JAMDA xxx (2014) 1e8

JAMDA

journal homepage: www.jamda.com

Original Study

Effects of a Multicomponent Home-Based Physical RehabilitationProgram on Mobility Recovery After Hip Fracture: A RandomizedControlled Trial

Anu Salpakoski MSc a,*, Timo Törmäkangas PhD a, Johanna Edgren MSc a,Mauri Kallinen MD, PhD b, Sanna E. Sihvonen PhD c, Maija Pesola MDd,Jukka Vanhatalo MD e, Marja Arkela PhD e, Taina Rantanen PhD a, Sarianna Sipilä PhD a

aGerontology Research Center and Department of Health Sciences, University of Jyväskylä, Jyväskylä, FinlandbDepartment of Medical Rehabilitation, Oulu University Hospital, Oulu, FinlandcUniversity of Applied Sciences, School of Health and Social Studies, Jyväskylä, FinlanddDepartment of Orthopedics and Traumatology, Central Finland Central Hospital, Jyväskylä, FinlandeDepartment of Physical and Rehabilitation Medicine, Central Finland Central Hospital, Jyväskylä, Finland

Keywords:Hip fracturerehabilitationmobilitydisablement processreenablement

The authors declare no conflicts of interest.* Address correspondence to Anu Salpakoski, M

Gerontology Research Center and Department of Hea40014 Jyväskylä, Finland.

E-mail address: anu.salpakoski@jyu.fi (A. Salpakos

1525-8610/$ - see front matter � 2014 - American Mhttp://dx.doi.org/10.1016/j.jamda.2013.12.083

a b s t r a c t

Objective: To investigate whether a home-based rehabilitation program for community-dwelling olderpeople with recent hip fracture is more effective than standard care in improving mobility recovery andreducing disability.Design: Randomized, controlled, parallel-group trial.Setting: Rehabilitation inparticipants’homes;measurements inuniversity-based laboratoryand localhospital.Participants: Clinical population of community-dwelling men and women (aged 60þ) recovering fromhip fracture. Participants were randomly assigned into control (n ¼ 41) or intervention (n ¼ 40) groupson average 42 � 23 days after discharge home.Intervention: A yearlong multicomponent home-based rehabilitation aimed at promoting mobilityrecovery and physical functional capacity after hip fracture. The intervention included evaluation andmodification of environmental hazards, guidance for safe walking, nonpharmacological pain manage-ment, a progressive home exercise program, physical activity counseling, and standard care.Measurements: Measurements were outlined according to the tiers of the disablement process, with theability to negotiate stairs as the main outcome. Prefracture ability to negotiate stairs was enquired at thehospital on average 10 � 5 days after fracture. Subsequently, current perceived ability to negotiate stairswas reported immediately before the intervention (on average 9 weeks after surgery) and 3, 6, and 12months thereafter. Other measurements included leg extension power deficit (LEP), functional balance(Berg Balance Scale) and lower extremity performance (Short Physical Performance Battery). Effects ofthe intervention were analyzed with generalized estimation equations and longitudinal repeated mea-sures mixture path models.Results: The intervention reduced perceived difficulties in negotiating stairs (interaction, group� time P¼.001) fromprefracture to12months comparedwith the control condition. Themixturepathmodel revealedthat less difficulty in negotiating stairs at 6 and12months correlatedwith better functional balance at 3 and6 months in the intervention group but not controls (group difference P ¼ .007 and P< .001, respectively).Conclusion: The individualized home-based rehabilitation program improved mobility recovery after hipfracture over standard care. To be efficacious in reducing or reversing disability after hip fracture,rehabilitation needs to be individualized, include many components, be progressive, and span a suffi-ciently long period. Current Controlled Trials (ISRCTN53680197).

� 2014 - American Medical Directors Association, Inc. All rights reserved.

Sc, University of Jyväskylä,lth Sciences, PO Box 35, FIN-

ki).

edical Directors Association, Inc. A

Mobility recovery after hip fracture is challenging. Major trauma,related surgery, and hospitalization lead to a dramatic decline inmuscle power on the side of the fracture and consequent asymmet-rical muscle power deficit.1 Poor muscle power and asymmetricaldeficit are associated with poor standing balance, slower walking

ll rights reserved.

A. Salpakoski et al. / JAMDA xxx (2014) 1e82

speed, and increased risk for injurious falls,2,3 thereby increasing therisk for unsuccessful mobility recovery and further mobility disability.

Different physical functions do not recover simultaneously afterhip fracture.4,5 For most routine mobility functions, like walkingindependently, with or without walking aids, the recovery occurs forthe most part within the first few months after discharge from hos-pital.5 The more challenging mobility functions, which require morestrength and balance than walking, such as negotiating stairs, requirea longer recovery period of up to 1 year post discharge.5 The presenceor absence of perceived difficulties in negotiating stairs describes aperson’s mobility ability in everyday life.5e8

Previous experimental studies investigating the effects of reha-bilitation on mobility recovery after a hip fracture have focused onsingle outcome measures only, rather than looking at the wholerecovery process, from the physiological to functional prerequisitesfor walking and, finally, to the ability or difficulty in walking in one’senvironment. To achieve an optimal rehabilitation outcome, it wouldbe logical to acknowledge the complexity of reenablement and totarget and follow up all the functional tiers in the disablementprocess9 during the recovery phase. To our knowledge, no previousstudies have investigated the effects of rehabilitation on mobilityrecovery in the context of functional tiers, as outlined in thedisablement process,9,10 following hip fracture.

Currently, as acknowledged by recent Cochrane Reviews11,12 andNational Institute for Health and Clinical Excellence guidance13

studies, there are no generally accepted rehabilitation guidelines formobility and functional capacity recovery after hip fracture. Most ofthe studies on this topic have used efficacy-driven research designswith relatively short interventions organized in outpatient clinicswith homogeneous groups of hip fracture participants. In many suchstudies, the persons likely to benefit the most from a programincluding physical exercise have been excluded. Traveling to orga-nized sessions on a weekly basis may be too demanding for manyolder people who are frail and suffer from pain and fear of falling aftera fracture. Therefore, a home-based individually tailored rehabilita-tion program may be the most promising strategy for mobilityrecovery and prevention of mobility disability after hip fracture.

Only a few home-based rehabilitation studies among community-dwelling older people recovering from hip fracture have been pub-lished.14e17 In those studies, self-reported or performance-basedmobility has been assessed as a secondary outcome only, andthe results obtained have been contradictory. However, amongcommunity-dwelling frail older people, home-based multicompo-nent rehabilitation reduced the progression of functional decline.18 Inthis study by Gill et al. (2002), rehabilitation included instructions forsafe moving and use of assistive devices, removing environmentalhazards, and a progressive exercise program. These encouraging re-sults suggest that a similar approach might also be efficacious amonghip fracture patients.

The aim of this randomized controlled trial was to investigate theeffects of an individually tailored multicomponent home-basedrehabilitation program on mobility recovery compared with stan-dard care in community-dwelling men and women older than 60. Thestudy began soon after discharge home from the hospital.

Fig. 1. Flow chart of the study.

Methods

Study Design and Participants

A detailed description of the design of this randomized controlledtrial and participant recruitment (ISRCTN53680197) has been pub-lished earlier.19 Briefly, staff of the physiotherapy department at thelocal hospital reviewed the medical records of all the ambulatory andcommunity-dwelling men and women aged 60 and older arriving for

surgery for a femoral neck or pertrochanteric fracture (ICD code S72.0or S72.1) who were residents in the city of Jyväskylä or 1 of the 9neighboring municipalities. All patients fulfilling the inclusion criteriawere informed about the study (n ¼ 296). Of those, 161 expressedinterest in participation and were further visited by a researcher.Finally, 136 persons were recruited to the study. Patients sufferingfrom severe memory problems (Mini Mental State Examination[MMSE] <18), alcoholism, a severe cardiovascular or pulmonarycondition or some other progressive disease, or suffering from severedepression (Beck Depression Inventory [BDI-II] >29) were excludedfrom the study. After exclusions, 81 patients with hip fractureparticipated in the study. The study flow chart is shown in Figure 1.

The pretrial power calculations were based on previously pub-lished longitudinal data on mobility recovery after a hip fracture.6 Inthat study, 45.3% of the community-dwelling participants on averagewere independent in 3 mobility tasks (chair rising, walking 1 block,and negotiating stairs) before the fracture. One year after hip fracture,20.7% of the sample on average had regained their prefracture level ofmobility. Based on those numbers, a minimum of 44 participants wereneeded in each group (in total 88 participants) to detect the expecteddifference between the study groups at a level of significance of a ¼0.05 and b ¼ 0.20. All participants gave their written informed con-sent before participating in the study and gave permission to reviewtheir medical records. The ethical committee of the Central FinlandHealth Care District approved the study protocol. This trial wasregistered with Current Controlled Trials Ltd. (ISRCTN53680197).

A. Salpakoski et al. / JAMDA xxx (2014) 1e8 3

Baseline measurements were conducted as soon as possible afterdischarge home, 44 to 239 days after fracture. Follow-up measure-ments were organized at 3, 6, and 12 months after the baselinemeasurements. The researchers who collected the outcome measureswere blinded to group allocation. After the baseline measurements,participantswere randomized into the intervention (n¼40, promotingmobility and standard care, ProMo) and control (n ¼ 41, standard careonly) groups. A computer-generated group allocation list was gener-ated by a statistician not involved in the recruitment or data collection.Randomization by sex and by surgical procedure (internal fixation vsarthroplasty) was performed in blocks of 10. With the present samplesize (40 and 41), our design has a power of 78%.

We assume that the multifactorial mobility recovery processincludes pathways similar to those in the disablement process,according to which pathology leads to impairments, functional limi-tations, and, further, to disability.9 In this study, hip fracture is thepathology under investigation, leg extension power deficit (LEPasymmetry) represents level of impairment, functional balance (BergBalance Scale [BBS]) and lower extremity performance (Short PhysicalPerformance Battery [SPPB]) represent level of functional limitationlevel, and perceived difficulties in negotiating stairs represents thedisability level.

Primary Outcomes

Asymmetrical leg extension power deficitLeg extension muscle power (Watt) was measured with the Not-

tingham Leg Extensor Power Rig.1,20 Muscle power is a product offorce and velocity, and it refers to the ability to produce force quickly.Power was measured at baseline and 3 months later. For all partici-pants, the nonfractured leg was assessed first. The measurement wasrepeated until no further improvement occurred and the bestperformance was used in the analyses. The deficit between legs wascalculated according to the formula: LEP nonfractured leg � LEP frac-tured leg. In our laboratory, test-retest precision with a 1- to 2-weekinterval is 8% � 7%.21

Lower extremity performance and functional balanceThe main outcomes at 6 months were the SPPB and BBS. The SPPB

includes habitual walking speed, chair rise, and balance tests, all ofwhich are essential prerequisites for mobility. The maximum score is12, and higher scores indicate better performance.22 The BBS evaluatesan individual’s ability to perform different tasks related to the skills ofsitting down, standing up, reaching, turning around, looking over one’sshoulder, and 1-foot standing. Themaximum score is 56, higher scoresindicating better functional balance.23 Assessments of SPPB and BBSwere performed at baseline, and 3, 6, and 12 months thereafter.

Mobility disabilityMobility-related disability was assessed by perceived difficulties

in negotiating stairs. The participants were asked “Are you able tonegotiate 1 flight of stairs (5 stairs)?” The response categories were(1) able to manage without difficulty, (2) able to manage with somedifficulty, (3) able to manage with lots of difficulty, (4) able to manageonly with the help of another person, and (5) unable to manage evenwith help. Perceived difficulty in negotiating stairs before fracturewas elicited at the hospital, on average 10 � 5 days postfracture andat baseline and 3, 6, and 12 months thereafter.

Clinical Characteristics

Information on the presence of chronic conditions, use ofprescription medications, fracture date and status, and type and dateof surgery, obtained by a prestructured questionnaire, current

prescriptions, and medical records, was confirmed during a medicalexamination. Contraindications for participation in the physical per-formance assessments24 and acute conditions, assessed by bloodcount, C-reactive protein, and hemoglobin analyses, were evaluatedby a physician. Body height and body weight were measured. Bodymass index was calculated as weight (kg) divided by height (m)squared (kg/m2). Cognitive status was assessed at baseline by theMMSE,25 and depressive mood by the BDI-II.26 Musculoskeletal painin the low back, hip, or knee region on both sides of the body duringthe past week was assessed with the Visual Analog Scale (VAS).27 Asummary index was calculated from all 6 VAS variables. Use ofwalking aids, indoors and outdoors separately, was elicited with thequestion “Do you use a walking aid indoors/outdoors in everydaylife?” The response options were “yes,” “no,” and “I don’t walk out-doors.” Participants who reported not walking outdoors were cate-gorized as using a walking aid outdoors. To analyze potential adverseevents of the intervention in more detail, we collected self-reporteduse of emergency room services at 3, 6, and 12 months after thebaseline assessments.

Standard Care

All participants received standard care after discharge home. Atbaseline, information on standard care after discharge was collectedby an interview. Seventy percent of the participants (68% vs 71%, P ¼.813 of the intervention and control groups, respectively) received awritten home exercise program. None of the participants were fol-lowed for compliance with the home exercises and the program wasnot updated.19

Intervention

Participants in the intervention group received standard care andthe Promotion Mobility rehabilitation program (ProMo) aiming torestore their mobility and physical functional capacity to their pre hipfracture level. ProMo was an individually tailored yearlong physicalactivity and rehabilitation intervention targeted at all levels of thedisablement process and implemented in the participants’ homes. Itbegan on average within 1 week of the baseline measurements andincluded 5 to 6 home visits by a physiotherapist.

During the first face-to-face session, an evaluation and modifica-tion of environmental hazards and guidance for safe walking28 wasperformed. Participants’ falls in relation to self-efficacy and satisfac-tion with assistive devices for walking were also discussed. An indi-vidual nonpharmacological pain management evaluation took placein the second face-to-face session and was repeated at 3 and 6months. It included a structured interview on the location andduration of offending musculoskeletal pain and discussion onparticipants’ knowledge of pain-relief strategies.29

The individual progressive home exercise program, comprisingstrengthening and stretching exercises for the lower limb muscles,balance training, and functional exercises, was implemented duringthe second home visit. The program was updated 4 to 5 times with amore demanding protocol. The functional exercises included walking,reaching, and turning in different directions, and stair climbing. Thestrengthening exercises included knee extension and flexion, hipabduction, plantar flexion, chair rising, and squatting. Workload wasprogressively increased with resistance bands of 3 different strengths.The standing balance exercises included weight shifting from one legto the other; stepping forward, to the side, and backwards; andstanding on 1 leg. The level of challenge was increased during thetraining period by reducing hand support and the base of support.The strengthening and stretching exercises were advised to be done 3times a week on the same day. The balance, walking, and functional

A. Salpakoski et al. / JAMDA xxx (2014) 1e84

exercises were to be done 2 to 3 times a week on the same day (notthe same day as the strengthening and stretching exercises). The totalexpected number of exercises was 156 for strengthening andstretching and 104 to 156 for balance. For the functional exercises,which were to be performed for the first 12 weeks only, the expectedaverage number was 30. All participants kept a daily exercise diary.

Individual motivational face-to-face physical activity counseling30

took place after 3 months in the participants’ homes. Counseling wasdelivered as a one-off session followed by phone calls at 4 and 8months, and a face-to-face meeting at 6 months to encourageparticipants’ engagement in physical activity.

Statistical Methods

All analyses were performed according to the intention-to-treatprinciple. First, we analyzed the effect of the intervention on thesingle outcome measures, using generalized estimating equation(GEE) models with interaction terms (group � time). In a case ofmissing data, the GEE methodology uses maximum-likelihood esti-mation. Mean change parameters from prefracture to follow-up timepoints were calculated based on the GEE model coefficients, and thedelta-method was used to obtain SEs. A chi-squared distributed teststatistic was used in comparing the mean-change parameters acrossthe intervention and control groups.

Second, a 2-group path model based on a modification of thecross-lagged panel design was constructed to analyze mobility re-covery with the hypothesized causal paths among impairment (LEPdeficit), functional limitation (SPPB and BBS), and disability(perceived difficulty negotiating stairs). The models were adjusted forage and number of chronic diseases, as these factors can either speedup or slow down the disability process. We assumed that the acuterecovery phase after hip fracture would be similar in both groups.

Table 1Baseline Characteristics of the Intervention and Control Groups

Intervention

n

DemographicsAge, y, mean � SD 40 80Women, n (%) 40 31Body height, cm, mean � SD 40 16Body weight, kg, mean � SD 40 65Body mass index, kg/m2, mean � SD 40 25No. of chronic diseases, mean � SD 40 3 �Blood pressure, diastolic, mm Hg, mean � SD 39 78Blood pressure, systolic, mm Hg, mean � SD 39 15Hemoglobin, g/L, mean � SD 40 12C-reactive protein, mean � SD 40 8 �Walking aid, indoors, n (%) 40 28Walking aid, outdoors, n (%) 40 30Type of surgery, n (%) 40Internal fixation 19Hemiarthroplasty 15Total hip replacement 6 (

Time from surgery to baseline, wk, mean � SD 40 9.3Lower body pain, mean � SD 40 12

Mobility at baselineLEP deficit, W, mean � SD 32 16SPPB, score, mean � SD 40 5.8Functional balance, score, mean � SD 37 40Stair climbing, n (%) 40No difficulties 20Some difficulties 10Great deal of difficulties 3 (Manage only with help 4 (Unable to manage even with help 3 (

LEP, leg extension power deficit; SPPB, Short Physical Performance Battery.*P value from median test.

After 3 months of the ProMo intervention, we expected the pathwayobserved in the intervention group to be different from that amongthe controls. We estimated the same model, but different path co-efficients, for the study groups. We assumed that the power deficit atbaseline and 3 months would be associated with functional balanceand physical performance at 3 and 6 months, which in turn would beassociated with perceived difficulties in negotiating stairs at 6 and 12months. We also added pathways from each variable to theirrespective measures in the following wave (LEP deficit, baseline /

LEP deficit, 3 months; functional balance, 3 months / functionalbalance, 6 months; negotiating stairs, 6 months / negotiating stairs,12 months) to separate the amount of variance in each variable due tothe previous measurements. In addition, the model enabled a directpathway from impairment level (LEP deficit) to disability level(negotiating stairs).

Descriptive statistics and group comparisons were carried outusing the basic methods, and significance tests using IBM SPSS Sta-tistics for Windows (version 19.0; IBM Corporation, Armonk, NY). Thepath models were analyzed with Mplus (version 6, 2011; Muthén &Muthén, Los Angeles, CA). A specific script was written to computemean changes in the R programming environment (version 2.15.2)using the parameter estimate and covariance matrices extracted fromthe GEE models. A P value of less than .05 was considered statisticallysignificant. Compliance with the intervention was calculated usingthe following formula: [amount of performed exercises]:[expectednumber of exercises] � 100.

Results

Baseline characteristics are summarized in Table 1. Most of theparticipants were women and the average age was 80.0 � 7.1 years.Mean time from surgery to the baseline measurements was 9.2 � 3.0

Control P Value

n

.9 � 7.7 41 79.1 � 6.4 .249(78) 41 32 (78) .9530.9 � 8.9 40 160.3 � 9.1 .785.8 � 11.9 41 65.9 � 11.3 .968.3 � 3.6 40 25.6 � 3.9 .7102 41 3 � 2 .581

.4 � 12.7 41 79.5 � 10.3 .6713.4 � 26.4 41 144.1 �20.9 .044*7 � 13 41 130 � 13 .35111 41 7 � 9 .715*

(70) 41 24 (59) .356(75) 39 35 (85) .276

41 .719(48) 19 (46)(38) 18 (44)15) 4 (10)� 2.3 41 9.2 � 3.6 .889

0.4 � 124.7 41 92.4 � 112.7 .292

.5 � 25.9 38 22.6 � 25.2 .967*� 2.5 41 6.6 � 2.2 .115

.5 � 10.4 41 43.6 � 8.5 .972*41 .380

(50) 24 (59)(25) 12 (29)8) 3 (7)10) 2 (5)8) 0 (0)

Fig. 2. Schematic drawing of the study design showing timetable of the measurements related to the pathway of the disablement process. P values represent the statistical dif-ferences between the study groups obtained from the mixture path models. Model is adjusted for age and number of chronic diseases.

A. Salpakoski et al. / JAMDA xxx (2014) 1e8 5

weeks. At the baseline measurements, the only statistically significantdifference between the study groups was in systolic blood pressure.

Figure 2 and Table 2 illustrate the pathway of the disablementprocess in the context of our study. The association between BBS at 3months and LEP deficit at baseline differed between the study groups.The coefficient for the intervention group was not significant,whereas the effect was positive in the control group, indicating thatthe large baseline deficit was related to better balance values. Lessdifficulty in negotiating stairs at 6 and 12 months correlated signifi-cantly with better functional balance at 3 and 6 months in theintervention group (P < .001), but not in the control group. Thedifferences in the associations between the 2 groups were statisticallysignificant: functional balance at 3 months to negotiating stairs at 6months (P ¼ .011) and from 6 to 12 months (P < .001).

We also analyzed the model with SPPB as a functional limitationmeasure. The associations between negotiating stairs at 12 monthsand SPPB at 6 months were significantly different between the studygroups (P < .001). Less difficulty in negotiating stairs correlated with abetter SPPB score among the intervention group (P < .001) but notamong the controls (P ¼ .550).

No statistically significant group � time interaction was observedin BBS or SPPB (Figure 3). However, the ProMo intervention reduced

Table 2Associations Between Different Tiers of the Disablement Process in the Intervention and

Intervention(Estimate, SE

Main pathwayLEP deficit (baseline) / Functional balance (3 mo) (�0.033, 0.03LEP deficit (3 mo) / Functional balance (6 mo) (�0.008, 0.02Functional balance (3 mo) / Negotiating stairs (6 mo) (�0.168, 0.03Functional balance (6 mo) / Negotiating stairs (12 mo) (�0.288, 0.03

Secondary pathwaysLEP deficit (baseline) / LEP deficit (3 mo) (0.506, 0.191Functional balance (3 mo) / Functional balance (6 mo) (0.861, 0.073Negotiating stairs (6 mo) / Negotiating stairs (12 mo) (�0.618, 0.27LEP deficit (baseline) / Negotiating stairs (6 mo) (�0.003, 0.01LEP deficit (3 mo) / Negotiating stairs (12 mo) (�0.018, 0.01

Estimate, path coefficient; LEP, leg extension power deficit.

perceived difficulties in negotiating stairs compared with controls(Figure 4, Table 3). The intervention group also reported lower valuesfor perceived difficulties after 12 months of rehabilitation comparedwith the prefracture level.

One participant dropped out before the 3-month follow-up forpersonal reasons. During the first 6 months, 4 participants weresuspended for a medical reason. Two returned to the intervention(revision operation, femoral fracture), however 2 were unable tocontinue (pneumonia þ new hip fracture, and pulmonary embolism).Five participants were suspended during the last 6 months of thestudy (pubic bone fracture, urinary tract infection, cerebral infarction,cardiac failure, and sacrum strain fracture) and none of themreturned. One participant died before the final measurements for anonintervention-related medical reason.

During the first 6 months, compliance was 61% for the strength-ening, 53% for the stretching, 65% for the balance, and 69% for thefunctional exercises. During the last 6 months, the correspondingvalues were 39%, 37%, and 43%. All but 1 participant attended the firstface-to-face physical activity counseling session, and 90% respondedto the first, 88% to the second, and 83% to the third phone contact.

Among controls, 4 revision operations were performed but nonew hip fractures occurred during the study. Two participants in the

Control Groups

) PControl(Estimate, SE) P

Intervention-ControlP

7) .369 (0.084, 0.027) .002 .0092) .721 (�0.038, 0.021) .062 .3369) <.001 (�0.042, 0.031) .170 .0116) <.001 (�0.026, 0.025) .296 <.001

) .008 (0.582, 0.089) <.001 .723) <.001 (1.019, 0.171) <.001 .3748) .026 (0.630, 0.351) .073 .0060) .763 (0.002, 0.013) .906 .7903) .156 (�0.002, 0.013) .866 .400

A B

Fig. 3. Short-term outcome. Changes in the functional limitation variables for the intervention and control groups; P value for the group by the time interaction term is shown frombaseline to 6 months: (A) BBS; (B) SPPB. In addition, the group � time interactions from baseline to 12 months were P ¼ .185 for BBS and P ¼ .414 for SPPB.

A. Salpakoski et al. / JAMDA xxx (2014) 1e86

control group dropped out for personal reasons, one before the3-month measurements and the other before the 6-monthmeasurements.

During the first 3 months of the study, 3% of the participants in theintervention group and 13% in the control group needed emergencyroom services. Thereafter, 8% of the participants in the interventiongroup and 20% of controls reported that they had used emergencyroom services. No significant difference was observed between thestudy groups.

Discussion

The yearlong multicomponent and progressive home-basedrehabilitation program improved the mobility recovery of patientswith hip fracture over standard care. Compared with their prefracturemobility, the intervention group had less and the control group more

Fig. 4. Changes in perceived difficulties in negotiating stairs, number of participants at eacfracture, baseline, and 3, 6, and 12 months; P value for the group � time interaction term.

difficulty in negotiating stairs 12 months after discharge. In addition,in the intervention group, the correlations between the different tiersof function, as outlined in the disablement process model, followedthe expected time line and the logic of the model over time: lessdifficulty negotiating stairs at 6 and 12 months correlated with betterbalance at 3 and 6 months. A similar pattern was not evident in thecontrols. The fact that the intervention improved mobility, even whencompared with the prefracture level, supports the idea ofreenablement.

Long-term home-based rehabilitation interventions have rarelybeen conducted with patients with hip fracture, and none of theexisting studies has had a design or outcomes comparable with thosereported here. Although increases in physical activity or short-termmobility improvements have been reported,15,16 significant long-term advantages of interventions for mobility recovery over stan-dard care have not been found. An important aspect of our

h time point, and response categories for the intervention and control groups at pre-

Table 3Mean Changes (SE) and Differences Between Study Groups in the Parameters forNegotiating Stairs in the Generalized Estimating Equation Model

Intervention Control Intervention-Control

MeanChange

SE MeanChange

SE c2 (df ¼ 1) P Value

Prefracture, baseline 0.547 0.103 0.353 0.085 2.07 .151Prefracture, 3 mo 0.741 0.110 0.458 0.105 2.96 .085Prefracture, 6 mo 0.795 0.063 0.423 0.085 10.49 .001Prefracture, 12 mo 0.710 0.075 0.307 0.072 12.96 <.001

The SE estimate was obtained using the delta method.

A. Salpakoski et al. / JAMDA xxx (2014) 1e8 7

intervention was the progressive home exercise program combinedwith pain management, home environment assessment, and physicalactivity counseling. We acknowledge that more complex and physi-cally demanding functions need longer recovery time than the moreroutinize functions requiring less muscle strength and balance. Theimmediate prerequisites of walking are muscle power andstrength1,31 and the ability to balance the body in an upright posi-tion.32,33 When postural balance is good, walking requires minimalmuscle strength.34 However, patients with hip fracture may havepersistent asymmetry in their lower extremity strength and power1

and severely impaired postural balance, especially among thosewhose strength was rather low to start with.2 Consequently, in ourstudy, the physical training program first addressed strength andpower deficits, and then sought to translate the improvements inmuscle power into improved balance and mobility. Finally, the aimwas to train specific balance and mobility skills while also increasingphysiological reserve capacity over that required for walking, therebyalso enabling more demanding tasks, such as stair climbing, to beaccomplished without difficulty, as was observed after 12 months ofintervention. Our intervention, in line with an earlier intervention14

did not appear to produce short-term benefits in balance or SPPB.The potential explanation for this is that during the first few monthsafter hip fracture, the processes of natural recovery are so strong thatit may be difficult to induce measurable benefits exceeding these.However, our study clearly showed that the benefits becomeapparent later and that the rehabilitation was beneficial.

We conducted the present trial with the aim of establishing theoverall effectiveness of a multicomponent rehabilitation, rather thanits individual components. In this study, pain management, motiva-tional physical activity counseling, and assessment of environmentalbarriers helped reduce the complexity of the factors influencing themain pathway in the disablement process by solving problems thatwould otherwise mask the association between the different func-tional tiers. The need to understand and prevent disablement after ahip fracture has also been acknowledged earlier.4 Although werecognize the existence of other widely accepted frameworks forfunctioning and disability, such as International Classification ofFunctioning,35 we chose to design our study on the basis of the tiersof the disablement process model.9 The language and terms used todescribe the disablement process have been widely used by geron-tologists and proven to be useful in describing the consequences ofdisease and injury.

Our intervention was well tolerated considering that the partici-pants were people who had recently experienced a catastrophic eventand who also had many vulnerabilities. Adverse events or the use ofhealth care services did not differ between the intervention andcontrol groups. In all cases of suspension of the intervention formedical reasons, the causes were unrelated to the intervention.Compliance with the home exercises was high and for the physicalactivity counseling compliance was excellent. The present rehabili-tation program was organized with a minimal number of home visitsand without specific training equipment. We conclude that the

program is feasible in the home setting. Our sample comprisedcommunity-dwelling patients with hip fracture and excluded thoseliving in an institution. The sample is thus not completely represen-tative of all patients with hip fracture, which may be considered astudy limitation. Of all the potential patients identified in therecruitment area, the participation rate was 27%, which is comparableto or somewhat above that earlier reported in randomized controlledtrials among clinical populations.36,37 However, our study might beslightly underpowered, as we managed to recruit only 81 partici-pants. A larger sample size might have yielded clearer results. Thestrengths of this study include the study design, the theory-basedapproach to the assessments and the intervention, and investiga-tion of a topic that has high societal and human relevance. In addi-tion, the innovative statistical modeling of the results may beconsidered a strength of the study.

In conclusion, among older clinical populations, it is important tofocus on restoring and supporting the functional independence ofpersons who may already have disabilities. Restoring the mobility ofpatients with hip fracture makes it possible for them to continueparticipating in and contributing to the community. Catastrophicmobility-related disability among community-dwelling patientswith hip fracture can be reversed by an individually tailoredmulticomponent home-based rehabilitation that is both progressiveand long enough to build up the capabilities needed also for morecomplex tasks.

Acknowledgments

We thank the physiotherapists in the Central Finland Health CareDistrict for their indispensable work in the recruitment of the par-ticipants and data collection. We are also grateful to Hanna-MaijaLapinkero and Veera Farin for their considerable efforts regardingthe study arrangements and for assistance with the data collection, toMervi Matero for help in arranging the health examinations, tophysician Pirkko Jäntti for conducting the medical examinations, andto all who assisted in the data collection. The Gerontology ResearchCenter is a joint effort between the University of Jyväskylä and theUniversity of Tampere.

References

1. Portegijs E, Rantanen T, Kallinen M, et al. Lower-limb pain, disease, and injuryburden as determinants of muscle strength deficit after hip fracture. J BoneJoint Surg Am 2009;91:1720e1728.

2. Portegijs E, Sipila S, Alen M, et al. Leg extension power asymmetry and mobilitylimitation in healthy older women. Arch Phys Med Rehabil 2005;86:1838e1842.

3. Portegijs E, Sipila S, Pajala S, et al. Asymmetrical lower extremity power deficitas a risk factor for injurious falls in healthy older women. J Am Geriatr Soc2006;54:551e553.

4. Magaziner J, Hawkes W, Hebel JR, et al. Recovery from hip fracture in eightareas of function. J Gerontol A Biol Sci Med Sci 2000;55:M498eM507.

5. Shyu YI, Chen MC, Liang J, et al. Predictors of functional recovery for hipfractured elders during 12 months following hospital discharge: A prospectivestudy on a Taiwanese sample. Osteoporos Int 2004;15:475e482.

6. Visser M, Harris TB, Fox KM, et al. Change in muscle mass and muscle strengthafter a hip fracture: Relationship to mobility recovery. J Gerontol A Biol Sci MedSci 2000;55:M434eM440.

7. Laukkanen P, Era P, Heikkinen RL, et al. Factors related to carrying out everydayactivities among elderly people aged 80. Aging (Milano) 1994;6:433e443.

8. Laukkanen P, Heikkinen E, Schroll M, Kauppinen M. A comparative study offactors related to carrying out physical activities of daily living (PADL) among75-year-old-men and women in two nordic localities. Aging (Milano) 1997;9:258e267.

9. Verbrugge LM, Jette AM. The disablement process. Soc Sci Med 1994;38:1e14.10. Nagi SZ. An epidemiology of disability among adults in the United States.

Milbank Mem Fund Q Health Soc 1976;54:439e467.11. Handoll HH, Cameron ID, Mak JC, Finnegan TP. Multidisciplinary rehabilitation

for older people with hip fractures. Cochrane Database Syst Rev;2009:CD007125.

A. Salpakoski et al. / JAMDA xxx (2014) 1e88

12. Handoll HH, Sherrington C, Mak JC. Interventions for improving mobilityafter hip fracture surgery in adults. Cochrane Database Syst Rev;2011:CD001704.

13. Ftouh S, Morga A, Swift C, Guideline Development Group. Management of hipfracture in adults: Summary of NICE guidance. BMJ 2011;342:d3304.

14. Tinetti ME, Baker DI, Gottschalk M, et al. Home-based multicomponent reha-bilitation program for older persons after hip fracture: A randomized trial. ArchPhys Med Rehabil 1999;80:916e922.

15. Orwig DL, Hochberg M, Yu-Yahiro J, et al. Delivery and outcomes of a yearlonghome exercise program after hip fracture: A randomized controlled trial. ArchIntern Med 2011;171:323e331.

16. Ziden L, Frandin K, Kreuter M. Home rehabilitation after hip fracture. A ran-domized controlled study on balance confidence, physical function andeveryday activities. Clin Rehabil 2008;22:1019e1033.

17. Tsauo JY, Leu WS, Chen YT, Yang RS. Effects on function and quality of life ofpostoperative home-based physical therapy for patients with hip fracture. ArchPhys Med Rehabil 2005;86:1953e1957.

18. Gill TM, Baker DI, Gottschalk M, et al. A program to prevent functional declinein physically frail, elderly persons who live at home. N Engl J Med 2002;347:1068e1074.

19. Sipila S, Salpakoski A, Edgren J, et al. Promoting mobility after hip fracture(ProMo): Study protocol and selected baseline results of a year-long random-ized controlled trial among community-dwelling older people. BMC Muscu-loskelet Disord 2011;12:277.

20. Bassey EJ, Short AH. A new method for measuring power output in a single legextension: Feasibility, reliability and validity. Eur J Appl Physiol Occup Physiol1990;60:385e390.

21. Tiainen K, Sipila S, Alen M, et al. Shared genetic and environmental effects onstrength and power in older female twins. Med Sci Sports Exerc 2005;37:72e78.

22. Guralnik JM, Simonsick EM, Ferrucci L, et al. A short physical performancebattery assessing lower extremity function: Association with self-reporteddisability and prediction of mortality and nursing home admission.J Gerontol 1994;49:M85eM94.

23. Berg KO, Wood-Dauphinee SL, Williams JI, Maki B. Measuring balance in theelderly: Validation of an instrument. Can J Public Health 1992;83:S7eS11.

24. American College of Sports Medicine. ACSM’s guidelines for exercise testingand prescription. 8th ed. Philadelphia: Lippincott Williams & Wilkins; 2010.

25. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”. A practical methodfor grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12:189e198.

26. Beck AT, Steer RA, Ball R, Ranieri W. Comparison of Beck DepressionInventories-IA and -II in psychiatric outpatients. J Pers Assess 1996;67:588e597.

27. Huskisson EC. Measurement of pain. Lancet 1974;2:1127e1131.28. Mänty M, Sihvonen S, Hulkko T, Lounamaa A. Iäkkäiden henkilöiden katu-

mistapaturmat -opas kaatumisten ja murtumien ehkäisyyn, Vol 8. Helsinki,Finland: National Public Health Institute; 2006.

29. Barry LC, Gill TM, Kerns RD, Reid MC. Identification of pain-reduction strategiesused by community-dwelling older persons. J Gerontol A Biol Sci Med Sci 2005;60:1569e1575.

30. Rasinaho M, Hirvensalo M, Tormakangas T, et al. Effect of physical activitycounseling on physical activity of older people in finland (ISRCTN 07330512).Health Promot Int 2012;27:463e474.

31. Portegijs E, Sipila S, Rantanen T, Lamb SE. Leg extension power deficit andmobility limitation in women recovering from hip fracture. Am J Phys MedRehabil 2008;87:363e370.

32. Sihvonen S, Kulmala J, Kallinen M, et al. Postural balance and self-reportedbalance confidence in older adults with a hip fracture history. Gerontology2009;55:630e636.

33. Portegijs E, Edgren J, Salpakoski A, et al. Balance confidence was associatedwith mobility and balance performance in older people with fall-related hipfracture: A cross-sectional study. Arch Phys Med Rehabil 2012;93:2340e2346.

34. Rantanen T, Guralnik JM, Ferrucci L, et al. Coimpairments: Strength and balanceas predictors of severe walking disability. J Gerontol A Biol Sci Med Sci 1999;54:M172eM176.

35. Ustun TB, Chatterji S, Bickenbach J, et al. The international classification offunctioning, disability and health: A new tool for understanding disability andhealth. Disabil Rehabil 2003;25:565e571.

36. Chang BH, Hendricks AM, Slawsky MT, Locastro JS. Patient recruitment to arandomized clinical trial of behavioral therapy for chronic heart failure. BMCMed Res Methodol 2004;4:8.

37. Frobell RB, Lohmander LS, Roos EM. The challenge of recruiting patients withanterior cruciate ligament injury of the knee into a randomized clinical trialcomparing surgical and non-surgical treatment. Contemp Clin Trials 2007;28:295e302.