Fear of Falling After Hip Fracture: A Systematic Review of Measurement Instruments, Prevalence, Interventions, and Related Factors

PROGRESS IN GERIATRICS

Fear of Falling After Hip Fracture A Systematic Review ofMeasurement Instruments Prevalence Interventions andRelated Factors

Jan Visschedijk MD MPHw Wilco Achterberg MD PhD Romke van Balen MD PhDz andCees Hertogh MD PhD

The objective of this review was to systematically describeand analyze fear of falling (FoF) in patients after a hipfracture focusing on measurement instruments for FoF theprevalence of FoF factors associated with FoF and inter-ventions that may reduce FoF Fifteen relevant studies werefound through a systematic literature review in which thePubMed Embase PsychINFO and CINAHL databaseswere searched Some of these studies indicated that 50 ormore of patients with a hip fracture suffer from FoF al-though adequate instruments still have to be validated forthis specific group FoF was associated with several negativerehabilitation outcomes such as loss of mobility institu-tionalization and mortality FoF was also related to lesstime spent on exercise and an increase in falls althoughknowledge about risk factors the prevalence over a longertime period and the exact causal relations with importanthealth outcomes is limited Most studies suffer from selec-tion bias by excluding patients with physical and cognitivedisorders Hence more research is required including inpatients who are frail and have comorbidities Only whenknowledge such as this becomes available can interventionsbe implemented to address FoF and improve rehabilitationoutcomes after a hip fracture J Am Geriatr Soc 581739ndash1748 2010

Key words hip fractures rehabilitation fear of fallingfalls efficacy elderly

Although the primary treatment of a hip fracture ismostly surgical the final functional result also depends

on multidisciplinary rehabilitation practices12 Several fac-tors have been associated with recovery after a hip fracturesuch as age sex marital state residence premorbid activ-ities of daily living (ADLs) walking ability cognition andnumber of comorbidities3ndash5 Despite much that is still un-known the importance of psychological factors has beenemphasized67 Fear of falling (FoF) in particular seems tobe an important psychological factor which may have aneven greater influence on functional recovery than pain ordepression8 FoF also reduces participation in exercisesduring the rehabilitation process910 Functional disabilitiescaused by FoF may restrict outcomes in the long term11

particularly because FoF is known to result in dependencyand poor functioning in older adults1213

FoF was first used in the context of the postfallsyndrome14 Several efforts have been made to operation-alize this concept particularly when measures were beingdeveloped Tinetti describes FoF as lsquolsquoa lasting concern aboutfalling that leads to an individual avoiding activities that heshe remains capable of performingrsquorsquo and has operationalizedFoF as a loss of self-efficacy to perform certain activitieswithout falling13 Others relate FoF to deteriorated posturalcontrol15 FoF has often been described more generally as abroader concept of intrinsic fear or worry about falling16

FoF is common in community-based older adults17 butmay be different in patients after a hip fracture becausethese patients have fallen and are suddenly restricted intheir activities In addition patients with a hip fracture havehigher levels of comorbidity and premorbid disability1819

Hence the objective of this review was to systematicallydescribe and analyze FoF in patients after hip fracture Theimportant questions to be addressed were

Which instruments are used to measure FoF in patientswith a hip fracture

What is the prevalence of FoF among patients with a hipfracture

Address correspondence to Jan Visschedijk Department of Nursing HomeMedicine Amsterdam VU University Medical CentreEMGO Institute forResearch in Extramural Medicine Van der Boechorststraat 7 1081 BTAmsterdam the Netherlands E-mail jvisschedijkvumcnl

DOI 101111j1532-5415201003036x

From the Department of Nursing Home Medicine Amsterdam VU Uni-versity Medical CentreEMGO Institute for Research in Extramural Medi-cine Amsterdam the Netherlands wZorggroep Solis Deventer theNetherlands and zDepartment of Public Health and Primary Care LeidenUniversity Medical Centre Leiden the Netherlands

JAGS 581739ndash1748 2010r 2010 Copyright the AuthorsJournal compilation r 2010 The American Geriatrics Society 0002-861410$1500

Which factors are associated with FoF after a hip fracture

Which interventions may reduce FoF after a hip fracture

A systematic review was conducted to answer these ques-tions All relevant studies related to FoF in patients with hipfractures were examined in this review

METHOD

Data Sources and Search Strategy

In March 2009 a literature search was conducted using fourdatabases PubMed (Medline) Embase PsychINFO andCINAHL The Cochrane Library was also consulted Fi-nally the reference lists of selected articles were scrutinizedfor relevant articles

The databases were searched using controlled terms(eg Medical Subject Headings in Medline) and free textwords These were customized to the database The fol-lowing search was used most frequently ((hip fracture)OR (proximal femur fracture)) AND ((fear of fall) OR

(concern of fall) OR (self-efficacy) OR (fear) OR (psy-chological factors))

Study Selection

All possible studies retrospective and prospective were in-cluded in the search Because the majority of hip fracturesoccur in people aged 65 and older no age limitation wasincluded Furthermore no restriction on the year of pub-lication of the article was made

The initial search resulted in 819 titles (Figure 1) InPubMed 362 titles were found to which 161 282 and 14new articles were subsequently added by searching EmbasePsychINFO and CINAHL respectively No additional stud-ies were found in the Cochrane Central Register Two inves-tigators (WA JV) screened the titles to find eligible studiesThe most important criterion was whether these articlescould describe studies related to FoF in patients with hipfractures Where there was any doubt the article was in-cluded One hundred fifty-one articles were selected and theabstracts read (WA JV) Articles were selected when theyprobably presented a study (not a review) that included FoF

Computerized searches

- Medline 362

- Embase + 161

- PsychINFO + 282

- CINAHL + 14

- Cochrane + 0

Manually screening JAGS + 0

Screening of 819 titles with possible studies by 2 independent reviewers

When doubtful articles were included

668 articles excluded

151 articles selected Abstracts reviewed by 2 reviewers

32 articles selected Full article read by 2 reviewers

119 articles excluded

18 articles excluded

14 articles selected

References of 14 articles reviewed 1

additional article included Full

article read by 2 reviewers

15 articles included

Figure 1 Strategy used for selection of published reports on fear of falling in patients with hip fracture

1740 VISSCHEDIJK ET AL SEPTEMBER 2010ndashVOL 58 NO 9 JAGS

or balance problems in patients with a hip fracture Further-more the full article needed to be available in EnglishGerman French or Dutch In addition the article needed todescribe a study and not a comment or personal opinion

Thirty-two articles met the above-mentioned criteriaTwo investigators (WA JV) read the full articles andassessed their ability to answer the research questionsQualitative studies and articles in which no analysis forpatients with hip fractures was provided were excludedFourteen articles were found providing relevant informa-tion for the research questions An additional article wasincluded after reviewing the references

Data Extraction and Synthesis

Appraisal tools that the Centre of Evidence-Based Medicineand other institutions provided were used to analyze thequality of the studies20ndash23 The articles were assessed inparticular on validity (Is there a well-defined study ques-tion Is the study design adequate to answer the studyquestion) importance of results (How great is the likeli-hood of the results How precise are the results) and theirapplicability to the rehabilitation process (Will the resultsbe helpful for the rehabilitation of our patients Are thebenefits worth the harms and the costs Do the results fitwith other available evidence) Using this format studieswere further analyzed and evaluated although it was notpossible to make adequate comparisons between the studiesand to provide a quality assessment because of the hetero-geneity of the studies in terms of design objectives vari-ables and outcome measures Statistical pooling of data(meta-analysis) was not possible either

RESULTS

The 15 studies that were found are summarized in Table 1All studies included measures for FoF Two studies ad-

dressed risk factors for FoF1116 and one compared differ-ent diagnostic measurements31 Eleven studies providedinformation about the association between FoF and othervariables Four intervention studies could be retrieved inwhich the effect of an intervention on FoF was assessed Thestudy features are summarized in Table 1 Two articles referto the same group of patients2435

Which Instruments Are Used to Measure FoF in Patientswith a Hip Fracture

All studies used at least one instrument to measure FoFThese instruments can be divided into two groups instru-ments intended to measure FoF directly and instrumentsfocusing on balance confidence or self-efficacy related tofalls The first group consisted mostly of single itemswhereas the second group usually included instrumentsconsisting of several items

The direct measures for FoF with single items weremostly answers to questions such as lsquolsquoDo you have fear offallingrsquorsquo or lsquolsquoAre you afraid of fallingrsquorsquo Two instrumentswere found that measure balance confidence or self-efficacyrelated to falls the Activity-related Balance Confidence(ABC) Scale and the Fall Efficacy Scale (FES) The items onthe ABC Scale increase in complexity from the beginning tothe end of the instrument The ABC Scale was used in fivestudies and the FES in eight Although these instruments are

used for patients with hip fracture no studies could befound in which the psychometric features of the instrumentshad been tested for this group of patients

Studies that had used or compared two or more instru-ments were of particular interest One cross-sectional studyused the FES (Swedish version FES(S)) and a direct measurefor FoF using a 4-point ordinal scale31 This study in whichpatients were assessed approximately 25 days after surgeryfound a significant relationship (Po001) between the twoinstruments The less fear a patient felt the higher their fall-related efficacy in different activities Patients who were neveror seldom afraid of falling had on average a 40 higher scoreon the FES(S) than patients who reported that they weresometimes or often afraid of falling A particular advantageof the FES(S) was that it indicated which daily activities thepatient perceived to be troublesome highlighting activities inwhich the patient might require further training

Another study found that perceived risk of further fallsand worry over further falls were significantly correlated(correlation coefficient 5 040 Po001) with each other16

When measured 5 to 8 days after surgery neither of thesemeasures was significantly associated with the FES whichmay indicate that they measure different constructs

Research also indicated that the FES was more sensitiveto change than the ABC Scale11 This is in line with findingsfrom earlier studies in which the FES was used in particularfor frail elderly whereas the ABC Scale which containsseveral complex activities is more often used for relativelyhealthy community samples40

What Is the Prevalence of FoF in Patients withHip Fracture

No studies were found that specifically focused on theprevalence of FoF in patients with hip fractures In additionno studies were found in which FoF was measured system-atically over a long period during the rehabilitation process

Some studies provided useful information about theprevalence of FoF after a hip fracture although differentinstruments were used and evidence-based cutoff pointswere missing In some studies the researchers themselvesdetermined the cutoff point When FoF was measuredwithin 1 week after surgery on a scale from 1 to 6 (1 5 nofear to 6 5 strongest fear) 50 (68135) of the patientsindicated that they were afraid of falling (score of 43)35

Another study in which FoF was measured on average 25days after surgery (range 6ndash80 days) revealed that 65(3655) of the patients had FoF sometimes or often31

In an intervention study FoF was measured on a scaleof 1 to 3 3 to 4 weeks after admission to a rehabilitationhospital after a successive training period of 12 weeks and3 months later30 In patients who followed a conventionalrehabilitation program the average FoF was 167 155and 178 respectively Therefore only small changes seemto appear over time Another author indicated an averagelevel of FoF of 22 (n 5 149) and 24 (n 5 166) on a scalethat ranged from 0 to 4 (0 5 no fear 4 5 strong fear) in twostudy cohorts 2 months after a hip fracture37

When using the FES(S) the mean score standard de-viation (SD) was 56 28 (range 0ndash10 0 5 no confidenceat all 10 5 full confidence) with higher scores reported foractivities such as personal grooming getting on and off the

FEAR OF FALLING AFTER HIP FRACTURE 1741JAGS SEPTEMBER 2010ndashVOL 58 NO 9

Table 1 Summary of Publications About Fear of Falling After Hip Fracture

Study Objective and Design Setting Sample

Measurement Instrument

for FoF

Becker et al24 Prognostic study to identify factorsthat predict mortality morbidityand admission to a long-term carefacility after hip fracture

Patients admitted to 5hospitals in southGermany

134 (home-dwelling) patientswith hip fracture 65 and oldermean age SD 803 76

Single question Do you havefear of falling Scale 1ndash6

Casado et al25 Prognostic study using data fromthe Baltimore Hip Study 5examining how social support forexercise by experts affected self-efficacy outcome expectationsand exercise behavior

Patients admitted to 9hospitals in Baltimore MD

164 community-dwellingwomen with hip fracture aged65 mean age SD810 69

Single question Can you rateyour fear of falling on a scale0ndash4 Range 0ndash426

Crotty et al27 Randomized controlled trial tomeasure effect of intervention(home rehabilitation after earlydischarge with therapists visitinghome focusing on negotiated setof goals)

Patients admitted to 3hospitals in AdelaideAustralia

66 patients of aged 65 withhip fracture 34 withaccelerated discharge withhome-based rehabilitation and32 allocated to control groupwith conventional careMedian age (quartiles)intervention group 835 (766855) control group 816(782 854)

ABC Scale 16 items range0ndash10028

FES 10 items range10ndash11029

Hauer et al30 Randomized controlled trial tomeasure effect of intervention(3-month physical training afterhip surgery)

Patients admitted to acutecare or inpatientrehabilitation because ofhip fracture or hipreplacementGermany

28 women with hip fractureaged 75 15 in interventiongroup 13 in control groupmean age SD 813 39

Single question Are you afraidof fallingRange 0ndash315

Ingemarsson et al31 Diagnostic cross-sectional studyto investigate relationshipbetween fall-related efficacy andtests of balance

Patients postoperativelycared for at the GeriatricClinic in Vasa HospitalGoteborg Sweden

55 patients operated on for hipfracture mean age SD823 68

FES - Swedish version 13items range 0ndash13032

Single question Are you afraidof falling Range0ndash3

Jones et al33 Intervention study to assess effectof community exercise program(focused on functional steppingand lower extremity strengtheningexercises)

Patients convalescing in arehabilitation unit in ateaching hospital OntarioCanada

25 patients aged 65 withhip fracture the first 17enrolled in the interventiongroup the next 8 controlsmean age SD 800 60

ABC Scale 16 items range0ndash100 confidence28

FES 10 items range10ndash10029

Kulmala et al34 Cross-sectional study toinvestigate association betweenself-assessed balance confidenceand functional balance and falls

Patients operated on atlocal hospital in Finland

79 patients operated on withhip fracture aged 60ndash85women aged 760 62 menaged 734 74

ABC Scale 16 items range16ndash16028

McKee et al16 Descriptive follow-up study todetermine whether FoF and fallsefficacy contribute to prediction ofhealth outcomes after hip fracture

Patients admitted to thehospital United Kingdom

82 patients with hip fractureaged 65 mean age SD802 73

Perceived risk of further falls inthe next 2 months 1 itemrange 1ndash6Worry over further falls in thenext 2 months 1 item range1ndash6FES 10 items range 10ndash6029

Muche et al35 Prognostic study to identify riskfactors for mortalityinstitutionalization and mobilitylimitations

Patients admitted to 5hospitals in Ulm southernGermany

135 patients with hip fractureaged 65 of 135 15 died infirst 6 months so data of 120patients used forinstitutionalization andmobility mean age SD803 76

Single question Do you havefear of falling Range 1ndash6

Oude Voshaar et al8 Prospective study to assess effectof factors such as paindepression and FoF on functionaloutcome part of a randomizedcontrolled trial to prevent and treatdepression after hip fracture

Patients admitted to one of4 orthopedic units inManchester UnitedKingdom

187 patients with hip fractureaged 60 mean age SD798 87

Modified FES 14 items range0ndash14036

(Continued )

1742 VISSCHEDIJK ET AL SEPTEMBER 2010ndashVOL 58 NO 9 JAGS

toilet getting in and out of a chair and getting in and out ofbed31 The FES(S) was administered 25 days on averageafter surgical repair of the hip fracture Another study re-ported an average score of 698 377 (range 0ndash140)(N 5 187) on the modified FES right after hip fracture8 Thewide confidence interval may be due to the heterogeneity ofthe patients which was also reflected in wide confidenceintervals for depression and pain scales in this study

Which Factors Are Associated with FoF After aHip Fracture

Associations between FoF and other variables were ex-plored in 11 studies810111624253134353738 The relevantvariables to which FoF is associated are listed in Table 2

Premorbid Factors

One study assessed premorbid factors that may have aninfluence on FoF16 The information was collected throughinterviews just after the fracture had occurred It was foundthat the FES had a strong association with prefall activityproblems and a weaker but significant association with his-tory of falls

Mortality

FoF may be a predictor of mortality This was explored intwo longitudinal studies from Germany that used the samepopulation sample2435 FoF was the third-best factor after

premorbid ADL and sex in this study but the first factor thatwas possibly modifiable

Institutionalization

The above-mentioned studies also found associations al-though not significant between FoF and institutionaliza-tion (admission to a nursing home within 6 months after hipfracture)2435

Physical Function Functional Recovery and Mobility

The majority of studies assessed the relationship betweenFoF and functional outcomes particularly mobil-ity81116243538 In two German studies FoF was a predic-tor for limited outdoor mobility (the capacity of goingoutdoor without personal assistance)2435

FoF and falls efficacy were assessed as independentvariables for the functional limitation dimension of theFunctional Limitation Profile (FLP)16 Functional limitationat 2 months was associated with perceived risk of furtherfalls (P 504) and FES score (P 5005) measured approx-imately 1 week after surgery These relationships were sub-sequently examined in multivariate models Withfunctional limitation as the outcome measure FES scoreand perceived risk of further falls did not add significantlyto the prediction of variance once length of stay falls his-tory and prefall activity problems had been controlled for

Table 1 (Contd)

Study Objective and Design Setting Sample

Measurement Instrument

for FoF

Petrella et al11 Prospective study to establishrelationship between physicalfunction and fall-related self-efficacy

Patients admitted torehabilitation programmefrom acute care settingOntario Canada

56 patients with hip fractureaged 65 mean age 797(range 65ndash95)

FES 10 items range 1ndash10(average of items)29

ABC 16 items range0ndash100 confidence28

Resnick et al (1)10 To describe through modellingselected intra- and interpersonalfactors that influence exercisebehavior in women after hipfracture who participated in theExercise Plus Programme

Patients from 6 hospitalsin greater Baltimore MD

209 female hip fracturepatients aged 65 165(79) of whom were availableat 2 months 169 (81) at 6months and 155 (75) at 12months mean age SD807 69

Single question Do you havefear of falling Range 0ndash4

Resnick et al(2)37 Cross-sectional study using datafrom BHS-4 and BHS-5randomized control trials

Women recruited from 3acute care facilities inBHS-4 and 9 acute carefacilities in BHS-5Baltimore MD

315 female patients with hipfracture aged 65 meanage SD BHS-4 825 69BHS-5 840 69

Single question Do you havefear of falling Range 0ndash4

Whitehead et al38 Prospective study to compare4-month outcomes of fallers andnonfallers and those with slow gaitspeed

Patients admitted toFlinders Medical CentreAustralia

73 community-dwellingpatients aged 60 whocompleted a rehabilitationprogram after hip fracturemean age SD 813 62

FES 10 items range0ndash10029

ABC Scale 16 items range0ndash100 confidence28

Ziden et al39 A randomized controlled study toinvestigate whether a homerehabilitation program canimprove balance confidencephysical function and dailyactivity level in the early phaseafter hip fracture

Patients admitted toSahlgrenska UniversityHospital GoteborgSweden

102 community-dwellingpatients with hip fracture aged65 48 enrolled in homerehabilitation program 54 incontrol group withconventional care meanage SD 819 68

FES Swedish version13 items range 0ndash13032

ABC 5 Activity-specific Balance Confidence Scale FES 5 Falls Efficacy Scale SD 5 Standard Deviation BHS 5 Baltimore Hip Study

FEAR OF FALLING AFTER HIP FRACTURE 1743JAGS SEPTEMBER 2010ndashVOL 58 NO 9

Table 2 Variables Associated with Fear of Falling (FoF) After Hip Fracture

Variable Study Associated Variable Association

Prefracture activity McKee et al16 Adapted ADL scale (self-assessed problemswith walking self-care indoor activitiesoutdoor activities)

FES associated with prefall activity problems (Po0001)Association between ADL scale and lsquolsquoworry over further falls innext two monthsrsquorsquo and lsquolsquoperceived risk of further falls in the nexttwo monthsrsquorsquo not significant

History of falls McKee et al16 Fall history (never fallen before fallen but notduring last year fallen in last year)

FES was associated with fall history (Po05) Worry over furtherfalls in next 2 months was associated with fall history (Po001)Association between fall history and lsquolsquoperceived risk of furtherfalls in the next two months was not significant

Mortality Becker et al24 Mortality within 6 months after surgery Multivariate logistic model FoF OR 5 422 for mortality 95CI 5 080ndash480

Muche et al35 Mortality within 6 months after surgery Percentage of patients who died was 177 for patients withstrong FoF and 45 for patients without (P 502)

Institutionalization Becker et al24 Living in nursing home 6 months after surgery Multivariate logistic model FoF for institutionalizationOR 5 223 95 CI 5 079ndash627

Muche et al35 Living in nursing home 6 months after surgery Percentage of patients who were institutionalized was 311 forpatients with strong FoF and 172 for patients without FoF(P 506)

Physical functionfunctionalrecovery balancemobility

Becker et al24 Ability to go outdoors without help of others Multivariate logistic model FoF for loss of mobility OR 5 19695 CI 5 080ndash480

Ingemarsson et al31 Functional reach balance tests on platform Significant relationship between subjective ability (FES) andobjectively measured balance (FR) (Po001) only a fewsignificant correlations between balance tests on platform andFES(S) and FR

McKee et al16 Functional recovery from injury physicallimitation dimension of the FLP

Physical limitation dimension at 2 months was associated withFES score (P 5005) physical limitation dimension at 2 monthswas associated with perceived risk of further falls (P 505)physical limitation dimension at 2 months was not significantlyassociated with worry over further falls

Muche R et al35 Ability to go outdoors without help of others Percentage of patients with mobility limitations was 375 forpatients with strong FoF and 188 for patients without FoF(P 502)

Oude Voshaar et al8 TUG gait speed FR activity subscale of self-report Sickness Impact Profile questionnaire

FoF to predict TUG at 6 months baseline OR 5 089 (P 504) andafter 6 weeks OR 5 075 (Po001)FoF to predict gait speed at 6 months baseline OR 5 093 (notsignificant) and after 6 weeks OR 5 073 (Po001)FoF to predict FR at 6 months baseline OR 5 106 (notsignificant) and after 6 weeks OR 5 132 (P 5006)FoF to predict Sickness Impact Profile at 6 months baselineOR 5 092 (P 511) and after 6 weeks OR 5 070 (Po001)

Petrella et al11 Physical function Functional IndependenceMeasure

No correlation was found between changes in the fall-relatedself-efficacy measures and the Functional IndependenceMeasure

Whitehead et al38 10-m walk test for gait speed Those with slower gait speed had lower self-efficacy (FES andABC) Patients with normal gait mean FES 713 229 meanABC 456 210 patients with slow gait mean FES786 338 mean ABC 755 166

Exercise Casado et al25 Outcome Expectations for Exercise ScaleSelf-Efficacy for Exercise Scale

Model indicated significant path between FoF and outcomeexpectations for exercise

Resnick et al10 Social Support for Exercise ScaleSelf-Efficacy for Exercise ScaleOutcome Expectations for Exercise ScaleStage of Change QuestionnaireExercise Time

At 2 months FoF was not significantly related to any of thevariables mentioned (in the table) At 6 months FoF was relatedto outcome expectations for exercise (path coefficient 023Po001) and indirectly related to exercise timeAt 12 months participants with less FoF had strong self-efficacyexpectations (path coefficient 025 Po001) FoF related alsoto outcome expectations (path coefficient 023 Po001)Through these FoF was related to time spent in exercise

Resnick et al37 Self-Efficacy for Exercise scaleOutcome Expectations for Exercise scaleYale Physical Activity Survey

The participants reported some FoF however no significantrelation between FoF and self-efficacy expectations and exercisebehavior

(Continued )

1744 VISSCHEDIJK ET AL SEPTEMBER 2010ndashVOL 58 NO 9 JAGS

The relationship between FoF and functional outcomeswas strongly established in another longitudinal study8 Inthe final multivariate model cognitive functioning and FoF(Modified FES) assessed 6 weeks after surgery consistentlypredicted functional recovery at 6 months measured usingthe Get Up and Go Test gait speed and functional reachAlso the overall multivariate models including all psycho-logical variables (cognition pain depression) consistentlyincluded FoF at 6 weeks as the most significant predictorafter correction for other factors such as age and level ofpremorbid functioning

Another study found no relationship between changesin physical functioning (Functional Independence MeasureADL mobility) during a rehabilitation program andchanges in fall-related self efficacy (FES and ABC)11 An-other author compared groups with different functionaloutcomes (those with normal walking speed vs thosewith low walking speed slower than 2 standard devia-tions (SDs) below the mean in a 10-m timed walking test)38

The mean of the FES and the ABC 4 months after surgerywere significantly lower for slow walkers than normalwalkers

Exercise

Data from two cohorts in the Baltimore Hip Studies (BHS-4and BHS-5) in which an intervention (Exercise Plus Pro-gramme) was tested were also used to assess FoF37 Whenwomen were tested at 2 months no significant relationshipsbetween FoF and participation in exercises could be dem-onstrated In another study using data from the BaltimoreHip Studies data were collected at 2 6 and 12 months andstructural equation models including FoF were tested10

Although FoF at 2 months was not significantly related at 6months it was related to exercise time In addition at 12months those with less FoF spent more time in exercise Amodel developed to analyze data from the BHS- 5 indicatedan association between FoF and exercise25

Falls

Three studies focused on the relationship between FoF andfalls163438 In a cross-sectional study 79 patients wereassessed who had undergone surgery for hip fracture 6months to 7 years before34 A lower ABC score was asso-ciated with recurrent falling and a lower Berg BalanceScore Participants with indoor falls had lower ABC scoresbut no difference in ABC score was found between outdoorfalls and no outdoor falls Another author found that lsquolsquonohistory of fallsrsquorsquo 2 months after hospital discharge was neg-atively associated with worry over further falls (P 5005)and positively with FES score (Po05)16

Finally the association between FoF and falls was con-firmed when differences between groups of fallers and non-fallers were studied Those who had fallen in the 4 monthsafter hip fracture had significantly lower FES and ABCscores at the 4-month follow-up38

Which Interventions May Reduce FoF After aHip Fracture

The effect of an intervention on FoF was assessed in fourstudies27303339 three of which were randomized con-trolled trials273039 Patients with severe comorbidity orcognitive disorders and patients who were not expected toreturn home were mostly excluded

One study27 evaluated a home-based rehabilitationprogram with early discharge After discharge therapistsvisited patients at home and negotiated a set of targetsPatients followed this program on average for 28 days As aresult of strict inclusion criteria only 66 of 188 patientswere included The study found that the mean FES score at4 months was significantly better for the interventiongroup The mean ABC score of patients was not signifi-cantly different between the intervention and controlgroup

Another study30 investigated a 12-week program ofambulatory training that started immediately after discharge

Table 2 (Contd)

Variable Study Associated Variable Association

Falls Kulmala et al34 Berg Balance Scale for functional balanceSelf-reported falls during previous 6 monthsFalls vs no fallsRecurrent falls vs occasional or no fallsIndoor falls vs no indoor falls Outdoor falls vsno outdoor falls

Lower ABC score was associated with recurrent falling and lowerBBS scoreMean ABC for no recurrent falls was 97 31 versus 68 51for recurrent falls Lower ABC scores were also related to indoorfalls Mean ABC score for no indoor falls was 100 32 versus72 35 for indoor falls Patients with outdoor falls did not differfrom those with no outdoor falls in ABC scores

McKee et al16 Falls in first 2 months after surgery (yesno) Not having fallen at 2 months was positively associated with FESscore (Po05)Not haven fallen was associated with worry over further falls(Po01)Not haven fallen was not significantly associated with perceivedrisk of further falls

Whitehead et al38 Fall history Those who had fallen had lower fall self-efficacy Fallers FESscore 617 226 ABC score 334 201 nonfallers FESscore 735 262 ABC score 535 230

ABC 5 Activity-specific Balance Confidence Scale ADL 5 activity of daily living BBS 5 Berg Balance Scale CI 5 confidence interval FES 5 Falls Efficacy Scale

FLP 5 Functional Limitation Profile FR 5 functional reach OR 5 odds ratio SD 5 standard deviation TUG 5 Timed Up and Go Test

FEAR OF FALLING AFTER HIP FRACTURE 1745JAGS SEPTEMBER 2010ndashVOL 58 NO 9

from the hospital The program included intensive trainingof relevant muscle groups and functional training to enhancebalance Measurements were taken 3 to 4 weeks after ad-mission to the hospital at the end of the training period and3 months later Although there was a clear improvement inFoF it was not significant The mean FoF score in the in-tervention group decreased from 150 071 to 078 083at the end of the training period 3 months later FoF was100 092 For the control group only a small decreasewas foundFfrom 167 10 to 155 088Fwhereasafter 3 months FoF increased to 178 067

A community exercise program focusing on functionalstepping and lower extremity strengthening exercises wasevaluated after a 4-month intervention period33 The first17 patients were enrolled in the intervention group and thenext 10 consecutive patients were controls The ABC scoreincreased in the intervention group from 766 218 to901 101 compared with an increase in the controlgroup from 808 191 to 943 61 FES increased in theintervention group from 839 150 to 936 66 com-pared with an increase in the control group from891 108 to 944 67 The differences were not sig-nificant between the intervention and control groups

In a study of a home rehabilitation program that had amaximum period of 3 weeks after discharge and was aimedto improve balance confidence physical function andADLs the intervention group reported significantly higherconfidence in performing daily activities as measured bythe FES39 The intervention group had a larger increase thancontrols in balance confidence on stairs and instrumentalactivities 1 month after discharge according to the FES Theimprovements in the means of the total score for the inter-vention and control groups were 306 and 135 respectively(Po001) the improvements in the means of the stairclimbing item for the intervention and control group were33 and 06 respectively (P 5002) and the improvementsin the means of the instrumental ADL items of the FES forthe intervention and control groups were 197 and 71respectively (Po001)

DISCUSSION

In this review 15 studies related to FoF in patients with hipfracture were evaluated The studies provided informationconcerning measuring FoF the prevalence of FoF associa-tions between FoF and other variables and interventions toimprove FoF

Measurement instruments can be divided into twogroups those that directly assess FoF using a single questionand those that particularly relate to keeping balance or self-efficacy in not falling during certain activities such as theABC Scale and FES The ABC Scale comprises many com-plex activities and has a greater responsiveness for peoplewith a higher degree of functioning than patients after hipfracture The FES was used in several modifications some-times focusing on the confidence someone has in not fallingwhen doing an activity and sometimes explicitly on the fearsomeone has about losing balance and falling during anactivity Modified versions of the FES have been developedbecause the FES probably has a ceiling effect39 (eg theinternational version (FES-I) to which more-difficult andsocial activities have been added) For frail elderly patients

after hip fracture the FES-I similar to the ABC may com-prise activities that are too complex and the ceiling effectmay be less relevant The FES(S) may be more suitable forpatients with hip fracture because it focuses on basic ADLswhich are relevant for patients with moderate to low func-tional ability32

No studies were found that assessed the psychometricfeatures of these instruments for patients with a hip frac-ture A systematic review of measurement instruments forthe psychological outcomes of falling evaluated the avail-able instruments for FoF40 Most of the instruments foundin the current review can also be found in that study whichidentified the same main categories (instruments that intendto measure FoF directly and those that focus on fall-relatedefficacy and confidence indicating that these are differentconstructs) In a few studies in which single-item instru-ments and FES instruments were included a correlationwas found It is likely that someone who has FoF also hasless confidence in performing certain activities that requirebalance Exactly how these constructs interact with eachother requires further research In addition other factorssuch as coping behavior motivation and outcome expec-tations may influence self-efficacy to execute certain activ-ities That study concluded that lsquolsquothe majority of researchreporting psychometric properties has focused on self-efficacy measures These instruments may prove superior toothers because of the strong and well-researched theoreticalbasersquorsquo Because almost all research has focused on healthycommunity-dwelling older adults evidence is lacking as towhether this statement can be extrapolated to all patientswith hip fracture

No studies were found that consistently assessed theprevalence of FoF after hip fracture over a long time periodMost studies used different instruments and the period be-tween hip fracture and measurement varied substantiallyTherefore it is difficult to compare these findings becauseFoF may not be stable over the rehabilitation period An-other limitation is that all studies excluded patients withcognitive and severe medical disorders which may give se-lection bias It is possible that particularly patients withcognitive and severe comorbidity suffer more often fromFoF A literature review reported that in community-dwell-ing older adults the prevalence of FoF varies between 21and 8517 The findings of the studies in this review arewithin these limits

Many factors have been associated with FoF in com-munity-based older adults17 Some of these were also foundin the current review Because most of the studies werecross-sectional the causality between these factors remainsunclear Only premorbid activity and history of falls wereshown to be risk factors for FoF after a hip fracture16 Fur-thermore this review reveals that FoF is a predictor of im-portant outcomes for the rehabilitation process such asmobility mortality and institutionalization Further re-search is needed to establish whether causal relationshipsexist with other factors FoF was related to falling but notwith outdoor falls34 It is possible that lack of FoF is a riskfactor for outdoor falls because patients with a low ABCscore are more reluctant to walk outside and are morecareful Patients with severe FoF may reduce their activitiesand spend more time indoors FoF may work protectivelyfor these older adults although the study may have some

1746 VISSCHEDIJK ET AL SEPTEMBER 2010ndashVOL 58 NO 9 JAGS

flaws due to recall bias (for falls) and because only a mi-nority of the potential participants consented to participatein the study

The finding that FoF may be related to exercise is par-ticularly important25 It may imply that FoF has to be ad-dressed throughout the rehabilitation process becauseexercise improves health outcomes2 One study found thatthe effect of FoF seemed to be strongest 12 months afterfracture rather than in the more-immediate postfractureperiod10 which lsquolsquosuggests that ongoing efforts might bemade to address the FoF well after their initial fracturersquorsquo Inaddition it has been speculated that lsquolsquothe level of fear offalling during rehabilitation is a more important predictorfor functional outcome than fear of falling directly aftersurgery by excluding patients who easily overcome theirinitial anxiety and including those who become aware oftheir fear during rehabilitationrsquorsquo8 More research is requiredto establish the precise (causal) relationship between FoFand important outcomes

Intervention studies have revealed that FoF can bemodified2739 but the studies have to be interpreted withcare because they included only relatively healthy patientspossibly causing a selection bias It is possible that patientswith more-severe medical and cognitive disorders have lessfavorable results because they are less trainable and moti-vated In one study30 14 of the 28 patients included un-derwent a total hip replacement which is a less commonprocedure for hip fracture and makes it cumbersome togeneralize these results to other populations In additionsample sizes of the studies were small and the follow-upperiods were mostly short In one study the small samplemay have caused the association not to be significant30 Inanother study the high number of nonconsenters and thestrict inclusion criteria may have caused selection bias33

Furthermore the control and intervention groups may nothave been comparable from the start as indicated by thedifferences between the groups in relation to the FES scoresat baseline In another study the difference in effect of theintervention on FoF may be even stronger with six patientsin the home-based rehabilitation program not receiving it(intention-to-treat principle) and several patients in theconventional care group receiving other types of treatmentafter discharge39 Because the intervention had only 1month of follow-up it is not clear whether these improve-ments will be sustained

Over the past years several interventions particularlyfor community-based older adults have been developed toreduce FoF4142 Different programs have been imple-mented some focusing more on exercise (balance trainingwalking tai chi) and others more on education (discussionsabout risk of falling adequate feeding habits and beingactive) Whether such programs are also useful for patientsafter hip fracture is largely unknown and requires furtherresearch

A major limitation of this review is the absence of asubstantial number of prospective studies Most studieswere cross-sectional which makes it impossible to describethe severity of FoF during the rehabilitation process and tofind causal relationships between FoF and relevant out-comes Prospective studies are necessary to bring moreclarity Another limitation relates to the inclusion of pre-dominantly relatively healthy older adults in the studies It

makes generalization of results to the whole population ofpeople with hip fracture cumbersome because a high pro-portion of patients with hip fracture suffer from chronicdiseases both physical and mental in nature1819 Finallythe studies included in this review had a wide variety ofdesigns and methodologies addressing FoF in differentmodalities This made comparison between studies andadequate rating not suitable

This review has shown that FoF in patients with hipfracture is common although adequate instruments stillhave to be validated for this specific group FoF is associatedwith several negative rehabilitation outcomes Knowledgeabout risk factors of FoF prevalence over a longer timeperiod and the exact causal relationship with importanthealth outcomes are still obscure This information isneeded to improve the outcomes of rehabilitation after hipfracture particularly for patients who also have additionalcognitive and medical disorders Based on this knowledgeadequate interventions can be developed that may reduceFoF and improve outcomes of rehabilitation after a hipfracture

ACKNOWLEDGMENTS

Conflict of Interest The editor in chief has reviewed theconflict of interest checklist provided by the authors and hasdetermined that the authors have no financial or any otherkind of personal conflicts with this paper

Author Contributions All authors participated in thestudy design Data collection WA and JV Data analysis andinterpretation WA JV and RB Drafting of the manuscriptJV and WA All authors assisted with revisions to themanuscript and approved the final version

Sponsorrsquos Role None

REFERENCES

1 Shabat S Hip fractures in elderly patients ndash perspectives towards the future

Disabil Rehabil 2005271039ndash1040

2 Chudyk AM Jutai JW Petrella RJ et al Systematic review of hip fracture

rehabilitation practices in the elderly Arch Phys Med Rehabil 200990246ndash

262

3 Balen VR Steyerberg EW Polder JJ et al Hip fracture in elderly patients

Outcomes for function quality of life and type of residence Clin Orthop Rel

Res 2001390232ndash243

4 Michel J Hoffmeyer P Klopfenstein et al Prognosis of functional recovery

1 year after hip fracture Typical patient profiles through cluster analysis

J Gerontol A Biol Sci Med Sci 200055AM508ndashM515

5 Osnes EK Lofthus CM Meyer HE et al Consequences of hip fracture on

activities of daily life and residential needs Osteoporos Int 200415567ndash574

6 Mossey J Mutran E Knott K et al Determinants of recovery 12 months after

hip fracture the importance of psychosocial factors Am J Public Health

198979279ndash286

7 Fortinsky RH Bohannon RW Litt MD et al Rehabilitation therapy self-

efficacy and functional recovery after hip fracture Int J Rehabil Res 2002

25241ndash246

8 Oude Voshaar RC Banerjee S Horan M et al Fear of falling more important

than pain and depression for functional recovery after surgery for hip fracture

in older people Pschol Med 2006361635ndash1645

9 Lees FD Clarck PG Nigg RN et al Barriers to exercise behavior among older

adults A focus group discussion study J Aging Phys Act 20051323ndash33

10 Resnick B Orwig D DrsquoAdamo C et al Factors that influence exercise activity

among women post hip fracture participating in the Exercise Plus Program

Clin Interv Aging 20072413ndash427

11 Petrella RJ Payne M Meyers A et al Physical function and the fear of falling

after hip fracture rehabilitation in the elderly Am J Phys Med Rehabil

200079154ndash160

FEAR OF FALLING AFTER HIP FRACTURE 1747JAGS SEPTEMBER 2010ndashVOL 58 NO 9

12 Wijlhuizen GJ Chorus AM Hopman-Rock Fragility fear of falling physical

activity and falls among older persons Some theoretical considerations to

interpret mediation Prev Med 200846A612ndash614

13 Tinetti ME Powell L Fear of falling and low self-efficacy A case of depen-

dence in elderly persons J Gerontol 19934835ndash38

14 Legters K Fear of falling Phys Ther 200282264ndash272

15 Maki BE Holliday PJ Topper AK Fear of falling and postural performance in

the elderly J Gerontol A Biol Sci Med Sci 199146AM123ndashM131

16 Mckee KJ Orbell S Austin CA et al Fear of falling falls efficacy and health

outcomes in older people following hip fracture Disabil Rehabil 200224327ndash333

17 Scheffer AC Schuurmans MJ Dijk van N et al Fear of falling measurement

strategy prevalence risk factors and consequences among older persons Age

Ageing 20083719ndash24

18 de Luise C Brimacombe M Pederson et al Comorbidity and mortality fol-

lowing hip fracture A population-based cohort study Aging Clin Exp Res

200820412ndash418

19 Liebson CL Tosteson ANA Gabriel SE et al Mortality disability and nursing

home use for persons with and without hip fracture A population-based study

J Am Geriatr Soc 2002501644ndash1650

20 Public Health Resource Unit [on-line] Available at httpwwwphrunhsuk

Accessed on May 8 2009

21 Centre for Evidence Based Medicine Critical Appraisal Tools 2006 [on-line]

Available at httpwwwcebmnet Accessed on May 8 2009

22 Centre for Health Evidence Usersrsquo Guides to Evidence-Based Practice [on-

line] Available at httpwwwcchenet Accessed on May 8 2009

23 Evidence Based Medicine Toolkit University of Alberta [on-line] Available at

httpwwwebmualbertaca Accessed on May 8 2009

24 Becker C Gebhard F Fleischer S et al Prediction of mortality mobility and

admission to long-term care after hip fractures Unfallchirurg 2003106

32ndash38

25 Casado BL Resnick B Zimmerman S et al Social support for exercise by

experts in older women post-hip fracture J Women Aging 20092148ndash62

26 Resnick G Daly MP Predictors of functional ability in geriatric rehabilitation

patients Rehabil Nurs 19982321ndash29

27 Crotty M Whitehead CH Gray S et al Early discharge and home rehabil-

itation after hip fracture achieves functional improvements A randomized

controlled trial Clin Rehabil 200216406ndash413

28 Powell LE Meyers The Activities-specific Balance Confidence (ABC) Scale

J Gerontol A Biol Sci Med Sci 199550AM28ndashM34

29 Tinetti ME Richman D Powell L Falls efficacy as an measure of fear of

falling J Gerontol 199045239ndash243

30 Hauer K Specht N Schuler M et al Intensive physical training in geriatric

patients after severe falls and hip surgery Age Ageing 20023149ndash57

31 Ingemarsson AH Frandin K Hellstrom K et al Balance function and fall-

related efficacy in patients with a newly operated hip fracture Clin Rehabil

200014497ndash505

32 Hellstrom K Lindmark B Fear of falling in patients with a stroke A reliability

study Clin Rehabil 199913509ndash517

33 Jones GR Jakobi JM Taylor AW et al Community exercise programme for

older adults recovering from hip fracture J Aging Phys Act 200614439ndash455

34 Kumala J Sihvonen S Kallinen M et al Balance confidence and functional

balance in relation to falls in older persons with hip fracture history J Geriatr

Phys Ther 200730114ndash120

35 Muche R Eichner Gebhard F et al Risikofaktoren und prognosemoglichkei-

ten fur mortalitat und soziofunktionelle Einschrankungen bei Alteren nach

proximalen Femurfrakturen Euro J Ger 20035187ndash194

36 Hill KD Schwarz JA Kalogeropoulos AJ et al Fear of falling revisited Arch

Phys Med Rehabil 1996771025ndash1029

37 Resnick B Orwig D Hawkes W et al The relationship between psychological

state and exercise behavior of older women 2 months after hip fracture

Rehabil Nurs 200732139ndash147

38 Whitehead C Miller M Crotty M Falls in community-dwelling older persons

following hip fracture Impact on self-efficacy balance and handicaps Clin

Rehabil 200317899ndash906

39 Ziden L Frandin K Kreuter M Home rehabilitation after hip fracture A

randomized controlled study on balance confidence physical function and

every activities Clin Rehabil 2008221019ndash1033

40 Jorstad E Hauer K Becker C et al Measuring the psychological outcomes of

falling A systematic review J Am Geriatr Soc 200553501ndash510

41 Jung D Lee Y Lee SM A meta-analysis of fear of falling treatment pro-

grammes for the elderly West J Nurs Res 2009316ndash16

42 Zijlstra GA van Haastregt JC van Rossum E et al Interventions to reduce fear

of falling in community-living older people A systematic review J Am Geriatr

Soc 200755603ndash615

1748 VISSCHEDIJK ET AL SEPTEMBER 2010ndashVOL 58 NO 9 JAGS