Regaining functional independence in the acute care setting following hip fracture.

Key Words: Acute care, Functional outcome, Hip fractures, Physical therapy, Rehabilitation.

Hip fracture is a health problem of high volume, high cost, and high risk. Medical expenses totaling billions of dollars are the result of over 200,000 hip fractures every year.[1] The dollar cost of medical treatment for hip fracture notwithstanding, the mortality and morbidity cost may represent an even larger burden to society. Approximately 20% of individuals with hip fractures will die within the first year after fracture.[2] Several authors[2-4] have suggested that as many as half of surviving patients do not recover their previous level of function within the first year after hip fracture. The challenge of returning elderly patients to their prefecture status has become even more daunting for rehabilitation specialists since the advent of prospective payment and managed care.[5] Elderly individuals who wish to remain in the community following hip fracture without 24-hour care must be able to move in and out of bed and ambulate indoors without human assistance, regardless of the daytime supports they may receive. These activities often serve as clinical "milestones" of functional recovery, particularly as an individual's level of independence influences discharge planning and placement in the first few days following fracture.

Given the prevalence of hip fracture and its societal burden, it is imperative to understand the trajectory of functional disability that ensues, particularly as the locus of rehabilitation shifts from the hospital to the home. Although a number of researchers have described longer-term functional outcomes such as ambulation and activities of daily living (ADL),[2] few have described functional recovery during the more immediate postfracture period and its potential relationship to discharge status.[6-9] Numerous personal characteristics and clinical factors have been studied for their associations with the functional outcomes of hip fracture, including age,[3,4,6-13] gender,[3,4,6-9,12] cognitive status,[3,6,7,9-11,13] preadmission living situation and social support,[3,4,6,7,10] history of previous fracture,[3,8,9] fracture site,[3,4,7-9] side of fracture,[8,9] surgical fixation,4S9 postoperative complications,[4,7] functional status before fracture or on admission to the acute care setting,[4,6,7,11-13] comorbidity,[3,4,6,7] depression,[3,7,11,13] muscle strength,[8,14] lower-extremity contracture,[8,14] physical therapy treatment,[3,8,9,12,14,15] and length of stay (LOS).[3,8] Using a broad array of sociodemographic, clinical, and functional status data on a hospital population, the primary purposes of this study were (1) to describe the achievement of independence in key functional milestones during the acute hospital stay among elderly patients admitted for hip fracture and (2) to document the factors that predicted independence in key functional milestones. The data also permitted an exploratory identification of the factors that predicted an elderly person's discharge to the home directly from the acute care setting in a premanaged care environment.

Method

Study Population

This study was conducted at Massachusetts General Hospital (MGH) in Boston, Mass, between November 1992 and December 1993. Data were collected prospectively on all patients with unilateral hip fractures who were treated by surgical reduction and referred for physical therapy. Study inclusion criteria were that the patient was 60 years of age or older and had been admitted from the home.

Measures

Based on the literature and our clinical experience, data were collected on five kinds of factors that could potentially affect postfracture independence in functional activities and discharge status: personal, medical, surgical, hospital, and acute care rehabilitation. Included among personal factors were gender and age, which was recorded as a continuous variable. Subjects were then divided into three groups: 60 to 74 years of age, 75 to 84 years of age, and 85 years of age or older. We classified each subject's self-reported ambulatory status 1 week prior to fracture according to five levels: independent without use of an assistive device, independent with use of an assistive device, dependent on human assistance only, dependent on human assistance and use of an assistive device, and nonambulatory. We later reclassified each subject as either dependent or independent with or without use of an assistive device. We also recorded prefecture living situation by indicating whether the subject lived at home alone, at home with a spouse, or at home with another person or was a resident of a hospital nursing home or other facility at the time of fracture.

Medical variables determined from patient records were whether there was a previous hip fracture, a myocardial infarction within 3 months prior to hip fracture, or a physician's diagnosis of stroke, depression, or dementia. Data on three surgical factors were collected: the side of fracture, the location of fracture (ie, femoral neck, intertrochanteric, or subtrochanteric), and the type of orthopedic procedure used (eg, hemiarthroplasty, compression screws, pins). We later collapsed these categories to compare patients who were treated with surgical procedures that delayed full weight bearing with those who were treated by hemiarthroplasty and who were full weight bearing earlier in their rehabilitation.

Information was gathered on two hospital factors: LOS (in days) and any occurrence of eight specific postoperative complications during the patient's hospital stay. These complications were wound infection, hip dislocation, refracture or loss of reduction, myocardial infarction, thromboembolic episode, pulmonary complication or pneumonia, urinary tract infection, or deep venous thrombosis. An occurrence of any other complications was also recorded. Due to the small number of occurrences of any type of complication, we divided subjects into two groups: those without complications and those with one or more complications.

We proposed that three factors related to the acute phase of a patient's rehabilitation might influence functional levels and contribute to decisions regarding placement of the patient after discharge from the acute care setting. The number of physical therapy treatments received over the LOS was recorded. To account for the increase in the number of treatments that might result merely from a longer LOS, we created a ratio of the number of treatments to the LOS (in days). The presence of joint contractures in either lower extremity was recorded by the treating physical therapist at the time of discharge. We regarded any loss of range of motion (ROM) at the hip, knee, or ankle of greater than 10 degrees as an indication that a contracture was present. Using a grading system adapted from the manual muscle test grades of Daniels and Worthingham,[16] the treating physical therapist noted five levels of muscle performance of the hip abductors of the affected limb at the time of discharge: no contraction, palpable contraction but no movement, movement with gravity eliminated but not through the full ROM, movement with gravity eliminated through the full ROM but without resistance, and movement with gravity eliminated through the full ROM with resistance. The last two levels of measurement were combined in data analysis to aggregate all subjects who could move through the full ROM with or without resistance.

In our investigation, there were two dependent variables. Based on a review of the literature[8,9,12,14,15,17] and our clinical experience, we identified seven key functional milestones for patients after a hip fracture: supine-to-sit transfer, sit-to-supine transfer, sit-to-stand transfer, independent ambulation on level surfaces with a walker, independent ambulation on level surfaces with crutches, independent ascent and descent of stairs with a railing and a crutch, and independent ascent and descent of stairs with crutches only. Independence in the first four milestones is essential if an elderly person is expected to be at home without 24-hour care. A single summary variable indicating that the patient achieved independence in all of these four milestones was generated from our data, and we used this as our first dependent variable.

For the second dependent variable, we recorded each subject's discharge placement as at home alone, at home with spouse, at home with others, or admission to another acute care hospital, rehabilitation facility, nursing home, or other institutional residence. Based on the small number of subjects who were discharged to the home in any of these categories, we later combined these categories to compare those subjects who were discharged to the home with all other subjects.

Procedure

Data were collected for all subjects by seven physical therapists who were assigned specifically to the Orthopaedics Team of Physical Therapy Services at MGH. All therapists had been trained in the data-collection process and had been given written instructions for data collection. Compliance with the prescribed data-collection procedures was monitored by one of the authors (TLF) using chart review, individual discussion, clinical observation, and group discussion at weekly team meetings. Data were extracted from the medical record, the physical therapist's initial evaluation, treatment notes, a milestone achievement record kept by the treating physical therapist for each subject, and discharge data. Physical therapy following operative repair of hip fracture is routinely prescribed for all patients at MGH.

In this study, a physical therapist evaluated the patients within 24 hours of being consulted by the surgeon and initiated therapeutic procedures to assist the patients in regaining function. Each patient was seen at the bedside approximately once per day by a physical therapist or physical therapist assistant to increase ROM, improve muscle strength and control, improve transfer ability in and out of bed and chairs, enhance endurance, develop balance, promote the ability to ambulate independently, and prevent complications. Each physical therapy session lasted at least one half hour, with frequency and duration of any particular exercise or activity set according to patient tolerance. Verbal and written communication were maintained with the nursing staff throughout the hospital stay to assist the patients at an appropriate level during ambulation and self-care, as determined by the physical therapist. Patients were progressed as their functional level improved. If patients required bathroom equipment such as a raised toilet seat or tub bench in order to be discharged to the home, this equipment was ordered by the physical therapist. An occupational therapist saw patients a day or two prior to discharge to the home to provide long-handled adaptive devices and to instruct patients in their use. Occupational therapy was deferred for patients who were being transferred from acute care to an inpatient rehabilitation facility and who therefore would see an occupational therapist in that setting.

Data Analysis

Simple descriptive statistics were used to describe the characteristics of the patients. Logistic regression analyses were used to determine the relative contributions of explanatory variables from each of the five factors influencing postfracture outcomes to each of the two dichotomous dependent variables: discharge to the home and independence in the four key functional milestones of household mobility. These multivariate analyses were conducted in two phases to account for the large number of variables that could be entered into the logistic regression model relative to the relatively small number of subjects available for our study. All variables within a particular factor were first entered into a regression model for each of the two dependent variables. Variables that reached statistical significance in these initial models were then entered into a final logistic model for each of the dependent variables. To facilitate interpretation of the statistical analysis, the logistic regression coefficients of these final models were converted into adjusted odds ratios (aORs) with 95% confidence intervals (CIs).

Results

Sample

There were 214 patients with unilateral hip fractures available for study during the period of our investigation. Of these, 5 patients died (inpatient mortality rate of 2.3%). One patient had an unusually long LOS of 280 days due to nonmedical issues regarding nursing home placement. Data from this patient and the deceased patients were not used in our analyses. We also excluded patients who were younger than 60 years of age or who were not living at home at the time of their admission to the hospital. The analyses, therefore, were limited to the 162 surviving patients who were 60 years of age or older at the time of hip fracture and who lived at home prior to admission. Descriptive characteristics of our study sample are presented in Tables 1 and 2.

[TABULAR DATA 1 OMITTED]

Table 2.

Discharge Characteristics for Independence in Functional Milestones

Functioned Milestone                      n      %
Supine-to-sit transfer                   59     36.4
Sit-to-supine transfer                   50     30.9
Sit-to stand transfer                    64     39.5
Ambulation with walker                   47     29.0
Ambulation with crutches                 12      7.4
Stair climbing with crutch/handrail      13      8.1
Stair climbing with crutches only         8      4.9

Personal Factors

The mean age of our sample was 78.6 years (SD=8.4, range=60-100), with a ratio of women to men of approximately 2:1. The majority of patients (90.2%) ambulated without human assistance prior to hip fracture, although 23.5% of these patients used an assistive device. A reasonably large proportion of patients (42.6%) lived alone prior to hospital admission for hip fracture, 32.7% lived with their spouse, and 24.7% lived with someone other than their spouse.

Medical Factors

A previous hip fracture had been sustained by 12.Y% of the patients. Dementia was evident by physician diagnosis in the medical record in 6.9% of the patients. A small proportion of patients (6.9%) were depressed, as documented in the medical record. Prior to the admission for hip fracture, 8.8% of the patients had experienced a stroke. Only 1.3% of the subjects had myocardial infarctions during the 3 months prior to hip fracture.

Surgical Factors

There were slightly more left-sided hip fractures (52.5%) than right-sided hip fractures (46.3%). Femoral neck fractures were most common (52.5%), and intertrochanteric fractures accounted for 44.4% of the sample. With respect to type of orthopedic fixation, 43.2% of the subjects had plate-and-screw fixation, 40.1% had hemiarthroplasty, and 12.3% had pin fixation.

Hospital Factors

The length of postoperative hospital stay ranged from 3 to 41 days. The mean LOS was 9 days, and the modal LOS was 6 days. A small percentage of subjects had any one of the complications specified in Table 1 during the postoperative period, the most common being urinary tract infection (n=10) and pneumonia (n=6). Slightly less than a third of the subjects had postoperative complications other than the eight complications listed in the data-collection instrument. Upon further review of the raw data, we were able to distinguish two broad categories of complications. There were a small number of postoperative events that often occur in older patients, such as postoperative delirium or exacerbation of preexisting medical conditions.[3,18,19] The larger category comprised transient cardiac or hemodynamic events following surgery that required additional monitoring, also not unusual in older patients.

After treatment at the acute care hospital, over three fourths of the subjects were discharged to inpatient subacute rehabilitation facilities and another 4.3% were transferred to longer-term nursing homes. Only 16.7% of the subjects were discharged directly to the home.

Acute Rehabilitation Factors

The presence of lower-extremity joint contractures was documented on the side of hip fracture in 13.2% of the patients at the time of discharge. On the noninvolved side, 8.5% of the patients had at least one contracture in a lower-extremity joint at the time of discharge.

Hip abductor muscle strength was measured in a gravity-eliminated position at the final physical therapy session prior to discharge. Although 86.3% of the noninvolved hips could be actively abducted throughout full ROM, only 29% of the fractured hips could be actively abducted with gravity eliminated throughout full ROM by the time of discharge.

The number of physical therapy treatments received during the postoperative acute hospital stay ranged from 2 to 19 treatments. The mean number of treatments was 6, the median was 5, and mode was 4. Only 12.3% of the subjects received an average of more than one treatment per day during the period in which they received physical therapy.

Very few patients progressed functionally beyond achievement of independence in bed mobility and ambulation with a walker during their acute hospital stay. Supine-to-sit transfers were achieved independently by 36.4% of the subjects, but fewer subjects (30.9%) could independently perform sit-to-supine transfers. Rising from a sitting position was the functional milestone that was most frequently performed independently during the acute hospital stay. Over one third of the subjects were able to move from a sitting to a standing position without human assistance. Independent ambulation with a walker was achieved by 29% of the subjects, but only 7.4% of the subjects were able to independently ambulate on level surfaces with crutches. Only 8% of the subjects could ascend and descend stairs using both a crutch and a handrail; even fewer subjects (4.9%) could negotiate stairs independently with crutches alone.

Independent Bed Mobility, Transfers, and Ambulation With a Walker

Using logistic regression analysis and converting regression coefficients into aORs, we found that being an independent ambulator prior to hip fracture, LOS, and a higher intensity of physical therapy improved the odds of achieving independence in the four milestones comprising bed mobility and ambulation with a walker (Tab. 3). Increasing age had a modestly negative effect on outcome. Patients aged 75 to 84 years had reduced odds of achieving independence in all four essential functional milestones (supine-to-sit transfer, sit-to-supine transfer, sit-to-stand transfer, ambulation on level surfaces with a walker) compared with younger patients.

[TABULAR DATA 3 OMITTED] Patients older than 84 years also had reduced odds of regaining independence in these basic functional mobility activities, but this finding closely missed reaching statistical significance, most likely due to the small number of patients in this age group in our sample. Patients who had the ability to ambulate independently prior to hip fracture had increased odds (aOR=4.4) of regaining independence in bed mobility, transfers, and ambulation with a walker. Patients who had longer LOSs had marginally increased odds (aOR=1.1) of achieving the first four functional milestones while in the acute care hospital. In contrast to LOS, patients who on average received more than one physical therapy treatment per day had substantially improved odds (aOR=3.98) of achieving independence in the first four functional milestones compared with the odds of patients who averaged one or less physical therapy treatments per day.

Discharge to the Home

Using logistic regression analysis and converting regression coefficients into aORs, we also found that predictors of discharge to the home were age, absence of postoperative complications, achieving independence in bed mobility and ambulation with a walker, and averaging more than one physical therapy treatment per day during the period in which they received physical therapy (Tab. 4). Patients aged 85 years or older had only 0.17 the odds of being discharged directly to the home from the acute care hospital compared with patients under 85 years of age. The odds of being discharged directly to the home were markedly lower among patients with postoperative complications (eg, urinary tract infection, pneumonia) compared with the odds of those patients who had no postoperative complications. The odds of being discharged to the home directly from the acute care setting were 5.7 times better for those patients who on average received more than one physical therapy treatment per day than for those patients who received one treatment or less per day. Patients who achieved independence in bed mobility and ambulation with a walker (de, the first four functional milestones) had substantially greater odds of discharge to the home (aOR= 16.4) compared with those patients who achieved independence in fewer than these four milestones.

[TABULAR DATA 4 OMITTED]

Discussion

Few attempts have been made to describe the factors that influence how patients regain functional independence in the early postoperative phase following hip fracture. This study used prospective natural history data, including the achievement of functional milestones, to ascertain factors that altered the odds of regaining independence in key functional milestones and being discharged directly to the home from the acute care setting.

Our findings are similar to those reported in the literature in several respects. Personal factor data (age, gender, prefecture living situation, and prefecture ambulatory status) and surgical factor data (side of fracture, type of fracture, type of internal fixation) bear a close resemblance to data described in most other studies on hip fracture[1,4,10,12,15,17,20]

The subjects in our study exhibited a lower prevalence of medical factors (eg, stroke, dementia, depression) compared with subjects described in other reports on hip fracture[10,11,13] It is possible that these conditions were underascertained by hospital physicians or not documented when found. This difference in our study is more likely due to the fact that we studied only patients who were admitted to the hospital from the home, and thus who would be expected to be relatively healthier than their institutionalized peers. Given the low prevalence of comorbidity, it is not surprising that we did not find a relationship between presence of medical conditions preadmission and postfracture functional outcome, as some other researchers[4,9,10] found.

The data of this study also suggest the potential impact of time frame and study site on the identification of factors that predict functional independence following hip fracture. Barnes and Dunovan,[14] in a study on elderly patients with hip fractures at a skilled nursing facility, found that decreased strength of the hip abductors on the fractured side and lower-extremity joint contracture were predictors of ambulatory dependence. Our procedure for collecting data on these variables was similar to that used by Barnes and Dunovan, but we found that these variables were not predictors of achievement of basic mobility or discharge to the home. In contrast to Barnes and Dunovan's study,[14] our data were collected in an acute care facility and our patients were community dwellers prior to fracture. Our findings suggest two possible interpretations of conflicting data: (1) These particular variables, hip abductor strength on the fractured side and lower-extremity joint contracture, do not predict functional outcome in all patients, and (2) predictors of function vary with the duration of the postoperative period. Noteworthy, for example, is that only 29% of our patients could abduct their operated hip actively through a full ROM (gravity eliminated) at the time of discharge from physical therapy in the acute care setting. Continued weakness over the course of recuperation may be a more important predictor of long-term outcome than impaired strength immediately after fixation. Thus, therapists may need to use different sets of predictors of outcome, depending on when the prediction is made during the course of recovery.

Some recent studies of patients with hip fractures who had been followed for 1 year or longer postfracture have demonstrated that ambulation status prior to fracture is associated with discharge to the home.[4,17] Our findings suggest that there is immediate clinical utility of this variable in predicting discharge status following hip fracture by virtue of its association with early functional recovery in the acute care setting.

Other studies of patients with hip fractures demonstrated that increasing age was negatively associated with the likelihood of discharge to the home.[6,12,15,17] In contrast to these studies, we found that age may have some more subtle associations with outcome. Although the statistical significance of our findings is hampered by sample size, the associations of age with functional independence or discharge to the home appear to be distinct, depending on the age group.

Intensity of physical therapy was a predictor of discharge to the home as well as a positive predictor of achieving independence in bed mobility and ambulation with a walker. We guardedly suggest that more than one physical therapy treatment per day during the first postoperative week improves the odds of regaining independent bed mobility and ambulation with a walker and of being discharged directly to the home.

A limitation of our study is that we were not able to discern the implicit and explicit judgments of rehabilitation potential that may have influenced the intensity of treatment. For example, it is possible that "healthier" patients were prioritized to receive more intense treatment in an effort to allow them to be discharged directly to the home from the acute care hospital. It is also possible that patients who were more frail received more intense physical therapy because their need for rehabilitation was perceived as greater. Our study was conducted prior to the implementation of critical pathways that specify in advance the postoperative day on which a discharge disposition is determined by the rehabilitation team or the expected day of discharge.

Given our modal LOS, the average treatment frequency for all of our subjects, and the fact that discharge decisions were typically made late in the course of hospitalization, it appears that a therapist would have had little opportunity to alter the frequency of treatment for those patients who were to be discharged directly to the home. Sample size and the limitations of the data we collected do not allow a more definitive interpretation of this finding, but our data support the need for an experimental study manipulating treatment frequency. We are also unable to determine whether premorbid health status and level of fitness may have affected outcomes. Future research should establish the degree to which more intense physical therapy decreases postoperative complications in this patient population, as complications such as those found in our study have been found to increase LOS.[10]

One of the most intriguing challenges facing physical therapy practitioners is the attempt to reduce acute care hospital LOS as a component of reducing the costs of health care. A decreased LOS is related to several clinical characteristics, including practice referral patterns. Practice patterns that influence LOS shift rapidly as managed care and capitation models spread throughout the marketplace, often even before "best practice" data are available. Length of hospital stay, therefore, remains an elusive object of study if the influence of the market on "best practice" in discharge planning cannot be controlled.

Interestingly, the LOS in our study was considerably lower than that of a study of patients with hip fractures done at the same institution a decade previously.[15] The mean LOS recorded a decade ago was 21 days, whereas the mean LOS recorded in our study was 9 days, with a mode of 6 days. The vast majority of patients with hip fractures in our study (77.8%) were transferred from the acute care hospital to a rehabilitation facility, whereas relatively fewer patients (16.7%) were discharged directly to the home. The finding that so few patients were discharged directly to the home from the acute care hospital is only partially explained by the fact that few patients achieved independence in the higher levels of functional mobility. At the time of this study, patient admission to an inpatient rehabilitation facility also depended on patient expectations and physician preference as well as on reimbursement in a premanaged care system. Although LOS is a major component of the costs of hip fracture, the data of this study do not allow us to determine whether decreased LOS in the acute care setting during the past 10 years has actually lowered the cost of treating patients with hip fractures. It is possible that costs associated with LOS have shifted from acute care settings to inpatient rehabilitation settings or home health care without any real reduction in the overall costs of treating patients with hip fractures.

Our analysis indicates that patients who regain independence in bed mobility and ambulation with a walker have substantially better odds of being discharged to the home than,patients who do not achieve similar independence. We also found that a longer LOS marginally improves the odds of achieving independence in these critical functional milestones. One or 2 extra days in the acute care setting might be sufficient to permit some patients to be discharged directly to the home, or to enhance safety and function when discharge to the home is mandated by reimbursement policy. Despite the common wisdom that acute care is the most expensive form of care, this may only be true in the aggregate. Physical therapy provided in the acute care setting may not be more expensive if it offsets the costs of care later in the course of the patient's recovery. Future research, particularly in studies that investigate the costs of hip fractures relative to LOS, should investigate this issue more fully over the continuum of care.

Conclusion

This study demonstrated that a modest proportion of patients with hip fractures treated by internal fixation regained independence in bed mobility, transfers, and ambulation with a walker during the early postoperative phase. Patients who ambulated independently prior to their hip fracture and patients who on average received more than one physical therapy treatment per day had increased odds of independence in performing supine-to-sit, sit-to-supine, and sit-to-stand transfers and ambulating with a walker prior to discharge from the acute care setting. Patients who achieved independence in these four functional mobility activities, patients who did not experience postoperative complications, and patients who on average received more than one physical therapy treatment per day had improved odds of discharge directly to the home from the acute care setting. Older patients were somewhat less likely to regain functional independence or be discharged directly to the home.

Acknowledgment

We thank the physical therapists of the Orthopaedics Team, Physical Therapy Services, MGH, for their assistance with data collection.

References

[1] Marottoli RA, Berkman LF, Leo-Summers L, Cooney LM. Predictors of mortality and institutionalization after hip fracture: the New Haven EPESE Cohort. Am J Public Health. 1994;84:1807-1812.

[2] Craik RL. Disability following hip fracture. Phys Ther. 1994;74:387398.

[3] Magaziner J, Simonsick EM, Kashner M, et al. Predictors of functional recovery one year following hospital discharge for hip fracture: a prospective study. J Gerontol. 1990;45:M101-M107.

[4] Koval KJ, Skovron ML, Aharanoff GB, et al. Ambulatory ability after hip fracture: a prospective study in geriatric patients. Clin Orthop. 1995;310:150-159.

[5] Campion EW, Jette AM, Cleary PD, et al. Hip fracture: a prospective study of hospital course, complications, and costs. J Gen Intern Med. 1987;2:78-82.

[6] Kiel DP, Eichorn A, Intrator O, et al. The outcomes of patients newly admitted to nursing homes after hip fracture. Am J Public Health. 1994;84:1281-1286.

[7] Marottoli RA, Berkman LF, Cooney LM. Decline in physical function following hip fracture. JAm Geriatr Soc. 1992;40:861-866.

[8] Barnes B, Dunovan K. Physical therapy discharge outcomes after hip fracture. Topics in Geriatric Rehabilitation. 1987;2(4):45-51.

[9] Barnes B. Ambulation outcomes after hip fracture. Phys Ther. 1984; 64;317-323.

[10] Ensberg MD, Paletta MJ, Galecki AT, et al. Identifying elderly patients for early discharge after hospitalization for hip fracture. J Gerontol. 1993;48:M187-M195.

[11] Mossey JM, Knott K, Craik RL. The effects of persistent depressive symptoms on hip fracture recovery. J Gerontol. 1990;45:M163-M168.

[12] Bohannon RW, Kloter KS, Cooper JA. Outcome of patients with hip fracture treated by physical therapy in an acute care hospital. Topics in Geriatric Rehabilitation. 1990;6(2):51-58.

[13] Mossey JM, Mutran E, Knott K, Craik RL. Determinants of recovery 12 months after hip fracture: the importance of psychological factors. Am J Public Health. 1989;79:279-286.

[14] Barnes B, Dunovan K. Functional outcomes after hip fracture. Phys Ther. 1987;67:1675-1679.

[15] Jette AM, Harris BA, Cleary PD, et al. Functional recovery after hip fracture. Arch Phys Med Rehabil. 1987;68:735-740.

[16] Hislop HJ, Montgomery J. Daniels and Worthingham's Muscle Testing: Techniques of Manual Examination. 6th ed. Philadelphia, Pa: WB Saunders Co; 1995.

[17] Thorngren K-G, Ceder L, Svensson K Predicting results of rehabilitation after hip fracture. Clin Orthop. 1993;287:76-81.

[18] Michelson JD, Myers A, Jinnah R, et al. Epidemiology of hip fractures among the elderly: risk factors for fracture type. Clin Orthop. 1995;311:129-135.

[19] Magaziner J, Simonsick EM, Kashner TM, et al. Survival experience of aged hip fracture patients. Am J Public Health. 1989;79:274-278.

[20] Lu-Yao GL, Baron JA, Barrett JA, et al. Treatment and survival among elderly Americans with hip fractures: a population-based study. Am J Public Health. 1994;84:1287-1291.