PST for Care Partners of Adults With Alzheimer's and Alzheimer'S-related Dementia

Over 65 million people in the United States currently provide informal (i.e. unpaid) care to an individual with a chronic illness or disability,1 including the nearly 6 million adults currently living with Alzheimer's disease (AD).2 Care partners of adults with AD and related dementias (AD/ADRD) often experience high caregiver burden, leading to emotional distress, depression, problematic alcohol use, health problems, isolation, and poor quality of life.3,4 Interventions to reduce care partner stress may reverse the negative impact of caregiver burden.5 The Global Impact of Dementia 2013-2050 policy brief specifically called for better education and skills training for care partners. Therefore, there is a critical need to provide evidence-based support to care partners so they can best manage their own lives and their care recipients' chronic health conditions.6

Problem-Solving Training (PST) is an evidence-based metacognitive strategy training approach that teaches a simple, systematic method for evaluating problems, generating and selecting solutions, developing specific goals and action plans, and evaluating and revising plans as needed.7-18 Individuals learn to set achievable goals under the coaching of a therapist and gain self-efficacy as they begin to see that the seemingly overwhelming problems are indeed solvable when approached in a stepwise, rational fashion. Few care partner interventions include problem-solving skills training to translate knowledge into meaningful action, leaving care partners struggling to effectively manage their own lives and their caregiving responsibilities. Additionally, as of 2018 Hispanic individuals make up roughly 39.6% of the population of Texas,19 but most care partner intervention studies exclude Spanish-speaking participants. Our team has translated and culturally adapted PST specifically for Spanish-speaking care partners,20 presenting the opportunity to close a critical health disparity among care partners in Texas. PST, in both English and Spanish, could be incorporated across clinical settings into existing education-based care partner interventions to fill the current gaps in clinical practice. PST for fcare partners may also improve care recipient outcomes as well.

Our long-term goal is to improve the health, well-being, and participation of care partners and individuals with AD/ADRD. Our multidisciplinary team has expertise in AD/ADRD and in delivering and evaluating PST. The investigators represent a multi-site collaborative leveraging the broader TARCC network. Our expertise and strong clinical and research environments ensure the rigor, clinical relevance, and feasibility of our proposed research. The investigators will conduct a multi-site randomized controlled trial with a factorial design of PST for care partners of adults with AD/ADRD to address the following specific aims:

Aim 1: Determine the optimal number of Problem Solving Training (PST) sessions and boosters to improve emotional symptoms and burden in English and Spanish speaking care partners of adults with AD/ADRD.

Hypothesis: Six PST sessions vs three sessions will be optimal for improving emotional symptoms and burden of both English and Spanish speaking care partners 6 months after baseline.

Hypothesis: Monthly brief booster sessions vs no boosters will be optimal for improving emotional symptoms and burden of both English and Spanish speaking care partners 6 months after baseline.

The investigators will compare a) care partner emotional symptom (depression) change and b) caregiver burden change from baseline to 6-months post-baseline, to assessed long-term effects of PST and booster sessions several months beyond completion of the training. The investigators will also measure secondary care partner outcomes, including caregiver self-efficacy, positive aspects of caregiving, and alcohol use. Finally, the investigators will assess care partners' satisfaction, confidence using the PST strategy, and engagement across all conditions.

Aim 2: Identify factors associated with response to treatment.

Hypothesis: Response to treatment will be associated with age, gender, engagement in and uptake of PST, working alliance, and goal attainment. To test this hypothesis, the investigators will assess the extent to which each of these factors predicts response to PST (change in outcome) using multiple linear regression.

Establishing specific guidelines for an evidence-based, implementable problem-solving intervention for both English and Spanish speaking care partners will result in direct clinical application to improve care partner health and well-being and to improve care recipient health and quality of life. The investigators will explore potential mechanisms of change for future prospective study. This proposal directly aligns with the mission of TARCC to take novel and innovative approaches to increase participation of diverse and underserved populations, promote collaboration across TARCC institutions, and contribute to advancement in therapeutic interventions to improve the lives of individuals with AD and related dementias.

Design. The investigators will conduct a randomized control trial of PST for care partners of adults with AD/ADRD to assess the effects of number of sessions (3 vs 6) and brief telephone booster sessions (boosters vs no boosters) on depressive symptoms (Patient Health Questionnaire-9) and caregiver burden (Zarit Burden Interview). The investigators will recruit n=180 participants (n=45 per arm) across our two enrollment sites and conduct assessments at baseline, end of intervention, and 6-months post-baseline. Booster sessions will begin 1 month after completion of PST and will occur monthly thereafter up to 6-months post-baseline conditions.

Randomization and Allocation. Randomization will be stratified by language and gender, then will occur in blocks of 16 (i.e. 4 participants allocated to each of the 4 conditions within each block). The investigators will uniformly sample (i.e. equal probability) random numbers from a set containing four 1's, four 2's, four 3's, and four 4's to generate the randomization table within each block.

Data Analysis \& Sample Size Determination: The investigators will conduct Bayesian mixed-effect ordered logistic regression models with subject-specific random effects to identify the optimal intervention group for our two primary outcomes at 6-months post-baseline (Aim 1). The investigators will apply the same statistical analysis approach to secondary outcome measures. The posterior probability that receiving 6-sessions decreases the odds ratio of higher PHQ-9 and ZBI scores and that receiving a booster decreases the odds ratio of higher PHQ-9 and ZBI scores will be reported in lieu of a p-value, which is inexistent in the Bayesian paradigm. In accordance with the approach used to power the current study, a significant result for any of the three treatments will be defined as a posterior probability greater than 98.33% in order to achieve an overall experiment-wise significance level of 5%. The investigators will apply the same statistical analysis approach to secondary care partner outcomes (caregiver self-efficacy, positive aspects of caregiving, coping, and alcohol use). To identify factors associated with PST response to intervention (Aim 2), the investigators will conduct the following exploratory analyses. Based on the PHQ-9 minimal clinically important difference (MCID) for individual change of 5 points,32 The investigators will categorize participants into one of three response to intervention groups: Response group (PHQ-9 score decrease by >5 points from baseline to 6-months), No Change group (PHQ-9 scores are within +5 points of baseline at 6-months), and Decliner group (PHQ-9 scores increase by >5 points at 6-months). The investigators will examine predictors of response to intervention group using multinomial logistic regression. The investigators will further explore predictors of change in primary and secondary outcomes from baseline to 6-months post-intervention using multiple linear regression.

Based on 1,000 iterations of a simulation of our pilot data (PST to improve PHQ-9 in care partners in a different clinical population14), the sample size necessary to detect a 30% decrease in the odds ratio for those who received six sessions or a 30% decrease in the odds ratio for those who received booster sessions with at most 5% significance level and at least 80% power was found to be 23 per arm (4 arms) per center (2) for a total of n=184 total (n=203 accounting for 10% attrition).