This study aims to develop and examine an innovative family-centered intervention model for managing cognitive decline, improving postoperative recovery of hip-fractured patients with cognitive impairment, and enhancing family caregivers' competence in dementia care. This care model is theoretically underpinned by: (a) the Progressively Lowered Stress Threshold Model, a component of Lawton's ecological model of aging, and (b) the concept of partnership with family caregivers to strengthen their competence in providing care. Training are provided to family caregivers to enhance their competence in caring for hip-fractured patients with cognitive impairment. The effectiveness of the care model has been evaluated in a randomized controlled trial. The study was approved by the Institutional Review Board of Chang Gung Memorial Hospital. A protocol of the family-centered care model was developed, and the research nurses were trained to provide the interventions. A checklist, consisting of postoperative care, rehabilitation exercises, nutritional health teaching, environmental modification suggestions, delirium care, and care issues for elders with cognitive impairment, as well as management of behavioral problems, was also developed and are recorded by the research nurses. This report is based on data collected from 149 dyads of participants who were recruited by September 2018 and randomly assigned to either an experimental group (n=73) or a control group (n=76). No significant differences are found between experimental and control group in their demographic and clinical variables including age, gender, diagnosis, surgery type, the length of hospital stay, the cognitive functioning, marital status, and educational level, as well as the age and gender of family caregivers. The refusal rate this year is 73.6%. The main reasons for caregivers not participating are not needed and too busy or afraid of being interrupted. No significant differences are found in demographic variables (ie, age, gender, diagnosis, surgery method, and length of hospital stay) between those who participated and those who refused. Causes of the attrition includes that participants refused to participate any more (n=25), died (n=12), moved to another location (n=6), and loss of contact (n=3). Older persons who quit participating in the study are more younger (p=.021) and more are diagnosed with inter-/sub-trochanteric fracture (p=.015) as well as more are receiving internal fixation (p=.029). Outcome variables including patients' cognitive function, clinical measures, self-care ability, family caregivers' competence and preparedness, health service utilization, quality of life, and cost of care. In addition to the clinical effectiveness of the family-centered care model will be evaluated by hierarchical linear models at the end of this study.

In this proposed clinical trial, the investigators intend to develop and examine an innovative family-centered intervention care model for managing cognitive decline, improving postoperative recovery of hip-fractured patients with cognitive impairment, and enhancing family caregiver's competence in dementia care. This study is innovative in two aspects: first, its focus on older persons with comorbid conditions of cognitive impairment and hip fracture, and second, the development of a family-centered care model for this population. Therefore, the goal of this proposed study is to examine the cost and effects of a family-centered care model that includes a family-centered interdisciplinary care component, and a training/support component for family caregivers of hip-fractured patients with cognitive impairment. The control group receives only usual care and the experimental group receives usual care and family-centered care.

The specific aims are:

1. To develop a family-centered care model for hip-fractured elders with cognitive impairment, consisting of family-centered geriatric assessment, continuous rehabilitation, and supported discharge planning, and a family caregiver training/support component to manage symptoms and behavioral problems associated with cognitive decline and enhance postoperative recovery of hip-fractured patients with cognitive impairment and to increase caregiver's competence in dementia care. The protocol is pre-tested within the context of a randomized intervention study subject to existing constraints of the clinical (i.e., CGMH) and community settings.
2. To evaluate the effectiveness of usual care and the family-centered care model for hip-fractured elders with cognitive impairment in a randomized control trial. Effects of the two care models will be evaluated by comparing the trajectories of selected outcome variables: patients' clinical outcomes, self-care ability, cognitive function, behavioral problems, health-related quality of life (HRQoL), and service utilization, and family caregivers' preparedness, competence, and HRQoL. Predictors of recovery trajectories for the usual care and family-centered care models will also be compared.
3. If the effectiveness of the family-center care is established, the costs associated with the usual care and family-centered care models will be analyzed. Cost items include: (1) costs for personnel time and home care visits, (2) costs of hospitalization, (3) costs of emergency or outpatient visits after hospital discharge due to a fall or re-fracture in the same location, and due to diagnosis/treatment for post-operative cognitive impairment, (4) patients' out-of-pocket costs for equipment such as walkers, crutches, nutritional supplements, or other necessities to improve their health conditions or to support walking, and (5) transportation expenses for patients and primary caregivers to visit hospitals or costs of time away from work.

Hypotheses

Based on results from previous studies on the effects of the interdisciplinary care model (Shyu et al., 2005, 2008) and of the family caregiver training program for patients with dementia (Huang et al., 2013; Kuo et al., 2013), following hypotheses are proposed. During the first year after hospital discharge, elderly hip-fractured patients with cognitive impairment who receive the family-centered care model will:

1. Have less cognitive decline, better clinical outcomes, self-care ability, HRQoL, and fewer behavioral problems than their counterparts who receive usual care.
2. Have more outpatient visits and fewer hospital readmissions and emergency room visits than their counterparts who receive usual care.
3. Their family caregivers will have better preparedness and competence in dementia care as well as HRQoL than caregivers of elderly hip-fractured patients with cognitive impairment who receive usual care.
4. Have fewer re-admissions and emergency room visits at a lower total costs than the usual care and their care givers will have lower productivity losses.

Background and Significance Dementia and hip fracture are both associated with substantial disability and mortality, often coexist, are increasingly common in older adults, and have shared risks (Friedman, Menzies, Bukata, Mendelson, & Kates, 2010). Cognitive impairment occurs in 31-88% of older persons after surgery for hip fracture (Holmes & House, 2000a, 2000b). Cognitive impairment is not only a major risk factor for falling and hip fracture, but also predicts postoperative complications, delirium, rehabilitation difficulties, recurrence of fracture, nursing home placement (Jalbert, Eaton, Miller, & Lapane, 2010; Lee et al., 2011; Yiannopoulou, Anastasiou, Ganetsos, Efthimiopoulos, & Papageorgiou, 2012), poor functional recovery, increased risk of mortality (Dubljanin-Racpopoć, Matanović, & Bumbasirević, 2010; Holmes & House, 2000b), and further cognitive impairment (Shyu et al., 2013a).

Hip fracture in Taiwan, as in many other countries, has become a serious health issue (Liou, Tsai, & Lin, 2002). Hip fractures are a significant cause of morbidity and mortality worldwide, especially in developed countries among people >50 years old (Johnell & Kanis, 2006). The number of osteoporotic hip fractures in 2000 was estimated to be 1.6 million worldwide (Johnell & Kanis, 2006), and this number was projected to increase by 2050 to 6.26 to 21 million (Parker & Johansen, 2006). By 2050 half of all hip fractures are estimated to occur in Asia (Dhanwal, Dennison, Harvey, & Cooper, 2011). The age-adjusted incidence rates (per 100, 000) for hip fracture in Taiwanese adults 50-100 years old from 1996 to 2000 were 225 in men and 505 in women, and these numbers were 10 to 20 times above the incidence rate for the general population (Chie, Yang, Liu, & Tsai, 2004). Excess mortality was reported to last 2 to 8 years after hip-fracture surgery (Johnston, Barnsdale, Smith, Duncan, & Hutchison, 2010; Robbins, Biggs, & Cauley, 2006). After hip fracture, around half of older persons who were independent before hip fracture become partly or completely dependent in self-care ability (Shyu, Chen, Liang, Wu, & Su, 2004; Wildner et al., 2002). Hip fracture represents a major and growing health care problem in Taiwan. Recovery from and management of hip fracture may be significantly complicated by coexisting dementia.

Few studies have explored the effects of intervention programs on postoperative cognitive impairment of older persons with a hip fracture. One study found that 37% of 256 elderly hip-fractured patients were cognitively impaired on admission, and 51% of them reached normal test scores during hospitalization (Strömberg, Lindgren, Nordin, Ohlén, & Svensson, 1997). They also found that cognitive impairment was associated with more postsurgical complications. A prior study (Shyu et al., 2013a) found that hip-fractured patients (N = 160) who received interdisciplinary care were 75% less likely to be cognitively impaired 6 months following discharge than those who received routine care (odds ratio = 0.25, P < 0.001).

However, it remains unclear whether older hip-fractured persons with dementia benefit from similar rehabilitation and geriatric consultation as those without dementia. Some studies found that patients with dementia, especially mild and moderate dementia, who suffered a hip fracture benefited from multidisciplinary geriatric assessment and rehabilitation by returning to the community (Huusko et al., 2000), improving functional recovery (Rolland et al., 2004), decreasing postoperative complications, especially delirium and regaining prior independence (Moncada et al., 2006; Stenvall et al., 2012). On the other hand, the investigators found that cognitively impaired elderly patients (n = 51) benefited from interdisciplinary care, consisting of geriatric consultation, continuous rehabilitation, and supported discharge planning, by regaining their prior walking ability, but the intervention did not prevent these elders' subsequent falls and decrease emergency room visits as it did for those without cognitive impairment (n =109) (Shyu et al., 2102). In other words, the interdisciplinary intervention effects were greater for those without cognitive impairment. Another longitudinal study of 231 older persons with hip fracture found that cognitively impaired patients did not retain rehabilitation gains in locomotion, transfers, self-care, and sphincter control at 1 year following post-acute rehabilitation discharge (Young, Xiong, & Pruzek, 2011). The authors suggested that these cognitively impaired patients would have benefited from routine monitoring of cognitive status, a high level of continuity, and a reorientation program.

In conclusion, of the few randomized controlled trials that have examined the effects of intervention programs on hip-fractured patients with cognitive impairment, most have analyzed the effects on a sub-sample of patients with cognitive impairment. Intervention programs have not yet been developed and tested specifically for hip-fractured patients with cognitive impairment. In addition, most studies on hip-fractured patients with cognitive impairment examined the short-term outcomes of rehabilitation, and few explored intervention effects on decreasing cognitive impairment and on recovery trajectories. To fill in these gaps in knowledge, this proposed clinical trial will develop and test a family-centered care model for hip-fractured patients with cognitive impairment and their family caregivers to facilitate postoperative recovery, decrease negative outcomes for patients, and support their family caregivers.

The care model in this study will be delivered using a family-centered approach based on the Progressively Lowered Stress Threshold Model (PLST) (Hall & Buckwalter, 1987) and the concept of partnership with family caregivers (Harvath et al., 1994) to strengthen family caregivers' competence in providing care to hip-fractured patients with cognitive impairment. The PLST model proposes that patients with dementia become anxious or agitated by external or environmental demands and internal stimuli due to increasing disabilities resulting from progressive cerebral pathology and associated cognitive decline. The affected person's behavior becomes increasingly dysfunctional and often catastrophic when the stimuli continue or increase. Thus, promoting dementia patients' adaptive behavior requires modifying and reducing environmental demands and stress. In this study, the PLST model will be tailored to the individual needs of hip-fractured patients with cognitive impairment and their family caregivers to enhance caregivers' preparedness and competence in dementia care.

At the same time, the concept of "partnership with family caregivers" emphasizes that caregivers possesses important knowledge about the patient under their care of and the care that patient receives (Harvath et al., 1994). The most useful and effective approach to dealing with caregivers' problems and concerns is to combine the knowledge of caregivers and nurses. Based on this notion, our family-centered care model focuses on nurse-caregiver collaboration. This collaboration combines the caregiver's and nurse's knowledge to facilitate in-home rehabilitation, identify the patient's behavioral problems and their causes, and to plan the patient's individual care, thus decreasing behavioral problems and improving postoperative cognitive function and recovery. The findings of this study can increase knowledge in geriatric medicine about handling comorbidities in older persons, especially those with cognitive impairment who are receiving treatment for other medical conditions.

Basic Design The cost and clinical/intervention effectiveness of the family-centered care model is evaluated using a randomized experimental design. The proposed project will take place over a 5-year period at CGMH.

The investigators compare a usual care model and a family-centered interdisciplinary care model, which consists of a family-centered approach to interdisciplinary care and a family caregiving-training component to enhance family caregivers' competence in providing post-operative care and handling behavioral problems and neuropsychiatric symptoms of adults with cognitive impairment. In this study, the investigators improved the original interdisciplinary care model, which included geriatric consultation, continuous rehabilitation, and discharge planning, by adding a family-centered approach by involving family caregivers in these main intervention components. For this trial, the investigators propose comparing one experimental group (usual care plus family-centered care) to a control group (usual care).

After a fall leading to hip fracture, patients are mostly sent directly to the hospital emergency room, although some patients may visit outpatient clinics and enter the hospital via the clinic. These patients are cared for by orthopedists and receive internal fixation or arthroplasty. Consultations for internal medicine care are occasionally made depending on the patient's condition. During the first 1-2 days after surgery, nurses teach patients how to exercise while still in bed, using caution while changing their position. Pain-relief medications and antibiotics are also administered (for 2-3 days). The first day after surgery, physical therapy usually starts only for those who received arthroplasty of hip replacement. In the first session, physical therapists train patients to use a walker and get in/out of bed. Around 3-4 days after surgery, patients are usually discharged from the hospital without home assessment. After hospital discharge, no in-home programs are provided for rehabilitation or nursing care. Patients usually come back to the clinic around 1-2 weeks, 1 month, 3 months, 6 months, and 1 year after hospital discharge. However, adherence to this follow-up schedule is poor. Telephone follow-ups are seldom used. The current routine practice for elderly hip-fractured persons in Taiwan lacks well-organized interdisciplinary care protocols, has no continuity of care, and does not provide specific care for hip-fractured patients with cognitive impairment.

Family-Centered Interdisciplinary Care The family-centered interdisciplinary care component. The interdisciplinary care model was developed and examined in prior studies, but with inconsistent findings (Shyu et al., 2008, 2010, 2013b). Prior analyses showed that the difference between usual care and the original interdisciplinary care model has narrowed due to improvements in usual care (Shyu et al., 2013b). At the same time, the effects of the original interdisciplinary care model were poorer for patients with cognitive impairment than without cognitive impairment (Shyu et al., 2012), and the model was not specifically designed from a family-centered perspective. Therefore, this study modified the original interdisciplinary care to involve family caregivers.

(A) Family-centered geriatric assessment. The geriatric consultation team has the tasks of (1) providing comprehensive geriatric assessment and medical supervision to detect any potential medical and functional problems, and (2) decreasing delays before surgery. In addition to receiving usual care, patients are contacted within the first day of admission by the geriatric nurse who completes the initial assessment and discusses it with the geriatrician. Family caregiver will be invited to participate in the assessment and provide information. This assessment collects information on medical and fall history, vital signs, physical examination, physical and cognitive functions, nutritional status, preoperative risks, current medications, and comorbidities. Based on this assessment/consultation, suggestions will be made to the primary surgeon regarding time of surgery, utilization of infection and thromboembolic prophylaxis, postoperative nutritional management, urinary tract management, and delirium management/prevention. The primary surgeon, geriatric nurse, and geriatrician will collaboratively develop a postoperative care plan.

In addition to receiving usual postoperative care, patients will be visited by the geriatric nurse on the first day after surgery to assess for signs of delirium, pain, and postoperative complications. If a patient presents any signs of delirium, cognitive impairment or any psychiatric symptoms, a geriatric psychiatrist will be consulted. If the patient has signs of other postoperative complications, the geriatrician will be consulted. Based on this geriatric consultation, suggestions are made to the primary surgeon on using infection and thromboembolic prophylaxis, postoperative nutritional management, urinary tract management, pain management, and delirium management. The health care team reviews the postoperative care plan to make any necessary changes.

In this study, the geriatric assessment will involve family caregivers. At the same time, the family members will be informed of the care plan and sign an agreement form.

(B) Family-centered continuous rehabilitation. The rehabilitation program will include (1) providing early postoperative rehabilitation to facilitate mobility, and (2) planning for hospital discharge with rehabilitation in the patient's usual environment. Postoperative rehabilitation starts on the first day after surgery and continues into the home setting after hospital discharge. During hospitalization, the geriatric nurse visits the patient once a day starting on the 1st day after surgery. For in-home rehabilitation, the geriatric nurse will visit patients to provide in-home rehabilitation training once every week during the first month, and once every 2 weeks for the second and third months following discharge. Both in-patient and in-home rehabilitation programs contain a hip fracture-oriented intervention and a general intervention for deteriorated physical fitness.

The hip fracture-oriented rehabilitation will focus on pain relief, range of motion, muscle strength and endurance, proprioceptive enhancement, and balance challenges. The general intervention to rehabilitate deteriorated physical fitness will emphasize exercises to increase physical fitness. Physical fitness is defined as "the ability to perform muscular work satisfactorily" (Shephard, 1978) and is assessed by five common parameters: aerobic capacity, anaerobic capacity, muscle strength and endurance, flexibility, and body composition (Skinner, Baldini, & Gardner, 1990). Therefore, as in our prior trial, the in-home muscle-strength training program will progress according to the patient's recovery condition in six stages: ankle pumping exercises, knee extension, gently bouncing jump with knee semiflexed and both feet on the floor, gently bouncing jump with knee semiflexed and single foot on the floor (Shyu et al., 2008, 2012), and ball-rolling activities (Shyu et al., 2013b). The last exercise uses lower extremities to enhance proprioception. Fast walking will be used at later stages to enhance aerobic capacity (Shyu et al., 2013b).

For the continuous rehabilitation program in this study, family caregivers will be taught the rehabilitation protocols as they are responsible for giving or managing patient care. Caregivers will be followed at each visit, progress of the rehabilitation will be monitored, and family caregivers will be asked to keep a diary to help monitor patients' rehabilitation adherence.

Family caregiver-training component. The family caregiver training will be a two-session in-home training program developed in our prior studies (Huang et al., 2013; Kuo et al., 2013) and delivered by trained registered research nurses. This training will be implemented during the 3rd and 4th home visits (2 and 3 weeks after discharge, respectively) when the discharge transition is more settled (Shyu, 2000). During the first caregiver training session (3rd home visit), the research nurse will work with the family caregiver on a structured guide to assess the condition of the hip-fractured patient with cognitive impairment, including habits, daily routines, preferences, behavioral problems and environmental safety and stimuli. The strengths, weakness, and resources of the family will also be assessed. The research nurse will work with the family caregiver to identify behavioral problems and symptoms to target. They will also use the PLST model to explore environmental stimuli (antecedents) and consequences of the targeted behavioral problems. They will then collaborate on a tentative plan to minimize these stimuli and decrease the targeted behavioral problems by modifying the daily schedule and environment.

During the second caregiver training session (4th home visit), the research nurse will confirm the behavioral problems and finalize the plan for handling specific behavioral problems. Specific suggestions will be re-emphasized and modified as needed to lower environmental stimuli of problem behaviors. The research nurse will also leave contact information for family caregivers if any problems arise between visits regarding delivery of the interventions.

Research nurses will make follow-up phone calls to family caregivers once a week for 2 weeks to determine if behavioral problems and symptoms are handled, whether the previously made suggestions are followed, and if there are any difficulties. Based on these discussions and evaluations, research nurses will give further suggestions. In the meantime, family caregivers in the control group will receive social contact follow-up phone calls at the same time as caregivers in the experimental group.

The sample size was estimated based on our interest in longitudinal changes in outcome variables. For sample estimation, the investigators used "Optimal Design" Software Version 3.01, free software with a graphical user interface to compute sample size for multi-level and longitudinal research (Spybrook, et al., 2011). Since the proposed intervention has never been conducted before, the investigators can only estimate the sample size based on prior clinical trials of an interdisciplinary care model used with a subsample of hip-fractured patients with cognitive impairment (Shyu et al., 2005, 2008) and a training program for family caregivers of patients with dementia (Huang et al., 2013; Kuo et al., 2013). To estimate patient outcome variables, outcome indicators of overall performance of ADLs and IADLs, cognitive function, and physical component summary scores (SF-36) were selected because they are the attributes most affected by hip fracture and most likely to change with intervention (Shyu et al., 2005, 2008). To estimate family caregiver outcomes, outcome indicators of preparedness, competence, and self-efficacy of managing behavioral problems were selected because they were the direct outcomes affected by our previous individualized, home-based family caregiver-training program (Huang et al., 2013). To estimate the sample size for examining the effects of the family-centered care model in this study, the alpha was set to 0.05 and sample size per group required for each indicator to achieve a power of 0.80. The majority of estimated overall sample sizes for the proposed study ranged from 30 to 116. Therefore, the study would require approximately at least 58 subjects for each group. Since fewer than 20% to 30% of cases were lost in our previous studies (Huang et al., 2013; Kuo et al., 2013), 76 patients in each group would be adequate for the proposed study. Based on our recent studies, It is expected to recruit 5 to 6 hip-fractured patients with cognitive impairment and at least 60 years old every month. Therefore, data collection will be conducted for 26 to 31 months and will follow patient subjects for 1 year after hospital discharge.

Procedure Subjects will be recruited by research assistants who will screen patient lists from the emergency room and hospital admissions daily to avoid losing potential cases. For patients who meet the study criteria, their primary physicians will be informed, and these patients will be invited to participate in this study. Elderly patients and their family caregivers who agree to participate will then be randomly assigned to the intervention group or the control group by a neutral party. Patients and their family caregivers in the experimental group will receive the family-centered care intervention, and those in the control group will receive usual care. Patients and families will not be aware of their group assignment and will be blinded to the care model they receive. Researchers who gather data will be independent of those who will deliver the interventions.

Patients and caregivers in both groups will be assessed five times longitudinally. The first assessment will be in the wards after surgery and before discharge. The second, third, fourth and fifth, assessments will be made at 1, 3, 6, and 12 months after discharge, respectively (Shyu et al., 2008, 2013b). Family caregivers will keep weekly diaries to record rehabilitation and cost information, which will be collected along with interview data. Diary recording will be promoted by providing rewards such as giving small gifts and constant feedback about patient's progress. Follow-up phone calls will be made for subjects who refuse further home visit but agree to give information over telephone calls. Those agree to accept telephone follow-ups will be telephoned to collect data on patients' self-care ability, service utilization, mortality, and family caregiver's preparedness and competence.

Training Program for Interveners and Data Collectors Interveners and data collectors will be trained based on our extensive experience from prior clinical trials. One registered nurse will be trained to deliver family-centered care, including the family-centered rehabilitation and discharge planning by the geriatric nurse specialist, Ching-Tzu Yang who is expected to graduate from her PhD program this year and become the post-doctoral fellow of this study. These nurses will also be trained by our co-PI, Dr. Huei-Ling Huang, who previously trained caregivers of patients with dementia (Hung et al., 2013, Kuo et al., 2013), to develop collaborative partnerships with family caregivers to identify environmental stimuli of and manage behavioral problems/symptoms of patients with cognitive impairment. The protocols of the care model will be outlined in a researcher-developed manual. The intervener nurses will participate in at least five training sessions: at least one introductory session to explain the study purpose, design, theoretical base, and procedures of the intervention program (2 hours), two sessions of simulated practice to become familiar with the manual (2 hours each), and two home visits supervised by the geriatric nurse specialist. After the intervener nurses achieve 100% consistency with the geriatric nurse specialist in care delivery, they can proceed independently. To ensure the quality of the intervention program, nurse interveners will attend regular team meetings with the surgeon, geriatrician, geriatric nurse, physical therapist, rehabilitation physician, and psychiatrist to discuss related issues and will be constantly supervised by the geriatric nurse specialist.

Another two registered nurses will collect data. These data collectors will have at least 3 years research experience and will receive two training sessions. In the first training session (2 hours), the PI or Co-PIs will explain the study purpose, design, instruments, and the geriatric nurse specialist will demonstrate a simulation interview with the family caregiver and patients. In the second training session (2 hours), the data collectors will practice collecting data from the patients and family caregivers under the supervision of the geriatric specialist. Data collectors will attend monthly research team meetings. Problems regarding data collection will be discussed and resolved on a regular, ongoing basis.

Data Analysis Analyses will be conducted under an intention-to-treat principle. Differences in baseline characteristics, including pre-fracture self-care ability, between the intervention group and control group will be assessed by one-way ANOVA or chi-square tests, with the significance level set to 0.05.

For Hypothesis 1, hip-fractured patients with cognitive impairment who receive the family-centered care model will have better clinical outcomes, self-care ability, cognitive function, HRQoL, and fewer behavioral problems than those who receive usual care. For Hypothesis 2, intervention participants will have more outpatient visits and fewer hospital readmissions and emergency room visits than those who receive usual care. Hypothesis 3, intervention participants' family caregivers will have better preparedness, competence, and HRQoL than those receiving usual care.

Changes in outcome variables (patients' cognitive function [MMSE, PRMQ, DRS-R-98], clinical outcomes, self-care ability, HRQoL [SF-36], and behavioral problems [CMAI], and family caregivers' preparedness, competence, and HRQoL) will be analyzed using hierarchical linear models (Raudenbush & Bryk, 2002). A series of multilevel linear models will be estimated using usual care (control group) as the reference category.

For binary outcome variables (dementia risk, delirium, subsequent falls, emergency room visit, hospital readmission, mortality, and complete independence in ADL [CBI = 100] and in CBI walking ability [walking ability item = 15]), a series of multilevel multinomial logit models will be estimated using usual care as the reference category. Doing so will allow us to estimate the odds for patient participants receiving family-centered care of being at risk for dementia or delirium, of using health services, of being completely independent in ADL and walking ability relative to the odds for the usual care group over time. When the investigators evaluate the non-linear function of changes with time, centering time will be at 3 and 6 months after discharge to minimize the possibility of multicollinearity because the most rapid improvement in recovery occurs during the first 6 months (Shyu et al., 2013b; Shyu, Chen, Liang, Wu et al., 2004). Finally, attrition will be accounted for with each dummy outcome variable, one identifying subjects who die during the 1-year follow-up period, and the other identifying those who drop out for other reasons.

Hypothesis 4, those received family-centered care will have fewer re-admissions and emergency room visits at a lower total costs than the usual care and their care givers will have lower productivity losses.

For the cost analysis, an "incremental cost-effectiveness ratio" will be calculated to demonstrate the additional costs for each elderly hip-fractured patient with better cognitive function outcomes and for clinical outcomes related to hip fracture within the 12-month period after hospital discharge. This ratio will then be used to determine whether the family-centered care model represents a more cost-effective care model for elderly hip-fractured patients. Additionally, a multi-way sensitivity analysis will be employed to test the robustness of the study results.

The study will be completed and all hypotheses will be tested. Moreover, the results will provide a reference for planning future care models for elderly hip-fractured persons with other comorbidities.

Inclusion Criteria:

• Subjects are:

  1. age 60 years or older,
  2. admitted to CGMH due to one-side hip fracture, and being diagnosed as needing surgery,
  3. assessed as having cognitive impairment by the Chinese Mini-Mental State Examination (CMMSE) (CMMSE score < 21 with < 6 years education, or CMMSE < 25 with ≥ 6 years education; Yip et al., 1992),
  4. having a primary family caregiver,
  5. living in northern Taiwan (i.e., greater Taipei area, Keelung, Taoyuan, or Shin-Ju province).

• Family caregivers:

  1. age 20 years or older,
  2. responsible for providing direct care to or supervising care received by the patient.

Exclusion Criteria:

• Subjects are

  1. cognitively intact by CMMSE,
  2. without a primary family caregiver,
  3. terminally ill,
  4. severe cognitive impairment such that they are completely unable to follow orders (CMMSE < 10; Yip et al., 1992).

All participants will be informed that there two groups of the study. They will be randomized and allocated to one group and the research assistant explains the study to you does not know which group you will be, nor does the outcome assessor.