Disability following hip fracture.

Key Words: Hip fractures, Outcome and process assessment (health care), Physical disablement, Rehabilitation.

When the incidence of hip fracture is examined in light of the high mortality rate and the low rate of return to functional ability, it becomes obvious quickly that a seemingly straightforward problem (eg, a fractured femur) poses a large interdisciplinary medical dilemma with extensive societal consequences. After the age of 50 years, the incidence rate rises dramatically, with the average age for hip fracture generally in the eighth decade.[1-3] Estimates of the incidence of hip fracture in persons over 65 years of age suggest that 79% of the hip fractures occur in women and 93% of the fractures occur among Caucasians.[4] The highest rates (350/100,000 or more) are reported in Norway, Sweden, Denmark, New Zealand, and the United States.[5] One prediction is that there will be 100% more trochanteric fractures among females and 133% among males between the years 1985 and 2000.[6] Although the total number of hip fractures grows as the entire population lives longer, the age-adjusted incidence rate is rising as well.

Hip fractures are a major public health problem because of the associated mortality and disability and the high economic cost. About $7.2 billion, an average of $29,800 per fracture, was spent on hip fractures in 1984 in the United States.[7] Using an extremely conservative estimate of a 3% rate of annual inflation, total annual costs of hip fracture will reach a minimum of $31 billion by the year 2020.

The alarming number of hip fractures and the prediction that the age-adjusted incidence is increasing have led to investigations into the etiology of fractures and strategies for hip fracture prevention. Efforts to decrease the number of fractures have not, however, been paralleled by equally intense efforts to prevent or reduce the disability associated with hip fracture. In the 21st century, a growing number of persons with significant functional disability following hip fracture will require ongoing medical and social services. Because it is unlikely that we can eliminate hip fractures, methods to maximize potential and diminish the level of disability should be a central focus of research and clinical efforts.

The purposes of this article are (1) to explore our knowledge of the result of hip fracture, with emphasis on functional disability; (2) to review the factors that influence a person's recovery and the success of hip fracture rehabilitation; and (3) to suggest avenues of future research.

Outcome

Outcome following hip fracture has been described using different measures and may reflect the specific interests of the investigators. Outcome in the late 1960s and 1970s was reported in terms of mortality, fracture malunion, aseptic necrosis, and later segmental collapse.[8-11] More recently, measures of the level of dependence in the home environment have received attention. Regardless of the variable selected, the purpose has been to measure the success or failure of some aspect of intervention.

Mortality

The reported mortality associated with hip fracture ranges from 2% to 63% depending on such factors as length of the follow-up, country studied, type of care, and age distribution of the sample.[3,12-17] Persons who have sustained a hip fracture show greater mortality postfracture than persons of similar age, race, and gender.[3,13] Although reported mortality rates over the past four decades may reflect improved medical management internationally,[17-20] approximately 20% of persons with hip fracture continue to die within 1 year postfracture.

Mortality rates change with the post-fracture interval. Compared with an age- and gender-matched cohort, the death rate is 15 times higher during the first month and 7 times higher during the second month after fracture.[15] Men die more often than women within the first 2 months. The mortality rate begins to parallel that for the matched cohort at approximately 6 months postfracture.[17,21,22] In long-term follow-up ranging from 1 to 10 years postfracture, the rate of death is similar for persons with and without hip fracture.[12,15,23] The mortality rate provides information related to the success of medical management, but it is an inadequate outcome measure for examining the quality of life in the survivors of hip fracture.

Disability

There appears to be more information about the residual disability following hip fracture than in any other medical population. Ambulatory status and the ability to perform various activities of daily living (ADL) are commonly used to describe the level of physical ability postfracture. The two variables indicate that most physical recovery occurs during the first 4 months, with additional but gradual improvement extending through the first year.[23]

Functional disability encompasses psychological and social recovery as well as physical recovery.[24] The four domains of functional disability are physical, mental, emotional, and social disability.

Ambulatory status. The range for the number of individuals achieving independence in ambulation 1 year postfracture is between 30% and 83%.[16,25-27] Reasons for the variability include the use of different methodologies, differences in the length of follow-up, and different definitions of recovery. For example, prefracture ambulatory status was used to assess postfracture performance in some studies,[27] whereas other investigators[26] describe ambulation postfracture without regard to prefracture ability.

The use of both prefracture and postfracture ambulatory status measures is illustrated by a study conducted by Laskin et al.[27] Like many studies, ambulation was scored on a scale that included ratings for mechanical or human assistance.[16,28] The highest score was assigned if the person received no assistance or supervision and used at most a cane. Mechanical aids included a quad cane, crutches, two canes, and a walker or a brace. The authors reported that 22% of patients with hip fracture returned to the prefracture ambulation category at 6 months postfracture. Using a postfracture ambulation criterion, 76% of the patients walked in some manner at 6 months postfracture.

To further illustrate the difference between the two methods of assessing ambulatory status, consider a person who is nonambulatory prior to the fracture and remains nonambulatory following the fracture. If prefracture status is the criterion for success, this person retains prefracture status. If, however, apostfracture criterion is used, the person will be included among individuals who did not recover ambulatory status following the hip fracture. The use of a prefracture criterion as the gold standard controls for the effect of premorbid conditions and may depict outcome following hip fracture more accurately. In lieu of prefracture information, however, the postfracture criterion at least provides data to compare performance among individuals with hip fracture and to estimate the extent of walking disability in individuals with and without hip fracture.

Regardless of whether investigators use a prefracture or postfracture criterion, results across studies indicate a long-term residual disability. For example, a small sample of patients 50 to 64 years of age was examined 2 years after hip fracture.[29] After 2 years, this young sample continued to walk more slowly, perceived their balance to be more impaired than it was prior to fracture, and continued to demonstrate more postural sway than did a group of nondisabled age-matched controls. Because hip fracture occurs most commonly in the eighth decade, the young age of this sample raises issues about the reason for falling. Walking was examined using a postfracture criterion, and the reference standard was a nondisabled cohort. It is, therefore, possible that the subjects had a balance or walking disorder prior to the hip fracture. If the sample was not "normal" prior to fracture, the postfracture results may also reflect prefracture impairment. Other data, however, such as the subjects' perception of balance prior to and following fracture, support the conclusion that additional disability occurred following the fracture; individuals did not believe that they had returned to prefracture status 2 years after the fracture.

For the most part, clinical ambulation scales merely indicate that the person can or cannot walk independently. Ambulation has also been studied in gait analysis laboratories to provide kinematic and kinetic information about walking performance following hip fracture.[30-33] Such studies provide more precise information about how level-surface ambulation differs for patients with hip fracture compared with age-matched controls or how the gait of a patient with hip fracture differs from other patients with walking disability. Emphasis has been on describing the kinematic and kinetic impairments in walking rather than developing a more global model of walking ability that explores the relationship between critical features of the-walking pattern and disability. Subjects have often been limited to small samples who have already achieved near-optimal recovery, which limits the ability to generalize the findings. Constraints in the laboratory often require independence in ambulation without physical support and the maximal mechanical support of a single cane. Ambulatory status is commonly assessed as the person walks over a level linoleum floor.

Walheim et al[34] demonstrated the utility of relating traditional measures of physical disability and data collected from a gait laboratory despite the limitations cited. Their walking capacity score includes nominal data using an ambulatory status score similar to one described by Katz and colleagues[16,28] and ordinal data collected in a gait analysis laboratory. Consistent with other literature, 52% of 149 patients with trochanteric fractures walked without assistive devices, 41% required at least two supports, and 7% of the patients were unable to walk at 6 months postfracture. The additional ordinal data reflect aspects related to the quality of ambulation. For example, maximal vertical force applied through the previously fractured limb during walking, expressed as a percentage of body weight, was >90% for 85% of the patients, indicating the ability to load the previously fractured limb. On the other hand, 26% of the individuals were unable to walk more than 100 m. Only 68% of the patients demonstrated temporal symmetry between limbs, as indicated by a quotient of the duration of the single-limb support phase on the operated limb compared with the nonoperated limb. At 6 months postfracture, therefore, although the clinical scale indicates that 93% of the sample walked in some manner, the additional information from the gait laboratory suggests a limitation in distance walking and the presence of a significant "limp."

The walking capacity score was also sensitive to change in the study by Walheim et al.[34] From 3 to 6 months postfracture, 27% of the patients improved, 59.5% were unchanged, and 3.5% deteriorated. The walking capacity score provides some indication of the quality of ambulation and enriches our understanding of walking ability. Note, however, that the results do not reflect the number of persons who returned to prefracture status. In addition, testing of walking performance was limited to the level surface.

Current descriptive statistics emphasize a profound long-term residual walking disability. The common deficit across studies is defined by the amount of physical or mechanical support required to walk some distance indoors, outdoors, on level surfaces, and on stairs. It is not clear, for example, that the traditional goal of independent ambulation affords the patient the ability to walk safely or efficiently in different environments (eg, across a street before the light changes, to the bus before it leaves., in a dimly lit room). The definition of gait ability has to be expanded to include some measure of the quality of the ambulation (eg, the distance traveled, the level of effort required). The description of gait outcome should go beyond the numbers and types of assistive devices to identify features which limit the success in accomplishing the task of walking. The relationship between impairment in neuromusculoskeletal status and ambulation may become apparent if the critical features of walking are identified.

Activities of daily living. Basic activities of daily living (BADL) include the ability to perform fundamental activities of daily life such as self-care or basic mobility.35 Activities such as meal preparation or shopping, considered more complex physical functions, are called instrumental activities of daily living (IADL).[36] Several tools have been developed to assess BADL and IADL. The Katz Index of ADL35 will be used to exemplify outcome. Like other tools, the Katz Index of ADL is limited to rating independence and dependence in performing routine activities. Bathing, dressing, toileting, transfer, continence, and feeding are scored by the clinician. Independence is defined as performing the task without supervision, direction, or active personal assistance.

Many studies use both measures of ambulatory status and the ability to perform routine ADL. No attempt has been made to equate level of difficulty recovering independence in ambulation to level of difficulty recovering independence in ADL. It is remarkable to note, however, the number of ADL functions that are compromised postfracture.

In a study by Cobey et al,[37] 95% of the subjects achieved an almost perfect prefracture score on a modified Katz Index of ADL.[37] This score was used to provide a grade of poor, partial, or full recovery following fracture. Poor recovery was defined as a significant deficit in recovery (ie, major change in style of living, no longer able to carry on household activities, not living alone or with family, confined to bed or chair). Partial recovery was defined as an almost complete recovery. Full recovery was defined as return to prefracture status. Even at 6 months postfracture, only 23.5% of the patients achieved full recovery 53% achieved partial recovery, and 23.5% achieved poor recovery.

Significant disability continues beyond 6 months postfracture. A modified version of the Katz Index of ADL was among the assessment tools used to examine persons 1 year postfracture in a study by Ceder et al.[38] Independence in ADL occurred if the person managed independently and completely with toileting, personal hygiene, and dressing and undressing. Of the patients living at home prior to fracture, 80% returned home within the first year. Of the individuals living at home 1 year postfracture, 82% achieved independence in ADL. The proportion of persons requiring home help at follow-up was greater than before fracture; an increased need for assistance occurred gradually during the follow-up year, particularly for patients living with someone.

Performance of ADL was among the variables used to describe hip fracture outcome in persons who were followed 9.5 years postfracture.[16] Some ability to perform ADL skills, as measured by the Katz Index of ADL, returned before independent ambulation skills. The likelihood of recovery of ADL was greatest during the first year. Of those who were completely independent in ADL prior to fracture, 55.4% returned to this level within 1 year. At 2 years postfracture, only 66% of the sample had returned to prefracture status in performing ADL. Additional recovery was not noted beyond 2 years postfracture.

The results of these two studies emphasize the duration of residual disability that remains in ADL postfracture. The discrepancy in outcomes between the two studies is explained partially because Ceder et al38 reported outcome only for the subjects who returned home, whereas Katz et al[16] reported outcome for subjects who remained institutionalized. Caution should be exercised, therefore, when comparing outcomes across studies. Understanding the prefracture level of disability and noting whether the results are reported for the entire sample or a selected subgroup are needed before studies can be interpreted.

Current measures of ambulatory status and assessment of ADL provide information about the ability to perform tasks independently, but they do not completely describe physical disability. A poor rating on the Katz Index of ADL, for example, indicates that the person is dependent in activities related to personal hygiene, dressing, or basic mobility. The score, however, does not indicate the types of problems that are typical in attempting to perform the ADL task. Additional information is needed about the quality of ADL performance. For example, which ADL tasks are the most difficult to achieve, and what problems does the person encounter while performing the tasks? Does it take too long to perform an ADL task, does the person lose attention, are there specific impairments that are limiting task performance, and so on?

Functional disability. The ability to walk and to perform ADL address the extent of physical disability but not the psychosocial disability that occurs following fracture. More global outcome measures such as the Functional Status Index (FSI)[39] and the Functional Index Measure (FIM)[40] include assessment of the person's ability to perform within society.

The FSI has been used to assess functional disability in a variety of diagnoses.[39] The shortened version of the FSI measures the level of independence and difficulty in performing 17 BADL tasks, LADL tasks, and social and role function. An emotional category is also assessed in the FSI by a validated scale developed for the Rand Health Insurance Study. This instrument provides measures of gross mobility including stair climbing, personal care, home chores, and interpersonal activities such as driving a car and attending meetings. The FSI defines function to include three dimensions: the ability to perform the task, the degree of dependence, and the degree of difficulty performing specific ADL tasks. This assessment therefore addresses some of the concerns raised about the more traditional ambulatory status or ADL scales.

The FSI was used to examine recovery in 75 patients with a mean age of 78 years.36 At 1 year postfracture, only 33% of the sample regained prefracture status in basic ADL. Only 21% and 26% of the sample, respectively, returned to or surpassed their prefracture levels of independence in instrumental ADL and social/role function. There was little change in emotional function over 1 year. The outcome described by the FSI appears to offer information that describes a much broader picture of the disability sustained following hip fracture and suggests avenues for intervention in the psychosocial as well as physical domains,

The Functional Index Measure (FIM) is another tool designed to measure functional disability.40 The FIM was developed by the American Congress of Rehabilitation Medicine, the American Academy of Physical Medicine, and others to create a national uniform data system for medical rehabilitation. The purposes of the FIM include rating severity of patient disability and defining the outcomes of medical rehabilitation. The FIM assesses self-care, sphincter management, mobility, locomotion, communication, and social cognition on a seven-level scale. The scale ranges from ratings of total assistance to complete independence and considers extent of assistance, supervision, and use of adaptive equipment. The data collected also include patient demographic characteristics, diagnoses, impairment groups, length of hospital inpatient stay, and hospital charges. A recent examination of records from 27,669 patients demonstrates that the FIM yields measures for two fundamental subsets of items: motor and cognitive function.40 Item difficulty varied slightly across impairment groups, which suggests that the instrument may reflect the impact of various kinds of impairments.

The authors of the FIM expect that the data will have immediate application in quality assurance and program evaluation efforts. The FSI appears to provide information related to the person's difficulty in performing certain tasks and may be useful to the clinician who is trying to develop alternate strategies for the patient to accomplish a task. Both instruments can be used to gather a more complete picture of the disability that accompanies hip fracture than the more traditional scales.

The results of all the studies reviewed emphasize that a significant level of residual disability remains postfracture. Very few individuals appear to return to prefracture ambulation; a majority of individuals are dependent in activities related to personal hygiene, dressing, or basic mobility and many individuals are unable to be active in the society.

Factors That Predict Recovery

When data are adjusted for age and comorbidity, the type of fracture or the type of repair do not appear to significantly influence in-hospital mortality or functional ability at 1 year postinjury.[15,41] The severity of fracture, including the amount of comminution or displacement, does not have a significant effect on outcome.[37] The level of recovery in ADL and walking ability are similar regardless of the site of the fracture or the repair method.[10,16,23,37] Although surgical repair is still not optimal because some restriction in weight bearing and range of motion is required, these limitations do not appear to account for the extent of residual disability. Other variables have been sought to account for the level of disability that follows hip fracture.

A myriad of physical, psychological, and social variables are related to returning home and independent management. Early ambulation and the ability to manage dressing and personal hygiene during the first 2 postoperative weeks, for example, predict return to the home.[10] A combination of three factors (ie, healthy at the time of fracture, living with someone before fracture, and the ability to walk within 2 weeks postsurgery) has been associated with living at home with good hip function 10 years after the fracture.[23,42] Institutionalization and loss of function are most likely in patients who have preexisting impairment of mental status, coexisting medical conditions, or functional disability prior to fracture. Depression and coping strategies have been suggested to influence the patient's motivation to recover.

It is difficult to rank order the list of variables that predict recovery because few studies have examined all of the variables in large samples. A review of the literature emphasizes, however, the importance of considering the impact of each variable in any attempt to maximize outcome following hip fracture.

Age, Gender, and Comorbidity

Using mortality as the outcome, a high mortality rate is predicted for men of advanced age who have at least one serious comorbid condition and marked delirium without dementia.[13,17,18] Me presence of at least one life-threatening condition increases the risk of mortality to 4.6 times at 3 months postfracture.[17] Organic brain syndrome and institutionalization following acute hospital stay are also associated with an increased risk of dying within 1 year of hip fracture.[17,22]

Using demographic and medical variables (ie, age, gender, and the number of associated diseases), earlier literature predicted return home with prefracture function.[15.16] The older the patient, the lower the chance of returning home.[10] One in 10 persons over the age of 90 years regain prefracture levels of ambulation.[4] Persons with four or more existing medical conditions at the time of hip surgery have a much poorer risk of recovering independent status than a patient with three or fewer conditions.[13,25] Arteriosclerotic heart disease, organic dementia, and cerebrovascular disease are considered the greatest threats to recovery from hip fracture. Other variables also have a significant impact on returning home and may be more predictive of residual disability.

Physical Impairment

Traditional clinical impairments are pain, limited hip joint range of motion (ROM), and diminished arm or hip strength.[25,34,43] Although it is difficult to compare results across studies, the inability to generate torque is most commonly cited as a predictor of poor recovery. Unfortunately, the relationship between physical impairment and recovery has been addressed to such a limited degree that we cannot conclude that muscle strength is a major determinant of recovery and that other impairments are unrelated to recovery.

Because normal strength testing using traditional manual muscle testing is contraindicated by the surgeon, Walheim et al[34] examined "normal" hip strength, defined as the ability to keep the straight leg lifted in flexion, extension, and abduction against gravity. Because straight leg raising is also contraindicated in some institutions, the results cannot be generalized. The improvement in the walking capacity score discussed previously was, however, closely related to improved strength of hip musculature. The reduction of pain was slightly related. A discriminant analysis using the variables of unstable fracture type, improvement in strength, and reduction in pain successfully predicted that 93% of the patients would achieve improved walking capacity. Age, gender, type of assistive device, status of fracture healing, and ROM did not predict walking capacity. The results of this study emphasize the potential sensitivity of some of the impairment measures in predicting ambulatory status and certainly highlight the need for additional research in this area.

The relationship between impairment in the hip musculature and outcome has been suggested in at least one other investigation.[25] A significant positive association between strength of the hip abductor muscles and independence in ambulation was reported. A lower-extremity contracture was considered a clinical sign of prefracture debilitation and was suggested as a predictor of poor rehabilitation performance. The sample size was limited to 65 patients, so additional research on a larger sample is necessary to generalize the findings.

Laboratory analysis of gait or other functional activities such as rising from a chair has focused on the impairment described by the kinematic and kinetic variables, but has failed to go beyond this level of description. Identified impairments include limited active ROM, inadequate torque development, asymmetrical timing and step length, and excessively slow movement time or velocity. There has been little attempt to relate traditional clinical impairment measures to the impairments measured under the more dynamic conditions in the laboratory. Too often the results of a sophisticated analysis yield a "laundry list" of impairments without regard to the level or quality of disability. In order to provide clinically relevant information, the biomechanical variables must be correlated with other medical, psychological, and social findings to determine their collective influence on functional ability.44

The need to search for this relationship is supported by the results of a study that examined 108 patients over 65 years of age.[37] The physical therapists' prediction of the patient's return of function showed the greatest correlation with the outcome and accounted for approximately 25% of the variance. The physical therapists' expectation of functional recovery prior to discharge was the best predictor of recovery at 6 months postfracture. Physicians' and nurses' ratings were not significant in predicting outcome. The physical therapists' ratings of motivation and mental clarity were also related to outcome at 6 months postinjury. Balance, motor coordination, and endurance were each predictive of return to functional status. These physical impairment measures are not considered in many routine assessments of persons with hip fracture. It is unclear whether these more dynamic and global measures of impairment are more sensitive indicators than the traditional measures of pain, ROM, and muscle strength.

Although the current literature suggests that muscle strength is related to successful outcome, this finding is based on limited research and should be used with caution. At this time, we can only speculate on which of the myriad of impairment measures predict recovery. No physical impairment has been identified as a determinant of recovery in a rigorous investigation. Although weak musculature is related to recovery, the specific muscles and the degree of weakness needed to alter physical ability are unknown. Moreover, the role of joint contracture, pain, and limited endurance has not been defined adequately.

The standard by which we assess physical impairments has been nondisabled function, This standard of ability is often based on performance of the young and ignores the effect of normal aging on muscle strength, joint flexibility, and motor control. We should develop thresholds or acceptable ranges for impairment measures necessary to guarantee physical ability. I contend that the relationship between function of the neuromusculoskeletal system and physical ability should be examined under dynamic rather than static conditions. Attention should be focused on which features of the system prevent ambulation or independence in dressing rather than on assessing the impairments in isolation of the function. Careful, system atic assessments using large samples may allow us to determine which data are unnecessary. The development of a standardized list of impairments that are sensitive to change and that are critically related to outcome are essential if we are going to reduce the current level of residual disability.

Mental Status and Depression

Unlike the tenuous relationship between physical impairment and recovery, the relationship between mental status and recovery following hip fracture is clearly indicated.[1,17,45] Acute confusional state, whether produced by organic or psychosocial factors, has a significant impact on the older patient with hip fracture.[1] Patients older than 84 years and patients with at least one serious medical condition are more likely to become disoriented following their fracture than younger and healthier patients.[17] The stay in the hospital setting is prolonged, and the number of postoperative complications such as urinary incontinence and pressure sores is increased.45 Cognitive impairment, particularly confusion, affects recovery.

A recent study by van der Sluijs and Walenkamp[43] examined 134 patients with hip fracture who were an average of 79 years of age. Variables examined included age, gender, medical condition, type of residence, social support, type of fracture and treatment, walking ability, and independence in ADL. Of the long list of variables, medical condition and mental status predicted rehabilitation success at discharge and at 1 year postfracture. At discharge, medical condition and mental status successfully classified 90% of the successful cases and 39% of the unsuccessful rehabilitations. At 12 months postfracture, mental status alone predicted outcome; 75% of the successes and 81% of the failures were classified correctly. The results of this study emphasize the need to expand our view beyond physical impairment measures and to consider mental status as an important variable when predicting recovery.

Patients who are depressed due to a disability also tend to need longer rehabilitation.[46-48] Depression impedes recovery processes if a person is not motivated to obtain adequate rehabilitation. The associations between depressive symptoms and functional disability and chronic conditions were investigated in an elderly cohort of 2,806 noninstitutionalized men and women.[46] Persons with both functional disability and chronic conditions were most likely to score in the depressed range of the Center for Epidemiologic Studies Depression Scale (CES-D). Mean CES-D scores ranged from 4.86 for those subjects without any disabilities to 13.51 for those with major functional disabilities. Functional disability was far more strongly associated with depressive symptoms than was age.

In another study, 196 white females were followed 2, 6, and 12 months after hip fracture.[47,49] Controlling for age, prefracture physical function, and cognitive status, persons consistently reporting few depressive symptoms using the CES-D were three times more likely than those with persistently elevated CES-D scores to achieve independence in walking. These persons were also nine times more likely to return to prefracture levels in at least five of seven physical function measures, and nine times more likely to be in the higher quartile of physical function. These findings emphasize the importance of persistently elevated depressive symptoms on recovery. There is a suggested need, therefore, to ensure that depression is not influencing recovery.

Psychosocial Factors

Traditionally, psychosocial factors have included positive emotional status as well as social functions (eg, shopping or visiting), not living alone, and the ability to manage household responsibilities or ADL.[10,50] The patients with the best chance of returning home are those who live with someone, who have social contacts outside the home, and who did their own shopping before the fracture.

The number of social supports were examined in persons who were at least 60 years of age, walked one block without aid before the fracture, and had some or no difficulty before the fracture.[42] Only 50% of those with five or fewer social supports regained this level of function 6 months after the fracture. Of those with 15 or more supports, 82% had recovered prefracture function. Arm strength and the number of social supports were associated with recovery of functional status. Depression and age were weakly correlated. Activity prior to the fracture, income, nutrition, number of coexisting medical illnesses, blood urea nitrogen levels, chest radiographs and electrocardiograms, type or severity of fracture, type of repair, or amount of physical therapy were not associated with functional recovery at 6 months. Although this study was limited to 65 patients with a mean age of 71 years and excluded persons with significant decline in mental status, it emphasizes the importance of nonmedical factors in predicting recovery.

The way a person views his or her health has also been implicated in subsequent health outcomes. Self-ratings of health, defined by responses to a single question such as "Compared with others your age, how do you rate your health?" are commonly used assessments in epidemiological and gerontological research.[51] Controlling for age, gender, life satisfaction, income, and urban versus rural residence, the risk of mortality for persons over 65 years of age whose self-rated health was rated as poor was 2.77 times that of those whose self-rated health was rated as excellent.[52] Individuals who had sustained a hip fracture and perceived their problem in a more mechanical fashion (ie, caused by the environment) were reported to show greater improvement in ambulation at 3 and 6 months postfracture relative to those with a different perception of disability.[52] Greater improvement in ambulation was also noted for individuals whose perception of disability was consistent with more autonomy, independence, and a sense of connection with the world around them.

Various types of coping strategies have also been shown to influence recovery. One study that examined older women with hip fractures suggested that women use a variety of coping strategies, with "seeking social support" being the most frequently used strategy.[53] Social support seeking may include the need for physical care, information, psychological support, and serving as an intermediary between the person and the health care system. The use of emotion-focused strategies was associated with poorer functional recovery. An example of an emotion-focused strategy is accepting responsibility for the fracture, which allows distress to be reduced but may not motivate the person to recover any further. This type of research provides greater insight for the practitioner into the reactions and behaviors of older adults and suggests methods to motivate patients to a higher level of functional recovery.

Despite the growing list of variables associated with recovery, the combined list does not account for the all of the variance between these factors and disability. Moreover, there is not consensus about which of the variables are the most sensitive indicators of recovery. This may be due in part to the different methodologies used across investigations, small sample sizes, different age groups studied, and perhaps a bias on the part of the investigators.

Another reason for the lack of a clear relationship between the multiple factors and recovery may be that the hip fracture affects a system already stressed by the process of aging, In an introductory course in biology, the idea of homeostasis is defined and examples are provided demonstrating how the body responds to disturbance or perturbation and returns to the control point. Metabolic, physiologic, and behavioral adjustments are used to maintain homeostasis. Because multiple degrees of freedom are available, the extent to which each factor contributes to homeostasis depends on the extent of the disturbance, the context, experience, and so forth. Arguing from a control systems perspective, the healthy person, therefore, can respond to disturbance and regain homeostasis by selecting from among a variety of options. Aging has been reported to result in composite changes in the body that in turn may limit the system's flexibility, degrees of freedom, or adaptability in maintaining homeostasis.54 The imposition of a hip fracture in an older person may, therefore, challenge the body's ability to maintain homeostasis. The presence of comorbidity, depression, confusion, and social isolation may further tax the compromised response to injury and result in functional disability greater than that expected to accompany a hip fracture. This hypothesis will be tested as our knowledge about the individual factors increases. The development of theoretical models that examine the interrelations among the variables and that predict their combined effect on recovery may help to clarify our understanding of recovery.

It is apparent from the literature just reviewed that the physical therapist should be part of a multidisciplinary team working to minimize factors that influence recovery. The most careful attention to exercises and ambulation training provides no guarantee that the patient who is disoriented, depressed, or socially isolated will return home to independent function. Teamwork to provide the patient appropriate support, realistic expectations about the length of their recovery, and help in developing more appropriate coping strategies may enable more effective physical training.

Intervention

Acute care hospitalization time for patients with hip fracture has decreased dramatically over the last 40 years.[22,38,55] In the early 1950s in Sweden, the patient was confined to bed for approximately 4 months and the hospital stay was more than 5 months.[23] The non-weight-bearing status was 4 months. More recent data from Sweden indicate that hospital stay has decreased to less than 16 days, non-weight-bearing status has decreased to 1 day in some instances, and the return rate home within 4 months has increased to 75%. Similar or improved trends are reported in the United States.[17,22] In the United States, approximately 40% to 50% of the patients are discharged to their previous accommodation directly from the acute care hospital.[19]

Medical reasons for the shorter hospitalization include prophylactic treatment for blood clotting and infection and better methods of operative treatment with fewer technical failures. Rehabilitation reasons for shorter stays include earlier mobilization and earlier weight bearing. Internal fixation has become the treatment of choice for trochanteric fractures, and either internal fixation or the use of a cemented or a noncemented prosthesis is now indicated for fractures of the neck of the femur.[56] These improvements have not, however, resulted in a significant decrease in the amount of residual disability following hip fracture.

Except for reports on different methods of surgical manage ment, there is a paucity of research on the influence of other types of intervention on hip fracture recovery. Jette et al[36] examined the effect of more intensive rehabilitation on functional recovery in a sample of 75 patients with a mean age of 78 years; 40 patients were assigned to a standard rehabilitation program, and 35 patients were assigned to a more intensive program. Standard treatment was breathing exercise and sets of quadriceps femoris muscle exercises on the first day progressing to daily active assistive exercises. Intensive rehabilitation included the standard just described in addition to 1 to 2 hours of individualized patient and family education about expectations, goal setting, and commitment to the rehabilitation program. Geriatric team evaluation focused on comprehensive assessment and follow-up of medical needs, weekly team meetings, a home visit by the physical therapist or social worker within 2 days of discharge to assess the home environment, and telephone calls by the physical therapists at least once per month. More intensive rehabilitation intervention did not alter recovery.

In contrast, Zuckerman et al[57] reported improved recovery in patients who underwent a comprehensive, interdisciplinary care program compared with matched groups of patients who did not undergo the program. The patients in the program had fewer postoperative complications, fewer transfers to the intensive care unit, improved ambulatory ability at discharge, and fewer discharges to nursing homes. The team included an orthopedic surgeon, a consulting internist or geriatrician, an anesthesiologist, a nutritionist, an opthamologist, a psychiatrist, a social worker, and physical therapists and occupational therapists. Physical therapy was scheduled twice each weekday and once each weekend. Occupational therapy evaluation for ADL function was performed before discharge. A case manager, who served as the resource person and patient advocate when postdischarge problems arose, was assigned prior to discharge.

Why the results of these two studies differ is not readily apparent. Jette et al36 compared 40 patients in the experimental program with 35 patients in the control program, whereas Zuckerman et al[57] included 431 patients in the intensive rehabilitation program compared with 60 control patients. Jette et al did not describe the number of physical therapy sessions, whereas Zuckerman et al described physical therapy sessions twice per day during the week and at least one time over the weekend. In addition, the interdisciplinary care team identified in the study by Zuckerman et al consisted of several nontraditional team members including a psychiatrist. The treatment afforded by the addition of these team members may have affected recovery.

Other studies emphasize that psychosocial intervention in the health care setting may improve patient well-being and the quality of care. The mean length of stay was reported, for example, to be reduced by more than 2 days with psychiatric liaison.58 Recent research has also suggested that the acute confusional state and depression that occur following the hip fracture are amenable to treatment.[17,48]

The results of the study conducted by Jette et al36 also raise the issue of treating the "oldest old." Older patients appear to require a longer time to recover from the hip fracture. More intensive rehabilitation and a longer period of rehabilitation appear to yield better recovery for this group. This issue was addressed in a study that examined 18 persons aged 90 or more years.[4] Physical therapy varied with patient and therapist and included ROM, strengthening exercises, gait training, and, in some instances, moist heat. The length of treatment ranged from 10 to 196 days, with a median of 78 days. Of the survivors, 62.5% gained independent ambulation and 44% returned home. One year posttreatment, 9 of the 10 ambulators continued ambulation, 8 used a walker, and 1 used no assistive device. Although no control data were offered, the results are within the range of predicted ambulation for significantly younger persons who sustain a hip fracture and support the hypothesis that additional treatment may result in a more independent status for the older patient.

An important issue is raised regarding this group of patients: Is age alone equated with a prolonged or poorer recovery? Does the older person require a longer time to recover because of a delayed return to homeostasis? Is age a demographic marker for unmeasured comorbidities? Is there a potential bias against older individuals on behalf of the health care providers who make decisions about type and duration of postfracture care? These questions must receive serious attention in face of the growing numbers of persons in their eighth decade who sustain hip fractures. Without sufficient information, the current medical climate may force clinicians to deny care when it can have an impact. If more intensive intervention can have an impact on improving outcome, as suggested in the study by Kauffman et al,[4] then the need for long-term institutional care may be reduced.

Intervention studies must be conducted to evaluate the efficacy of current physical and psychosocial treatment methods and to justify care to third-party payers for all individuals who sustain hip fractures. There is growing pressure to reduce or eliminate care that has not been shown to be effective. The dearth of studies that examine the effect of intervention on altering outcome is disturbing in view of the current climate that demands evidence of treatment efficacy. Without this type of information, we can only speculate about the factors that influence recovery and we have little or no data to justify intervention.

Additional Avenues for Research

The use of the "gold standard" of nondisabled performance to measure outcome is based on an assumption that return to prefracture function is possible. The significant gap, however, between the goal of nondisabled function and the reality of outcome following hip fracture challenges the basis for this assumption. It is not clear whether the goal of return to prefracture status should be abandoned or whether current intervention has failed to overcome the significant mortality and residual functional disability among survivors.

The time has come to move beyond descriptive epidemiology to identify modifiable risk factors (ie, factors that are amenable to treatment). The list of variables that predict recovery emphasize the complexity of the recovery process but do little to guide the development of comprehensive treatment programs. Until major determinants of recovery are identified, intervention will continue to be based on the untested assumption that there is a relationship between some impairment and disability. Does the 90-year-old man with emphysema who used a walker to walk every day to the YMCA prior to fracture have a poorer chance of recovery than the 69-year-old woman who functions independently in her home and is mourning the recent death of her husband? How disabled is the person who walks independently with a walker at 0.3 m/s using limited joint motion and asymmetrical step length? What impairments contribute to this performance? Is it pain, poor endurance, weak musculature, poor motivation, depression or a combination of these factors? How many of these variables can be modified by treatment? If the impairments are modified, is the residual disability reduced?

One avenue to address these questions on small samples is the use of sophisticated laboratories. Such research must explore the relationship between musculoskeletal impairment and functional disability. If the research goes beyond the biomechanical analyses of movement to include medical and social factors, the relationship among physiological, biomechanical, and other factors may become more clear. The theoretical basis for physical therapy intervention must be based, in large part, on the relationship between musculoskeletal impairment and disability. Assessment should focus on modifiable variables, and appropriate treatment should follow.

Other avenues for identifying the determinants of recovery include prospective national intervention studies encompassing a large, heterogeneous sample of patients and careful documentation of individual histories of recovery using qualitative research techniques. The research should be designed to provide specific information about how medical and social variables relate and predict disability.

A classification scheme is critical to identify major determinants of recovery in prospective intervention studies. If developed appropriately, the classification scheme can also guide goal setting and treatment planning. In my opinion, the classification should include demographic and medical variables, physical and psychosocial impairment measures, and functional ability measures. Prefracture status is critical. The FSI or the FIM can be used to describe the level of functional ability prior to fracture. Age, comorbid disease, social supports, and presence of depression should be noted. Careful assessment of upper- and lower-extremity integrity should be made. Although this may be viewed as a "shotgun approach," a schema that neglects any of the factors associated previously with recovery will not reveal the complete relationship among the factors and their ability to predict recovery. The same data must be collected throughout any research investigation. Collecting information only at the beginning of the study to classify prefracture status or only at the completion of the study to describe outcome is inadequate.

Predictions of outcome following hip fracture are based on studies that control for physical therapy intervention (ie, all patients have received physical therapy intervention). There are very few studies that exclude physical therapy intervention or that include different types of physical therapy intervention.

One of the challenges for physical therapy in participating in the types of research just described is to determine whether we are offering treatment that is optimal. Compared with the recent literature that has emphasized exploring new avenues for intervention related to depression[50] or improved coping strategies,[53] for example, there is a dearth of literature that indicates any significant change in physical therapy intervention. In addition, limited research is available that compares different interventions used.

In a study conducted by Mossey et al,[47] home visits were made by a physical therapist at 2 and 6 months postfracture in a sample of women who were active community dwellers prior to hip fracture. I collected anecdotal information during the follow-up visits. At the 6-month follow-up, the women continued to be compliant with the hip strengthening exercises that were prescribed at the time of hospital discharge. The hip abduction exercises were still performed in a gravity-eliminated position. Many of the women continued to use long-handled reachers to retrieve objects from the floor and to take sponge baths because they did not want to reinjure their hip or because of fear of falling.

Without data to refute such anecdotal reports, it can be assumed that physical therapy interventions are similar to those offered 20 years ago (ie, early ambulation training, muscle strengthening, and active ROM). It is difficult to believe that the level of functional disability cannot be reduced. As a minimum, follow-up visits into the home to progress exercise programs and to ensure progress in ADL might reduce the current level of residual disability. The effect of aquatic programs and the use of electrical stimulation to promote muscle strength are being explored by clinicians, but there are no reports of research using any of the newer approaches.

In current clinical practice, it is usual for a physical therapy department to have a specific protocol or a critical path for treatment of patients with hip fracture.[59] There is no literature that details the rationale for a differential care delivery system. In today's political and economic climate, it is unrealistic to offer the same level of care to all patients. Maintenance may be an appropriate goal for some patients., whereas recovery is the appropriate goal for others. From the current literature, it is not clear who fits into each category of care.

Demopoulos et al[60] argue that rehabilitation programs should be based on a classification scheme such that those persons with excellent predictors for recovery of prefracture status receive more intensive rehabilitation. Hielema[61] Suggests, on the other hand, that the clinician must decide whether to intervene to prevent a predicted poor outcome. What evidence does the clinician have to help decide which factors predict recovery regardless of intervention, and which of the predictive factors are amenable to change? Without evidence to support that intervention is successful in mediating poor predicted outcome, differential care delivery may be established with an inadequate theoretical basis. in light of current knowledge of functional recovery, for example, rehabilitation may be denied for the socially isolated oldest persons with significant comorbid disease. Although Kauffman et al[4] demonstrated that additional and patient-specific rehabilitation in persons over the age of 90 years resulted in functional recovery that far exceeded what had been predicted previously, this is a single study conducted on a small sample without a control group.

Careful documentation of the success of intervention in achieving particular goals with different patient categories and performed on a national basis will enable comparisons across treatment facilities. Such information will be useful in comparing the effects of different types of intervention and may lead to justification for continued or differential care to further enhance recovery in specific types of patients. Discussions for who receives what type of intervention should be based on data that support the efficacy of treatment with a particular group of patients. If physical therapists do not contribute to this literature, someone else in the health care system may decide how to mete out treatment.

Summary

Hip fractures place a tremendous burden on the injured person, significant others, our health care system, and society. There is an abundant body of literature regarding hip fracture. Research has focused attention on the etiology of fracture and strategies to prevent falls. Improved surgical repair technology and better medical management have significantly reduced the length of hospital stay. A decrease in mobility and activity and a change in the quality of life continue to be reported despite improved surgical repair.

Attention in the clinical setting must shift from biomedical factors that limit recovery to include the myriad of other medical and psychosocial factors. It is not clear, at this time, how these factors relate to produce the large discrepancy between optimal performance and the actual limitations in functional ability. Although the bulk of the research implies a vital role for physical therapy in achieving recovery, the impact of physical therapy intervention on further reducing the discrepancy between surgical success and functional disability is also unknown at this time. Much additional research is needed to identify the major determinants of recovery and to develop interventions that successfully reduce the residual disability. The physical therapist must become actively engaged in a national research effort to decide the efficacy of current treatment methods and to develop innovative treatments in an economic climate that demands more optimal recovery in a shorter period of treatment time.

Acknowledgments

I would like to thank Kristin Stotts, FIT, for helping to gather the information used to prepare this manuscript and Elizabeth Clark for her support in preparing this manuscript. I would also like to thank Dr Jana Mossey.

References

[1] Porter R, Miller C, Grainger D, et al. Prediction of hip fracture in elderly women: a prospective study. Br Med J. 1990;301:638-641. [2] Rockwood P, Horn J, Cryer C. Hip fractures: a future epidemic? J Orthop Trauma. 1990;4: 388-393. [3] Kelsy J, Hoffman S. Risk factors for hip fracture. N Engl J Med. 1987;316:404-406. [4] Kauffman T, Albright L, Wagner C. Rehabilitation outcomes after hip fracture in persons 90 years old and older. Arch Phys Med Rehabil. 1987;.68:369-371. [5] Maggi S, Kelsey J, Litvak J, et al. Incidence of hip fractures in the elderly: a cross-national analysis. Osteoporosis Int. 1991;1:232-241. [6] Nilsson R, Lofman O, Berglund K, et al. Increased hip-fracture incidence in the county of Ostergotland, Sweden, 1940-1986, with forecasts up to the year 2000: an epidemiological study. Int J Epidemiol. 1991;20:1018-1024. [7] Cummings S, Rubin S, Black D. The future of hip fractures in the United States: numbers, costs, and potential effects of postmenopausal estrogen, Clin Orthop. 1990;252:163-166. [8] Ceder L, Svensson K, Thorngren K. Statistical prediction of rehabilitation in elderly patients with hip fractures. Clin Orthop. 1980; 152:185-190. [9] Lowell J. Results and complications of femoral neck fractures. Clin Orthop, 1980;152:162-172. [10] Ceder L, Thorngren KG, Wallden B. Prognostic indicators and early home rehablitation in elderly patients with hip fractures. Clin Orithop. 980;152:173-184. [11] Halpin P, Nelson C. A system of classification of femoral neck fractures with special reference of choice of treatment. Clin Orthop. 1980;152:44-48. [12] Lewinnek GE, Kelsey J, White AA, et al. The significance and comparative analysis of the epidemiology of hip fractures. Clin Orthop. 1980;152:35-43 [13] Kenzora JE, McCarthy RE, Lowell JD, et al. Hip fracture mortality. Clin Orthop, 1984;186: 45-56. [14] Ray W, Griffin M, Baugh D. Moriality following hip fracture before and after implementation of the prospective payment system. Arch Interm Med. 1990;150:2109-2114. [15] Dahl E. Mortality and life expectancy after hip fractures. Acia Orthop Scand 1980;51:163-170. [16] Katz S, Heiple K, Downs T, et al. Long-term course of 147 patients with fracture of the hip. Surg Gynecol Obstet. 1967;124:1219-1230. [17] Magaziner J, Simonsick E, Kashner TM, et al. Survival experience of aged hip fracture patients. AJPH. 1989;79:274-278. [18] Myers A, Robinson E, van Natta M, et al. Hip fractures among the elderly: factors associated with in-hospital mortality. Am J Epidemiol 1991;134:1128-1137. [19] Fitzgerald JF,. Dittus RS, Institutionalized patients with hip fractures: characteristics associated with returning to community dwelling. J Gen Intern Med. 1990:.5:298-303. [20] Fitzgerald JF, Moore PS, Dittus RS. The care of elderly patients with hip fracture: changes since implementation of the prospective payment system. N Engl J Med, 1988;319: 1392-1397. [21] Cummings S, Kelsey J, Nevitt M, et al. Epidemiology of osteoporosis and osteoporotic fractures. Epidemiol Rev. 1985;7:178-208. [22] Schroder HM. The cost of hospitalizing hip fracture patients has increased despite sholler hospitalization time, Injury. 1991;22:135-138. [23] Borgquist L, Ceder L, Thorngren K. Function and social status 10 years after hip fracture. Acta Orthop Scand 1990;61:404-410. [24] Jette A. Functional disability and rehabilitation of the aged. Topics in Geriatric Rehabilitation. 1986;1:1-7. [25] Barnes B, Dunovan K. Functional outcomes after hip fracture. Phys Ther. 1987;67:1675-1679. [26] Miller C. Survival and ambulation following hip fracture. J Bone Joint Surg [Am]. 1978; 60:930-933. [27] Laskin R, Gruber M, Zimmerman A. Intertrochanteric fractures of the hip in elderly. Clin Orthop. 1979;141:188-195. [28] Katz S, Ford A, Heiple K, et al. Stages of illness in the aged: recovery after fracture of the hip. J Gerontol. 1964;19:285-293. [29] Jarnlo G. Hip fracture patients: background factors and function. Scand J Rehabil Med Suppl 1990;24:1-31. [30] Crowninshield RD, Brand RA, Johnston RC. The effects of walking velocity and age on hip kinematics and kinetics. Clin Orthop. 1978;132: 140-144, [31] Krebs DE. Elbaum L, Riley PO, et al. Exercise and gait effects on in vivo hip contact pressures. Phys Ther. 1991;71:301-309. [32] Murray MP, Gore DR, Brewer BJ, et al. Joint function after total hip arthroplasty: a four-year follow-up of 72 cases with Charnley and Muller replacements. Clin Orthop. 1981;157:119-124. [33] Yamamoto S, Suto Y, Kawamura H, et al. Quantitative gait evaluation of hip diseases using principal component analysis. J Biomech. 1983;16:717-726. [34] Walheim G, Barrios C, Stark A, et al. Post-operative improvement of walking capacity in patients with trochanteric hip fractures: a prospective analysis 3 and 6 months after surgery. J Orthop Trauma. 1990;4:137-143. [35] Katz S, Ford AB, Moskowitz RW, et al. Studies of illness in the aged. JAMA. 1963; 185:914-919. [36] Jette A, Harris B, Cleary P, et al. Functional recovery after hip fracture. Arch Phys Med Rehabil. 1987;68:735-740. [37] Cobey J, Conant L, Weil U, et al. Indicators of recovery from fractures of the hip. Clin Orathop. 1976;117:258-262. [38] Ceder L, Ekelund L, Inerot S, et al. Rehabilitation after hip fracture in the elderly. Acta Orthop Scand. 1979;50:681-688. [39] Jette A. Functional Status Index: reliability of a chronic disease exaluation instrument. Arch Phys Med Rehabil. 1980;61:395-401. [40] Heinemann AW, Linacre JM, Wright BD, ct al. Relationships between impairment and physical disability as measured by the Functional independence Measure, Arch Phys Med Rehabil. 1993;74:566-573. [41] Evans JG, Prudham D, Wandless I. A prospective study of fractured proximal femur: factors predisposing to survival. Age Ageing. 1979;8:246-250. [42] Cummings S, Phillips S, Wheat M, et al. Recovery of function after hip fracture: the role of social supports. J Am Geriatr Soc. 1988; 36:801 806. [43] van der Sluijs J, Walenkamp G. How predictable is rehabilitation after hip fracture? A prospective study of 134 patients. Acta Orthop Scand. 1991;62:567-572. [44] Schultz A. Mobility impariment in the elderly: challenges for biomechanics research. J Biomech. 1992;25:519-528. [45] Brannstrom B, Gustafson Y, Norberg A, et al. ADL performance and dependency on nursing care in patients with hip fractures and acute confusion in a task allocation care system. Scand J Caring Sci. 1991; 5:3-11. [46] Berkman L, Berkman C, Lasl S, et al. Depressive symptoms in relation to physical health and functioning in the elderly. Am J Epidemiol 1986;124:372-388. [47] Mossey JM, Mutran E, Knott K, et al, Determinants of recovery 12 months after hip fracture: the importance of psychosocial factors. AJPH. 1989;72:279-286. [48] Mossey JM, Knott K, Craik RL. The effects of persistent depressive symptoms on hip fracture recovery. J Gerontol. 1990;45:M163-M168. [49] Kellie SE, Brody JA. Sex-specific and race-specific hip fracture rates. AJPH. 1990;80:326-331. [50] Cole E. Assessing needs for elders' networks. Journal of Geriatric Nursing. 1985;11: 31-34. [51] Mossey JM, Shapiro E. Self-rated health: a predictor of mortality among the elderly. AJPH. 1982;72:800-808. [52] Borkan J, Ouirk M, Sullivan M. Finding meaning after the fall: injury narratives from elderly hip fracture patients. Soc Sci Med. 1991;33:947-957 [53] Roberto K. Coping strategies of older women with hip fractures: resources and outcomes. J Gerontol. 1992;.47:21-26. [54] Welford AT. Between bodily changes and performance: some possible reasons for slowing with age. Exp Aging Res. 1984;. 10: 73-88. [55] Ceder L, Lindberg L, Odberg E. Differentiated care of hip fracture in the elderly: mean hospital days and results of rehabilitation. Acta Orthop Scand 1980;51:157-162 [56] Wykman A, Olsson E, Axdorph G, et al. Total hip arthroplasty: a comparison between cemented and press-fit noncemented fixation. J Arthroplasty. 1991;6:19-29. [57] Zuckerman J, Sakales S, Fabian D, et al. The challenge of geriatric hip fractures. Bull NY Acad Med. 1990;66:255 265. [58] Strain J, Lyons J, Hammer J, et al. Cost off-set from a psychiatric consultation-liaison intervention with elderly hip fracture patients. Am J Psychiatry. 1991;148:1044-1049. [59] Gatti L, Scott CM. Physical therapy management of total joint replacements at Pennsylvania Hospital. Physical Therapy Practice. 1992;1: 54-68. [60] Demopoulos JT, Pinter I, Crocetti AM, et al. Prediction of outcome and rehabilitative management after surgical operations on the hip. Bull NY Acad Med. 1976;52:761-769. [61] Hielema FJ. Epidemiology of hip fracture: review with implications for physical therapists. Phys Ther. 1979;59:1221-1225.