Smart Care for Older Persons Recovering From Hip-fracture Surgery

Specific Aims and Hypothesis:

The proposed research aims to examine the effects of a Smart Care Model (SCM) on the health outcomes of older patients recovering from hip-fracture surgery (hip-fractured older patients) and on their caregivers' caregiving burden. The SCM entails real-time activity monitoring and feedback using smart clothing with sensors embedded in a computer network.

The proposed study has the following specific aims.

1. To pretest, finalize and evaluate the protocol for the SCM.
2. To evaluate the effectiveness of usual care and the SCM for hip-fractured elders 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, subsequent falls, health-related quality of life (HRQoL), adherence, and service utilization as well as family caregivers' preparedness, perceived balance between competing needs, depressive symptoms, and HRQoL. During the first year after hospital discharge, hip-fractured elderly patients who receive the SCM will 2-1. Have better clinical outcomes, self-care ability, adherence and HRQoL and fewer subsequent falls than those who receive usual care. 2-2. Have more outpatient visits and fewer hospital readmissions and emergency department visits than those who receive usual care. 2-3. Family caregivers of hip-fractured patients receiving the SCM will have better preparedness, perceived balance between competing needs, and HRQoL as well as fewer depressive symptoms than caregivers of patients receiving usual care.
3. If the effectiveness of the SCM is established, its cost-effectiveness will be analyzed from the perspectives of the health care system and society. The "incremental cost-effectiveness ratio" will be calculated to show the incremental costs for an additional quality-adjusted life year (QALY) of elderly hip-fractured patients receiving the SCM.
4. To qualitatively explore the experiences of receiving the SCM for older hip-fractured persons and their family caregivers with in-depth interviews. The qualitative and quantitative components of the proposed study will be integrated to finalize the typical protocol, context and cases for implementing the SCM for older hip-fractured persons.

Basic Design:

Mixed methods,a concurrent embedded strategy in which quantitative and qualitative data are collected at the same time, with the quantitative component being predominant and the qualitative component collected from a sample subset , will be implemented. For the quantitative component, a randomized experimental design will be used to examine the effectiveness of the Smart Care Model (SCM) over a 5-year period. For the qualitative component, we will concurrently collect and compare qualitative data from a sample subset of patients and family caregivers who receive the SCM and those who receive usual care.Finally, the qualitative and quantitative data will be merged to provide comprehensive evidence and finalize protocols and typical cases for the SCM for hip-fractured older persons. The mixed-method approach has the advantages of confirming findings, validating and deepening understanding of the effects of care models, and broadening insights into the different issues underlying care models.

Intervention Models Participants in this trial will receive two care models: usual care and the Smart Care Model (SCM). The SCM will have two components: fall risk monitoring with an alarm, and activity-level monitoring with feedback. These SCM components are designed to enhance caregiver competence, decrease patient risk for subsequent falls, and improve postoperative recovery for hip-fractured patients with cognitive impairment. We have integrated the technology of remote monitoring with smart clothing into continuous rehabilitation and discharge planning, along with geriatric consultation. We will compare the effects of the SCM and usual care in this trial.

Participants Subjects will be recruited from the trauma wards of CGMH at Linkou and New Taipei Tucheng hospital.For the quantitative component, the sample size was estimated based on our interest in longitudinal changes in outcome variables. We plan to obtain at least 126 participants (63 for each group). Since fewer than 20% of cases were lost in our previous study, we estimated that 158 participants (79 in each group) would be adequate for the proposed study.

For the qualitative component, a subset sample of participants and family caregivers who received the SCM will be interviewed one-on-one to explore their experiences of receiving the SCM and compared with a subset of participants who received routine care. Ten patients and their family caregivers from the experimental group and 10 from the control group will be recruited.

Procedure As shown in Figure 5, potential subjects will be recruited by research assistants who will screen patient lists from the emergency department and hospital admissions twice a day to avoid losing potential cases. For patients who meet the study criteria, their primary physicians will be informed, and these potential subjects and their family caregivers 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 or control group.

Randomization to groups will follow a randomization table created by our university statistician. The project manager will create sequentially numbered sealed envelopes containing group assignment. Each consenting patient and their family caregiver will receive the envelope in the order in which they are enrolled. The project manager will have access to the spreadsheet that indicates the randomization order, but not the research nurses who assess outcomes. After randomization, patients and their family caregivers in the experimental group will receive the SCM, and those in the control group will receive usual care. Patients and families will not be aware of their group enrollment and will be masked to the care model they receive. Research nurses who collect data will be independent of those who deliver the interventions. For the experimental group, the geriatric nurse will install sensors with the engineer during the home-environment assessment before hospital discharge and will spend around 60 minutes to help caregivers and patients use the smart clothing at home. Then a registered nurse will visit participants' homes once a week for 1 month after discharge, then once each month to help solve problems in using the smart-care system, download data, along with in-home rehabilitation. Each visit will take 30 to 60 minutes.

Patients and caregivers in both groups will be assessed 8 times longitudinally, each time around 30 minutes. The first assessment will be within 24 hours after hospital admission in the emergency department, and the second assessment will be in the wards after surgery and before discharge. The third, fourth, fifth, sixth, seventh and eighth assessments will be made 1, 3, 6, 12, 18 and 24 months after discharge, respectively. Families will keep weekly diaries to record rehabilitation and cost information. The incentive for diary recording will be enhanced by providing rewards. Qualitative data on experiences of using the SCM will be collected through 20-minute face-to-face interviews with a subset of at least 10 SCM-group and 10 routine-care patients and their family caregivers. They will be told that the interview questions have no right or wrong answers.

Data Analysis Quantitative Data Quantitative data will be analyzed under an intention-to-treat principle. This analysis will test three hypotheses about participants who receive the SCM vs. those who receive usual care. Hypothesis 2-1: elderly hip-fractured patients who receive the SCM will have better clinical outcomes, self-care ability, adherence and HRQoL and fewer subsequent falls; hypothesis 2-2: elderly hip-fractured patients who receive the SCM will have more outpatient visits and fewer hospital readmissions and emergency department visits; and hypothesis 2- 3: family caregivers of those who receive the SCM will have better preparedness, perceived balance between competing needs, and HRQoL, as well as fewer depressive symptoms than caregivers of those who receive usual care.

Changes in continuous outcome variables will be analyzed by hierarchical linear models (HLM), with usual care (control group) as the reference category. For binary outcome variables, a series of multilevel multinomial logit models will be estimated using usual care as the reference category. When we evaluate non-linear changes with time, the possibility of multicollinearity will be minimized by centering time at 3 and 6 months after discharge since the most rapid improvement in recovery occurs during the first 6 months. Finally, attrition will be accounted for by two dummy variables, one identifying subjects who died during the 1-year follow-up period, and the other identifying those who dropped out for other reasons. In addition, family caregiver characteristics (gender, age, education background and relationship with the hip-fractured person) will be compared between the experimental and control groups. Characteristics that differed significantly will be controlled in the HLM data analysis.

For specific aim 3, the "incremental cost-effectiveness ratio" will be calculated to demonstrate the incremental cost for an additional QALY of elderly hip-fractured patients receiving the SCM. Additionally, the robustness of the study results will be tested using multi-way sensitivity analysis.

Qualitative Data Transcripts and field notes will be coded line by line, with data collection and analysis concurrent. Data analysis will use three types of codes: descriptive, interpretative, and pattern. Descriptive codes imply little interpretation.Interpretive codes analyze or infer meaning from participants' words. Pattern codes are more inferential and explanatory. Data will be analyzed both within and across cases. Data analysis and collection will be discussed periodically among the research team. Memos will be written concurrently with data collection, coding and categorizing to record researchers' thoughts during data collection and analysis.