Team-based Versus Primary Care Clinician-led Advance Care Planning in Practice-based Research Networks

Background and Significance:

Many people die after living with chronic conditions that have a broadly predictable course. This should give patients and families the opportunity to consider what is most important to them and plan for the care they wish to receive in their final months and at the end of life. However, in many cases health care defaults to more testing and treatments, creating a potential mismatch between healthcare services and what patients and families want.

Serious illness care planning occurs once a patient with advanced illness reaches a life expectancy of one to two years. Patients face a wide range of choices that can profoundly affect their quality of life. Planning for the last months and days may improve quality of life, by assuring health care supports patient goals and that preferences for life-sustaining interventions honored. For many patients, this planning should occur in the primary care setting, but primary care practices are over-extended, clinicians and staff are often uncomfortable discussing prognosis, and patients and families often do not ask.

The Serious Illness Care Program (SICP) provides a comprehensive implementation strategy, training modules, and a structured conversation guide to promote realistic and compassionate discussions with seriously ill patients in any health care setting. SICP is widely used, having trained hundreds of clinicians and used with thousands of patients. It contributes to the quadruple aims of optimizing healthcare performance through improving patient experience and appropriate utilization of resources, increasing population health, and improving the work life of clinicians. The challenge is that it is unclear whether whether targeting clinicians or entire teams is more efficient and effective when implementing SICP in primary care.

Study Aims:

The study will fill this critical gap by comparing two models of SICP in primary care: clinician-focused SICP and team-based SICP. In the clinician-focused model, a patient's primary care clinician is responsible for serious illness care planning, while in the team-based model, tasks are purposefully shared across roles. This cluster-randomized trial has the following aims:

• AIM 1. Assess the comparative effectiveness in primary care of team-based vs. primary clinician-focused SICP to achieve the long-term outcomes of concordance of care with patient goals and more time at home for patients with prognoses of two years of life or less.
• AIM 2. Assess the comparative effectiveness in primary care of a team-based SICP vs. clinician-focused SICP on place of death and family bereavement in cases where the patient dies during the study period, as well as additional secondary short term and intermediate outcomes including patient, family and clinician satisfaction with communication and decisions and quality of life.
• AIM 3. Identify determinants of successful implementation of the two different models of SICP across varying settings, with a focus on the comparison of practices in the U.S. and Canada and on practice-level characteristics.

Overall study design A cluster-randomized controlled trial (cRCT) conducted in primary care practices recruited from seven practice-based research networks (PBRNs) in the U.S. and Canada. The practices will be randomly assigned to implement either the primary care clinician-focused model or the team-based model. A cRCT is proposed because the two models cannot be implemented simultaneously in the same practice. SICP requires changes in work flow and different training is required for the clinician-focused and team-based models.

Main components of the intervention and comparators The clinician-focused model provides training and standard infrastructure support within the practice, similar to what would be provided for any primary care clinician activity. The team-based model splits the serious illness care conversation and planning into its components and these are shared across team members and may be spread over time. The composition of the teams and how they divide their tasks will be allowed to vary across practices and patients; this is necessary and realistic as staffing patterns and resources vary across practices and ACP should be customized to patient needs.

Study population The target population is adults living in the community with serious illnesses and their families. The study will be conducted in a minimum of 36 primary care practices in seven primary care PBRNs (five in the U.S., two in Canada). Each PBRN will recruit a minimum of six practices to enroll 130 patients for a total of 1260 enrolled, so that with attrition, over 750 will be available for analysis.

Primary and secondary outcomes The primary outcomes are patient-centered. These are: a) care that corresponds to patient goals and b) time spent at home. Assuring goal-concordant care is the overarching purpose of the SICP. Patient outcomes will be measured at enrollment, six months, and one year of follow-up. If the patient dies during the study period, place of death will be recorded and a family member contacted to complete a bereavement survey.

Secondary outcomes will also include intermediate and proximal or short-term outcomes for patients and families including patient and caregiver quality of life, depression and anxiety, satisfaction with communication, decision quality, engagement and acceptability of the SICP program and model to patients.

Analytic methods Analytic models will be refined with the input of the Research Project Partnership (Steering Committee), including patient and clinician stakeholder representatives as well as experts in biostatistics. Analysis will be conducted under an intention to treat assumption with multiple imputation procedures and selection models for missing data. Multilevel modeling will be used to account for clustering at the practice (cluster) and individual level, and interaction terms will be included in the analysis to identify potential heterogeneity of treatment effects.