The prevalence of Alzheimer disease and related dementias (ADRDs) among the Hispanic and Latinx populations is 1.5 times higher than among non-Hispanic White populations [1]. Hispanic and Latinx adults exhibit an earlier onset of ADRDs and more severe symptoms at the time of diagnosis than other ethnic and racial groups [2,3]. Moreover, the presence of comorbidities, such as diabetes, heart disease, and hypertension, among Hispanic and Latinx individuals with ADRDs introduces further health complexities, exacerbating ADRD-related symptoms and accelerating functional decline [4]. Despite facing worse health outcomes and increased disability, nursing home admissions are lower among older Hispanic and Latinx individuals, suggesting increased reliance on informal caregiving [5].

Notably, Hispanic and Latinx ADRD caregivers exhibit poorer mental health outcomes compared with non-Hispanic White ADRD caregivers [6-9]. However, a limited understanding of the daily stressors and dynamic predictors of changes in mental health over time exists. Prior investigations into the mental health of Hispanic and Latinx ADRD caregivers have reported inconsistent results, likely due to varying quality of evidence across studies, which highlights the need for more in-depth research [10]. This inconsistency may arise from dynamic interactions between contextual, individual-level, and unique cultural factors contributing to within-person variability.

Conventional between-group designs and cross-cultural comparisons may not fully capture such within-group variability, thus potentially masking important factors driving mental health outcomes over time. Previous studies on caregiver mental health have found daily variations in mental health [11]. The emerging evidence of daily mental health experiences exemplifies the necessity to examine day-to-day variations in these experiences—especially among populations that may be disproportionately impacted by ADRDs and caregiving-related stress.

To address this gap, the Nuestros Días (Spanish for “our days”) study implements a daily diary (DD) approach, offering a novel and comprehensive method to understand the day-to-day stress and mental health experiences of Hispanic and Latinx ADRD caregivers. Daily diaries provide granular data that can determine intricate patterns and nuances that might be otherwise missed in broader survey methods [12]. Examining daily experiences can reveal how specific events or triggers may lead to heightened stress or emotional arousal, allowing for a more holistic understanding of the factors influencing an individual’s well-being. As such, when studying the mental health of Hispanic and Latinx ADRD caregivers, DD studies can provide invaluable insights and further directions for ecologically relevant and culturally tailored interventions to support the well-being of Hispanic and Latinx ADRD caregivers.

This multi-time scale study will test the hypothesis that contextual, individual-level, and cultural factors influence daily and long-term mental health patterns among Hispanic and Latinx ADRD caregivers. Participants (N=500) will be asked to complete surveys measuring contextual, individual-level, and cultural factors at 3 time points (ie, enrollment, 6 months, and 12 months). Immediately after completing each survey, participants will complete 21 days of DD surveys evaluating caregivers’ day-to-day caregiving experiences and mental health (for a total of 63 diary surveys across the 3 sampling periods).

The study procedures were approved by the University of Alabama at Birmingham institutional review board (IRB-300008710). The study was deemed to be minimal risk, and a waiver of written consent was approved. Participants will be provided a study information sheet with details about the study’s purpose, procedures, risks, and contact information. Participants are informed that all data will be kept deidentified and that no questions regarding immigration status will be asked.

The target population for the study is English- and Spanish-speaking Hispanic and Latinx ADRD caregivers across the United States. The sample eligibility criteria are guided by AARP’s definition of family caregivers [18] and the sampling frame for family caregivers used by the National Institute on Aging–funded REACH trials [19].

Inclusion criteria are as follows: (1) the caregiver is 18 years or older; (2) the caregiver identifies as Hispanic or Latinx descent; (3) the caregiver provides informal care to a spouse/common-law partner, parent, or grandparent (or in-law; 60 years or older), living with dementia; (4) the caregiver assists with at least 2 instrumental activities of daily living or 1 activity of daily living; and (5) the caregiver lives with or shares cooking facilities with the care recipient.

The target sample includes any type of dementia (ie, Alzheimer and frontotemporal dementia) reported by the caregiver, and this will be accounted for in the analytic models. The AD8 questionnaire is being used to determine dementia eligibility status [20], as this is a valid and reliable measure with excellent discriminant validity to distinguish between normal cognition and mild cognitive impairment or dementia. A cutoff score of 2 or greater (ie, cognitive impairment is likely to be present) on the AD8 will be used to determine eligibility. Participants will have the option of completing the study in the language of their choice (ie, either English or Spanish).

Exclusion criteria are as follows: (1) admission to a nursing home or institutional care for the care recipient planned within the next 6 months, (2) the caregiver is terminally ill, and (3) the caregiver does not have reliable internet access.

Several strategies are in place to facilitate participant recruitment (Textbox 1). All of the recruitment materials have been developed in English and Spanish. As part of our analytical strategy, we will examine the effectiveness of each recruitment strategy including participant characteristics, reach, number of participants enrolled via each method, and costs associated with each strategy. We will also examine demographic differences that arise from the different recruitment methods used to identify possible selection biases.

All data will be exported from REDCap and analyzed using various statistical software platforms, for example, SAS (SAS Institute Inc), SPSS (IBM Corp), R (R Core Team), and STATA. (StataCorp). First, univariate and bivariate approaches will be used to examine outstanding issues with the DD survey data including outliers and missing data. Descriptive statistics and data distributions will be evaluated to ensure values are within reasonable limits and evaluate potential issues of nonnormality and collinearity. A false discovery rate approach [57] will be used to adjust for multiple inferences when appropriate. To ensure the validity and reliability of the proposed measures in our sample, we will perform psychometric testing on all measures as part of our analytical strategy including measures that were adapted for the purposes of this study. The main analytical approaches for each of the study’s specific aims are detailed below.

For aim 1, the objective is to identify daily protective and risk factors that impact the daily odds of depression and anxiety symptoms among Hispanic and Latinx ADRD caregivers. Generalized linear mixed modeling (GLMM) accounting for the nesting of diary survey data within participants will be used to examine associations between contextual, individual-level, and cultural factors and daily mental health (ie, depression and anxiety-related symptoms).

In aim 2, the main goal is to investigate the developmental course of depression and anxiety symptoms among Hispanic and Latinx ADRD caregivers over 12 months. To achieve this, we will use group-based trajectory modeling (GBTM) [58] to identify clusters (ie, trajectory groups) of daily mental health symptom occurrence over time and determine personal and cultural factors that predict group membership. We will follow current best practices in GBTM [59] for model selection.

For aim 3, the purpose is to examine the association between developmental trajectories of depression and anxiety symptoms with levels of resilience and distal health outcomes. We will use regression modeling with the trajectories identified in aim 2 and posterior probabilities to make prognostic predictions of overall physical health and clinically significant depression and anxiety at 12 months. We will also examine levels of resilience at baseline, 6- months, and 12 months as a time-varying covariate for these relationships [58].

For the planned aim 1 analyses, the design proposed in this study will collect a maximum of 31,500 data points across 500 participants with 63 observations per participant (21 days x 3 sampling periods). Assuming a 0.2 uniform probability of dropout over the 12-month study period and an average of 80% diary completion, the expected datapoint collection is 22,680 datapoints (500×16.8+450×16.8+400×16.8). Further assuming, 2 equally sized groups, an intraclass correlation of 0.4 (ie, average correlation between any 2 repeated measurements on the same individual), a standardized outcome (eg, stress and depression), and an adjusted significance level of 0.025, the expected effective sample provides 80% power to detect a time-averaged between-group difference (d) of 0.19 [60]. Under similar assumptions for intraclass correlation and significance level, regarding within-cohort changes in standardized outcomes between sampling periods (eg, baseline vs 6 months, each with up to 21 observations per participant) the expected effective sample size provides 80% power to detect a within-cohort change (d) of 0.04.

For aims 2 and 3, there are no established guidelines for assessing power or determining appropriate sample sizes for GBTM as these are exploratory analyses and the number of resulting trajectories is not known a priori. In previous studies [61] evaluating depression symptoms, discrimination, and ethnic identity among Latinos, a sample of 91 participants was sufficient to detect variations in depressive symptoms. In another study [62] using GBTM, a sample of 94 caregivers of individuals with malignant brain tumors was sufficient for a 2-group trajectory model of caregiver depression and anxiety. Considering that precision is expected to increase with a greater number of data points [63], we expect the target sample size of 500 participants will yield enough precision and power for our planned analyses. Further, GBTM falls under the umbrella of growth mixture models. Previous literature on growth mixture models suggests that a sample size of 500 is sufficient for the Bayesian information criterion and sample-size adjusted Bayesian information criterion to accurately identify the correct number of latent growth trajectories [64].

This study addresses a critical gap in the literature by implementing an innovative design and analytical strategy to test the hypothesis that contextual, individual-level, and cultural factors influence daily and long-term patterns of mental health outcomes among Hispanic and Latinx ADRD caregivers. With cases of ADRDs rapidly increasing among Hispanic and Latinx individuals [8,65,66], the number of Hispanic and Latinx caregivers is consequently growing. Given that Hispanic and Latinx adults develop ADRDs at a younger age and experience complex comorbid conditions, as well as more severe symptomatology than other racial or ethnic groups, Hispanic and Latinx caregivers are uniquely susceptible to increased caregiving stress that can negatively impact mental health outcomes. Thus, a better understanding of the dynamic interactions between risk and protective factors will contribute to the identification of modifiable targets to inform ecologically valid and culturally responsive interventions that will better support Hispanic and Latinx ADRD caregivers.

To our knowledge, this is the first study to provide an in-depth examination of how daily and enduring caregiving-related stress and burden influence the mental health of Hispanic and Latinx ADRD caregivers. This examination is conducted based on the novel application of the protective factor model of resilience. In this model, promotive resources (eg, coping strategies and cultural values) can either reduce risk or improve other protective factors [67,68], and resilience is defined as the capacity for positive adaptation despite adversity [69-71]. Thus, resilience theory provides a strengths-based framework to understand why some caregivers report better mental health outcomes than others despite similar risk exposures. While most studies to date have primarily relied on the stress-process model [72] to inform clinical recommendations, small effect sizes have been repeatedly reported in interventions based on this model [73,74]. The measures included in this study align with domains of the protective-factor model of resilience frameworks including promotive resources/assets and risk factors, that decrease vulnerability to caregiving-related stress and increase resilience over time. Therefore, the results from this study will help inform the development of novel resilience-based interventions to support the well-being of Hispanic and Latinx ADRD caregivers.

In addition, the application of advanced analytical strategies, such as GBTM, allows for a thorough evaluation and understanding of caregivers’ mental health trajectories over time in this understudied population. Further, a subsequent evaluation of the predictive utility of the identified trajectories will allow us to identify which caregivers, under what circumstances, are at risk of developing clinically significant depression and anxiety over time, and which caregivers are resilient. Finally, findings from the Nuestros Días study will translate into key guidance for targeted and culturally responsive behavioral interventions to mitigate the negative effects of caregiving-related stress on Hispanic and Latinx ADRD caregiver mental health.

Although results from the Nuestros Días study will provide compelling directions for future research and tailoring of clinical interventions for caregivers’ mental health, several limitations should be taken into consideration.

First, concerns may be raised regarding the potential burden of completing daily surveys. Although we acknowledge that the study design may appear intensive, it provides a rich in-depth examination of day-to-day changes to the variables of interest, and previous studies [15] using this methodology report good compliance along with unsolicited positive feedback from participants. The daily surveys currently take an average of 10-15 minutes to complete, and the study is purposefully designed to minimize participants’ burdens while supporting the successful completion of the study aims. However, we acknowledge that there could be a potential tendency for participants to provide random answers to DD assessments due to inconvenience. To identify these scenarios, research team members will monitor the DD data to preliminarily identify randomly completed questionnaires.

Second, in the case of low variability in daily mental health measures, we will explore alternative analytical approaches as appropriate (eg, using the single-item questions instead of converting them to a count of symptoms). It is possible to observe low variability in caregivers’ depression and anxiety scores due to either cultural perceptions of mental health or consistently elevated rates of these symptoms at the baseline assessment. To address this concern, we will create a symptom burden score for each period based on the number of days that each participant reported having anxiety and depression symptoms.

Third, the possibility of missing data is a ubiquitous limitation, particularly in longitudinal quantitative research; however, we will use analytical methods [75] (ie, GLMM and GBTM) robust for missing data and will appropriately evaluate patterns of missingness to prevent model misspecification. We also use maximum likelihood estimation to estimate the parameters of the models. Maximum likelihood estimation accounts for data missing at random by estimating the marginal probability of observing a variable given the available data [76]. Finally, we will examine nonparticipant issues (eg, technical failures) and compare the means of observed variables for responders and nonresponders across days, periods, and participants.

Fourth, given the longitudinal nature of the study design, attrition is an additional limitation to consider. To address this concern, we have implemented strategies to keep participants engaged between sampling periods. For example, we mail birthday and holiday cards to participants enrolled in the study. We also send out reminder letters 1 month prior to the 6-month and 12-month data collection waves and place reminder phone calls 1 week before data collection begins. If any mail is returned as undeliverable, we will follow up with an alternate contact provided by the participant upon enrollment. Similarly, after 1 week of 3 unsuccessful attempts to reach the participant at the 6- and 12-month timepoints, we will follow up with the alternate contact. Previous research [77,78] has found that such strategies increase retention over single strategies. If these strategies are unsuccessful, we will identify participants lost to attrition, reopen recruitment, and match the new participants to the lost participants on several demographic variables. Finally, we will use Leigh et al’s approach [79] of specifying an instrumental variable to test whether the likelihood of attrition influences the relationship between key independent measures and mental health outcomes.

Previous studies on Hispanic and Latinx ADRD caregivers have identified important factors that influence the risk of poor mental health outcomes (eg, physical health, family support, and acculturation) [8,80-83]. However, the dynamic relationships among these individual, contextual, and cultural factors, as well as their effects on caregivers’ mental health progression over time, remain relatively unknown among Hispanic and Latinx caregivers. A recent study [62] examining psychological distress among family caregivers of individuals with malignant brain tumors found a trajectory of decreasing psychological distress over 12 months following diagnosis, suggesting psychological adaptation and increased flexibility over time. Such mental health patterns over time are unknown among Hispanic and Latinx ADRD caregivers, which presents a missed opportunity to determine when caregivers are at most risk for developing depression and anxiety.

In addition, while numerous studies [6-10] have observed elevated levels of depression among Hispanic and Latinx caregivers, findings are not consistent across studies, and these discrepancies may be due to the dynamic nature of mental health outcomes, as well as temporal variations in risk or protective factors, yet unexamined in Hispanic and Latinx ADRD caregivers. Further, previous research [61,84-87] has found a high degree of variation in the effects of cultural factors on mental health outcomes. These inconsistencies may be due to the heterogeneity within the culturally, genetically, and regionally diverse Hispanic and Latinx subgroups (eg, Mexican, Puerto Rican, Cuban, and Central American), thus emphasizing the importance of implementing innovative design and analytical strategies to disentangle the within-group variability in this population.

The Nuestros Días study overcomes these limitations and addresses these gaps through the multi-time scale design described and the combination of GLMM and GBTM strategies. Results from this study will identify modifiable intervention targets and at-risk groups, thereby providing foundational knowledge to inform evidence based on the care of Hispanic and Latinx ADRD caregivers.

In summary, the Nuestros Días study aims to address gaps in the knowledge about the influence of contextual, individual-level, and cultural factors on daily and long-term patterns of mental health of Hispanic and Latinx ADRD caregivers through the implementation of an innovative design and analytical approach. Results from this study will contribute to the development and testing of culturally responsive, resilience-based behavioral interventions to support the well-being of Hispanic and Latinx ADRD caregivers.