Study was not funded. No participants were enrolled.

Dementia is a disabling, progressive major neurocognitive disorder. Although cognitive symptoms drive the diagnosis of dementia, neuropsychiatric symptoms (NPS), such as depression, agitation/aggression, and psychosis, are common and associated with a number of significant adverse outcomes. There are currently extremely limited FDA (Food and Drug Administration)-approved treatments for NPS of dementia; there remains a critical need for safe and effective interventions for NPS that can be easily administered and monitored in typical clinical settings. As cognitive impairment progresses, persons living with dementia or cognitive impairment (PLWD/CI) increasingly rely on surrogate decision-makers, such that the quality of life for PLWD/CI is directly impacted by the decisions made by a surrogate. Discrepancy between preferences of PLWD/CI and proxy assessments by surrogate decision-makers is common. Our previous work has shown that such discrepancies in everyday preferences are associated with a higher burden of NPS for PLWD/CI. We hypothesize that discrepancy in everyday preferences assessment between PLWD/CI and care partners creates a set of "unmet needs" that increase the likelihood or severity of NPS; our preliminary data support this model of unmet needs as drivers of NPS. Specifically, as the neurodegenerative process disrupts the ability of PLWD/CI not only to provide for their own needs but also to communicate needs and preferences effectively to others, NPS may emerge in response. Thus, our hypothesis fits within a larger conceptual framework for understanding etiology of NPS: i.e., the mismatch between personality habits, physical/mental states, and environmental factors drive NPS as a means to resolve or communicate unmet needs. Building from this conceptual framework and our own preliminary data, the proposed pilot project addresses major gaps in the availability of safe, effective, and accessible strategies to reduce NPS by developing and testing the feasibility, acceptability, fidelity, and mechanistic target engagement of a templated, clinician-facilitated intervention to align everyday living preferences assessment between PLWD/CI and their care partners (n=20 dyads) to reduce NPS. The study is unique in developing a bioethics-driven intervention for NPS that is effective, portable, and easily transferrable to diverse settings. The crux of this intervention is meeting the needs of PLWD/CI by aligning everyday decision-making with priorities that matter most.

A. SIGNIFICANCE

A1. Definition, prevalence, impact of dementia and neuropsychiatric symptoms. Dementia is a disabling, progressive major neurocognitive disorder that affects approximately 35 million people worldwide; the number of affected people is expected to increase to 135 million worldwide by 2050.1,2 Although cognitive symptoms drive diagnosis of dementia, neuropsychiatric symptoms (NPS), such as depression, irritability, agitation, and psychosis are commonly experienced and associated with increased morbidity and mortality, increased care partner distress, and earlier institutionalization.3 These symptoms are debilitating and experienced by more than 90% of persons living with dementia. The occurrence of NPS among persons with cognitive impairment may precede the diagnosis of dementia by years; there are currently extremely limited FDA (Food and Drug Administration)-approved treatments for NPS.4-6 There remains a critical need for safe and effective interventions for the full spectrum of NPS that can be easily administered and monitored in typical clinical settings.

A2. Neuropsychiatric symptoms and surrogate decision-making. As cognitive impairment progresses, persons living with dementia or cognitive impairment (PLWD/CI) increasingly rely on surrogate decision-makers; quality of life for PLWD/CI is directly impacted by decisions made by surrogates.7 PLWD/CI, even those with more advanced illness, can still articulate values and preferences reliably.7-9 Indeed, a guiding ethical principle in the care of older adults is alignment of decision-making with the priorities that matter most for each individual older adult.10 Such an approach is exemplified by the Patient Priorities Care framework by Tinnetti et al., in which older adults with multiple chronic medical conditions articulate their priorities for care and work with their clinical teams to align care options with those priorities in an iterative fashion.11,12

For older adults with cognitive impairment, however, previous studies have shown lack of alignment is common between care preferences of PLWD/CI and proxy assessments made by surrogate decision-makers.13-16 Outside of care options, potential impacts of lack of alignment between PLWD/CI and care partners in everyday living preferences has been relatively underexplored. Our data showed that significant discrepancy exists between importance ratings of "social engagement" preferences among PLWD/CI and the proxy ratings made by their care partners.17 We observed that this discrepancy is prospectively associated with a higher burden of NPS among PLWD/CI over time.18,19 We hypothesize that misalignment in everyday preferences assessment between PLWD/CI and care partners creates a set of "unmet needs" that increase the likelihood or severity of NPS; our prior data support this model of unmet needs as drivers of NPS. As the neurodegenerative process disrupts ability of PLWD/CI not only to provide for their own needs but also to communicate preferences effectively to others, NPS may emerge in response.20

Our hypothesis fits within a larger conceptual framework for understanding etiology of NPS: mismatch between personality habits, physical/mental states, and environmental factors drive NPS as a means to resolve or communicate unmet needs.20 Recommended first-line treatments for NPS are non-pharmacological, such as the DICE approach.3 In this model, problematic behaviors are Described, potential underlying causes are Investigated, treatment plans are Created, and strategies are Evaluated for effectiveness.3 The challenges in implementing such interventions, however, include lack of training for clinicians, lack of training in the assessment of NPS, and limitations on time for implementation in standard clinical practice.3

A3. Significance of proposal. In line with bioethics standards for decision-making for older adults as well as the "unmet needs" etiologic hypothesis of NPS, our conceptual model posits that misalignment in surrogate decision-makers' understanding of the preferences of PLWD/CI with PLWD/CI's preferences may give rise to NPS. The proposed pilot project will lay the foundation to address major research gaps in the development of safe, effective, and accessible interventions to reduce NPS and NPS-related care partner distress by testing the feasibility, acceptability, fidelity, and mechanistic target engagement of a novel behavioral intervention that aims to align everyday preferences assessment.

B. INNOVATION. This study includes several features that increase its significance. First, it uses an innovative therapeutic and methodological approach by focusing on PLWD/CI and moving beyond healthcare preferences to look at everyday experiences. Second, there is novel use of existing evidence-based frameworks for developing behavioral interventions for PLWD/CI and care partners. Third, it addresses major therapeutic gaps for NPS and related care partner distress: NPS are associated with hospitalization, morbidity/mortality, institutionalization, high costs of care and adverse impact on care partners and PLWD/CI; thus, we will develop a new intervention with potential for high clinical impact. Fourth, the study is unique in developing a bioethics-driven intervention for NPS that is effective, portable, and easily transferrable to diverse settings. The crux of this intervention is meeting the needs of PLWD/CI by aligning everyday decision-making with priorities that matter most.

C. APPROACH

C1. Prior work. The previous work of our study team has established our experience, expertise, and administrative capacity in exploring relationships between NPS and alignment in everyday preferences between PLWD/CI and their care partners. We have a track record of recruiting dyads of PLWD/CI and their care partners for cross-sectional analyses,17 as well as longitudinal analyses assessing changes in preferences alignment and NPS over time.18,19 We have expertise in collecting and analyzing NPS data using the Neuropsychiatric Inventory brief Questionnaire (NPI-Q),21 as well as everyday preferences for PLWD/CI and proxy assessments by care partners using the Preferences for Everyday Living Inventory (PELI), a validated preferences assessment tool for older adults.22 Our prior work has established the PELI as a valid and reliable instrument across time for measuring preferences of PLWD/CI as well as measuring alignment in proxy assessments, which we have identified as our mechanistic intervention target.18 In support of the rationale for the current study, our prior work found that higher levels of discrepancy in assessment of "social engagement" preferences between PLWD/CI and proxy ratings by care partners was associated with a higher average burden of NPS (as measured by NPI-Q) across follow-up time.19 Consistent with the unmet needs hypothesis, these results suggest that developing novel interventions to align care partner appreciation of the everyday preferences of PLWD/CI may have a high potential to reduce the burden of NPS.

C2. Study population/procedures for participant recruitment and selection. We plan to enroll 20 dyads comprising community-dwelling persons with a primary diagnosis of clinically significant cognitive impairment (mild cognitive impairment (MCI) or dementia) and their informal care partners (i.e., relatives or friends). As approximately 10-15% of persons with MCI progress to dementia per year, including persons with MCI as well as those with dementia will allow us to capture an important transitional phase of cognitive decline to dementia.23 Participants will be recruited from the Geriatric Outpatient Services at McLean Hospital. Referrals will come from attending clinicians on these services, following standard IRB procedures and maintaining confidentiality per HIPAA standards. Specifically, referring clinicians will obtain consent from participants or legally authorized representatives (i.e., family with power of attorney or who serve as the patient's health care proxy) to make the referral. Then participants and/or families will be approached by research staff per standard IRB-approved procedures. The McLean Geriatric Outpatient Services have approximately 7000 visits/year, 600-900 unique patients, over 40% of whom carry a diagnosis of neurocognitive disorder, and over 60% of whom are women. Participants will be recruited without regard to gender, race, or ethnicity. The proposed investigation is designed to include participants from racial or ethnic minority backgrounds, which is estimated to comprise approximately 15% of the total sample. The PI is responsible for identifying and recruiting participants.

Eligible participants. Community-dwelling persons 55 years or older with a primary diagnosis of MCI or mild to moderate dementia will be eligible (Montreal Cognitive Assessment Score, MoCA, > 10).24,25 PLWD/CI must have an identified care partner: a family member or close friend who knows the person well with contact at least three times/week. Informed consent will be obtained; if PLWD/CI are not able to provide informed consent, healthcare proxies will be asked to provide informed consent, with assent provided by PLWD/CI. The study clinician (board-certified geriatric psychiatrist) will be available to participants if unstable NPS develop during the study period. Participants must be fluent in written and oral English. PLWD/CI who are not experiencing any NPS at baseline (i.e., NPI-Q score of zero at baseline) will be excluded. PLWD/CI living in long-term care environments (e.g., nursing home, assisted living facility) will be excluded. Persons with comorbid active primary psychiatric illness (e.g., major depressive or bipolar mood disorders, psychotic disorders) will be excluded. We will recruit 20 dyads comprising PLWD/CI and the care partner.

Participant enrollment. There will be a pre-screening procedure in which demographic and diagnostic information is obtained. For potential participants who express interest and appear to meet inclusion criteria based on the pre-screening phone call, an informed consent/screening visit will be arranged. The informed consent will be obtained in the office of the PI with both the participant with cognitive impairment and the care partner. Given the study measures used, non-English speakers will not be enrolled in this pilot project. Given the study population is persons with cognitive impairment, cognitive impairment may affect decision-making capacity. Capacity to consent will be assessed by the PI, a licensed, board-certified geriatric psychiatrist. Capacity will be assessed by investigating the participant's ability to understand risks and benefits of the study itself, how those apply to the participant's situation, and to manipulate the information to arrive at a consistent decision. Further, the consent form will be reviewed by the participant's study partner, who will also sign the consent form to participate. Given that the cognitive impairment is expected to be on the mild end, reviewing consent with the care partner adds an extra layer of caution in reviewing capacity to consent. In minimizing undue influence, no current patients of the PI will be recruited. This is a single-arm, open label pilot so there will be no randomization.

C3. Description of the psychoeducational intervention. Our intervention is based on the framework of the Patient Priorities Care (PPC) initiative, a manualized multi-step approach in which: a facilitator works with an older adult with multiple chronic health conditions to clarify values and preferences (step 1); these priorities are then communicated to the clinician for further review and consideration (step 2); then, the older adult and clinician work to align care options with those priorities (step 3).26 We have developed our intervention to be a similar three-step, templated, clinician facilitated-approach between a PLWD/CI and his/her care partner to align everyday preferences. This approach is novel and innovative in focusing on PLWD/CI and moving beyond preferences of healthcare to look at everyday experiences.

In step 1, PLWD/CI will complete the PELI (hereafter, "assessment PELI-CI" for Cognitive Impairment) to articulate their preferences. The PELI is a validated preferences assessment tool for older adults with current preferences rated on a 4-point Likert scale ranging from 1 (very important) to 4 (not important at all).22 Our previous work used exploratory factor analysis to derive "domains" embedded within the 55-item PELI. We identified four domains: "autonomous choice," "social engagement," "personal growth," and "keeping a routine."17 In our previous work, we found a significant discrepancy between self-ratings by PLWD/CI and proxy ratings by care partners only within the "social engagement" domain; this discrepancy was noted at both cross-sectional baseline assessment as well as longitudinal assessment approximately one year later.17,18 Our prior work found only discrepancy in "social engagement" preferences was associated with NPS across time; there were no such associations with the other domains of the PELI.19 Based on these data, we identified alignment in "social engagement" preferences as our mechanistic intervention target. Given these prior data, we will use a modified PELI featuring only the items within the "social engagement" preferences domain as a valid and reliable measure of preferences assessment and alignment with proxy assessment. In step 2, care partners will consider the preferences of PLWD/CI; care partners will complete concurrent but separate proxy PELI assessments from the perspective of the PLWD/CI (i.e., as if acting as surrogate decision-makers; hereafter "assessment PELI-CP" for Care Partner).

In the Intervention (step 3), the PELI-CI will first be reviewed with the PLWD/CI and care partner, highlighting the highest rated preferences (preference items with a score of 1 or 2) of the PLWD/CI. The study clinician will then ask the PLWD/CI to identify up to 3 "social engagement" preferences that matter most to them. Second, in discussion with the PLWD/CI and care partner, the study clinician will review areas where there is a difference in ratings between the PELI-CI and PELI-CP (preference items with a non-zero difference). The study clinician will then ask the PLWD/CI to identify up to 3 items of disagreement that are most important for the care partner to know about. After this discussion, a 1-page Preferences Priorities document will be completed listing the preferences identified as most important to the PLWD/CI and the areas of disagreement most important for the care partner to know about; this templated document will be kept by the PLWD/CI and care partner for their reference. Assessments PELI-CI and PELI-CP will be completed again at weeks 4 and 8. Follow-up visits will be completed in person, via secure video platforms (e.g., Zoom), or over the phone per participant and care partner preference.

C4. Assessment. We will assess the feasibility, acceptability, and fidelity of the intervention through a variety of measures, including: number of dyads screened per month; number of dyads enrolled per month; reasons for ineligibility; time from screening to enrollment; proportion of screened eligible dyads who enroll; proportion of dyads enrolled who complete at least one follow-up assessment; retention rates for enrolled dyads; reasons for drop-out; proportion of missing data from assessments/study measures; proportion of planned assessments that are completed; time to complete assessments/study measures; any unexpected adverse events. Prior to the intervention and again after completion of all study visits, PLWD/CI will be asked if care partners understand their preferences and if they feel able to discuss their preferences with their care partners; care partners will be asked if they understand the preferences of PLWD/CI and if they feel able to discuss the preferences of the PLWD/CI (yes, no, don't know, response format for all responses). PLWD/CI will be asked how confident they are that the care partner understands their preferences and how confident they are that they are able to discuss their preferences with their care partner; care partners will be asked how confident they are that they understand the preferences of the PLWD/CI and how confident they are that they are able to discuss the preferences of the PLWD/CI (1: not at all confident, 2: not very confident; 3: fairly confident; 4: very confident, response format for all responses). Severity of the NPS burden and care partner distress will be quantified using the NPI-Q. The NPI-Q features 12 symptom items: delusions, hallucinations, agitation/aggression, depression, anxiety, euphoria, apathy, disinhibition, irritability/lability, motor disturbance, nighttime behaviors, and appetite/eating. Symptom severity is rated on a 3-point scale and related care partner distress is rated on a 5-point scale. The NPI-Q is completed by the care partner. The NPI-Q will be administered at baseline, week 4 and week 8.

C5. Analytic methods. All analyses will be conducted using SAS Version 9.4.

Statistical analysis Aim 1: Data regarding feasibility, acceptability, and fidelity of the intervention will be analyzed via means, frequencies, and percentages, employing t-tests or Wilcoxon rank sum test and chi-square or Fisher exact tests as appropriate.

Statistical analysis Aim 2: As in our prior work,18 to calculate a discrepancy score for each dyad (i.e., PLWD/CI and care partner), a mean domain score for both assessments PEL-CI and PELI-CP will be calculated as the average of all non-missing data for each assessment. The discrepancy score will then be obtained by subtracting the mean domain score for assessment PELI-CP from the mean domain score for assessment PELI-CI for each dyad. To assess changes in baseline data (week 0) and to account for repeated measures over time, we will use longitudinal mixed effects models across weeks 4 and 8 with time trend parameter(s) as fixed effects, with a dyad-level random intercept. Assessing changes in discrepancy score across time in the trial will allow for further clarification of mechanistic target engagement for the intervention.

Statistical analysis Exploratory Aim: For the NPS dependent variables, we will use longitudinal mixed effects models across weeks 4 and 8 with time trend parameter(s) as fixed effects, with a dyad-level random intercept, to assess changes in baseline data (week 0) and to account for repeated measures over time. We will assess changes over time in total NPI-Q score as well as changes over time in total care partner distress score on the NPI-Q in separate models. We will exploit the co-synchronous assessment of the NPS and discrepancy measures across time to check for evidence of directional causality between these two variables as well as the time lag across which the effects occur, employing variations of cross-lagged analyses, as developed by team Biostatistician Dr. Locascio.27 A mixed effects model with NPS as the dependent variable and discrepancy scores as a fixed effect time-varying predictor synchronous with NPS across time will be run. The same model will then be run with discrepancy scores lagged 4 weeks previous to each NPS, i.e., baseline discrepancy will be paired with 4-week NPS, and 4-week discrepancy paired with 8-week NPS. Finally, a cross-sectional analysis of baseline discrepancy paired with 8-week NPS will be run. These same models with the roles of NPS and discrepancy reversed will also be conducted. To rule out the possibility that any association found in these analyses between NPS and discrepancy is spurious due to each variable separately changing across time but being otherwise independent, static correlations of NPS and discrepancy will be run at each timepoint individually; we will also run analyses parallel to those above but where a time measure itself is covaried.

As this is a pilot study, power was not calculated.

C6. Implications, alternatives, and next steps. These data will directly inform the development of further research protocols and applications to assess the pure efficacy of the intervention in research settings and then the real-world efficacy in community settings, following the revised NIH stage model for behavioral intervention development.28 Strengths of this study include comprehensive assessment of everyday social engagement preferences, longitudinal analyses, and use of gold-standard cognitive and neurobehavioral assessments. Study limitations include a relatively small sample size and anticipated limited racial/ethnic diversity. Given the primary aim of the study is to assess feasibility and acceptability of the intervention, we are aware of the compromise of a relatively smaller sample size and subsequent reductions in power in assessing efficacy of the intervention itself. Next steps will be to carry the intervention forward into more well-powered embedded pragmatic clinical trials to measure the efficacy of the intervention in real-world community settings. Alternative approaches to addressing NPS of dementia include psychopharmacologic management, and the inherent side effects that come with medication management, as well as non-pharmacological approaches such as the DICE model, described above. Interventions like the DICE model may attend to other potential etiologies of NPS including pain, sensory changes, medical comorbidity, functional limitations, which are not directly assessed in our approach.

Community-dwelling persons 55 years or older with a primary diagnosis of MCI or mild to moderate dementia will be eligible (Montreal Cognitive Assessment Score, MoCA, > 10).24,25 PLWD/CI must have an identified care partner: a family member or close friend who knows the person well with contact at least three times/week. Informed consent will be obtained; if PLWD/CI are not able to provide informed consent, healthcare proxies will be asked to provide informed consent, with assent provided by PLWD/CI. The study clinician (board-certified geriatric psychiatrist) will be available to participants if unstable NPS develop during the study period. Participants must be fluent in written and oral English. PLWD/CI who are not experiencing any NPS at baseline (i.e., NPI-Q score of zero at baseline) will be excluded. PLWD/CI living in long-term care environments (e.g., nursing home, assisted living facility) will be excluded. Persons with comorbid active primary psychiatric illness (e.g., major depressive or bipolar mood disorders, psychotic disorders) will be excluded. We will recruit 20 dyads comprising PLWD/CI and the care partner.