The DIVERT-CARE (Collaboration Action Research & Evaluation) Trial

BACKGROUND

Context:

Public home care services are delivered to at least 6% of Canadians age 65-74, 15% age 75-84 and 32% age 85 or older. These clients are medically complex, access care across settings, have very high emergency department utilization rates, and have relatively poor access to effective chronic disease management. Their frequent emergency department use is not aligned with chronic disease management or geriatric care principles.

Effective chronic disease management models employ multiple components delivered by a coordinated multidisciplinary team. According to the 'chronic disease management model', home care plays a complementary function to medical practitioners. Clinical and self-care support as well as case management are among the most effective components in chronic disease management. Self-care education and support has been shown to improve health outcomes across chronic diseases. Sustained follow-up by nurses or other non-medical staff is also effective.

Canadian home care providers, historically focused on the delivery of personal support services, have started to develop supportive chronic disease management capacity (e.g. specialist nurse monitoring). However, most trials exclude frail seniors and are not specific to home care, which leaves little evidence to inform chronic disease management practices.

Previous Work:

Effective chronic disease management in home care has been limited by insufficient targeting of clients at most need or most likely to benefit. The investigators developed and validated a prognostic case-finding tool for home care known as the Detection of Indicators and Vulnerabilities of Emergency Room Trips (DIVERT) Scale that has been recommended for home care. It can be derived in real time from the Resident Assessment Instrument-Home Care (RAI-HC) standardized home care assessment used in 9 Canadian provinces as well as Estonia, Finland, Hong Kong, Iceland, Ireland, Italy, Japan, the Netherlands, New Zealand, Singapore, Spain, Switzerland, and some U.S. states. Cardio-respiratory symptoms and conditions are prominent predictive elements of the DIVERT Scale.

Based on evidence-based guidelines developed - in part - by the investigator team, the investigators piloted the combined use of the DIVERT Scale with a multi-component cardio-respiratory management model. The pilot pragmatic cluster randomized trial included over 200 clients across six home care caseloads. The pilot was recognized on the 2015 Ontario Minister's Medal Honour Role for Excellence in Health Quality.

OBJECTIVES The main objective is to evaluate the effectiveness and preliminary cost-effectiveness of a targeted, person-centered cardio-respiratory management model.

Research questions:

1. Among home care clients experiencing cardio-respiratory symptoms (objectively targeted using the DIVERT Scale), can a guideline-based, feasible, multi-component/complex, cardio-respiratory management model, compared to regular care, reduce/postpone unplanned emergency department (ED) visits, improve patient activation, reduce cardio-respiratory symptoms, reduce unplanned hospital use, or improve health-related quality of life for clients, over a 6-month follow-up period from baseline?
2. If effective, is there evidence for cost-effectiveness?

METHODS

Intervention:

From evidence-based guidelines developed - in part - by this team, extensive client profiling, and input of clients/families as well as health professionals, the investigators developed a person-centred, multi-component cardio-respiratory management model containing the following components: scheduled nurse-led self-management support (based on a training program, and tool-kit), advanced care and goal planning, clinical pharmacist medication reconciliation, team case rounds, SBAR (the Situation, Background, Assessment and Recommendation) communication protocol with primary care, and a standardized transition package. Each component has a specific objective within the model, however the manner in which it is delivered may be adapted. The approach is based on evidence of effective implementations in other fields, and includes all elements for 'person-centred care'.

Design:

The investigators will conduct a pragmatic, cluster-randomized trial that leverages secondary electronic patient records. Sixty-six (66) geographic areas (home care caseloads) will be randomized using a 1:2 (intervention: control) ratio.

Population:

Approximately 1,080 long-stay home care clients living in a non-institutional setting from 66 distinct geographic home care caseloads in Canada. Inclusion criteria: clients who are in the DIVERT Scale target groups (9,10, 14, 15) will be included (i.e. at least one cardio-respiratory symptom (chest pain, dyspnea, dizziness, irregular pulse) and at least one cardiac condition (congestive heart failure or coronary artery disease)). The investigators will utilize the DIVERT Scale as an objective measure of eligibility. Clients who are receiving palliative care (i.e., prognosis of less than six months to live, K8e from Resident Assessment Instrument-Home Care (RAI-HC)) or dialysis (P2g from RAI-HC) at baseline assessment will be excluded.

Recruitment:

Eligible clients will be identified by each case manager during regular home care enrollment and reassessment using the RAI-HC assessment (standard practice) for between 6 to 8 months. Eligible clients will be automatically included into the intervention or 'regular care' control on an intent-to-treat basis. Each home care provider's process for attaining consent to intervention will apply. Study informed consent will not be sought given that the cardio-respiratory management model is considered best practice care and is offered - whole or in part - at the full clinical discretion of the home care provider as per existing practice. The cardio-respiratory management is accepted - whole or in part - at the full clinical discretion of the clients as per existing practice. Trial investigators have no part in the intervention pilot data collection, individual care decision-making, or records management during the study period beyond providing guidance and access to resources. All trial data will be extracted from secondary data and anonymized prior to transmission by home care provider staff.

Outcomes and Follow-up:

Primary Outcomes:

• Difference in days to first unplanned emergency department visit within 6 months of baseline
• Presence of cardio-respiratory symptoms over 6-month follow-up
• Cost of all home care services within 6 months of baseline

Secondary Outcomes:

• Change in patient activation (based on 13 item patient activation measure) over 6-month follow-up
• Change in health-related quality of life (based on Minimum Data Set Health Status Index) over 6-month follow-up
• The difference in the number of unplanned emergency department visits per days at risk (alive, not institutionalized) within 6 months of baseline
• Number of unplanned inpatient (medical) hospital days per days at risk (alive, not institutionalized) within 6 months of baseline

A 6-month follow-up will be used to reflect the time at greatest risk among home care clients.

Sample calculations/Assumptions:

Primary outcome: Time to first unplanned emergency department visit within 6 months of baseline (alive, not institutionalized). Ability to detect a hazard ratio of 0.75.

Mean size of a home care caseload: 120 clients Mean prevalence of DIVERT target group in each caseload: 30% Expected recruitment over 6 months per caseload: 30 Mean cluster size: 30 Allocation: 1:2 (intervention: control)

Simulations using retrospective secondary data sources were undertaken to explore the power of a hypothetical DIVERT-CARE trial conducted in the Hamilton-Niagara-Haldimand-Brant (HNHB) health region of Ontario from December 2014 to June 2015. The simulations found that 60 HNHB home care caseloads randomized at a 1:2 intervention to control ratio could expect to enroll 1,809 patients across seven months. The simulation linked patients to their actual emergency department utilization records, and extended the time to first ED visit figures for patients in 20 randomly selected intervention caseloads to achieve a hazard ratio of 0.75, a figure chosen to be more conservative than the pilot study but still clinically significant. The overall event rate was 35.5% in the intervention group and 44.8% in the control. The median time to first visit was 88 days in the intervention group and 75 days in the control. The power of the simulated DIVERT trial was 94.12% with a two-sided alpha of 0.05. Simulations with a hazard ratio of 0.80 yielded a power of 79.96% The intracluster correlation coefficient (ICC) was estimated to be 0.005.

Main Analyses:

The primary hypothesis will be assessed through a multilevel discrete-time proportional hazards model. The dependent variable will be days until first emergency department visit, censored at date of home care discharge for any reason. Caseload and partner site will be included as nested random effects. The hazard ratio, 95% confidence interval, and p-value will be reported for the treatment group effect and covariates when applicable. A two-sided alpha level of 0.05 will be used to judge statistical significance.

The economic evaluation will examine total care costs, controlling for length of stay, between treatment groups to compare incremental costs to incremental effects (i.e. cost per ED visit averted).

The secondary hypotheses will be evaluated by a multilevel general linear model. Six-month change in health-related quality of life, number of symptoms, patient activation, and number of emergency department visits will be the respective dependent variables. Caseload and home care provider site will be included as nested random effects. The unit change in each measure, 95% confidence interval, and p-value will be reported for the treatment group effect and covariates when applicable. A two-sided alpha level of 0.05 will be used to judge statistical significance.

Main Secondary Datasets:

RAI-HC: The Resident Assessment Instrument Home Care (RAI-HC) is a standardized comprehensive assessment containing approximately 200 items, has been found to document major domains of health reliably. The Canadian Institute for Health Information (CIHI) collects and reports this data from publically funded home care programs. CIHI conducts its own quality assurance procedures to ensure accuracy and completeness. CIHI also provides a document describing the interpretation of all included variables for the RAI-HC dataset.

The Client Health and Related Information System: The Client Health and Related Information System (CHRIS) is an online patient management system for Community Care Access Centre (CCAC) use. It includes patient assessments, documents, provider and vendor contracts, billing of services information, and medical supplies and equipment rental costs. CHRIS continues to maintain the quality and integrity of its data independently. Cost data from sites outside of Ontario will be estimated as per the services rendered.

National Ambulatory Care Reporting System: The National Ambulatory Care Reporting System (NACRS) contains data from all hospital-based and community-based ambulatory care. It is collected and maintained by CIHI, which conducts its own quality assurance procedures to ensure accuracy and completeness. CIHI also provides a document describing the interpretation of all included variables for this dataset.

Discharge Abstract Database: The Discharge Abstract Database (DAD) captures administrative, clinical and demographic information on hospital discharges (including deaths, sign-outs and transfers). It is collected and maintained by CIHI, which conducts its own quality assurance procedures to ensure accuracy and completeness. CIHI also provides a document describing the interpretation of all included variables for this dataset.