2003; 51: 306-313.

20 Jensen J, Nyberg L, Rosendahl E, et al. Effectsof a fall prevention program including exercise

on mobility and falls in frail older people livingin residential care facilities. Aging Clin Exp Res2004; 16: 283-292.

21 Ray WA, Taylor JA, Meador KG, et al. A rand-omized trial of a consultation service to reducefalls in nursing homes. JAMA1997; 278: 557-562.

22 Schnelle JF, Kapur K, Alessi C, et al. Does anexercise and incontinence intervention savehealthcare costs in a nursing home population?J Am Geriatr Soc2003; 51: 161-168.

23 Kerse N, Butler M, Robinson E, Todd M. Fallprevention in residential care: a cluster, rand-omized, controlled trial. J Am Geriatr Soc2004;52: 524-531.

(Received 13 Aug 2007, accepted 12 Mar 2008)

validasi metode skrining pasien risiko jatuh

Documents