Caregiver Assessment of Socioemotional Functioning in People Living With Neurodegenerative Disease (Part 2): Concurrent Validity and Test–Retest Reliability of a New Measure

Abstract

Background and Objectives

Care recipients with neurodegenerative disease can have socioemotional impairments that are challenging for their caregivers. We examine the concurrent and incremental validity of a novel informant-rated measure of care recipient functioning, the Caregiver Assessment of Socioemotional Functioning (CASEF), by examining links with clinician-rated behavioral problems, neuropsychiatric symptoms, and disease severity, as well as caregivers’ reported burden and relationship quality. We also examine test–retest reliability of the CASEF.

Research Design and Methods

Informal caregivers (N = 177) of individuals with neurodegenerative disease reported on their care recipients’ socioemotional functioning and completed structured interviews with clinicians to assess care recipients’ disease severity and neuropsychiatric symptoms. Care recipients’ behavioral symptoms were evaluated with clinician rating scales. Another sample of caregivers (N = 61) repeated the CASEF 4–6 months following an initial administration.

Results

CASEF subscales evidenced concurrent validity with other measures of care recipients’ socioemotional functioning, disease severity, neuropsychiatric symptoms, and behavioral symptoms. CASEF subscales also evidenced incremental validity by accounting for significant additional variance in caregiver burden and relationship quality after accounting for care recipients’ neuropsychiatric and behavioral symptoms. Finally, CASEF subscales demonstrated adequate to excellent test–retest reliability.

Discussion and Implications

Findings highlight the validity and reliability of the CASEF. Coupled with findings from Part 1 of this series (which established CASEF’s construct validity), these findings support using the CASEF in clinical and research settings to assess changes in care recipient socioemotional functioning over time and to examine links between care recipient socioemotional changes and caregiver burden, health, and well-being.

Providing care for a close loved one living with a neurodegenerative disease can be a fulfilling experience (Yu et al., 2018). However, caregiving can also cause an intense burden and negatively alter the quality of the relationship between caregivers and their care recipients with neurodegenerative disease (Watson et al., 2019). Theories suggest that increased burden and diminished relationship quality promote physical and mental health problems for caregivers (Enright et al., 2020). A number of factors are thought to contribute to caregiver burden and the deterioration of relationship quality, including care recipients’ cognitive impairments and disease severity (Mioshi et al., 2013). However, the behavioral and psychological symptoms of dementia (BPSDs) are among the most burdensome and complex manifestations of dementia for caregivers (Kales et al., 2015).

Behavioral and Psychological Symptoms of Dementia

BPSDs are often described as signs of disturbed perception, thoughts, mood, or behavior (Kales et al., 2015). The term BPSDs is often used interchangeably with neuropsychiatric symptoms, which have been defined as the noncognitive symptoms of dementia that occur in clusters or syndromes like depression, apathy, anxiety, or psychosis (Frédéric & Cummings, 2002). One or more neuropsychiatric symptoms are likely to affect nearly all people with neurodegeneration over the course of their illness (Lyketsos et al., 2011). Unfortunately, the complexity of BPSDs makes them difficult to measure and categorize (Kales et al., 2015). After decades of intensive research, neuropsychiatric symptom clusters can now be assessed by standardized instruments such as the Neuropsychiatric Inventory with reasonable levels of reliability and validity (Cummings et al., 1994). Such standardized tools are extremely useful and allow clinicians and scientists to correlate particular symptoms with neuroimaging, neuropathological, or genetic data, and assess their associations with caregiver outcomes (Balthazar et al., 2014).

Traditional measures of neuropsychiatric symptoms, however, do not typically capture changes in specific domains of socioemotional functioning such as (a) negative emotional reactivity, the ability to generate a negative emotional response (e.g., sadness, anger), (b) positive emotional reactivity, the ability to generate a positive emotional response (e.g., amusement, joy), (c) emotion recognition, the ability to recognize others’ emotions accurately. From a clinical affective science perspective, or (d) emotion regulation, the ability to change or modulate one’s emotional response. These core aspects of emotional functioning may underpin broader neuropsychiatric symptom clusters (Levenson et al., 2014). For example, a lack of positive emotional reactivity could underpin the lack of interest in pleasurable activities observed in apathy.

In contemporary research on neurodegenerative diseases, the construct validity of informant report measures used to assess care recipient socioemotional functioning have primarily been established in neurotypical populations where they are used as self-report rather than informant report measures (for a review of measures, see Rankin, 2021). Typically, these informant report measures capture a single domain of socioemotional functioning. For example, the Interpersonal Reactivity Index (Davis, 1983) was validated as a self-report measure of aspects of empathy in neurotypicals, but has been used as an informant report measure in care recipients (e.g., Rankin et al., 2006). The field lacks a well-validated caregiver report measure that captures multiple key aspects of socioemotional functioning (i.e., negative and positive emotional reactivity and emotion recognition).

Care Recipient Socioemotional Functioning, Caregiver Burden, and Relationship Quality

Prior research has linked poor performance on laboratory-based assessments of care-recipient socioemotional functioning to adverse consequences for caregivers (Brown et al., 2018; Chen et al., 2017; Otero & Levenson, 2017). These laboratory assessments, which typically assess socioemotional functioning in response to standardized stimuli and often involve multimodal measures (e.g., physiology, behavior, subjective experience), require extensive resources, time, and experience (Levenson et al., 2014). To begin to disentangle associations among care recipient socioemotional functioning, other BPSDs or neuropsychiatric symptoms, and caregiver outcomes across time, the field would benefit from a validated measure of socioemotional functioning that can be gathered relatively quickly and easily. Because neurodegenerative diseases are progressive, measures should also be evaluated for their appropriateness for longitudinal studies.

In our companion article (Part 1 of this series), we established the construct validity (structural, convergent, and discriminant validity) for the three facets of socioemotional functioning (negative and positive emotional reactivity, and emotion recognition). Changes in each of these socioemotional domains could burden caregivers and erode relationship quality with the care recipient. Positive emotional reactivity with a partner is thought to support connection and satisfaction within relationships (Brown & Fredrickson, 2021; Brown et al., 2022; Otero et al., 2020) and reductions in a care recipient’s positive emotional reactivity could lead caregivers to feel less satisfied with their relationship and more burdened. Negative emotional reactivity could promote shared negative affect within relationships, which is detrimental to relationship quality and could increase burden (Brown et al., 2021; Levenson & Gottman, 1983). Finally, emotion recognition skills are thought to build social bonds and help reduce partners’ distress (Brown et al., 2020a, 2023). Thus, declines in a care recipient’s emotional recognition could negatively impact the quality of relationships and increase caregivers’ burden. Additional research is needed to understand how distinct facets of socioemotional functioning relate to caregiver burden and relationship quality. This research could help clinicians develop and disseminate more targeted and individualized interventions for caregiver-care-recipient dyads.

The Current Study

The current study is the second in a two-part series focused on the Caregiver Assessment of Socioemotional Functioning (CASEF), a newly developed caregiver-report measure that assesses care recipients’ socioemotional functioning. In Part 2, we focus on additional psychometric properties of the three CASEF subscales (negative emotional reactivity, positive emotional reactivity, and emotion recognition) whose construct validity was established in Part 1 of this series. To evaluate concurrent and incremental validity of the three subscales, we examine the associations among the CASEF subscales and (a) caregiver ratings of care recipient depressed affect, positive affect, and sensitivity to the expressed behavior of others; (b) clinician ratings of behavioral problems, disease severity, and neuropsychiatric symptoms; and (c) caregiver ratings of burden and relationship quality. Finally, we examine test–retest reliability of the CASEF subscales over a 4–6-month time period. We do not include an additional CASEF subscale that measured emotion regulation because this subscale was not related to the laboratory measures of emotion regulation in Part 1.

Method

Participants

The study was approved by the Committee for the Protection of Human Subjects at the University of California, Berkeley. Study participants included caregivers and care recipients with neurodegenerative disease enrolled as part of an ongoing research collaboration between Berkeley and the Memory and Aging Center of the University of California, San Francisco (UCSF) studying socioemotional functioning in dementia care recipients and their caregivers. The current manuscript includes participants from two samples: (a) Sample 1 (N = 219) was collected from 2007 to 2012 and focused primarily on care recipients, and (b) Sample 2 (N = 177) was recruited from 2013 to 2020 and focused both on care recipients and caregivers (data collection ended due to COVID-19). See Tables 1 and 2 for demographic and clinical characteristics of the samples (and the companion article for extensive details). Data from Sample 2 were used to evaluate the concurrent and incremental validity of the CASEF because data were available for the CASEF as well as the full set of caregiver-report and clinical-report measures described later. Sample 1 participants did not complete the full set of these measures; however, a subsample (N = 61) completed the CASEF on two occasions separated by 4–6 months, thus providing data for evaluating test–retest validity (Samples 1 and 2 data were not combined for any analyses).

Procedures

Caregivers from Sample 2 completed a battery of questionnaires at Berkeley. Clinician ratings of care recipients functioning (described later) were conducted at UCSF within 3 months of the Berkeley visit. All caregivers in Sample 1 were given the opportunity to repeat the CASEF questionnaire following their initial survey. Sixty-one caregivers from Sample 1 mailed in CASEF data from their homes 4–6 months after they initially completed the questionnaire at Berkeley (see companion article, Part 1 of this series, for additional sample details).

Measures

Caregiver Reports of Care Recipient Functioning

Caregiver assessment of socioemotional functioning.

Three CASEF subscales demonstrated high levels of construct validity based on factor analyses and associations with laboratory-based tasks of socioemotional functioning (see Part 1 of this series): (a) negative emotional reactivity, (b) positive emotional reactivity, and (c) emotion recognition. To complete the CASEF, caregivers reported on the extent to which their partner expressed (emotional reactivity, e.g., “My partner expresses anger”) and recognized (emotion recognition, e.g., “My partner recognizes and understands anger”) each of 10 specific emotions reflecting a range of negative (i.e., anger, disgust, embarrassment, fear, guilt, sadness, shame) and positive (i.e., amusement, joy, pride) emotions over the past month. Responses ranged from 0 (not at all) to 4 (a lot).

Care recipient depressed affect and positive affect.

Caregivers completed the 20-item Center for Epidemiological Studies Depression Scale (CESD), a well-validated measure of depressive symptoms (Santor, 2004). Caregivers rated how often care recipients had symptoms over the past week. Responses ranged from 0 (none of the time) to 3 (most or all of the time). Prior research on the CESD’s factor structure has revealed 4 factors including depressed affect, positive affect, somatic symptoms, and interpersonal symptoms (Shafer, 2006). The current study used the 7-item depressed affect factor (e.g., “felt sad,” “could not shake blues”) and the 4-item positive affect factor (e.g., “felt hopeful about the future,” “felt happy”), with items reverse scored where appropriate such that higher scores indicate greater depressed affect and lower positive affect. We expected depressed and positive affect to be related to CASEF negative emotional reactivity and positive emotional reactivity respectively.

Care recipient sensitivity to others’ expressed behavior.

Caregivers rated their care recipient’s sensitivity to others’ expressed behavior using a 6-item “sensitivity to others’ expressed behavior” subscale (e.g., “My partner is often able to read people’s true emotions through their eyes.”) of the well-validated Revised Self-Monitoring Scale (RSMS; Lennox & Wolfe, 1984). Higher scores indicate greater sensitivity to other’s expressed emotions. We expected these scores to be related to CASEF emotion recognition.

Clinician Ratings

Care recipient behavioral symptoms.

At UCSF, a masters-level neuropsychology trainee or licensed neuropsychologist rated 10 behavioral symptoms that were present during an approximately one-hour neuropsychological testing period (i.e., agitation, stimulus boundedness, perseveration, decreased initiation, motor stereotypies, distraction, lack of social engagement, impulsivity, social inappropriateness, and impaired or fluctuating attention). The UCSF neuropsychological battery includes a variety of standardized assessments of memory, executive function, and other domains of cognition (Kramer et al., 2003). Each behavioral symptom behavior was rated using a 4-point scale: 0 (none), 1 (mild), 2 (moderate), 3 (severe) behaviors. Higher scores indicate more severe behavioral symptoms.

Care recipient neuropsychiatric symptoms.

Care recipients’ neuropsychiatric symptoms were assessed by a trained clinician at UCSF using the well-validated Neuropsychiatric Inventory (NPI; Cummings et al., 1994). The NPI is an informant-rated structured clinical interview that assesses the presence, severity, and frequency of 12 neuropsychiatric symptoms commonly observed in neurodegenerative diseases (i.e., delusions, hallucinations, agitation, depressed mood, anxiety, elation, apathy, disinhibition, irritability, aberrant motor disturbances, nighttime sleep disturbances, and appetite or eating changes). Higher scores indicate greater overall neuropsychiatric symptoms.

Care recipient dementia severity.

Dementia severity was assessed using the well-validated Clinical Dementia Rating Scale (CDR; Morris, 1997). CDR ratings were based on a semistructured interview with the caregiver administered by a trained nurse clinician at UCSF and information obtained from the care recipient. We used the CDR Box Score, which assesses levels of impairment in several domains impacted by dementia (i.e., memory, problem-solving, orientation, judgment, community affairs, home and hobbies, and personal care). Higher scores indicate greater dementia severity.

Caregiver Reports of Burden and Relationship Quality

Caregiver burden

. Caregivers rated their subjective stress using the well-validated 12-item self-report version of the Zarit Burden Interview (ZBI; Bedard et al., 2001; e.g., “Do you feel strained when you are around your relative?”). Higher scores indicate greater caregiver burden.

Caregiver relationship quality.

Spousal caregivers (not nonspousal) completed a well-validated 15-item self-report measure of marital relationship quality, the Locke–Wallace Marital Adjustment Scale (Locke & Wallace, 1959; e.g., “When disagreements arise, they usually result in (a) husband giving in, (b) wife giving in, and (c) agreement by mutual give and take.”) Higher scores indicate greater relationship quality.

Statistical Analyses

All statistical analyses for the current manuscript were conducted in IBM SPSS. To assess the concurrent validity of the CASEF subscales (negative emotional reactivity, positive emotional reactivity, emotion recognition) with caregiver reports of care recipient functioning, we conducted three linear regression models. In each regression, one of the three caregiver reports (depressed affect, positive affect, and sensitivity to others’ expressed behavior) served as the dependent variable and the predictor variables were the three CASEF subscale scores.

To assess the concurrent validity of the CASEF subscales with clinician ratings, we conducted three linear regressions. In each regression, one of the three clinician ratings (behavioral symptoms, dementia severity, neuropsychiatric symptoms) served as the dependent variable and the predictor variables were the three CASEF subscale scores.

To examine concurrent validity of the CASEF subscales with caregiver burden and relationship quality, we conducted two linear regressions. In each regression, one of the caregiver scores (burden, relationship quality) served as the dependent variable and the predictor variables were the three CASEF subscale scores.

To examine incremental validity, we conducted two additional regressions. In each regression, one of the caregiver scores (burden, relationship quality) served as the dependent variable. On the first step, we entered clinician ratings of care recipient dementia severity, behavioral symptoms, and neuropsychiatric symptoms as predictors. On the second step, we added the three CASEF subscale scores as predictors. We examined the change in R² associated with adding these CASEF scores as a measure of incremental validity and also determined which predictors were statistically significant in the regression model associated with this second step.

To examine the test–retest reliability of the CASEF subscales, we used the data from the 61 participants from Sample 1 who repeated the CASEF on two occasions 4–6 months apart. Cronbach’s alpha was computed as a measure of the consistency of subscale scores over time. Because this statistical test is correlational in nature, a repeated measures ANOVA was also computed to determine the stability of scores over time.

Results

Correlations among primary variables are presented in Table 3. Detailed results from validity analyses for CASEF subscales are found in Table 4 (caregiver reports of care recipient functioning), Table 5 (clinician ratings of care recipient functioning), and Table 6 (caregiver reports of burden and relationship quality).

Associations with Caregiver Reports of Care Recipient Functioning

Care Recipient Depressed Affect

The regression model examining concurrent validity of CASEF reactivity subscales with caregiver reports of care recipient depressed affect revealed that greater CASEF negative emotional reactivity was positively associated with care recipients’ depressed affect, β = 0.51, t(172) = 8.29, p < .001, whereas CASEF positive emotional reactivity was negatively associated with care recipients depressed affect, β = −0.33, t(172) = −4.70, p < .001. CASEF emotion recognition was not significantly associated with care recipient depressed affect.

Care Recipient Positive Affect

The regression model examining concurrent validity of CASEF reactivity subscales with caregiver reports of care recipient positive affect revealed that greater CASEF positive emotional reactivity was positively associated with care recipients’ positive affect, β = 0.52, t(172) = 7.46, p < .001, whereas CASEF negative emotional reactivity was negatively associated with care recipients positive affect, β = −0.22, t(172) = −3.55, p < .001. CASEF emotion recognition was not significantly associated with care recipient positive affect.

Care Recipient Sensitivity to Other’s Expressed Behavior

The regression model examining concurrent validity of CASEF emotion recognition with caregiver reports of care recipients’ sensitivity to others’ expressed emotion revealed that greater CASEF emotion recognition was positively associated with greater sensitivity to others’ expressed behavior, β = 0.76, t(170) = 13.64, p < .001. CASEF negative emotional reactivity and positive emotional reactivity scores were not significantly associated with sensitivity to others’ expressed behavior.

Associations with Clinician Ratings of Care Recipient Functioning

Care Recipient Behavioral Symptoms

The regression model examining the concurrent validity of CASEF subscales with clinician-rated behavioral symptoms revealed that greater CASEF emotion recognition was significantly associated with lower care recipient behavioral symptoms, β = −0.24, t(115) = −2.32, p = .022. CASEF negative emotional reactivity and positive emotional reactivity were not significantly associated with care recipient behavioral symptoms.

Care Recipient Neuropsychiatric Symptoms

The regression model examining concurrent validity of CASEF subscales with clinician ratings of care recipients’ neuropsychiatric symptoms revealed that greater CASEF negative emotional reactivity, β = 0.20, t(124) = 2.55, p = .012, and lower CASEF emotion recognition, β = −0.46, t(124) = −5.19, p < .001, were significantly associated with greater care recipients’ neuropsychiatric symptoms. CASEF-positive emotional reactivity was not significantly associated with care recipients’ neuropsychiatric symptoms.

Care Recipient Dementia Severity

The regression model examining the concurrent validity of CASEF subscales with care recipients’ dementia severity revealed that greater CASEF positive emotional reactivity, β = −0.19, t(138) = −2.25, p = .026, and greater CASEF emotion recognition, β = −0.36, t(138) = −4.13, p < .001, were significantly associated with lower dementia severity. Negative emotional reactivity on the CASEF was not significantly associated with dementia severity.

Associations with Caregiver Burden and Relationship Quality

Caregiver Burden

The regression model examining the concurrent validity of CASEF subscales with caregiver burden revealed that greater CASEF negative emotional reactivity, β = 0.30, t(131) = 4.04, p < .001, lower CASEF positive emotional reactivity, β = −0.16, t(131) = −2.01, p = .046, and lower CASEF emotion recognition, β = −0.36, t(131) = 4.47, p < .001, were all significantly associated with greater caregiver burden (see Supplemental Table 1 in the Supplementary Material).

The regression model examining incremental validity of CASEF subscales for predicting caregiver burden after accounting for behavioral symptoms, neuropsychiatric symptoms, and dementia severity, revealed that CASEF scores added significant explanatory variance in the prediction of caregiver burden beyond these other variables, R²_change = 0.060, F_change (3, 80) = 2.915, p = .039. In the final step of this model, greater CASEF negative emotional reactivity, lower behavioral symptoms, lower neuropsychiatric symptoms, and lower dementia severity were each independently associated with greater caregiver burden (see Table 6).

Caregiver Relationship Quality

The regression model examining concurrent validity of CASEF subscales with caregiver relationship quality revealed that greater CASEF emotion recognition, β = .51, t(139) = 6.51, p < .001, and lower CASEF negative emotional reactivity, β = −0.20, t(139) = −2.89, p = .004, were associated with greater caregiver relationship quality, whereas CASEF positive emotional reactivity was not significantly associated with caregiver relationship quality in this model.

The regression model examining incremental validity of CASEF subscales for predicting caregiver relationship quality after accounting for behavioral symptoms, neuropsychiatric symptoms, and dementia severity, revealed that CASEF scores added significant explanatory variance in the prediction of caregiver relationship quality beyond these other variables, R²_change = 0.20, F_change (3, 92) = 10.25, p < .001. In the final step of this model, greater CASEF emotion recognition and lower neuropsychiatric symptoms were each independently associated with caregiver relationship quality (see Table 6)

Test–Retest Reliability

Tests of score consistency over the 4–6 month period were high for CASEF emotion recognition (alpha = 0.90) and positive reactivity (alpha = 0.85) and somewhat lower for negative reactivity (alpha = 0.66). The analysis to determine stability over time revealed no significant changes in CASEF emotion recognition, F(1,46) = 0.57, p = .45, positive reactivity, F(1,60) = 0.52, p = .47, or negative reactivity, F(1,57) = 0.24, p = .23 (see the Supplementary Material for means).

Discussion

The current study examines concurrent and incremental validity, and test–retest reliability of the CASEF, a newly developed informant report measure of socioemotional functioning in neurodegenerative disease. Part 1 of this series identified three subscales (negative emotional reactivity, positive emotional reactivity, emotion recognition) that had high construct validity based on factor analyses and associations with laboratory tests of emotional functioning. The current study (Part 2) examined additional aspects of the validity (concurrent and incremental) of these three subscales using: (a) caregiver reports of care-recipient functioning, (b) clinician ratings of care-recipient functioning, and (c) caregiver reports of burden and relationship quality. We also evaluated test–retest reliability over a 4–6 month period. Overall, the CASEF showed high levels of concurrent and incremental validity as well as high test–retest reliability. Combining results from Part 1 and Part 2 of this series, the CASEF shows promise for providing researchers and clinicians with a reliable, valid, and relatively easy-to-administer informant report measure of care recipients’ socioemotional functioning.

Validity and Reliability of CASEF Subscales

In the present study, evidence for the concurrent validity of the CASEF subscales was found in three areas. First, the CASEF subscales were associated with conceptually similar subscales on other caregiver-rated scales. Specifically, higher CASEF emotion recognition scores were related to higher informant reports of care recipients’ sensitivity to others’ expressed behavior. In addition, higher CASEF negative and positive reactivity scores were related to higher informant reports of care recipients’ depressed and positive affect respectively. Second, the CASEF subscales were associated with clinician ratings of care recipients’ dementia severity. Specifically, lower CASEF positive reactivity and emotion recognition scores were associated with higher levels of dementia severity. Third, the CASEF subscales were associated with caregiver reports of burden and relationship quality. Specifically, higher CASEF negative reactivity and lower CASEF positive reactivity were associated with higher caregiver burden and lower CASEF emotion recognition scores were associated with higher caregiver burden. In addition, higher CASEF negative reactivity and lower emotion recognition scores were associated with lower relationship quality.

Although most findings supported the concurrent validity of the CASEF, there were several instances where the CASEF subscales did not show expected relationships with other measures. For example, although CASEF positive reactivity and emotion recognition scores were related to dementia severity, CASEF negative reactivity scores were not. This may reflect the nature of our sample, which had relatively low levels of severe dementia. Associations may differ in populations with greater dementia severity given that greater negative affect, such as agitation and aggression, may be more common in more severe disease pathology (Gottesman & Stern, 2019). In addition, CASEF negative and positive reactivity scores were not related to clinician ratings of care recipient behavioral symptoms. This might be explained by the nature of the items in the scale. The ten-item behavioral symptom scale only included one purely affective symptom—agitation.

The incremental validity of the CASEF subscales was supported in two important domains of caregiver functioning. The CASEF subscales (and particularly negative reactivity) accounted for significant additional variance in caregiver burden beyond that accounted for by clinician-rated measures of care recipient functioning (dementia severity, behavioral symptoms, neuropsychiatric symptoms). A similar pattern of findings was obtained for caregiver relationship quality. Again, the three CASEF subscales collectively accounted for significant additional variance in relationship quality beyond the measures of care recipient functioning, but in this case, CASEF emotion recognition emerged as the most important. Findings are important given the substantial roles that caregiver burden and relationship factors play in caregiver mental and physical health problems (Clyburn et al., 2000), caregiver and care recipient mortality (Schulz et al., 2020), and care recipient institutionalization (Gaugler et al., 2011).

Finally, tests of the reliability of the CASEF subscales were also encouraging. The three CASEF subscales demonstrated moderate to excellent test–retest reliability over a 4–6-month period using Cronbach’s alpha, which captures the consistency of participants’ pairs of ratings across time. We also found that mean scores on the CASEF subscales demonstrated stability over this brief time period.

Using the CASEF in Research and Clinical Settings

Given the profound effects that dementia can have on care recipients’ emotional functioning and on caregiver health (Brown et al., 2020b, c) there is a need for a reliable, valid, easy-to-administer informant-report measure of care recipient emotional functioning that can be used in research and clinical settings. By including multiple aspects of emotional functioning (negative reactivity, positive reactivity, emotion recognition) and multiple specific emotions, the CASEF can better characterize the emotional functioning of individual care recipients in research settings via an easily administered questionnaire.

Better measurement, understanding, and taxonomization, of care recipients socioemotional functioning, may enable health care providers to create better and more personalized treatments (Kales et al., 2015). For example, our incremental validity analyses also indicated that care recipients negative emotional reactivity was particularly associated with caregiver burden. This suggests that care recipients who regularly express negative emotions (e.g., fear, anger, and sadness) may be especially burdensome for their caregivers. In contrast, we found that care recipient emotion recognition was particularly associated with relationship quality. This is consistent with research emphasizing the important role of emotion recognition in building and maintaining social relationships (Levenson et al., 2017). The inability to recognize a partner’s emotions might lead to an inability to downregulate a partner’s negative emotions or share in a partner’s positive emotions, which are critical components of happy relationships and good health (Bloch et al., 2014; Brown & Fredrickson, 2021; Brown et al., 2020a; Otero et al., 2019).

Identifying care recipients with high negative emotional reactivity or low emotion recognition scores could help clinicians design more person-centered treatments. For example, a caregiver caring for someone scoring low on emotion recognition on the CASEF might benefit from caregiver psychoeducation about the effects that neurodegeneration can have on empathy (Brown, Hua, et al., 2020), which could facilitate reappraisal of problematic care recipient behavior as a symptom of a disease instead of emotional indifference or a problem in their relationship. Moving beyond psychoeducation, person-centered interventions could be developed and tailored to help caregivers manage the specific behavioral challenges of the care recipient.

Strengths, Limitations, and Future Directions

The current study has several strengths including evaluating multiple types of validity (concurrent, incremental) and reliability using data obtained from caregivers and clinicians. Additionally, the relatively large sample included care recipients diagnosed with a variety of neurodegenerative diseases that can affect cognitive, affective, language, and motor functioning. The study has several limitations as well. Despite heterogeneity in our sample in dementia diagnoses, generalizability to racially and sociodemographically diverse populations may be limited due to participants being predominately white and well-educated spousal caregivers. Future work should validate the CASEF in nationally representative samples. Another limitation is that we did not establish whether the CASEF is sensitive to change in the caregiver recipient's emotional functioning over time. Our 4–6 month test–retest interval, although appropriate for establishing test–retest reliability, seems insufficiently long for significant changes in emotional functioning to occur. Future studies should administer the CASEF repeatedly over longer time periods. If the CASEF is sensitive to changes over time, then including it in population-based longitudinal studies could examine whether socioemotional decline contributes to health care costs, institutionalization of care recipients, and other public health concerns.

In terms of future directions, additional longitudinal research is needed to determine the predictive validity of the CASEF and the relationship between CASEF subscales and specific behavioral and neuropsychiatric symptoms. Such work would help researchers and clinicians better understand how emotional reactivity and recognition relate to particular neuropsychiatric symptom clusters. For example, although our results indicated that positive reactivity was not associated with overall neuropsychiatric symptoms as measured by the NPI, future research could examine how changes in positive emotional reactivity relate to more specific symptoms such as euphoria or apathy. Finally, additional psychometric work with the CASEF may help produce versions with fewer items for measuring the three subscales and help develop a better fourth subscale for measuring emotion regulation, an important aspect of emotional functioning that was not well measured in the current version (see Part 1).

Conclusion

To date, most validated informant-rated measures of functioning in neurodegenerative populations have primarily assessed cognitive and/or functional impairment. Although cognitive and functional changes are crucial indicators of care recipient decline, socioemotional changes are equally clinically significant and are among the first symptoms of various forms of neurodegenerative disease (Santamaría-García et al., 2016). With our companion article (Part 1) that demonstrated construct validity, present findings support the concurrent and incremental validity and test–retest reliability of the CASEF for assessing socioemotional functioning in neurodegenerative diseases. Care recipient socioemotional functioning, as measured by the CASEF, has important implications for understanding caregivers’ burden and relationship quality. In addition, the CASEF may be a useful adjunct for producing more tailored interventions that could help preserve the health and well-being of both caregivers and care recipients.

Funding

This work was supported by the National Institutes of Health [R37AG017766 to RWL]. Preparation of this manuscript was supported by National Institute on Aging [P01AG019724 to RWL, R00AG073617 to CLB], and the Department of Veterans Affairs Office of Academic Affiliations Advanced Fellowship Program in Mental Illness Research and Treatment and the VISN 21 MIRECC to MCO.

Conflict of Interest

None.

Data Availability

The studies in the current manuscript were not pre-registered, and all de-identifiable data, materials, and code are available from the authors upon request.

Acknowledgments

The researchers would like to thank past and present members of the Berkeley Psychophysiology Lab, research coordinators at the UCSF Memory and Aging Center, and members of the Georgetown CARES Lab. Researchers would also like to thank care recipients and their caregivers for their participation and feedback.

References

Author notes