An interdisciplinary intervention for older Taiwanese patients after surgery for hip fracture improves health-related quality of life

Inclusion criteria for subjects were (1) ≥ 60 years, (2) admitted to hospital for an accidental single-side hip fracture, (3) receiving hip arthroplasty or internal fixation, (4) able to perform active movement against gravity and some resistance or full resistance, and with pre-fracture Chinese Barthel Index (CBI) score > 70, and (5) living in northern Taiwan. Patients were excluded if they were (1) severely cognitively impaired and completely unable to follow orders (determined by a score < 10 on the Chinese Mini-Mental State Examination [MMSE, 23]), or (2) terminally ill.

Muscle power of the unaffected limb at admission (post hip fracture) was assessed by a trained research nurse. The pre-fracture CBI score was obtained from participants or/and carers by the research nurse. Using the Barthel Index to retrospectively assess pre-fracture physical functioning has been suggested by the UK National Health Service [24]. Although outcomes of cognitively impaired elders have been improved by intensive rehabilitation programs [25‐27], this study included only participants with mild and moderate cognitive impairment. This decision was based on our reasoning that those with severe cognitive impairment (MMSE score < 10) are disoriented to time, place, and persons, have lost their learning ability due to severe memory deficit, and have difficulty following directions [28], which might require different protocols for the rehabilitation intervention. To avoid complicating the study design, we included only participants with mild and moderate cognitive impairment.

The sample was recruited and followed from September 2001 to November 2003 (Figure 1). Of those who met the criteria and agreed to participate (n = 162), 80 were randomly assigned to the experimental, and 82 to the control group. Patients who declined to participate (n = 134) and those who agreed (n = 164) did not differ significantly in age (p = 0.322) and gender (p = 0.517). Based on our pilot study data, a power of 0.80, and an alpha of 0.05, we estimated a sample size of 65 subjects in each group to obtain a median effect size of 0.50 [29] for improved performance of activities of daily living (ADLs) measured by the CBI (experimental vs control = 38.5 vs 31.5) and 61 subjects in each group for improved physical functioning measured by the physical function scale of the Taiwan version of the Medical Outcomes Study Short-Form 36 (SF-36) [30] from post-surgery to the third month after discharge (experimental vs control = 22.9 vs 11.1) [21, 22]. To allow for 18% to 24% potential dropouts, we therefore aimed to recruit around 80 subjects in each group.

Current practice in caring for elders with hip fracture in Taiwan lacks well-organized interdisciplinary care protocols and continuity of care. After a fall incident, patients usually are sent directly to the hospital emergency room (ER) and are cared for by orthopedists. Elderly patients with a femoral neck fracture of subcapital and transcervical types are treated with hip hemiarthroplasty and the basal neck type is treated with closed reduction using dynamic hip screw fixation. Before surgery, referrals are occasionally made for internal medicine care. Routine examinations before surgery include EKG, blood chemistry tests and cell counts, and X-rays. In the first 2 to 3 days after surgery, patients receive antibiotics and pain medication, and are taught to exercise with caution while still in bed. Physical therapy usually starts on the second or third day after removal of the hemovac, which is used for postoperative wound drainage. The usual hospital stay following surgery is around 7 days and no in-home programs are provided. Adherence to clinical follow-ups at 1, 3, 6 and 12 months after surgery is poor. Patients are encouraged to ambulate with protected weight bearing for 3 months. Using a walker and touching the ground lightly are recommended.

The intervention program developed for this study included three components: geriatric consultation service, rehabilitation program, and discharge-planning service [21, 22].

Unlike routine care that provides no geriatric assessment, the intervention program offered geriatric consultation that was delivered by a geriatric nurse and a geriatrician. Before surgery, the geriatric nurse contacted the patient and completed the initial assessment to obtain medical and fall history, vital signs, physical examination, physical and cognitive functional assessment, nutritional status, preoperative risk assessment, current medications, and comorbidities. This information was reviewed by the geriatrician, who conducted geriatric assessments. The geriatric assessment was conducted for all subjects in the experimental group, and clinical suggestions were made for patients ≥ 80 years old, with high operative risk, poor nutritional status, cognitive impairment or disorientation, or those with unstable comorbid conditions (Figure 2). Based on this assessment, the geriatrician made suggestions to the surgeon regarding time of surgery, antibiotics and thromboembolic prophylaxis, postoperative nutritional management, urinary tract management, and delirium management/prevention. The geriatrician's suggestions were generally followed except the use of anticoagulants for thromboembolic prophylaxis underutilization rates. On the first day after surgery, the geriatric nurse visited the patient to assess for signs of delirium, pain, and postoperative complications. Based on this assessment, the geriatric nurse revised suggestions for the postoperative care plan to the surgeon. Each pre-surgical nursing assessment and geriatrician visit lasted around 60 minutes and the post-surgical nursing visit lasted 30 minutes.

Unlike routine care in which postoperative rehabilitation did not start until 2-3 days following surgery and no in-home rehabilitation was provided, the rehabilitation program in this study emphasized providing early postoperative rehabilitation and in-home rehabilitation. This rehabilitation program was delivered by the geriatric nurse, physical therapist, and rehabilitation physician. Both the in-hospital and in-home components of the rehabilitation program contained a hip fracture-oriented intervention and a general intervention for deteriorated physical fitness. The hip fracture-oriented rehabilitation emphasized pain relief, range of motion, muscle strength and endurance, proprioceptive enhancement, and balance challenges. The general intervention for deteriorated physical fitness rehabilitation emphasized exercises to increase physical fitness, including aerobic capacity, anaerobic capacity, muscle strength and endurance, and flexibility.

During hospitalization, rehabilitation started on the first day post-surgery and was delivered once a day by the geriatric nurse. According to patient's condition, the exercise protocol progressed from ankle exercises in bed to knee and hip joint flexion and extension exercises, to walking, and then climbing up and down stairs using a walker. During this period, around 4 rehabilitation sessions were provided by the geriatric nurse. In addition, the physical therapist made 2 visits to assess the patient's condition and provide rehabilitation sessions. The rehabilitation physician also made one visit to provide consultation.

For in-home rehabilitation, according to the patient's condition, the exercise protocol emphasized ankle dorsiflexion with knee extension, isometric full knee extension, gently bouncing vertical jump with knee semiflexed and foot on the floor, and ball rolling activities to enhance proprioception. During the first month, in-home rehabilitation was delivered by nurses once per week. During the second and third months, in-home rehabilitation was delivered by nurses once every 2 weeks. In addition, the physical therapist conducted one assessment within the first week after discharge, and at the third week and third month after discharge.

In addition to the geriatric consultation and rehabilitation components, a discharge service component was delivered by geriatric nurses. Unlike routine care, in which discharge planning is not provided to all patients and does not include home assessment, the geriatric nurse in this study assessed patients at discharge for caregivers' competence, resources, family function, patient self-care ability, patient and family caregiver needs for continuing health and social services, and made necessary referrals during hospitalization. The geriatric nurse also made a home visit before discharge to assess the home environment and suggested environmental modifications. The nurse also made phone calls to remind patients about follow-up visits to clinics.

Differences in baseline characteristics between the experimental and control groups were assessed by two-sample t-tests or chi-square tests. Effects of the interdisciplinary intervention were evaluated using a generalized estimating equation (GEE) approach to account for correlations in repeated observations over time. For a given outcome variable, the GEE model includes the following predictors: treatment (1 = experimental group, 0 = control group), and three dummy variables representing measurements made at 3, 6, and 12 months after hospitalization (with 1 month after discharge as the reference). GEE analyses were carried out using SAS Win 8.0.

All analyses were undertaken according to the intention-to-treat approach [40]. Missing data due to attrition (i.e., mortality, loss to follow-up, and refusal to participate) after randomization were imputed using multiple imputation [41, 42]. For instance, we imputed missing data on performance of ADLs for subjects who dropped out or refused to participate after randomization by using baseline data (e.g., age, gender, education, functional status before fracture, functional status at discharge, and repeated observations of HRQOL if available).

For those who died during the trial, no imputation was made after death except for one subject in the experimental group who died before assessment at the first month after discharge. This decision was based on the principle of intent to treat, i.e., all randomized subjects should be included in the analysis [40]. In addition, the imputed value can be regarded as the outcome shortly before death. Furthermore, it should be noted that GEE allows partial information to be used in the analysis, i.e., data obtained before subjects' death can still contribute to estimating parameters.

Three complete data sets were imputed using NORM software developed by Schafer [43], and each set was analyzed. Estimates were then averaged across the three imputations to derive a single point estimate.

Of the 162 participants in the final sample, the majority (68.5%) were female, with an average age of 78.16 years (SD = 7.76). Around half the participants were married (51.9%) and illiterate (48.8%), 63% received internal fixation, and 37% received arthroplasty. The mean pre-fracture CBI was 96.08 (SD = 6.47), representing good independence in performing ADLs, and 84.6% could walk independently before the fracture. The experimental and control groups did not differ significantly in baseline characteristics (i.e., gender, age, marital status, education, type of surgery, pre-fracture performance of ADLs, body mass index [BMI], and walking ability) or timing of surgery and 1-year mortality (Table 1). Eighty-seven hip fractures were treated with internal fixation and followed-up for 1 year. Seven fractures failed to heal, and the union rate was 91%. However, only 80% of patients regained ambulatory ability. Fifty-one hip fractures were treated with arthroplasty and followed-up for 1 year. However, 90% of patients regained ambulatory ability. Two patients' hip prostheses were dislocated (2/51) and reduced with a closed technique [44]. The two groups were similar in weight bearing status, and were encouraged to ambulate with protected weight bearing for 3 months.

Regression coefficients for overall effects by time and intervention are presented in Table 2. Outcomes for the experimental and control groups are compared according to the GEE approach using the first month and control group as baseline (Table 3). The intervention had a significant effect on subjects' bodily pain, vitality, mental health, physical function, and role physical (Table 2). After controlling for time, the experimental group had significantly better overall outcome in bodily pain (β = 9.38, p = 0.002), vitality (β = 9.40, p < 0.001), mental health (β = 8.16, p = 0.004), physical function (β = 16.01, p < 0.001), and role physical (β = 22.66, p < 0.001) than the control group at any time point during first year after discharge (Tables 2 and 3). β can be interpreted as representing the intervention effect on variables of HRQOL over the 12-month period after hospital discharge. In other words, the benefits of the interdisciplinary intervention program on bodily pain, vitality, mental health, physical function, and role physical lasted beyond 3 months after discharge. Furthermore, the benefits of the interdisciplinary intervention appeared to be greater for physical-related health outcomes such as physical function (Figure 3) and role physical (Figure 4).

For the time effect, bodily pain, social function, role emotional, physical function, and role physical were significantly better at the third, sixth, and twelfth months after discharge than at the first month after discharge. Mental health at the sixth month after discharge was significantly better than at the first month after discharge (Table 2).