Amyloid-β PET Scan Results Disclosure and Care-Partner Emotional Well-Being Over Time

Abstract

Background:

Diagnostic tests, such as amyloid-β positron emission tomography (PET) scans, can increase appropriate therapeutic management for the underlying causes of cognitive decline. To evaluate the full utility of this diagnostic tool, information is needed on whether results from amyloid-β PET scans influence care-partner outcomes.

Objective:

This study examines the extent to which previous disclosure of elevated amyloid (suggestive of Alzheimer’s disease (AD) etiology) versus not-elevated amyloid (not suggestive of AD etiology) is associated with changes in care-partner wellbeing.

Methods:

The study used data derived from a national longitudinal survey of Medicare beneficiaries (n = 921) with mild cognitive impairment (MCI) or dementia and their care-partners. Care-partner wellbeing outcomes included depressive symptoms (PHQ-8), subjective burden (4-item Zarit burden score), and a 3-item measure of loneliness. Change was measured between 4 (Time 1) and 18 (Time 2) months after receiving the scan results. Adjusted linear regression models regressed change (Time 2-Time 1) in each outcome on scan result.

Results:

Care-partners were primarily white, non-Hispanic, college-educated, and married to the care recipient. Elevated amyloid was not associated with statistically significant Time 1 differences in outcomes or with statistically significant changes in depressive symptoms 0.22 (−0.18, 0.61), subjective burden 0.36 (−0.01, 0.73), or loneliness 0.15 (−0.01, 0.32) for care-partners from one time point to another.

Conclusion:

Given advances in AD biomarker testing, future research in more diverse samples is needed to understand the influence of scan results on care-partner wellbeing across populations.

INTRODUCTION

In 2020, 11.2 million care-partners assisted 6.2 million US adults diagnosed with Alzheimer’s disease (AD) [1]. Yet, AD negatively impacts care-partner emotional wellbeing. Rates of depression among AD care-partners exceed 30% [2], nearly 50% experience subjective burden [2], and 65% report moderate to severe loneliness [3]. Care-partner emotional wellbeing may also deteriorate as patient function declines [4–6] and caregiving intensifies [7]. Care-partner distress and burden have serious implications for people living with AD and may be associated with nursing home admission and elder abuse [8].

Although treatments for AD have limited efficacy for long-term benefits, diagnostic tests—including amyloid-β positron emission tomography (PET) scans [9, 10]—can facilitate appropriate therapeutic management of the underlying cause of cognitive decline [11]. One of the neuropathological changes associated with AD is the accumulation of amyloid-β plaque. While presence of amyloid-β plaque (hence referred to as amyloid) is not definitive of a diagnosis of AD [12], it is one of the few biomarkers available to identify potential etiologies of memory loss and other related symptoms. Therefore, an AD diagnosis informed by an amyloid-β PET scan may enable individuals with elevated amyloid to access newly approved treatments and enter clinical studies testing experimental treatments [13]. PET scans are used in research settings, but are not yet widely available in clinical settings [1]. Medicare does not currently offer reimbursement for amyloid-β PET scans as a diagnostic procedure for AD, but studies are underway to understand the impact of these tests on patients and their care-partners to generate evidence for recommendations about more widespread use.

Findings from a small randomized controlled study of receiving an amyloid-β PET scan finds minimal emotional effects of elevated amyloid scan results on care-partner emotional distress [14]. However, given the high rates of AD care-partner depression and subjective burden, there is a need to examine this relationship in larger samples. Evidence about whether amyloid-β PET scans influence care-partner wellbeing could help to identify opportunities to support care-partners. Care-partners often interact with the health care team when care recipients are being evaluated for AD. Therefore, divulging amyloid-β PET scan results may serve as an information intervention for care-partners. If providers are aware of how amyloid-β PET scan results impact care-partner wellbeing, they might be prompted to evaluate care-partner wellbeing to connect them with needed supports.

We examine the extent to which prior disclosure of elevated versus not-elevated amyloid-β PET scan results are associated with changes in care-partner depressive symptoms, subjective burden, and loneliness over 18 months. We hypothesize that an AD diagnosis informed by an elevated amyloid-β PET result may impact care-partner emotional wellbeing through two opposing pathways. First, the AD diagnosis may offer the care dyad a perceived level of certainty that may improve emotional wellbeing for care-partners over time by allowing them to make sense of their situation, develop coping skills, and better plan for the future [15, 16]. However, an AD diagnosis could also increase distress [17] and loneliness as care-partners begin to anticipate and experience the decline of their family member and the growing need for caregiving that will likely follow [18].

METHODS

Study design and population

The sample derives from the CARE-IDEAS (Caregivers’ Reactions and Experience, a supplemental study of the Imaging Dementia Evidence for Amyloid Scanning Study) study [19], an add-on component to the original IDEAS study. IDEAS examined how the results of amyloid-β PET scans influence clinical management of Medicare beneficiaries aged ≥65 with mild cognitive impairment (MCI) or dementia [11]. Of the 3,717 IDEAS study participants who agreed to be contacted for CARE-IDEAS, 1,028 dyads consisting of participants with cognitive impairment and their care-partners participated in telephone-based surveys at two time-points; at; approximately 4 months after their scan or Time 1 (T1) and again at 18 months after their scan or Time 2 (T2). Care recipients and care-partners were interviewed separately. All participants provided verbal consent and answered a parallel set of questions regarding their demographics; physical, emotional, and cognitive wellbeing; and long-term planning. Care-partners were asked about average hours per week spent caring for the care recipient.

Dyads were excluded from this analysis if they had an uninterpretable amyloid-β PET scan result (n = 3), indicated that the care-partner was not their primary caregiver (n = 57), or were missing scores for any of the study’s three outcomes (n = 47). The resulting sample comprised 921 dyads. The Brown University and Duke University Institutional Review Boards approved the CARE-IDEAS study (#1606001534, #00076890).

Study variables

Amyloid-β PET scans were performed at participating facilities as part of the IDEAS study prior to data collection for CARE-IDEAS. A radiologist or nuclear medicine specialist interpreted the amyloid-β PET scan, and a neurologist shared the results with care recipients. Scan results were categorized by interpreting clinicians as “elevated” versus “not-elevated”.

The three care-partner wellbeing outcomes were assessed at two time points: ~4 months (T1) and ~18 months (T2) following the scan. Depressive symptoms were measured using the Patient Health Questionnaire-8 (PHQ-8) [20]. We used a non-standard PHQ-8 scale, as three of the items were broken into sub-questions to make them easier to understand, including the questions about trouble sleeping, appetite, and energy level. To score the PHQ-8 scale, we included the score from the highest scoring sub-question in the total scale score. Exploratory analyses suggested that this approach did not meaningfully change the overall PHQ-8 scores. The PHQ-8 score ranges from 0 to 24 and a score of 10 or higher represents current depression. Loneliness was assessed as the sum of three, three-level questions ranging from 3–9 about lacking companionship, feeling left out, and feeling isolated [21]. Subjective care-partner burden was assessed with a 4-item ‘screening’ version of the Zarit Burden Interview [22]. The total score ranges from 0 to 16 with a suggested cutoff of 8 for high burden [22]. For all scales, higher scores indicate higher levels of the measured construct.

Care-partner characteristics assessed at T1 included age (in years), sex, race, ethnicity, education (recoded to bachelor’s degree or higher versus other), medical literacy (consisted of two questions about filling out medical forms), self-reported general health (1 = excellent, 5 = poor), marital status, relationship with care recipient (e.g., spouse or significant other) and residential status (e.g., living together or separately), weekly time spent caring for the care recipient (0–5 hours, 6 or more hours), work status (retired, not working, working more, working less), and the domain scores from the CAregiver Perceptions About CommunIcaTion with Clinical Team members (CAPACITY) instrument [23], which measures care-partner perceptions of communication and integration with the care team. We assessed cognitive function using a 35-point version of the Modified Telephone Interview for Cognitive Status (TICS-m), with higher scores indicating better function [24]. Some T1 variables were not included as model covariates due to multi-collinearity. Codebooks and model syntax are available at https://doi.org/10.26300/gprt-kw18.

Statistical analysis

Missing covariate responses were imputed using the mode for categorical and mean for continuous variables, with values taken from the study sample. To test for differences in covariates between individuals with elevated versus not-elevated amyloid, we used chi-square tests for categorical and Kruskal-Wallis tests for continuous variables. To investigate the association between care recipient scan result and change in care-partner emotional wellbeing, separate multivariable linear regression models regressed change (T2-T1) in each outcome (depression, loneliness, and subjective burden) on scan results (elevated versus not-elevated), adjusting for relevant non-collinear T1 covariates. All statistical analyses were performed using the statistical software package SAS version 9.4.

RESULTS

Table 1 describes T1 care-partner and care recipient characteristics overall and by scan result. Care-partners were primarily female (68.5%), non-Hispanic white (94.4%), married to the care recipient (90.7%), and college graduates (62.6%). Care recipients were primarily male (62.3%), also non-Hispanic white (93.7%), received elevated amyloid-β PET scan results (68%). The median age was 74 for care recipients and 71 for care-partners. Bivariate differences of most T1 characteristics were not statistically significant between dyads categorized as elevated versus not-elevated for amyloid on the amyloid-β PET scan (Table 1). T1 sample mean (standard deviation) for the PHQ-8 was 2.4 (3.0), loneliness was 3.9 (1.4), and subjective burden 4.0 (3.1). Differences in outcomes at T1 between elevated and not-elevated groups ranged from 0 to 0.3 and were not statistically significant. Figure 1 shows the mean score, interquartile range, and outliers for each outcome and differences between the elevated and not-elevated groups across time points.

Table 1.

Care-partner and care recipient characteristics, by Amyloid-β PET scan result

Variable	Overall	Elevated	Not-elevated	p
N dyads	921	623	298
Care-Partner Characteristics
Age, Mean (SD) (range 29–91)	70.2 (8.9)	70.7 (8.6)	69.2 (9.4)	0.02
Male	290 (31.5%)	206(33.1%)	84 (28.2%)	0.14
Care recipient is spouse or significant other	835 (90.7%)	566 (90.9%)	269 (90.3%)	0.78
Married or living with partner	885 (96.1%)	602 (96.6%)	283 (95.0%)	0.22
Number of people in household, Mean (SD)	2.2 (0.7)	2.2 (0.7)	2.3 (0.9)	0.41
Non-Hispanic white*	869 (94.4%)	592 (95.0%)	277 (93.0%)	0.20
Bachelor’s degree or higher	577 (62.6%)	392 (62.9%)	185 (62.1%)	0.81
Time providing care for the care recipient				0.21
5 hours or fewer a week	320 (34.7%)	208 (33.4%)	112(37.6%)
6 or more hours a week	163 (17.7%)	119 (19.1%)	44 (14.8%)
Don’t know/refused/missing	438 (47.6%)	296 (47.5%)	142 (47.7%)
Extremely confident filling out medical forms	586 (63.6%)	392 (62.9%)	194 (65.1%)	0.52
Gets help with medical forms	236 (25.6%)	164 (26.3%)	72 (24.2%)	0.48
General health status (self-assessed), Mean (SD) (range 1–5)	2.3 (0.9)	2.3 (0.9)	2.4 (0.9)	0.39
Change in work status because of care recipient				0.95
Retired	240 (26.1%)	161 (25.8%)	79 (26.5%)
Not working, no change	459 (49.8%)	310(49.8%)	149 (50.0%)
Working, same or different hours	222(24.1%)	152 (24.4%)	70 (23.5%)
Living in the same house as care recipient	860 (93.4%)	582 (93.4%)	278 (93.3%)	0.94
Cognitive functioning (TICS-M), Mean (SD) (range 4–41)	28.2 (4.7)	28.2 (4.7)	28.2 (4.6)	0.92
Overall CAPACITY score, Mean (SD)	2.3 (0.6)	2.3 (0.6)	2.3 (0.6)	0.21
CAPACITY: communication domain score, Mean (SD) (range 1–4)	3.1 (0.6)	3.1 (0.6)	3.1 (0.6)	0.30
CAPACITY: capacity domain score, Mean (SD) (range 1–4)	1.6 (0.7)	1.6 (0.8)	1.5 (0.7)	0.24
Care-partner outcomes at Time 1
Depression (PHQ8), Mean (SD)	2.4 (3.0)	2.4 (2.9)	2.4 (3.1)	0.79
Loneliness, Mean (SD)	3.9 (1.4)	3.9 (1.3)	4.0 (1.4)	0.18
Subjective Burden (4-item ZBI), Mean (SD)	4.0 (3.1)	4.1 (3.1)	3.8 (3.2)	0.20
Care Recipient Characteristics
Age, Mean (SD) (range 64–92)	74.3 (5.5)	74.4 (5.5)	73.9 (5.5)	0.10
Male	574 (62.3%)	377 (60.5%)	197 (66.1%)	0.10
Non-Hispanic white	863 (93.7%)	584 (93.7%)	279 (93.6%)	0.95
T1 General health status (self-assessed), Mean (SD) (range 1–5)	2.5 (1.0)	2.4 (1.0)	2.7 (1.0)	<0.001
Dementia diagnosis (versus MCI)	217 (23.6%)	165 (26.5%)	52 (17.4%)	0.003

^*Race and ethnicity were recoded to binary due to the skewed population demographics. Approximately 7% of Care Recipients and ~6% of Care Partner’s self-identified their race as Hispanic-white, non-Hispanic Black or African American, Asian, American Indian or Alaska Native, mixed race, or did not respond to the question.Within the ~10% of Care Partners who were not a spouse or significant other to the Care Recipient, ~7% were children (including in-law and step), and the remaining 3% were other family members (including in-law and step), friends or other. TICS-m, Modified Telephone Interview for Cognitive Status; CAPACITY, CAregiver Perceptions About CommunIcaTion with Clinical Team members; MCI, mild cognitive impairment. The surveys were collected at approximately 1 month (T1) and 18 months (T2) following the Amyloid-β PET scan. All characteristics were taken from the T1 survey. To test for elevated versus not-elevated group differences, we used chi-square tests for categorical variables and Kruskal-Wallis tests for continuous variables.

Fig. 1.

Box and whisker plots of each outcome at Time 1 and Time 2 by elevated and not-elevated groups, median, interquartile range, and outliers.

For all three care partner outcomes assessed at Time 1 (T1), scores worsened over time though mean differences in scores by scan result were not statistically different: loneliness 0.15 (−0.01, 0.32), subjective burden 0.36 (−0.01, 0.73), or depression 0.22 (−0.18, 0.61) (Table 2, Fig. 2). Coefficients represent a unit difference in change scores between elevated and not-elevated groups and therefore a positive coefficient is indicative of higher loneliness, burden, and depression. Supplementary Table 1 provides covariate model parameters for all outcomes.

Table 2.

Mean change (Time 2 – Time 1) in Care-Partner outcomes by Amyloid-β PET scan result, adjusted for Time 1 covariates

Dependent	Elevated	Not-elevated	Mean difference (Elevated-Not-elevated)
Depression (PHQ-8) (range 0–24)	0.46 (−0.27, 1.19)	0.24 (−0.54, 1.02)	0.22 (−0.18, 0.61)
Loneliness (range 3–9)	−0.23 (−0.54, 0.08)	−0.38 (−0.71,−0.05)	0.15 (−0.01,0.32)
Subjective Burden (ZARIT-4) (range 0–16)	0.36 (−0.33, 1.05)	0.00 (−0.73, 0.73)	0.36 (−0.01, 0.73)

CP, care-partner. Standard deviations shownin parentheses.Time 1 covariates included:CAPACITY (Communication and Capacity domains), Care-partner (CP) education, Living status (CP lives with care recipient or not), CP household size, CP work status, CP time spent caring for care recipient, CP cognitive status assessed through the TICS-m, CP confidence filling out medical forms (yes/no), CP gets help with medical forms (yes/no), CP and care recipient general health (self-reported), care recipient dementia diagnosis (versus mild cognitive impairment), care recipient and CP age, CP sex, care recipient and CP race and ethnicity, and CP marital status.

Fig. 2.

Adjusted changes in outcome means over two time points by elevated and not-elevated groups, means and 95% confidence intervals.

DISCUSSION

Our study examined whether results from an amyloid-β PET scan were associated with changes in care-partner emotional wellbeing over time. Contrary to our scenario 1 hypothesis and consistent to findings from Lingler et al. [14], we found no significant between group differences at T1 or changes over 18 months in depressive symptoms, subjective burden, or loneliness at the 0.05 level. While mean changes in outcomes were also small and not clinically significant, scores trended towards worse emotional wellbeing over time and more so among the elevated amyloid group. While this trend could be related to the potential for more severe symptoms among the elevated group, change in care partner mood and distress is likely a more complex process. A recent study using qualitative data from a subsample of CARE-IDEAS study care partners (n = 196) found that increases in care partner anxiety after the scan results were higher for care partners of patients with MCI than for care partners of patients with dementia. Study findings suggest that reactions to scan results might change based on whether care partners expect the results to be elevated (i.e., for patients with dementia) or not (i.e., for patients with MCI) [25]. Furthermore, the uncertainty of not having a diagnosis for cognitive impairment-related symptoms could lead to higher levels of depression, burden, and loneliness. In this case, we may have found a non-significant effect if the scan result produces distress for both those with elevated and non-elevated amyloid. Another explanation for our results could be attributed to a floor effect in levels of distress. Compared with other populations, including care-partners of older Veterans [26], care-partners in this sample had relatively lower rates of depression and subjective burden. This finding suggests our sample may have had access to resources to manage the emotional challenges of receiving an amyloid-β PET scan result. A prior study showed that care-partners of Veterans provided care for an average of 6.5 years [26]. Therefore, it is possible that wellbeing is influenced by longer duration of caregiving, and we may have observed more negative changes had we followed our sample for more than 18 months.

Vast evidence supports the benefit of effective caregiver support interventions [27]. Yet, use of supportive services among this population is persistently low [28, 29]. Family members are often engaged when care recipients are being assessed for AD. However, only 40% of AD care-partners report that care recipients’ providers assess their caregiving needs [30]. Exploratory work suggests that models of care addressing family priorities may decrease feelings of isolation [31]. Therefore, the point in time at which providers relay amyloid-β PET scan results to families could also serve as an opportunity for clinical teams to evaluate care-partners’ needs for supports and make referrals for services, such as respite care and skills training (e.g., coping, nursing tasks).

Limitations of this study include the high proportion of white, college-educated individuals. Results may not generalize to minority groups or those with fewer resources. Also, the T1 survey was administered after families received amyloid-β PET scan results; therefore, we were not able to assess changes in pre-scan emotional wellbeing to post-scan time points. As the median time between receipt of amyloid-β PET scan results and T1 survey administration was 4.5 months, it is possible that changes in care-partner emotional wellbeing were not attributed to learning about the scan results. Another limitation was loss to follow-up; respondents who did not participate in the T2 survey had higher T1 depressive symptoms and this may explain why depressive symptoms did not change significantly over time. Finally, given our use of a non-standard PHQ-8 scale, results pertaining to depressive symptoms might not be comparable to other samples. Nonetheless, this study used a novel dataset containing amyloid-β PET scan results that were linked to longitudinal care-partner reported data inclusive of validated emotional wellbeing measures.

While observed changes in care-partner wellbeing were not statistically or clinically significant, care-partners whose family members receive an AD biomarker informed diagnosis of AD might still benefit from supportive services to enhance existing resilience and coping behaviors. Amyloid-β PET scans are a new diagnostic tool for AD being considered for inclusion as a Medicare-covered service and other new technologies, including blood-based AD biomarkers, are showing promise to detect amyloid levels [32]. If these tools are incorporated into routine clinical practice, accompanying changes in clinic workflow could improve support for AD care-partners. For example, assessing care-partner needs and providing referrals to community-based services when disclosing scan results are opportunities for health systems to proactively offset potentially negative outcomes for care recipients with AD and their care-partners. However, to understand mechanisms between biomarker results and care-partner outcomes as well as inform subsequent intervention, there is an immediate need for additional research on amyloid-β PET scan results disclosure. This is especially important among groups with greater racial, ethnic, and socioeconomic diversity as there may be cultural-related variation in meaning assigned to AD biomarker testing and how biomarker results influence subsequent care-partner distress [25].