The purpose of this pragmatic study is to investigate whether implementation of a comprehensive post-acute stroke service model that integrates Early Supported Discharge (ESD) and Transitional Care Management (TCM) for stroke survivors discharged home improves functional outcomes post-stroke, reduces caregiver stress, and reduces readmission rates.
Stroke mortality is 20-40% higher in North Carolina (NC) than in the overall United States. After discharge, stroke patients are at high risk for complications. Although a model of stroke post-care (early supported discharge) exists in Europe and Canada, it has not been adapted for and tested in the US, although patients and stakeholders attest that post-acute care does not meet their needs. Transitional care services from hospital to home are now reimbursed by Centers for Medicaid and Medicare Services (CMS), but only for 30 days after discharge. The study team proposed a pragmatic, cluster randomized trial of 41 NC hospitals to determine the effectiveness of COMprehensive Post-Acute Stroke Services (COMPASS), a patient-centered intervention uniting transitional care management services and elements of early supported discharge in stroke patients discharged directly home.
The study team will build on the successful North Carolina Stroke Care Collaborative (NCSCC) registry, a prospective stroke database in which 51 (of 113) hospitals in NC enroll patients. In preparation for COMPASS, the study team engaged these hospitals via webinars. Over 80% of NCSCC hospitals demonstrated interest in participation and provided letters of support.
The main question of this pragmatic trial is: Does implementation of COMPASS for all stroke patients discharged directly home improve functional outcomes as measured by the Stroke Impact Scale-16 (SIS-16) at 90 days post stroke? The primary aim is to: compare the COMPASS model versus usual care on stroke survivors' self-reported functional status at 90 days post-stroke. The secondary aims are to determine if the COMPASS model: 1) reduces caregiver strain (Modified Caregiver Strain Index, MCSI) 90 days post-stroke; 2) reduces all-cause 30- and 90-day readmissions; 3) affects mortality, health care use, continuity of care, and use of transitional care management billing codes, using claims data up to one year after stroke hospitalization; and 4) differentially affects primary and secondary outcomes by race, sex, age, stroke severity, and insurance status. Patients over age 18 admitted to a NCSCC hospital with a diagnosis of ischemic or hemorrhagic stroke or transient ischemic attack and discharged from acute care hospitalization to home will be included (about 6,000 patients/year).
NCSCC participating hospitals will be randomized (stratified by stroke volume and primary stroke center status) to receive COMPASS or usual care (control group) in Phase 1. In Phase 2, usual care hospitals will cross over to COMPASS, while the early intervention hospitals sustain the intervention using hospital-based resources. The trial has three integrated intervention components: 1) COMPASS, which combines transitional care services provided by advanced practice providers (APPs) and early supported discharge services coordinated by the Post-Acute Coordinators (PAC); 2) COMPASS-funded post-acute care coordinators who will engage patient and stakeholder communities to improve post-acute stroke comprehensive stroke services; and 3) development of a stroke metrics score card for NCSCC hospitals and primary care providers. Well-trained APPs and coordinators will have access to online learning and ongoing support/consultation from WFBH personnel and board-certified vascular neurologists.
The study team will assess 90-day and 1-year outcomes. Outcomes at 90-days will be assessed by telephone surveyors blinded to patient's group assignment. Patients will be informed about COMPASS in the hospital, and can opt out of 90-day phone follow-up at that time. Those who agree to be surveyed will be asked to provide informed consent at 90 day phone call to collect outcomes data and link their data to claims. The primary outcome will be patient-reported functional status (SIS-16). Secondary outcomes at 90 days include: caregiver stress (MCSI); unadjusted 30- and 90-day all-cause readmissions captured via claims data; proportion of patients with 7- and 14-day post-stroke hospitalization; mortality; continuity of care; use of transitional care management billing codes; medication adherence; and health care use (emergency department visits, number of hospital admissions and inpatient days, and admissions to skilled nursing and inpatient rehabilitation facilities). Tertiary analysis include: cognitive status; blood pressure management; depression; use of community resources; recurrent stroke; physician follow-up; falls; satisfaction with care; use of TCM billing; fatigue; modified Rankin; and physical activity.
As the primary endpoint is a continuous variable, the study team will use a mixed model (including fixed and random effects) to compare COMPASS to control. The intraclass correlation, ρ, obtained from data collected at eligible NCSCC hospitals, is estimated as 0.036 based on preliminary data. Conservatively doubling this to 0.072 there is 90% power to detect a 0.259 standard deviation difference between COMPASS and control groups for the outcomes of interest. Additionally, there is 90% power to detect a 0.305 SD difference in any subgroup at least 20% of the overall sample.
This proposal is led by three highly experienced researchers as co-principal investigators. The team includes expertise in stroke care, large clinical trials, biostatistics, managing clinical registries, survey and acquisition of patient or proxy-reported outcomes, community-based practice improvement, building community coalitions to reduce readmissions, claims analyses, registry management, translating evidence into practice with large multi-site collaboratives, and engaging patients and stakeholders in research. The planning phase of this project has been guided by patients and stakeholders. Each community will form a coalition to advise and support implementation of COMPASS, provide feedback to the team, and help create sustainability. If the COMPASS model shows effectiveness, engaged patients and stakeholders will be key partners to disseminate and implement COMPASS throughout the state and beyond.
- COMPASS Intervention Other
Intervention Desc: A Post-Acute Coordinator (PAC) will visit each patient prior to discharge from the hospital. Patient will receive a follow-up telephone call two days after having been discharged. 7-14 days after discharge, the patient will attend post-acute stroke clinic visit and receive an assessment from an Advanced Practice Provider (APP), a brief patient-reported functional assessment to generate an individualized Care Plan, and referrals from an APP. The patient's primary caregiver will be assessed to ensure caregiver availability and ability to support the patient and the caregiver's ability to cope with the new challenges of caregiving. Patient will receive a call at 30 and 60 days post-discharge for follow-up of functional status, recovery, risk factor management and their access or utilization of recommended services. ARM 1: Kind: Experimental Label: COMPASS Intervention Description: Participating hospitals randomized to the intervention will change the structure and process for delivery of post-acute stroke care.
- Allocation: Randomized
- Masking: Single Blind (Outcomes Assessor)
- Purpose: Health Services Research
- Endpoint: Efficacy Study
- Intervention: Parallel Assignment
|Type||Measure||Time Frame||Safety Issue|
|Primary||Stroke Impact Scale (SIS-16)||measured 90 days post-stroke||No|
|Secondary||Modified Caregiver Strain Index||measured 90 days post-stroke||No|
|Secondary||All-cause hospital readmissions at 30- and 90-days post-stroke||measured 1 year after index discharge||No|
|Secondary||Mortality||measured 1 year after index discharge||No|
|Secondary||Number of Patient Emergency Department Visits||measured 1 year after index discharge||No|
|Secondary||Number of Patient Hospitalizations||measured 1 year after index discharge||No|
|Secondary||Number of Patients with Medication Adherence as measured by the Morisky Medication Adherence Scale (MMAS-4)||measured 1 year after index discharge||No|
|Secondary||Number of Patient Admissions to Skilled Nursing Facilities||measured 1 year after index discharge|
|Secondary||Number of Patient Admissions to Inpatient Rehabilitation Facilities||measured 1 year after index discharge|
|Secondary||Self-rated general health||measured 90 days post-stroke|
|Secondary||Blood pressure management effectiveness||measured 90 days post-stroke|