Culturally Tailoring an Advance Care Planning Intervention for American Indians

Research Design. A quasi-experimental waitlist study design will be used to test the hypothesis that the intervention participants will experience increased ACP self-efficacy, readiness, and completion rates. It is also expected that the intervention participants will experience increased facilitators and decreased barriers to ACP. A control sample of waitlist participants will be used.

Based on the odds of completing an ACP indicated in previously implemented MY WAY interventions and using 90% power as a threshold, it is estimated that will be a need for at least an n = 60 to find effects. The intervention group, Cohort 1, will be those who participated in the first round of both the Community Educational Sessions and Sharing Sessions. The control group, Cohort 2, will be those who participated in the second round of both the Community Education Sessions and Sharing Sessions.

Measurement. Process Measures that are a part of the 9-item ACP Engagement Survey that includes measurement of self-efficacy and readiness, identified from Behavioral Change Theory, will be used. Self-efficacy and readiness subscales have Likert responses of "not at all, a little, somewhat, fairly, extremely" and have good reliability and discriminant validity. Barriers and Facilitators to ACP Questionnaire used in the MY WAY randomized controlled trial will be used. Pretest interviews with these measurements will be conducted by telephone or in person at the time of pre-registration for both cohorts by the Project Coordinator and team members.

Cohort 1 will be interviewed five times: Baseline pretest, Community Information Session Evaluation; Sharing Session Evaluation, 8 week-post test and 6 month post-test.

Cohort 2 will be interviewed five times: Baseline pretest, 2nd baseline control pretest, waitlist Community Information Session evaluation, and waitlist Sharing Session evaluation, 8 week post test.

The first interview at registration will serve as baseline for the waitlist group, Cohort 2. Cohort 2 will be contacted again approximately 6 months later and two-weeks prior to the second Community Education Session to remind them of their registration and to re-interview them for their intervention pretest. Posttest interviews will also be conducted by the Project Coordinator eight weeks post Sharing Session for both groups to capture ACP self-efficacy, readiness, and completion rates.

Statistical Analyses. Demographic characteristics will be examined using descriptive statistics (frequencies, measures of central tendency, measures of variation) and stratified analyses (e.g., delineating measures by demographics, attendance, ACP agreement etc.) to examine the association between demographics and program outcomes. ACP Engagement and the Barriers and Facilitators to ACP Questionnaire (ACP measures) will be compared between those who do and do not engage in Sharing Sessions and between Cohort 1 and Cohort 2. The tribal community as well as location of Community Education Session will be included in all analyses as nested participant variables to ensure that differences between tribal communities and education delivery do not significantly change results. To account for these nested variables, hierarchical linear modeling or a similar method will be used in comparison analyses. It is likely that there will be significant attrition from Community Education Sessions to Sharing Sessions. A Heckman Selectivity analyses will be used to determine if there are differences between those who complete the full intervention and those who complete only baseline measures or Community Educational Sessions. Due to the nature of the study, there may also be attrition due to death. If there are differences in attrition, the inverse Mill's ratio will be used in all further analyses. Comparison of baseline measures from pre-registration interviews will serve to ensure that there are no significant differences between the cohorts.

After these preliminary analyses, change in Cohort 1's ACP measures can be compared to changes from baseline to the waitlist pretest measures from Cohort 2. Since Cohort 2 received no intervention in the first 6 months, the analysis will control for natural variations over time and any measures which show significantly larger change in Cohort 1 can be attributed to the intervention. Cohorts 1 and 2 can then be combined to examine intervention effects (i.e., change from pretest to posttest) in the full sample of participants who experienced the Sharing Sessions; baseline results for Cohort 1 will be combined with waitlist pretest results for Cohort 2 and compared to the combined group of 70 Sharing Sessions' posttests. Finally, the combined cohorts will be compared to a subsample of matched participants who experienced the Community Education Sessions only to determine how each step in MY WAY affects the outcome ACP measures.