HIP FRACTURE INCIDENCE increases substantially

Transcription

1 246 REVIEW ARTICLE (META-ANALYSIS) Systematic Review of Hip Fracture Rehabilitation Practices in the Elderly Anna M. Chudyk, MSc, Jeffrey W. Jutai, PhD, Robert J. Petrella, PhD, Mark Speechley, PhD ABSTRACT. Chudyk AM, Jutai JW, Petrella RJ, Speechley M. Systematic review of hip fracture rehabilitation practices in the elderly. Arch Phys Med Rehabil 2009;90: Objective: To address the need for a research synthesis on the effectiveness of the full range of hip fracture rehabilitation interventions for older adults and make evidence based conclusions. Data Sources: Medline, PubMed, EMBASE, CINAHL, and the Cochrane Central Register of Controlled Trials were searched from 1980 to 2007 for studies published in English. The terms rehabilitation and hip fracture were exploded in order to obtain related search terms and categories. Study Selection: In the initial search of the databases, a combined total of 1031 articles was identified. Studies that did not focus on hip fracture rehabilitation, did not include persons over the age of 50 years, and/or did not include measures of physical outcome were excluded. Data Extraction: Only studies with an Oxford Center for Evidence-Based Medicine Levels of Evidence level of I (randomized controlled trial, RCT) or II (cohort) were reviewed. The methodologic quality of both types of studies was assessed using a modified version of the Downs and Black checklist. Data Synthesis: There were 55 studies that met our selection criteria: 30 RCTs and 25 nonrandomized trials. They were distributed across 6 categories for rehabilitation intervention (care pathways, early rehabilitation, interdisciplinary care, occupational and physical therapy, exercise, intervention not specified) and 3 settings (acute care hospital, postacute care/ rehabilitation, postrehabilitation). Conclusions: When looking across all of the intervention types, the most frequently reported positive outcomes were associated with measures of ambulatory ability. Eleven intervention categories across 3 settings were associated with improved ambulatory outcomes. Seven intervention approaches were related to improved functional recovery, while 6 intervention approaches were related to improved strength and balance recovery. Decreased length of stay and increased falls self-efficacy were associated with 2 interventions, while 1 intervention had a positive effect on lower-extremity power generation. Key Words: Elderly; Evidence-based medicine; Hip fractures; Rehabilitation by the American Congress of Rehabilitation Medicine HIP FRACTURE INCIDENCE increases substantially with age, rising from 22.5 and 23.9 per 100,000 population at age 50 years to and per 100,000 population by age 80 years, for men and women, respectively. 1 In 2000, about 12.5% of the Canadian population was age 65 years and older; this figure is expected to rise to 25% by ,3 In accordance with this aging population trend, the annual number of proximal femoral fractures is projected to increase from 23,375 in 1993/1994 to 88,124 in 2041, requiring an estimated 1.8 million acute care hospital days. 3 Even with successful repair, persons who sustain hip fracture exhibit high mortality and often demonstrate permanent disability and dependency. 4 Despite the high social costs associated with hip fractures, scientific knowledge about best practices is not applied systematically or expeditiously to rehabilitative services. The purpose of this review is to conduct a critical examination of the literature in the area of rehabilitation after hip fracture, to identify practices that have strong evidentiary bases as well as areas needing further research. When the term rehabilitation is applied to the hip fracture, it encompasses a wide variety of practices across time points after fracture. Reviews published to date have focused on studies within a domain of rehabilitation practice, such as physical therapy or multidisciplinary care, and have not compared strength of evidence across domains. This review examined all practices included in the hip fracture rehabilitation continuum, from clinical pathways (beginning with preoperative assessment) through inpatient and outpatient care, to home-based rehabilitation. METHODS The internet databases Medline, PubMed, EMBASE, CINAHL, and the Cochrane Central Register of Controlled Trials were searched using similar search strategies focusing on treatment outcomes in hip fracture rehabilitation. The terms rehabilitation and hip fracture were exploded in order to obtain related search terms and categories. The searches were limited to studies published in English between the List of Abbreviations From the Department of Epidemiology and Biostatistics, University of Western Ontario (Chudyk, Speechley); and Aging, Rehabilitation and Geriatric Care Program, Lawson Health Research Institute (Petrella), London, ON; Faculty of Health Sciences, University of Ottawa, Canada. Supported by the Canadian Institutes of Health Research (grant no. MIA79781), the Frederick Banting and Charles Best Canada Graduate Scholarships Master s Award, and the Joseph A. Scott Studentship in Aging and Mobility. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated. Reprint requests to Anna M. Chudyk, MSc, 801 Commissioners Rd E, London, ON, N6C 5J1, Canada, achudyk@uwo.ca /09/ $36.00/0 doi: /j.apmr ADL CI CP IADL IRF LOS NRT OR OT PRT PT RCT SNF activity of daily living confidence interval clinical pathway instrumental activity of daily living inpatient rehabilitation facility length of stay nonrandomized trial odds ratio occupational therapy progressive resistance training physical therapy randomized controlled trial skilled nursing facility

2 HIP FRACTURE REHABILITATION IN THE ELDERLY, Chudyk 247 years 1980 and 2007 and carried out in human subjects. Both prospective and retrospective studies were considered, as were studies that used either experimental or nonexperimental designs, but unpublished data or studies were not included. Reference Manager 11.0 was used for database management. In the initial search of the databases, a combined total of 1031 articles were identified. After combining the database results using Reference Manager 11.0, the abstracts of the articles were reviewed, and studies that did not contain hip fracture patients as their study population, did not perform separate analyses for hip fracture subjects, or did not focus on the effects of physical rehabilitative practices on hip fracture outcomes were excluded. A second reviewer confirmed the decision to exclude an abstract. The abstracts of the remaining studies were re-reviewed, and hard copies of studies that did not contain an abstract, or whose abstract did not give a clear insight into the nature of the study, were obtained. Articles that did not contain measures that focused on physical outcomes, contained a study population younger than 50 years old, or focused on dietary interventions were further excluded. Bibliographies of applicable reviews and meta-analyses were checked in order to identify other relevant articles that were not found through the database searches. Studies were assigned the following evidence levels, determined in accordance with an adapted version of the Oxford Centre for Evidence-Based Medicine Levels of Evidence: (I) RCT, (II) cohort, (III) case control, (IV) case series, and (V) expert opinion. Articles whose evidence level was below (II) were not included in this review. The methodologic quality of both experimental and nonexperimental studies was assessed using the Downs and Black 5 checklist. The tool was modified slightly for use in this review. Specifically, the scoring for question 27 dealing with statistical power was simplified to a choice of awarding either 1 point or 0 points, depending on whether there was sufficient power to detect a clinically important effect. Downs and Black 5 score ranges were grouped into the following 4 quality levels: excellent (26 28), good (20 25), fair (15 19), and poor ( 14). Although this approach provides an evidence level and quality level for each study, it does not take account of the number of studies and the consistency of results across different studies of the same type. Accordingly, the following strength of evidence levels were adapted from methods used by the authors of the Evidence- Based Review of Stroke Rehabilitation (appendix 1). In total, 55 studies were included in this review (tables 1 6). They are organized into 6 different rehabilitation intervention approaches, including a category for studies in which the approach was not well defined. The tables summarizing the studies and their outcomes are also organized by intervention approach. RESULTS The selected studies investigated the effectiveness of hip fracture rehabilitation using 6 different approaches, in 3 types of setting, as summarized in appendix 2. Clinical Pathway A CP is a multidisciplinary tool designed to promote a consistent approach to outcome-focused care, within a predefined time, by reducing unnecessary variation in practice. It describes routine interventions for a group of patients with similar needs and includes the expected outcomes at each step. Within an acute care setting intensive occupational therapy and/or physical therapy exercises. Improved functional recovery 6,7 and decreased LOS 6-8 Evidence Level 2a (Moderate)* More favorable discharge destination 6 Evidence Level 2b (Limited) Four studies 6-9 investigated the roles of multidisciplinary CPs whose focus on PT and/or OT exercises was more intensive than that of the standard groups used for comparison. Cameron et al 6 conducted an RCT in which different care paths were established for nursing home and nonnursing home patients, with care extending posthospital discharge and patients receiving twice-daily PT. Although there was no difference between the 2 groups in functional recovery on discharge or 4 months postfracture, the study found that among subjects with limited pre-existing disability, subjects in the intervention group had superior functional recovery 2 weeks and 1 month after surgery. In the randomized controlled study by Swanson et al, 7 a multidisciplinary approach characterized by early surgery and mobilization, as well as intensive, twice-daily PT and daily OT, was associated with higher mean functional levels at discharge, as well as a shorter LOS in the hospital. Koval et al 8 found that the initiation of a multidisciplinary CP in which OT and PT were initiated on the first postoperative day, and carried out at least twice daily, did not result in differences in recovery of ambulatory ability, but did result in shorter LOS, compared with data gathered before initiation of the pathway. Jette et al 9 conducted a cohort study investigating the effects of a CP in which daily rehabilitation exercises were organized by a multidisciplinary team. Despite the increased focus on rehabilitative exercises, subjects in the CP displayed patterns of 12-month recovery in physical and social functioning that were similar to those of subjects who had received standard care. Within an acute care setting early mobilization. Improved functional recovery 10 Evidence Level 2c (Weak) Discharge location and LOS 10,11,13 Conflicting Evidence Four cohort studies with large sample sizes and broad inclusion criteria evaluated the effects of implementation of evidence-based CPs, with early mobilization as part of their pathway procedures. No differences were found before and after the implementation of the CPs in terms of hospital readmission 11,13 and mortality 11,12 ; moreover, Beaupre et al 10 noted differences in functional recovery only after accounting for levels of social support among both groups. Early Supported Discharge Within an acute care setting. Improved falls efficacy 14 and short-term functional recovery 14,15 Evidence Level 2b (Limited) LOS 14,16,17 Conflicting Evidence Results from 2 NRTs 8,9 of fair quality show that CPs with intensive OT and/or PT exercise do not have an impact on discharge destination Evidence Level 2c (Weak). Results from 1 RCT 7 of good quality show that there is Limited Evidence (Level 2b) that CPs with intensive OT and/or PT exercise do not have an impact on mortality, while results from 1 nonrandomized trial 8 of fair quality show that there is Weak Evidence (Level 2c) that CPs in this category are related to decreased mortality. In subjects undergoing care in a CP with early mobilization as part of its pathway procedures, compared with subjects with low social support receiving standard care. Results from 1 NRT of fair quality 17 show that early supported discharge may be associated with decreased functional recovery at hospital discharge, with no differences present after discharge from home care Evidence Level 2c (Weak).

3 248 HIP FRACTURE REHABILITATION IN THE ELDERLY, Chudyk Table 1: Key Features of the Studies Selected for Evaluation Clinical Pathway Clinical Pathway Cameron et al 6 Evidence levels: RCT, 22 CG: standard care Swanson et al 7 Evidence levels: RCT, 20 CG: standard care PT daily Koval et al 8 trial, 19 CG: care prior to pathway initiation Jette et al 9 trial, 14 CG: standard care early mobilization, exercises daily Beaupre et al 10 trial, 17 CG: care prior to pathway initiation Beaupre et al 11 trial, 18 CG: care prior to pathway initiation March et al 12 trial, 17 CG: standard care Choong et al 13 trial, 19 CG: standard care Intervention/intensity: early mobilization; PT/ twice a day; mobility training; discharge planning; PT continuing on hospital discharge Intervention/intensity: PT twice daily; daily OT and social worker visits; discharge planning; multidisciplinary team; early mobilization Intervention/intensity: ambulation training and strengthening exercises (via OT and PT) minimum twice daily; discharge planning Intervention/intensity: standard care plus breathing exercises and progressive daily exercises; discharge planning; individualized education; multidisciplinary team Intervention/intensity: diet/high fiber, high calorie; early ambulation; lower-extremity exercises twice daily Intervention/intensity: multidisciplinary team; treatment components based on published evidence where possible Intervention/intensity: early mobilization; protein supplementation if needed; active and early rehabilitation Intervention/intensity: early mobilization; discharge planning; documentation specifying medical responsibilities by discipline and time frame Functional recovery: among subjects with limited pre-existing disability, IG had * Barthel Index scores 2wk and 1mo after injury (CI, and ). LOS: + in IG (CI,.63.98d). Discharge destination: + IG living outside nursing homes preinjury discharged to nursing homes or died (CI, 2.5% to 30.5%). All persons from nursing homes returned there after discharge. Hospital readmission: N 4mo after injury. LOS: + in IG (17 vs 24d). Mortality: N over 6mo. Functional recovery: * mean MBI functional levels at discharge in IG. Dynamic balance: N at discharge. LOS: + in IG (13.7 vs 21.6d). Mortality: + inpatient mortality (OR 3.7; CI, ) and 1-y mortality (OR 1.7; CI, ) in IG. Discharge location and decline in ambulatory abilities: N. Self-reported physical and social function (shortened Functional Status Index), postfracture living situation: N. Functional recovery: in baseline, 3-mo or 6-mo MBI scores. New institutionalizations, median total LOS: N 6mo postfracture. After risk adjustment, CG subjects with low social support were predicted to have lower functional outcome and increased odds of institutionalization compared with CG subjects with high social support and IG subjects with high or low social support. Mortality: N during the 6mo postfracture. LOS, readmissions and discharge location: N in LOS or readmissions to hospitals or rehabilitation. * IG discharged home (51 vs 26) or to long-term care (183 vs 166). Mortality, and new nursing home placement 4mo postdischarge: N. LOS: N for nonnursing home patients (9 vs 10d), but + for nursing home patients in the IG (5 vs 6d). * Time spent in rehabilitation for IG (26 vs 21d). LOS: + in IG (6.6 vs 8.0d). Hospital readmission: N. NOTE. Number listed under evidence level refers to study s score on the Downs and Black checklist. 5 Abbreviations: CG, control group (subjects); Dynamic balance, time taken to walk 10m and back; IG, intervention group (subjects); MBI, Modified Barthel Index; *, significantly greater/significant improvement; +, significantly decreased/significantly less; N, no significant difference; where P.05 considered significant between-group difference unless otherwise noted.

4 HIP FRACTURE REHABILITATION IN THE ELDERLY, Chudyk 249 Table 2: Key Features of the Studies Selected for Evaluation Early Supported Discharge Early Supported Discharge Crotty et al 14 Evidence levels: RCT, 24 CG: standard multidisciplinary care and in-hospital rehabilitation Crotty et al 15 Evidence levels: RCT, 18 CG: See 14 Van Balen et al 16 trial, 17 CG: standard care median hospital stay of 18 d Jaglal et al 17 trial, 15 CG: secondary data received standard care Intervention/intensity: discharged from acute care within 48h of randomization; mean number of 13.6 multidisciplinary team home visits before discharge from intervention See 14 Intervention/intensity: early discharge to a special rehabilitation ward in a nursing home; median hospital stay of 11d Intervention/intensity: aimed to discharge subjects on postoperative day 5 to an enhanced 7-d service plan. Ambulation and SF-36: N 4mo after randomization on TUG and SF-36. Balance (ABC, BBS), falls efficacy (FES, LHS): * in FES scores in IG at 4mo postrandomization. N Otherwise. Functional recovery: * in MBI scores from baseline to 4mo postrandomization in IG. LOS: + in IG during acute care, but *, in rehabilitation overall. Follow-up to, 14 12mo postrandomization: Functional recovery and ambulation: N. Both groups achieved * in MBI and TUG test scores and + SF-36 physical scores. Ambulation (RAP), quality of life (NHP, DCOOP), mortality, total LOS in an institution, and hospital readmissions: N. Four months postfracture, 1 of 5 of subjects regained their previous ADLs level. Discharge destination: 1mo postfracture, * number IG were in a nursing home. N 4mo postfracture. Ambulation (TUG) and functional recovery (FIM): * hospital discharge FIM and TUG scores in CG. N In home-care discharge TUG and FIM scores. Hospital LOS: + in IG. NOTE. Number listed under evidence level refers to study s score on the Downs and Black checklist. 5 Abbreviations: ABC, Activities Specific Balance Confidence Scale; BBS, Berg Balance Scale; CG, control group; DCOOP, Dartmouth COOP Functional Health Assessment Charts; FES, Falls Efficacy Scale; IG, intervention group; LHS, London Handicap Scale; MBI, Modified Barthel Index; NHP, Nottingham Health Profile; RAP, Rehabilitation Activities Profile; SF-36, Medical Outcomes Study 36-Item Short-Form Health Survey; TUG, Timed Up and Go test; *, significantly greater/significant improvement; +, significantly decreased/significantly less; N, no significant difference; where P.05 considered significant between-group difference unless otherwise noted. The effects of early supported discharge on outcomes in patients with hip fracture were investigated by 2 RCTs and 2 nonrandomized trials. A trial by Crotty et al 14 and its 12-month follow-up 15 compared conventional care to an intervention consisting of early discharge to home-based therapy. Patients in the intervention group had greater increased functional recovery and falls efficacy 4 months after the start of the trial, but at 12 months, there was no difference between the groups in functional recovery or mobility. Van Balen et al 16 found that early discharge to a rehabilitation ward in a nursing home had no benefit over standard care in terms of mortality rates, ambulation, total LOS in an institution, and health-related quality of life. A pilot study was carried out by Jaglal et al 17 in which the intervention aimed to discharge patients out of acute care on postoperative day 5, after which they received an enhanced 7-day service plan. At hospital discharge, patients in the intervention group performed worse on measures of ambulation and functional recovery than subjects who had received standard care, but after discharge from the home-based rehabilitation program, differences in ambulation and functional recovery were no longer present between the groups. Interdisciplinary Care All 11 articles, of which were RCTs 18,19,21-27 and2of which were nonrandomized trials, 20,28 exploring the benefits of interdisciplinary care contained sample sizes of over 100 subjects and a control group that received standard care. However, the exact nature of standard care varied between the studies, and the descriptions of what standard care constituted were often lacking. All but 3 studies 21,22,26 involved an intervention group receiving postoperative management of hip fracture from a geriatrician and an orthopedic surgeon, with support from a multidisciplinary team composed of varying health and community-support professions; in the other 3 studies, the intervention group consisted of an interdisciplinary team monitored by a geriatrician. Within an acute care setting joint postoperative management. Ambulatory outcomes, 24,27,28 LOS, 18-20,23,24,27,28 functional recovery 18,20,23,24,27 Conflicting Evidence Of the studies involving joint postoperative management, only 3 found some evidence of increased functional 23,27 and/or ambulatory 27,28 recovery in the intervention group compared with the control group. At discharge, Zuckerman et al 28 found that the intervention group had significantly better ambulatory ability, while Kennie et al 23 reported that the intervention group had better functional recovery. Although significantly more subjects in the intervention group reported functional and ambulatory recovery 3 months postfracture in the study by Vidan et al, 27 these between-group differences in functional outcome were no longer present 6 and 12 months postfracture. Within an acute care setting interdisciplinary team monitored by a geriatrician. Improved ambulation, 26 quadriceps strength 26 Evidence Level 2c (Weak)

5 250 HIP FRACTURE REHABILITATION IN THE ELDERLY, Chudyk Table 3: Key Features of the Studies Selected for Evaluation Interdisciplinary Care Interdisciplinary Care Fordham et al 18 Evidence levels: RCT, 17 CG: standard care from the orthopedic specialty Gilchrist et al 19 Evidence levels: RCT, 17 CG: standard care on an orthopedic ward Hempsall et al 20 Evidence levels: nonrandomized trial, 23 CG: standard care Huusko et al 21 Evidence levels: RCT, 17 CG: postoperative discharge to standard hospital care (median of 4.5 PT sessions/wk) Huusko et al 22 Evidence levels: RCT, 17 CG: standard hospital postoperative care Kennie et al 23 Evidence levels: RCT, 22 CG: standard care on an orthopedic unit Naglie et al 24 Evidence levels: RCT, 25 CG: standard care Reid and Kennie 25 Evidence level: RCT, short follow-up to 23 so not rated further CG: see 23 Shyu et al 26 Evidence levels: RCT, 19 CG: standard care amount of PT depended on personal insurance policies, no provisions made for home rehabilitation Intervention/intensity: collaborative care between orthopedic surgeons and geriatricians; begun a mean of 18 d postfracture fixation Intervention/intensity: postoperative collaborative care between orthopedic surgeons and geriatricians; PT; OT; social worker Intervention/intensity: postoperative collaborative care between orthopedic surgeons and geriatricians Intervention/intensity: postoperative care from a geriatric team; PT/median 7.3 sessions a week while in hospital (for about 2wk) plus 10 home visits postdischarge; daily ADLs exercises with nurses; OT as needed Intervention/intensity: see 21 Intervention/intensity: collaborative postoperative care between orthopedic surgeons and geriatricians; PT; OT Intervention/intensity*: interdisciplinary care from an internist-geriatrician Intervention/intensity: see 23 Intervention/intensity: geriatric consultation service on admission preoperation; PT/ daily for 4d, followed by assessment at 1wk, 3wk, and 3mo postdischarge; rehabilitation physician/1 visit; geriatric nurse/8 home visits over 3mo postdischarge LOS, functional recovery (author s test, therapists prognosis), and return to community at discharge: N. Mortality, LOS, and discharge location (for patients admitted from home): N. Mortality, change in residential status, and functional recovery (CAPE): N at discharge and 6 mo. The median change in CAPE score from hospital admission to 6-mo follow-up was 1 in both groups. LOS: + in IG by 9.5d (CI, d). Functional recovery (Katz Index, Lawton, and Brody IADLs dependency scale): N from baseline to 3mo and 12mo postsurgery. Overall, both groups had a decline in independence in IADLs from baseline to 1y (score change of 1, IQR 2.0 in the IG and 1, IQR 3.0 in the CG), and no median change in independency in ADLs (score change of 0, IQR 1.0 in both groups). LOS and mortality: N. *LOS: N among patients with normal scores or with severe dementia. + in IG who had mild or moderate dementia, when compared to CG with mild or moderate dementia. Mortality: N. Change in residence: 3mo postsurgery, * number of IG with mild or moderate dementia still living independently. N 12mo postsurgery. Functional recovery (Katz index): * independence in IG at discharge. LOS: + median stay in IG (CI, 2 25d). Return to community: IG had + discharges to nursing care and * number of discharges to patients own homes. LOS: * in IG (22.6 vs 20.9d) in hospital, but N in days spent in an institution over a 6-mo period. Mortality, proportion of patients alive with no decline in ambulation, transfers, or residential status, 6-mo health care utilization, functional recovery (MBI, Lawton and Brody IADLs): N 3 and 6mo after hip surgery. Baseline versus 1 y later: Functional recovery (Katz index): * independence in IG. Quality of life (life satisfaction index): N. Discharge location: * number IG had better or no change in their place of residence. Strength (hip flexion, peak forces of fractured limb s quadriceps), ambulation: * number IG recovered their previous walking ability and had a better hip flexion ratio by 1mo, and recovered their previous walking ability and generated higher peak quadriceps forces 3mo after hospital discharge. LOS, readmission rates, institutionalization, falls, mortality: N.

6 HIP FRACTURE REHABILITATION IN THE ELDERLY, Chudyk 251 Table 3 (Cont d): Key Features of the Studies Selected for Evaluation Interdisciplinary Care Interdisciplinary Care Vidan et al 27 Evidence levels: RCT, 20 CG: standard orthopedic care Zuckerman et al 28 Evidence levels: nonrandomized trial, 16 CG: standard care Intervention/intensity: joint postoperative orthopedic and geriatric care; geriatrician visits daily; therapy planned by a rehabilitation specialist; social worker visits Intervention/intensity: postoperative PT twice per weekday and once each weekend; OT before discharge; case manager as a resource person for 3 to 6mo postdischarge Functional recovery (Lawton and Brody IADLs): * in ADLs in IG 1 and 3mo after discharge, but N in IADLs. Quality of life (SF-36): 3mo postdischarge, * in SF-36 measures of bodily pain, vitality, social functioning, general mental health, physical functioning, and role limitations because of physical health problems in IG. LOS: N. In-hospital mortality: + in IG. Functional recovery (modified Katz index) and ambulation (FAC*): * in IG 3mo postfracture; N 6 and 12mo postfracture. In-hospital mortality and LOS: N. Ambulation: * number of IG able to ambulate independently with or without assistive devices and/or minimal noncontact supervision, and able to ambulate for greater distances at discharge. N In need of ambulatory aids. Discharge destination: N. NOTE. Number listed under evidence level refers to study s score on the Downs and Black checklist. 5 Abbreviations: CAPE, Clifton Assessment Procedure for the Elderly; CG, control group; FAC, Functional Ambulation Classification; IG, interventions group; IQR, interquartile range; SF-36, Medical Outcomes Study 36-Item Short-Form Health Survey; *, significantly greater/ significant improvement; +, significantly decreased/significantly less; N, no significant difference; where P.05 considered significant between-group difference unless otherwise noted. *Subjects placed into 1 of 4 categories according to their score on the Mini-Mental State Examination administered 10 d postsurgery: score of 0 to 11 was defined as suspected severe dementia; 12 to 17 defined as suspected moderate dementia; 18 to 23 defined as suspected mild dementia; 24 to 30 defined as normal. Significantly more IG than CG subjects received care from an OT (140 vs 10), social worker (140 vs 75), or dietician (34 vs 19). Among subjects receiving PT and OT, those in the IG received significantly more hours of care per patient (14.2 vs 5.7 and 10.8 vs 3.3, respectively). Shorter LOS 22 and short-term improvements in residential status, 22 Evidence Level 2c (Weak) Functional recovery of ADLs 21,26 Conflicting Evidence An RCT by Huusko et al 21 found that intensive geriatric rehabilitation resulted in improved IADLs recovery at 3 months, but not 1 year, compared with standard care. In a subanalysis of this study, Huusko et al 22 found that patients with mild or moderate dementia who were treated by the pathway had a shorter total LOS than the control group, and more were living independently at 3 months; however, at 12 months postsurgery, differences in independent living were no longer present. In a pilot study carried out by Shyu et al, 26 a geriatric consultation service in addition to discharge planning and rehabilitation resulted in better ADLs recovery, ambulation, quadriceps strength, and quality of life 3 months postdischarge compared with standard care. Occupational Therapy/Physical Therapy Within an acute care setting high-frequency occupational therapy/physical therapy. Increased likelihood of early ambulation, 29 functional recovery 30 Evidence Level 2c (Weak) Hoenig et al 29 conducted a study exploring the timing of surgical intervention and frequency of OT/PT on hip fracture Among individuals with mild-to-moderate dementia. Based on within-group differences. patient outcomes. When frequency of OT/PT was looked at as the single independent variable, receiving more than 5 sessions a week of OT/PT was determined to be a significant predictor of early ambulation. Early ambulation, in turn, was associated with return home on discharge. An outcome study by Jones et al 30 found that patients undergoing intensive PT (1.5h/d, 5/wk) and OT (1h/d, 5d/wk) experienced significant within-group increases in total FIM scores and motor FIM scores at discharge compared with baseline. Within an acute care setting additional occupational therapy combined with physical therapy. Improved short-term functional recovery of ADLs 31 Evidence Level 2b (Limited) Improved ambulation 32,33 Evidence Level 2c (Weak) Three studies investigated the role of additional OT, combined with PT, in the functional recovery of patients with hip fracture in acute care. Hagsten et al 31,32 conducted 2 studies investigating whether receiving OT in addition to PT had an impact on outcomes that was greater than the impact of receiving PT and conventional nursing care. Despite evidence of some short-term improvements in ADLs performance in intervention group subjects, after 2 months, the mean percentage of recovered ADLs and IADLs ability was almost 100% for subjects in both the intervention and control groups. 31 In a study by Roberts et al, 33 the implementation of an integrated care pathway in which patients received twice as much OT as the standard care pathway was related to a statistically significant increase in ability to walk alone at discharge.

7 252 HIP FRACTURE REHABILITATION IN THE ELDERLY, Chudyk Table 4: Key Features of the Studies Selected for Evaluation OT/PT OT/PT Hoenig et al 29 trial, 18 CG: less than 5 sessions a wk of OT/PT Jones et al 30 trial, 18 CG: none Hagsten et al 31 Evidence levels: RCT, 20 CG: standard PT, standard care from nursing staff Hagsten et al 32 Evidence levels: RCT, 16 CG: standard postoperative rehabilitation from nurses Roberts et al 33 trial, 19 CG: received one half OT contacts of the IG Koval et al 34 trial, 17 CG: subjects discharged to an outside rehabilitation facility Petrella et al 35 trial, 19 CG: none Mendelsohn et al 36 trial, 16 CG: none Intervention/intensity: highfrequency PT and OT/more than 5 sessions a wk of PT/OT Intervention/intensity: begun postoperatively PT/1.5h/d, 5 d/ wk; OT/1h/d, 5d/wk (both until discharge) Intervention/intensity: begun 3 to 4d postoperation standard PT; OT/45 60min each weekday morning until discharge Intervention/intensity: begun 3 4d postoperation, OT/45 60 minutes each weekday until discharge home plus home assessment prior to discharge Intervention/intensity: twice as much OT as controls (same amount of PT) postoperatively until discharge Intervention/intensity: intensive inpatient rehabilitation: PT 2h/d; OT 1h/d; family involvement in therapy sessions encouraged; discharge planning Intervention/intensity: inpatient OT and PT sessions after acute care/ 80-min sessions 3 to 5 times a wk for 3 6wk; inpatient referral services as needed Intervention/intensity: inpatient OT and PT sessions after acute care/ 45 60min each weekday for 4wk LOS and return to the community: N. Ambulation: 1.76 fold (CI, ) increased likelihood of earlier ambulation in IG. Functional recovery (FIM, motor FIM subscale): within-group * in total FIM score and motor FIM score between admission and discharge. Efficacy and efficiency (MRFS efficacy score): within-group * in FIM MRFS efficacy and efficiency, as well as motor FIM MRFS efficacy and efficiency scores between admission and discharge. Functional recovery (Klein-Bell scale; modified DRI with 2 items added to investigate fear of performing ADLs/IADLs, and pain level): at discharge, IG had * ADLs abilities in the areas of dressing, toilet visits, and bathing/hygiene as measured by the Klein-Bell scale. N At 2mo. Quality of life (SWED-QUAL): N at discharge and 2-mo follow-up; at 2mo, CG generally regained self-reported prefracture status in 6 of 12, and IG in 10 of 12 SWED-QUAL subscales. Functional recovery (modified DRI): N overall, 2mo postdischarge, but there was a * in IG in moving around indoors. LOS: N in mean total hospital LOS, but IG had * mean LOS in the orthopedic unit by 6.5d (CI, ). Ambulation: 1.6 fold (CI, ) increased likelihood of being able to walk alone at discharge from orthopedic unit in IG. Discharge destination, 30-d mortality, readmission within 30d of discharge: N. Functional recovery (modified Katz index; modified Lawton and Brody IADLs; return to prefracture level of home assistance) and ambulation (recovery of prefracture walking ability): 3mo postoperation, + recovery of prefracture independence levels in BADLs in IG. Otherwise, N 3, 6, and 12mo postoperation. Hospital discharge status, discharge destination, mortality: N at 3-mo, 6-mo, and 12-mo follow-ups. LOS: since initiation of the rehabilitation program, + LOS for acute care, but * in total hospital LOS. Functional recovery (FIM) and quality of life (Vitality Plus Scale): within-group * from admission to discharge. Falls (FES), and balance (ABC): within-group * from admission to discharge. Absolute angular error: within-group * from admission to discharge in the injured knee and hip. Balance (BBS), functional recovery (FIM), ambulation (gait speed, TUG), strength (30-s chair stand): within-group * from admission to discharge.

8 HIP FRACTURE REHABILITATION IN THE ELDERLY, Chudyk 253 Table 4 (Cont d): Key Features of the Studies Selected for Evaluation OT/PT OT/PT Mitchell et al 37 Evidence levels: RCT, 19 CG: received the same PT as the IG, but no QT Lauridsen et al 38 Evidence levels: RCT, 18 CG: PT/15 30-min sessions every weekday Tinetti et al 39 trial, 14 CG: none Tsauo et al 40 Evidence levels: RCT, 18 CG: at hospital discharge, instructed to continue practicing PT program received during acute care Binder and Brown 41 Evidence levels: RCT, 17 CG: 6 mo of low-intensity home exercise focusing on flexibility and prohibited from engaging in weight training Intervention/intensity: begun a median of 15d postsurgery PT 20min/d, 5 d/wk for 6wk; bilateral QT 2 times/wk with progressively increased intensity for 6wk Intervention/intensity: rehabilitation ward PT/2-h sessions 3 times/wk for a median of 14 sessions Intervention/intensity: home-based PT progressive exercises/ 3 times/wk at first, decreasing to 1 3 times/mo, median of 12wk; functional therapy/1 2 hourly sessions/wk for a median of 5 sessions Intervention/intensity: progressive home-based PT after acute care/ 1 session/wk over span of 12wk Intervention/intensity: after completion of standard PT/PT aerobic exercise for the first 3 months, then addition of progressive resistance training for the next 3mo/45 90-min sessions, 3 times/wk, for 6mo Leg extensor power (Nottingham Power Rig), ambulation (Elderly Mobility Scale; gait speed, TUG), and balance (FR): * in IG in all but TUG and gait speed at the end of the intervention and 10wk later. Functional recovery (Barthel Index): * in general mobility in IG at the end of the intervention. N Ten weeks after end of intervention. Quality of life (Nottingham Health Profile): * in energy 10wk after the end of intervention in IG; otherwise, N on any NHP subscores. Thirty-seven (24 IG and 13 CG) of 88 patients discontinued the intervention prematurely. Duration of physical rehabilitation until patient was able to perform 5 functional capacity objectives unaided: N. LOS: N when an intention-to-treat analysis was performed; + in IG when a per protocol analysis was performed. Balance (modified BBS), ambulation (5 items from the gait component of POMA), upper-extremity and lower-extremity strength (1 RM for triceps and knee extensors): within-group * from baseline at 6-mo follow-up. Functional recovery (ability to perform stairs, transfers, outdoor gait): N (within-group) from baseline at 6-mo follow-up. *Hip ROM (goniometer), hip and knee strength (dynamometer), walking velocity: N 1, 3, and 6mo after acute care discharge. Harris Hip Score: * in Harris Hip scores and Harris pain scores in IG 1 and 3mo after acute care discharge. * in Harris total-pain scores in IG 3 and 6mo after acute care discharge. Health-related quality of life (WHOQOL-BREF): * psychologic domain scores at 1 and 3mo and * physical health domain scores at 3mo in IG; otherwise, N. Functional recovery (PPT, OARS): * PPT score in IG between baseline and 3 and 6 mo. N OARS from baseline to 6mo. Functional Status Questionnaire, SF-36 Social Function subscale: * in IG between baseline and 3mo. Fat-free mass and bone mineral density: N from baseline to 6 mo. Strength of fractured side (maximum strength for knee extension), balance (BBS; single-limb stance), ambulation (fast walking speed), SF-36 physical function subscale, modified Hip Rating Questionnaire: * in IG between baseline and 3 and 6mo. NOTE. Number listed under evidence level refers to study s score on the Downs and Black checklist. 5 Abbreviations: ABC, Activities Specific Balance Confidence Scale; BADLs, basic activities of daily living; BBS, Berg Balance Scale; CG, control group; DRI, Disability Rating Index; FES, Falls Efficacy Scale; FR, Functional Reach Test; IG, intervention group; MRFS, Montebello Rehabilitation Factor score; OARS, Older American Resources and Services Functional Assessment Questionnaire; PPT, modified Physical Performance Test; POMA, Performance Oriented Mobility Assessment; QT, quadriceps training; RM, repetition maximum; SF-36, Medical Outcomes Study 36-Item Short-Form Health Survey; SWED-QUAL, Swedish Health-Related Quality of Life Survey; TUG, Timed Up and Go test; WHOQOL-BREF, abbreviated WHO Quality of Life Questionnaire; *, significantly greater/significant improvement; +, significantly decreased/ significantly less; N, no significant difference; where P.05 considered significant between-group difference unless otherwise noted. *More than 50% of the study participants dropped out by the 6-mo postdischarge follow-up.

9 254 HIP FRACTURE REHABILITATION IN THE ELDERLY, Chudyk Table 5: Key Features of the Studies Selected for Evaluation Exercise Exercise Baker et al 42 Evidence levels: nonrandomized trial, 13 CG: standard care gait retraining using a frame Sherrington et al 43 Evidence levels: RCT, 21 CG: none comparing 2 interventions Lamb et al 44 Evidence levels: RCT, 21 CG: received placebo stimulation Sherrington et al 45 Evidence levels: RCT, 17 CG: received no intervention Sherrington and Lord 46 Evidence levels: RCT, 16 CG: not clearly described Intervention/intensity: use of a treadmill for gait retraining in hospital after fracture Intervention/intensity: begun on rehabilitation ward I1: usual PT plus progressive WBE every weekday for 2wk; I2: usual PT plus progressive NWBE every weekday for 2wk Intervention/intensity: PNMS begun about a week after fracture fixation/ stimulus intensity minimum required for visible muscle contraction, worn daily for 3h (total duration of 84h) Intervention/intensity: I1: PT plus progressive WBE every weekday for 4mo I2: PT plus NWBE every weekday for 4mo Intervention/intensity: progressive WBE begun an average of 7mo after hip fracture at least once daily for 1mo LOS: N. *Ambulation (gait velocity, cadence, stride length, double-support phase, stance) and strength (hip flexion, hip abduction, and knee extension): in the subgroup analysis, * in IG on hip flexion strength and knee extension strength in the unaffected limb, and hip abduction strength in the affected limb. All outcomes assessed 2 wk after initiation of interventions: Strength (1 RM for knee extensor, isometric force generation of knee extensor, hip abductor and hip muscles, lateral step-up): N hip flexion (9.3N; CI, ) and * hip abduction (6.5; CI, ) in the nonaffected leg in I1; moreover, * in ability to do a lateral step-up in the affected leg with none or 1 hand support (OR 3.4; CI, ) in I1. Otherwise N. Balance (postural sway, FR, step test) and functional performance (PPME): N. Ambulation (time to walk 6m, walking aids): * in walking ability with 1 stick or no aid in I1. * number of I1 required a less supportive walking aid. Self-reported balance: * in I1. Self-reported health, fall risk, and sleep quality: N. Ambulation (timed tests of mobility, recovery of walking ability): IG had * in timed mobility tests between 7-wk and 13-week follow-ups. IG had * in recovery of indoor walking at 13wk. Postural stability (tandem stand): * in IG 7wk after fixation, but not at 13wk. Lower-limb muscle power (Nottingham Leg Extensor Power rig): N. Strength (1RM for knee extensor, isometric force generation of hip abductor and hip flexor using hand-held dynamometer) and ambulation (time to walk 6m, number of steps taken in 6m walk): * knee extension strength in I1 when compared to CG at 4 months. Otherwise N 1 and 4mo after end of interventions. Balance (postural sway, FR, step test): 4mo after completion of the interventions, * in step test in both affected leg (2.8 steps; CI, ) and nonaffected leg (2.6 steps; CI, ) and on the FR (5.9cm; CI, ) in I1 compared with CG. Also, * step test in both affected (2.9 steps; CI, ) and nonaffected leg (3.1 steps; CI, ) and on the FR (7.1cm; CI, ) in I1 compared with I2. N Otherwise. Functional recovery (PPME, timed supine-to-sit, timed sit-to-stand 5): at 4mo, the I1 group performed significantly better than the CG on the timed sit-to-stand (8.1s; CI, ), as did the I2 group (5.8s; CI, ). N Otherwise. Strength (quadriceps, ability to perform WBE test without hand support): IG had * in quadriceps strength in the affected leg at the end of the trial. Otherwise N. Balance (postural sway, FR): N at the end of the trial. Ambulation (walking velocity, cadence): * in walking velocity in IG at trial end. Otherwise N.

10 HIP FRACTURE REHABILITATION IN THE ELDERLY, Chudyk 255 Table 5 (Cont d): Key Features of the Studies Selected for Evaluation Exercise Exercise Elinge et al 47 Evidence levels: RCT, 17 CG: received no intervention Mangione et al 48 Evidence levels: RCT, 19 CG: received biweekly mailings on a variety of nonexercise topics and asked not to begin any exercise program Peterson et al 49 Evidence levels: RCT, 14 CG: usual care Hauer et al 50 Evidence levels: RCT, 18 CG: met 3 times/wk for motor placebo activities Host et al 51 Evidence levels: RCT, 15 CG: not described Intervention/intensity: begun 106 to 194 days postfracture group learning program regarding osteoporosis, falls prevention, ADLs, and physical activity plus a home training program and WBE 2h/wk for 10wk Interventions/intensity: after PT related to hip fracture 2 sessions a week for 2mo, followed by 1 session a wk for 1mo of either of the following: I1: home-based PRT/3 sets of 8 repetitions for various exercises; I2: aerobic exercises/20-min sessions Intervention/intensity: begun 6 to 8wk postsurgery high-intensity circuit training focusing on strength, gait and balance/2, hour-long sessions a week, biweekly for 8wk Intervention/intensity: after hospital discharge high-intensity PRT and progressive functional balance training /three 2.25-hour sessions a wk for 3mo Intervention/intensity: begun within 16wk of discharge from PT after hip fracture exercise program in an indoor facility/45 90min a session, 3 sessions a wk; first 3mo targeting flexibility, balance, coordination, movement; second 3mo: shortened version of first phase s exercises plus PRT Functional recovery (Barthel index and perceived difficulty in carrying out the index s items): N at the end of the intervention and 12mo later. The perceived number of ADLs items performed with difficulties only decreased within the IG over the 12mo postintervention. Quality of life (modified Branholm interest checklist and perceived impact of hip fracture on ability to participate in the activities): immediately after cessation of the intervention, * perceived ability to participate in social life in IG. N Otherwise. Power (force production): * at completion of the trial in I1 and I2 compared with CG. Ambulation (6-min walk distance, free gait speed) and functional recovery (SF-36 physical function subscale): N at completion of the trial. Strength (1RM for hip flexors, extensors, quadriceps, and hamstring muscle groups), ambulation (6-min walk test; TUG; observational gait analysis) and balance (FR): N on trial completion. Only 17 people completed any of the fourth assessment 1y after hospital discharge. N Even when a mixed-effects ANOVA was performed to correct for varied assessment times between the subjects. Strength (leg-press, leg-extensor, leg flexor, ankle plantar flexion, hand grip, chair rise, stair rise): * In IG in leg press in both legs and affected and unaffected sides, and leg extensors on the affected and unaffected sides at the end of the training period and 3mo later. * In IG in leg flexor strength in both legs and the affected side, ankle plantar flexion in both legs and the nonaffected side, and chair and stair rise at the end of the intervention period, but only leg flexor strength on the affected side remained significant at the subsequent 3-mo follow-up. N Otherwise. Functional recovery (POMA, physical activity questionnaire for older people, Barthel ADLs, Lawton IADLs), balance (modified balance test, FR), and ambulation (walking velocity, TUG, walking steadiness): * In IG on the POMA, box-step in the nonaffected leg, walking velocity, TUG, and involvement in activities overall, as well as sports activities, following the end of the intervention period. N Three months after the end of the intervention. Falls (FHI, fear of falling): * in FHI in IG at 3mo after the end of the intervention. Otherwise N. Strength (knee extension, knee flexion, ankle plantar flexion, and leg press): within-group * in isokinetic peak torque values in the fractured limb from baseline to end of the first phase and end of the second phase of the intervention; similar trend in the nonfractured limb, except no differences were noted for the knee extensors at 180 /s. Withingroup * in 1RM for knee extension and leg press exercise after the end of the second phase of the intervention compared with the first.

11 256 HIP FRACTURE REHABILITATION IN THE ELDERLY, Chudyk Table 5 (Cont d): Key Features of the Studies Selected for Evaluation Exercise Exercise Jones et al 52 Evidence levels: nonrandomized trial, 16 CG: conventional homecare services Intervention/intensity: begun 74 27d postfracture community exercise program involving aerobic stepping exercise and progressive WBE/two 45-min sessions a wk for 16wk Functional recovery (modified PPT, timed stair climb) and ambulation (preferred walking speed, fast walking speed): within-group * after the second phase of the intervention, compared with baseline. All outcomes assessed after 16wk: Functional recovery (Yale Physical Activity Survey), ambulation (TUG, self-paced stepping test, pedometer), balance (BBS, FR, ABC), falls efficacy (FES): * self-reported physical activity, TUG scores, and BBS scores in IG. N Otherwise. Strength (1RM for knee extensors and hip abductors using dynamometer): * knee extensor strength in both the affected and unaffected leg, as well as in overall lower extremity strength in IG. N Otherwise. NOTE. Number listed under evidence level refers to study s score on the Downs and Black checklist. 5 Abbreviations: ABC, Activities Specific Balance Confidence Scale; ANOVA, analysis of variance; BBS, Berg Balance Scale; CG, control group; FES; Falls Efficacy Scale; FHI, Falls Handicap Inventory; FR, functional reach; IG, intervention group; I1, intervention 1; I2, intervention 2; NWBE; nonweight-bearing exercise; PNMS, patterned neuromuscular intervention; PPME, Physical Performance and Mobility Examination; PPT, Physical Performance Test; RM, repetition maximum; TUG, Timed Up and Go Test; SF-36; Medical Outcomes Study 36-Item Short-Form Health Survey; WBE, weight bearing exercise; *, significantly greater/significant improvement; +, significantly decreased/significantly less; N, no significant difference; where P.05 considered significant between-group difference unless otherwise noted. *Significant increase in within-group variance from admission to discharge led to a subgroup analysis containing 6 pairs of subjects matched for number of predictors of poor outcome. Data included only for participants who completed at least 30 sessions in each exercise phase. Although the presence of a CG was mentioned in the study design, outcomes were presented as within-group comparisons for the IG. Inpatient setting occupational therapy and physical therapy on a rehabilitation ward. Increased hospital LOS 34 Evidence Level 2c (Weak) Improved balance 35,36 and falls self-efficacy 3 Evidence Level 2c (Weak) # Improved functional recovery, 35,36 ambulation, 36 and strength 36 at discharge #vii Evidence Level 2c (Weak) Three studies investigated the role of rehabilitation ward programs emphasizing various intensities of OT and PT. Koval et al 34 found that an intensive (2 hours of PT and 1 hour of OT daily) rehabilitation program resulted in an increase in total LOS in an institution and decreased basic ADLs recovery 3 months postoperation, with no differences in hospital discharge status, discharge location, or functional measures at 6-month and 12-month postoperation follow-ups compared with a control group. Petrella et al 35 and Mendelsohn et al 36 conducted studies investigating the benefits of combined OT/PT sessions of various intensities and durations. Both studies reported significant within-group improvements in measures of functional recovery and balance from admission to discharge. Inpatient setting quadriceps training; high-intensity PT. Improved power, 37 balance, 37 and short-term general mobility 37 Evidence Level 2c (Weak) The addition of quadriceps training to 20-minute sessions of standard inpatient PT, taking place 5 days a week, was found to result in greater improvements in leg extensor power, functional reach, the Elderly Mobility Scale, and the quality of life measure of energy, compared with standard PT only, 10 weeks after the end of the 6-week intervention. 37 Increasing inpatient PT intensity, while maintaining the same duration of training # Based on within-group differences. period, was not found to result in a quicker ability to perform functional objectives. 38 Outpatient setting programs with a progressive homebased PT component. **Improved balance 39 Evidence Level 2c (Weak) Tsauo et al 40 conducted a study investigating the effects of a home-based PT intervention. On acute care discharge, patients in the treatment group were instructed to continue their inpatient PT regimen, in addition to being visited at home by a physical therapist 8 times in a 12-week period. Six months after acute care discharge, more than 50% of the study participants dropped out, at which point the intervention group had better outcomes on the Harris total-pain Scores, but not measures of range of motion, strength, or ambulation. Tinetti et al 39 found that a program consisting of a progressive exercise PT component, as well as OT focusing on ADLs, lasting a median length of 12 weeks, resulted in significant within-group improvements in balance, strength, and ambulation 6 months after the start of the intervention. Outpatient setting aerobic training followed by aerobic exercise and PRT in an indoor facility. Improved strength, 41 balance, 41 ambulation, 41 and functional recovery 41 Evidence Level 2c (Weak) The effects of 6 months of extended outpatient rehabilitation, led by a physical therapist in an indoor facility, were examined in patients who had been community-dwelling prior to hip fracture. 41 On completion of standard PT, subjects were Based on within-group differences. Results from 1 NRT of fair quality 40 did not find differences in strength and ambulatory outcomes between a program with a progressive home-based PT component and a control group, while another nonrandomized trial 39 of fair quality found that a program with a progressive home-based PT component was associated with within-group increases in strength and ambulatory outcomes.