Communication Patterns and Characteristics of Family Caregivers and Persons Living With Dementia: Secondary Analysis of Video Observation

Abstract

Background:

It is essential to characterize communication patterns for better health outcomes for family caregivers and persons living with dementia.

Objective:

This study aimed to examine the relationships between communication patterns and the characteristics of dyads of family caregivers and persons living with dementia.

Methods:

A secondary analysis was conducted using 75 video-recorded home care observations from 19 dyads. Participant characteristics and caregiver burden, depression, and sense of competence were collected from the parent study. The video-recorded dyadic communication patterns were assessed using a coding scheme developed based on Communication Accommodation Theory and Classical Test Theory. The relative frequency of the communication patterns was compared between groups.

Results:

Overall, 8311 caregiver and 8024 care recipient communication behaviors were observed. Caregiver communication patterns were categorized as facilitative, disabling, and neutral. Care recipient communication patterns were categorized as engaging, challenging, and neutral. Caregiver gender, care recipient gender, care recipient education level, dementia diagnosis length, types of dementia, dyadic gender difference, burden, depression, and competence of caregiver, and types of communication were significantly associated with caregiver communication. Dementia diagnosis length, caregiver competence, dyadic gender difference, and types of communication were significantly associated with care recipient communication.

Conclusions:

The findings demonstrated different communication patterns depending on individual and dyad characteristics and evidence for dyadic communication support to promote meaningful interaction for persons living with dementia. Further analysis is needed to identify mediating factors and causal relationships.

In 2023, an estimated 6.7 million older adults aged 65 and older in the United States were living with Alzheimer’s disease and other dementias. ¹ Family caregivers provide most of the unpaid care for persons living with dementia in the community and often perceive communication with persons living with dementia as stressful and problematic. ¹ Caregiver burden, depression, and competence in dealing with burden are significantly related to characteristics of communication patterns between family caregivers and persons living with dementia. For example, communication problems and challenging behaviors of persons living with dementia impede maintaining social interaction with and task completion for persons living with dementia, leading to increasing caregiving burden and depression. ² On the contrary, engaging in meaningful communication is critical to maintaining co-mmunication competence of family caregivers in care. ³ Communication Accommodation Theory (CAT) focuses on why and how people modify their communication patterns during interpersonal interaction and the outcome of those adjustments. ⁴ Communication Accommodation Theory differentiates accommodative and nonaccommodative communication, which are either positive or negative to support emotion and relationship in interpersonal communication. ⁴ In this research, CAT was used to conceptualize communication behaviors in dementia family dyads and the roles that caregivers play in engaging persons living with dementia. Communication behaviors captured by video observations were categorized as facilitative or disabling for family caregivers and engaging or challenging for persons living with dementia.

Communication Patterns in Dementia Family Care

Micro-level verbal and nonverbal communication refers to the smallest distinguishable verbal and nonverbal behavior unit to which a coder can assign a classification, such as an eye gaze, head nodding, and an utterance. ⁵ Micro-level assessment allows for a more detailed and accurate characterization of communication, which can be assessed using a video observation. ⁵ Prior studies have investigated the relationships between caregiver-care recipient dyadic communication patterns and caregiving experience. For example, Petrovsky and colleagues reported caregivers’ harsh communication (eg, yelling, screaming) was positively associated with caregiver burden and depression among 250 family caregiver-care recipient dyads. ² Yet, several limitations are noted in prior studies, such as small sample sizes (N < 50), ⁶ focusing on one type of communication (eg, conflicted or task-related),^2,6 and the use of self-reported questionnaires that may be inaccurate due to recall bias.^2,6,7 Most of all, these studies did not differentiate between verbal and nonverbal communication, which is critical for understanding and improving communication in dementia care.^6,7 Addressing these limitations is necessary for understanding and improving communication in dementia care.

To fill this gap, we developed the Dyadic Communication Observational coding scheme in DEmentia care (DCODE) to characterize patterns of video-recorded dyadic communication between family caregivers (eg, spouses, adult children) and persons living with dementia. ⁸ The DCODE is a coding system specifically designed to measure micro-level verbal and nonverbal communication behaviors in family caregiver-care recipient dyads during task-related and social informal interactions based on Classical Test Theory (CTT) and CAT. The DCODE has strengths over other instruments to assess family caregiver-care recipient dyadic communication. First, the DCODE was developed based on the conceptual framework for use with both spousal and adult-child caregiver-care recipient communication during task-related and social communication contexts, which allows assessment of a broader range of communication situations. Other dyadic instruments found in the literature were developed without a conceptual framework and without consideration of type of communication. ⁹ Moreover, the DCODE assesses 3 categories of communication, including positive, negative, and neutral communication. Communication is a dynamic process with specific context-dependent behaviors (ie, neutral communication). The DCODE allows more flexible understanding of family dyadic communication by separating context-dependent communication from positive or negative communication.

Purpose

This study aimed: (1) to characterize the frequency of communication patterns between family caregivers and care recipients in video-recorded in-home care interactions and (2) to examine the relationships between individual characteristics of family caregivers and persons living with dementia (independent variables) and positive, negative, and neutral communication patterns (dependent variables).

Methods

Study Design

This cross-sectional study was a secondary behavioral analysis of in-home, video-recorded observations between family caregivers and persons living with dementia. Video-recorded observations were obtained from an extensive repository of the Supporting Family Caregivers with Technology for Dementia Home Care study. ¹⁰ The parent study was a randomized clinical trial that tested the effectiveness of an in-home video monitoring telehealth intervention for family caregivers on behaviors of persons living with dementia. ¹⁰ University Institutional Review Board (IRB) approval was obtained for the parent study and this study.

In the parent study, caregivers were eligible if they: (1) were the spouse, a child, the spouse of a child, a friend or neighbor, or paid paraprofessional caregiver and (2) provided in-home care to the care recipient weekly. Care recipients were eligible if they were diagnosed with Alzheimer’s disease or other dementias. ¹⁰ The participants were recruited from 2 Midwest areas in the United States by utilizing advertisements through local newspapers and magazines and giving presentations to civic groups, organizations, and regional conferences in the communities. ¹⁰ Caregivers took control over recordings and the selection of videos during care situations to submit those which they wanted feedback on from the experts who reviewed the videos each week during the study period. All participants provided informed consents/assents by written format to participate in the parent study. In this study, only spouse and adult children (child or child in-law) caregivers were included.

Sample and Setting

In the parent study, a total of 284 videos were recorded from 84 dyads (83 family caregivers and 71 care recipients). Caregivers had a mean age of 64.2 years. Most caregivers were spouses (66.3%), female (71.2%), and non-Hispanic white (92.8%), and received a college or higher level of education (59.0%). Care recipients had a mean age of 75.7 years. Most care recipients were male (59.2%), non-Hispanic white (95.8%), and received education less than college (54.9%). Care recipients were diagnosed with dementia for 4.2 years on average with half having moderate severe dementia (50.7%).

The videos were screened based on the following inclusion criteria: (1) video includes dyadic interaction limited to one consented caregiver and one assented care recipient; (2) video includes at least 1 verbal and/or nonverbal behavior from both the caregiver and the care recipient; (3) video includes at least 10 seconds where the caregiver and the care recipient are visible; (4) video length is at least 30 seconds; and (5) the quality of the video is good enough so behaviors can be observed. Videos were excluded if one dyad partner was not visible or if no communication behaviors were observed.

Among the 284 videos, 75 videos from 19 unique caregiver-care recipient dyads were eligible for this study. Fifty videos recorded communication in spousal dyads and 25 videos in adult-child dyads. Among the 75 videos, each dyad recorded 4 videos on average (range = 2-10). The mean length of the 75 videos was 8.46 minutes (SD = 4.39 minutes, range = 0.71-22.37 minutes). The types of communication in the videos were social informal communication (n = 35, 46.7%) and task-related communication related to assistance with activities of daily living (ADL; ie, mealtime, dressing, medication administration, n = 31, 41.3%) or activities or games (n = 9, 12.0%).

Interventionist notes from the parent study were reviewed for each video to determine if communication-related feedback was provided during the weekly feedback sessions. Based on that information, 41 videos were recorded before the caregiver received any communication-related intervention (pre-intervention), and the other 34 videos were recorded after receiving any communication-related intervention (post-intervention). Of the 19 caregivers, 9 caregivers only recorded videos before they received the communication-related intervention, 5 caregivers only recorded videos after they received the intervention, and 5 caregivers recorded videos both before and after they received the intervention. We made efforts to include as many videos as we could, and it was confirmed that there was no significant relationship between receiving the in-home video monitoring telehealth intervention and the use of communication patterns by both caregiver and care recipient. The parent study implemented an in-home video monitoring telehealth intervention that did not specifically focus on improving communication skills or strategies, but rather aimed to provide caregiver education on managing behavior problems, ADL-related care, safety, and caregiver support, such as respite care.

Data Collection and Variables

Participant demographic information (ie, age, gender, kinship status, education, type of dementia, dementia diagnosis length, caregiving length, and Functional Assessment STaging [FAST] score for dementia staging), Modified Zarit Burden Scale (MZBS), ¹¹ Center for Epidemiologic Studies Depression Scale (CES-D), ¹² and Short Sense of Competence Questionnaire (SSCQ) ¹³ at the baseline assessment from the parent study were used.

Care recipient and caregiver communication were coded from the 75 videos using the DCODE. ⁸ The development process, detailed item definitions, and rationales are available elsewhere. ⁸ The items of the DCODE are shown in Supplementary Table 1. The item definitions and the coding rationale of the DCODE were developed from existing instruments and the literature describing communication behaviors in dementia populations.^8,9 The pool of items was initially developed through a comprehensive literature review. The items were reviewed by 6 experts specialized in nursing, communication, dementia, and family caregiving to evaluate item relevancy, clarity, specificity, and feasibility to establish face and content validity. Based on the feedback and qualitative review gathered from the experts, the items were further refined.

The final items were confirmed after undergoing pilot testing and psychometric testing. First, based on the CAT, caregiver and care recipient communication were initially categorized as accommodative or nonaccommodative based on: (1) whether caregiver communication behavior can facilitate/disable the engagement of care recipient in communi-cation and (2) whether care recipients show engaging or challenging communication as a response of caregiver communication. The DCODE differentiates accommodative and nonaccommodative dyadic communication necessary to ch-aracterize and evaluate family communication interventions in dementia. Furthermore, after a reiterative refinement process and the literature review, we found context-dependent communication behaviors, classified as either accommodative or nonaccommodative depending on the situation and context of communication. ³ These behaviors were classified as neutral communication. For example, “clarification-qu-estion” was classified as caregiver neutral communication because it has been reported as a facilitative behavior to identify the target word by care recipient ¹⁴ ; however, it has been reported as disabling when the care recipient is agitated. ³

Second, the DCODE was psychometrically tested and showed adequate evidence of reliability for internal consistency (Cronbach’s α = 0.73), intra-rater reliability (Cohen’s kappa = 0.83, rate of agreement = 83.88%), and inter-rater reliability (Cohen’s kappa = 0.81, rate of agreement = 81.75%). Its validity was acceptable in content validity (item content validity index [I-CVI] = 0.93, scale-level content validity index [S-CVI/UA] = 0.71, scale content validity index/average proportion [S-CVI/Ave] = 0.93), convergent validity (r_s = 0.46 between caregiver facilitative and care recipient engaging communication, r_s = 0.51 between caregiver disabling and care recipient challenging communication), and predictive validity (Wald χ² = 2.92, 95% confidence interval [CI] = 0.04, 1.26; odds ratio [OR] = 0.21, p = .09 between caregiver burden and caregiver disabling communication, Wald χ² = 2.94, 95% CI = 0.73, 103.84; OR = 8.73, p = .09 between caregiver burden and care recipient challenging communication). The DCODE also showed feasibility and ease of use in video observational analysis of dyadic communication in dementia family populations. ⁸

The final items for caregiver communication consist of 24 verbal and 19 nonverbal behavior items (Table 1). Caregiver facilitative communication was defined as supportive and collaborative communication to maintain interaction with the care recipient (19 items). For example, back-channeling was categorized as facilitative, because it was defined as “responsive to care recipient’s communication,” as nonresponsive behaviors are considered negative that may hinder engaging care recipient in communication. Caregiver disabling communication was defined as nonresponsive, demanding, and critical communication, leading to care recipient disengagement (13 items). For example, withdrawing was defined as verbal expression of disengagement from the interactions, showing backing off or shutting down. Caregiver neutral communication was defined as communication that is depending on the context (11 items).

Table 1.

Descriptive Caregiver (CG) Communication Behaviors (N = 8311).

Communication category	n (%)	Relative frequency (%)	Relative duration (%)
Facilitative nonverbal
Minimize or eliminate distractions	45 (3.5)	0.13 (2.7)	0.71 (81.6)
Intimate spacing	101 (7.8)	0.23 (4.8)
Eye focused on the activity	633 (48.9)	1.47 (30.7)	0.16 (18.4)
Head nodding	109 (8.4)	0.20 (4.2)
Express positive affection nonverbally	221 (17.1)	0.43 (9.0)
Comfort/empathetic touch	24 (1.9)	0.08 (1.7)
Providing assistance	160 (12.4)	2.24 (46.9)
Facilitative verbal
Address person by the person’s name	12 (0.4)	0.04 (0.7)
Address person by endearment	41 (1.4)	0.10 (1.7)
Back-channeling	374 (12.5)	0.92 (15.1)
Distraction	79 (2.6)	0.14 (2.3)
Statement-clarification	283 (9.5)	0.50 (8.2)
Question-permission	17 (0.6)	0.02 (0.3)
Statement-permission	9 (0.3)	0.02 (0.3)
Suggestion	604 (20.3)	1.28 (21.1)
Praise or positive feedback	186 (6.3)	0.46 (7.6)
Reassurance	91 (3.0)	0.17 (2.8)
Providing information	1137 (38.2)	2.24 (36.8)
Use humor	145 (4.9)	0.19 (3.1)
Disabling nonverbal
Interacting in loud noise	18 (2.4)	0.06 (3.2)	0.22 (27.5)
Third person speaking	169 (22.1)	0.42 (22.4)	0.52 (65.0)
Walk away from CR	78 (10.2)	0.19 (10.2)
Eye stares at space	434 (56.9)	0.68 (36.4)	0.06 (7.5)
Eye closed a	0 (0.0)	0.00 (0.0)	0 (0.0)
Express negative affection	64 (8.4)	0.52 (27.8)
Disabling verbal
Check CR’s memory	366 (35.1)	0.50 (31.3)
Domineering/controlling	42 (4.0)	0.08 (5.0)
Correcting	216 (20.7)	0.33 (20.6)
Verbalize negative affection	245 (23.5)	0.37 (23.1)
Scolding	81 (7.7)	0.17 (10.6)
Withdrawing	19 (1.8)	0.02 (1.3)
Speak louder	75 (7.2)	0.13 (8.1)
Neutral nonverbal
Direct eye gaze to CR	857 (53.3)	1.60 (55.7)	0.10 (100.0)
Shaking head	63 (3.9)	0.12 (4.2)
Attention/guided touch	34 (2.1)	0.09 (3.1)
Hand gestures: illustration/demonstration	250 (15.5)	0.36 (12.6)
Hand gestures: showing an object	115 (7.2)	0.22 (7.7)
Hand gestures: pointing	289 (18.0)	0.48 (16.7)
Neutral verbal
Speak slowly	11 (1.8)	0.02 (1.6)
Question-clarification	284 (45.4)	0.62 (50.0)
Redirect/replace	72 (11.5)	0.15 (12.1)
Paraphrases, rephrase	148 (23.7)	0.27 (21.8)
Repeats	110 (17.6)	0.18 (14.5)

Abbreviation: CR: care recipient.

^aNot observed and excluded in the analysis.

The final items for care recipient communication consist of 24 verbal and 17 nonverbal behavior items (Table 2). Care recipient engaging communication was defined as responsive and collaborative communication as a sign of engagement (22 items). Care recipient challenging communication was defined as nonresponsive, resistive, withdrawing, and critical communication showing disengagement (12 items). Care recipient neutral communication was defined as the behavior not classified as either engaging or challenging depending on the context (7 items).

Table 2.

Descriptive Care Recipient (CR) Communication Behaviors (N = 8024).

Communication category	n (%)	Relative frequency (%)	Relative duration (%)
Engaging nonverbal
Intimate spacing	34 (1.4)	0.11 (2.2)
Direct eye gaze to CG	891 (37.4)	1.82 (37.3)	0.08 (29.6)
Eye focused on the activity	722 (30.3)	1.62 (33.2)	0.19 (70.4)
Head nodding	200 (8.5)	0.33 (6.8)
Express positive affection nonverbally	414 (17.4)	0.67 (13.7)
Comfort/empathetic touch	8 (0.3)	0.02 (0.4)
Perform activity independently/with partial assistance	112 (4.7)	0.31 (6.4)
Engaging verbal
Address person by the person’s name	5 (0.2)	0.01 (0.3)
Address person by endearment	8 (0.4)	0.02 (0.6)
Back-channeling	565 (25.9)	0.97 (29.4)
Ask for help	108 (5.0)	0.22 (6.7)
Ask for reassurance	86 (3.9)	0.17 (5.2)
Question-clarification	242 (11.1)	0.24 (7.3)
Statement-clarification	398 (18.3)	0.40 (12.1)
Question-permission	3 (0.1)	0.01 (0.3)
Statement-permission	12 (0.5)	0.04 (1.2)
Suggestion	132 (6.1)	0.15 (4.5)
Praise or positive feedback	119 (5.5)	0.20 (6.1)
Reassurance	31 (1.4)	0.05 (1.2)
Providing information	387 (17.8)	0.71 (21.5)
Use humor	80 (3.7)	0.11 (3.6)
Speak slowly	2 (0.1)	0.00 (0.0)
Challenging nonverbal
Walk away from CG	26 (4.7)	0.10 (8.4)
Eye stares at space	527 (95.3)	1.09 (91.6)	0.08 (100.0)
Eye closed a	0 (0.0)	0.0 (0.0)	0 (0.0)
Challenging verbal
Repeats (more than 3 times)	12 (2.0)	0.02 (1.3)
Incoherent communication	58 (9.7)	0.08 (5.3)
Unrecognizable communication	78 (13.0)	0.69 (45.4)
Correcting	99 (16.5)	0.14 (9.3)
Self-talk	12 (2.0)	0.03 (2.0)
Verbalize negative affection	197 (32.8)	0.29 (19.1)
Scolding	22 (3.6)	0.04 (2.6)
Withdrawing	82 (13.6)	0.15 (9.8)
Speak louder	41 (6.8)	0.08 (5.2)
Neutral b
Shaking head	112 (21.7)	0.20 (22.5)
Express negative affection nonverbally	236 (45.7)	0.38 (42.7)
Attention touch a	0 (0.0)	0.0 (0.0)
Hand gestures: illustration/demonstration	51 (9.9)	0.09 (10.1)
Hand gestures: showing an object	30 (5.8)	0.05 (5.6)
Hand gestures: pointing	72 (14.0)	0.14 (15.7)
Perform activity with full assistance	15 (2.9)	0.03 (3.4)

Abbreviation: CG: caregiver.

^aNot observed and excluded in the analysis.

^bThis category includes only nonverbal items because none of the verbal items fit this category.

Two coders coded the videos from November 6, 2020 to January 7, 2021, using the DCODE with a behavior data analysis program, Noldus The Observer^® 14.0 (Noldus Information Technology Inc, Leesburg, Virginia). A detailed video coding process was described elsewhere. ⁸ Two coders underwent a 2-week training period to learn coding using the coder manual. They had 3 weekly meetings to compare coding results of the videos and address any ambiguities using coding notes. The coding scheme and procedure were updated during meetings. Due to the large number of items and the second-by-second coding procedure, the coders initially coded caregiver items, then care recipient items through a second pass on each video to accurately capture each person’s behavior. Each interaction in each video was coded second by second. Each utterance was coded as a verbal behavior unit. Each observed nonverbal behavior was coded as a nonverbal unit. Due to the varied length of each video, the sum of the relative frequency of each item was used to summarize the communication patterns. Seven caregiver and 4 care recipient nonverbal behavior items (see Tables 1 and 2) were coded for the duration of the observed time.

Data Analysis

All statistical analysis was performed in IBM SPSS Statistics 27.0 software (SPSS Inc, Chicago, IL). The level of significance was p = .05. Mean, standard deviation, and percentage were used to summarize participant characteristics. All characteristics were dichotomized for group difference analysis, because this study had a small sample size and continuous data representing participant characteristics were non-normally distributed. This approach ensured that an adequate number of videos were available in each group, allowing for meaningful comparisons of communication patterns based on the characteristics of caregiver and care recipients. All caregiver-reported characteristics (ie, burden, depression, sense of competence) were dichotomized based on the cut-off value of each instrument. The MZBS, CES-D, and SSCQ total scores were binarily categorized based on the cut-off value for MZBS (0 = no subjective burden and 1 = subjective burden), the cut-off value for CES-D (0 = no clinical depression and 1 = clinical depression), and the median value for SSCQ (0 = low competence, 1 = high competence), respectively.^11-13,15

The relative frequency of each item was calculated as the total frequency of each item divided by the total length of actual interaction time of each video. The relative duration was calculated as the total duration of each duration item (ie, eye gaze and adjustment of the environment) by the total length of actual interaction time of each video. The relative frequency of caregiver and care recipient patterns (ie, the sum of the relative frequency of all items in each pattern) was not normally distributed based on the Shapiro-Wilk test. ¹⁶ Ranked analysis of covariance (ANCOVA) was used to examine the association between participant characteristics and communication patterns using the residual maximum likelihood estimation after controlling for the dyad-level clustering effects as a covariate. ¹⁷

Results

Characteristics of dyads are shown in Table 3. Videos were recorded from caregivers who were most often care recipients’ spouses (n = 50, 66.7%), 61 years or older (n = 52, 69.3%), educated college or higher (n = 57, 68.4%), more burdened (n = 61, 81.3%), and less depressed (n = 46, 61.3%). Care recipients in most videos were 76 years or older (n = 46, 61.3%); were diagnosed with dementia for less than 5 years (n = 47, 62.7%), and had moderate severe/severe stage dementia (n = 63, 84.0%). The caregiver-care recipient dyad in 61 videos had different gender (81.3%).

Table 3.

Characteristics of Care Recipients and Caregivers (N = 38).

	Care recipient (n = 19)		Caregiver (n = 19)
	n (%)	Mean (SD)	n (%)	Mean (SD)
Age (years)		76.35 (6.61)		63.45 (11.75)
Gender
Female	10 (52.6)		12 (63.2)
Male	9 (47.4)		7 (36.8)
Relationship
Spouse			11 (57.9)
Daughter			4 (21.0)
Son			3 (15.8)
Daughter in law			1 (5.3)
Education
Less than college	12 (63.2)		6 (31.6)
College	3 (15.8)		9 (47.4)
Higher than college	4 (21.0)		4 (21.0)
Type of dementia a
Alzheimer’s	10 (52.6)
Vascular	1 (5.3)
Lewy body	2 (10.6)
Parkinsonism	1 (5.3)
Unknown	6 (31.6)
Dementia diagnosis length (year)		4.72 (3.31)
Caregiving length (year)				3.67 (3.31)
FAST score
Mild	3 (15.8)
Moderate	1 (5.3)
Moderate severe	14 (73.6)
Severe	1 (5.3)
SSCQ score b
Low (< 24)			8 (44.4)
High (≥ 24)			10 (55.6)
MZBI score b
Low (< 16)			2 (11.1)
High (≥ 16)			16 (88.9)
CES-D score b
Low (< 16)			12 (66.7)
High (≥ 16)			6 (33.3)

Abbreviations: FAST: Functional Assessment STaging; SSCQ: Short Sense of Competence Questionnaire; MZBI: Modified Zarit Burden Inventory; CES-D: Center for Epidemiologic Studies Depression Scale.

^aSome subjects have more than 1 type of dementia, the rate is over 100% due to rounding.

^bN = 18 due to 1 missing value.

Table 1 describes caregiver communication behaviors. A total of 8311 caregiver behaviors were counted: facilitative nonverbal = 1293 (15.6%), facilitative verbal = 2978 (35.8%), disabling nonverbal = 763 (9.2%), disabling verbal = 1044 (12.6%), neutral nonverbal = 1608 (19.3%), and neutral verbal = 625 (7.5%). The sum of relative frequency for all caregiver behaviors was 18.44 (frequency/min): facilitative nonverbal = 4.78 (25.9%), facilitative verbal = 6.08 (33.0%), disabling nonverbal = 1.87 (10.1%), disabling verbal = 1.60 (8.7%), neutral nonverbal = 2.87 (15.6%), and neutral verbal = 1.24 (6.7%). The longest duration of observed behavior was “minimize or eliminate distractions” (mean = 0.71 minutes/min).

Table 2 describes care recipient communication behaviors. Care recipient behaviors totaled 8024: engaging nonverbal = 2381 (29.7%), engaging verbal = 3561 (44.4%), challenging nonverbal = 553 (6.9%), challenging verbal = 1013 (12.6%), and neutral nonverbal = 516 (6.4%). The sum of relative frequency of all care recipient behaviors was 12.22: engaging nonverbal = 4.88 (39.9%), engaging verbal = 3.61 (29.5%), challenging nonverbal = 1.19 (9.8%), challenging verbal = 1.65 (13.5%), and neutral nonverbal = 0.89 (7.3%). The longest duration of observed behavior was “eye focused on the activity” (mean = 0.19 minutes/min).

The bivariate associations between participant characteristics and communication patterns are shown in Tables 4 and 5. There were no statistically significant differences in communication by dyadic relationship (spouse vs adult children), caregiver education (less than college vs college and higher), caregiving length (< 4 years vs ≥ 4 years), receiving the communication-related intervention (pre-intervention vs post-intervention), care recipient age (< 76 vs ≥ 76 years), or stage of dementia (mild/moderate vs moderate-severe/severe).

Table 4.

Relationships Between CG Communication Patterns and Characteristics (N = 75 Videos).

	CG facilitative			CG disabling			CG neutral
	Mean (SD)	F	p	Mean (SD)	F	p	Mean (SD)	F	p
CG gender
Female (n = 42)	7.86 (4.64)	5.090	.027	3.42 (2.31)	2.285	.135	3.36 (2.33)	2.901	.093
Male (n = 33)	10.28 (5.00)			3.55 (3.60)			5.02 (3.48)
Relationship with CR
Spouse (n = 50)	8.98 (4.82)	0.124	.726	3.66 (3.32)	1.073	.304	4.03 (3.14)	1.796	.184
Adult children (n = 25)	8.81 (5.21)			3.13 (1.90)			4.21 (2.74)
CG age
< 61 years (n = 23)	9.11 (5.31)	0.002	.961	3.30 (1.88)	3.129	.081	4.28 (2.84)	2.191	.143
≥ 61 years (n = 52)	8.84 (4.79)			3.56 (3.30)			4.00 (3.08)
CG education
Less than college (n = 18)	8.77 (4.94)	0.007	.933	3.93 (3.08)	0.045	.833	3.71 (2.92)	1.413	.238
College and higher (n = 57)	8.97 (4.95)			3.34 (2.89)			4.21 (3.03)
Caregiving length
< 4 years (n = 36)	9.44 (4.64)	1.696	.197	4.28 (3.45)	0.033	.857	5.15 (3.22)	2.733	.103
≥ 4 years (n = 39)	8.45 (5.17)			2.74 (2.12)			3.11 (2.41)
Communication intervention
Pre-intervention (n = 41)	8.44 (4.44)	0.751	.389	3.55 (2.97)	0.107	.744	4.05 (3.29)	0.863	.356
Post-intervention (n = 34)	9.52 (5.46)			3.39 (2.91)			4.14 (2.63)
CG competence score a
Low (< 24, n = 40)	8.48 (5.40)	2.340	.131	2.17 (1.82)	10.254	.002	3.00 (2.07)	3.264	.075
High (≥ 24, n = 33)	9.39 (4.13)			5.18 (3.20)			5.45 (3.46)
CG burden score a
Low (< 16, n = 12)	9.11 (2.92)	0.785	.379	7.45 (3.42)	5.040	.028	7.90 (3.27)	6.395	.014
High (≥ 16, n = 61)	8.85 (5.17)			2.76 (2.13)			3.36 (2.36)
CG depression score a
Low (< 16, n = 46)	9.47 (4.31)	4.124	.046	3.83 (3.25)	1.118	.294	4.72 (3.13)	2.004	.161
High (≥ 16, n = 27)	7.90 (5.62)			3.03 (2.28)			3.06 (2.55)
CR gender
Female (n = 41)	10.00 (4.80)	5.531	.021	3.36 (3.13)	2.000	.162	4.76 (3.22)	3.867	.053
Male (n = 34)	7.63 (4.81)			3.63 (2.69)			3.28 (2.51)
CR age
< 76 years (n = 29)	8.95 (4.76)	0.068	.796	4.66 (3.68)	2.627	.109	5.03 (3.69)	1.562	.215
≥ 76 years (n = 46)	8.91 (5.07)			2.74 (2.04)			3.50 (2.31)
CR education
Less than college (n = 45)	8.00 (4.28)	2.554	.114	4.50 (3.08)	5.484	.022	4.68 (3.24)	0.313	.578
College and higher (n = 30)	10.32 (5.53)			1.95 (1.85)			3.21 (2.37)
Dementia diagnosis length
< 5 years (n = 47)	8.34 (4.90)	2.216	.141	4.10 (2.98)	20.658	<.001	4.33 (3.34)	1.002	.320
≥ 5 years (n = 28)	9.90 (4.88)			2.43 (2.54)			3.68 (2.28)
Type of dementia
Alzheimer’s (n = 44)	7.73 (4.63)	7.617	.001	3.91 (3.20)	2.088	.153	4.16 (3.38)	0.415	.521
Others (n = 31)	10.62 (4.89)			2.86 (2.40)			3.98 (2.38)
Stage of dementia
Mild/moderate (n = 12)	9.96 (3.60)	2.169	.145	4.29 (3.25)	0.016	.900	3.84 (2.12)	0.085	.772
Moderate severe/severe (n = 63)	8.73 (5.13)			3.33 (2.86)			4.14 (3.14)
CG and CR gender difference
Identical (n = 14)	9.21 (5.55)	0.003	.959	3.95 (1.98)	7.335	.008	4.38 (2.86)	2.132	.149
Non-identical (n = 61)	8.86 (4.81)			3.37 (3.10)			4.02 (3.04)
Type of communication
Task-related (n = 40)	8.74 (5.25)	0.726	.397	2.52 (2.26)	7.497	.008	3.36 (2.60)	3.553	.063
Social informal (n = 35)	9.17 (4.58)			4.58 (3.23)			4.93 (3.22)

Mean: relative frequency, total frequency per minute divided by total length of video; used ranked ANCOVA test controlling dyad.

Abbreviations: CG: Caregiver; CR: care recipient.

^aN = 73 due to the missing value.

Table 5.

Relationships Between CR Communication Patterns and Characteristics (N = 75 Videos).

	CR engaging			CR challenging			CR neutral nonverbal
	Mean (SD)	F	p	Mean (SD)	F	p	Mean (SD)	F	p
CG gender
Female (n = 42)	8.06 (4.58)	0.046	.832	2.15 (1.64)	0.198	.658	0.90 (0.92)	0.136	.713
Male (n = 33)	8.28 (5.53)			2.26 (2.05)			0.88 (0.99)
Relationship with CR
Spouse (n = 50)	7.94 (4.05)	0.028	.867	2.12 (1.87)	2.225	.140	0.79 (0.89)	3.435	.068
Adult children (n = 25)	8.61 (6.55)			2.36 (1.74)			1.09 (1.03)
CG age
< 61 years (n = 23)	9.09 (6.61)	0.733	.395	2.37 (1.79)	2.270	.136	1.11 (1.07)	3.395	.069
≥ 61 years (n = 52)	7.75 (4.09)			2.12 (1.85)			0.79 (0.87)
CG education
Less than college (n = 18)	7.88 (6.81)	1.430	.236	2.65 (2.04)	0.562	.456	0.98 (0.99)	0.019	.890
College and higher (n = 57)	8.25 (4.33)			2.06 (1.74)			0.64 (0.93)
Caregiving length
< 4 years (n = 36)	8.73 (3.95)	1.432	.235	2.58 (1.62)	1.281	.261	1.02 (0.95)	0.118	.732
≥ 4 years (n = 39)	7.63 (5.78)			1.85 (1.94)			0.77 (0.93)
Communication intervention
Pre-intervention (n = 41)	7.62 (5.29)	2.354	.129	2.25 (2.16)	0.622	.433	1.02 (0.95)	2.338	.131
Post-intervention (n = 34)	8.81 (4.59)			2.14 (1.33)			0.73 (0.91)
CG competence score a
Low (< 24, n = 40)	7.10 (4.14)	2.472	.120	1.50 (1.30)	9.518	.003	0.64 (0.79)	5.061	.028
High (≥ 24, n = 33)	9.54 (5.68)			3.11 (2.01)			1.24 (1.03)
CG burden score a
Low (< 16, n = 12)	9.66 (2.98)	1.629	.206	3.02 (1.19)	1.179	.281	1.39 (1.02)	1.006	.319
High (≥ 16, n = 61)	7.92 (5.29)			2.07 (1.91)			0.82 (0.91)
CG depression score a
Low (< 16, n = 46)	7.85 (3.96)	0.254	.616	2.44 (1.96)	0.025	.874	0.86 (0.85)	0.738	.393
High (≥ 16, n = 27)	8.81 (6.48)			1.87 (1.57)			0.99 (1.09)
CR gender
Female (n = 41)	8.09 (4.15)	0.117	.733	2.34 (1.99)	0.127	.723	0.87 (0.94)	0.085	.771
Male (n = 34)	8.24 (5.91)			2.03 (1.60)			0.92 (0.95)
CR age
< 76 years (n = 29)	8.57 (5.67)	0.137	.712	2.42 (1.62)	1.547	.218	1.08 (1.06)	0.584	.447
≥ 76 years (n = 46)	7.90 (4.55)			2.06 (1.94)			0.77 (0.84)
CR education
Less than college (n = 45)	8.55 (5.54)	0.002	.968	2.40 (1.49)	1.066	.305	1.13 (1.03)	3.680	.059
College and higher (n = 30)	7.58 (4.03)			1.90 (2.22)			0.53 (0.64)
Dementia diagnosis length
< 5 years (n = 47)	8.93 (5.28)	4.596	.035	2.42 (1.89)	6.073	.016	1.07 (0.98)	7.551	.008
≥ 5 years (n = 28)	6.86 (4.22)			1.82 (1.67)			0.60 (0.80)
Type of dementia
Alzheimer’s (n = 44)	7.96 (4.30)	0.000	.994	1.93 (1.44)	1.282	.261	0.88 (0.90)	0.005	.946
Others (n = 31)	8.43 (5.89)			2.57 (2.22)			0.91 (1.01)
Stage of dementia
Mild/moderate (n = 12)	9.16 (3.72)	1.318	.255	2.94 (1.34)	3.389	.070	0.87 (0.86)	0.098	.755
Moderate severe/severe (n = 63)	7.97 (5.19)			2.06 (1.87)			0.89 (0.96)
CG and CR gender difference
Identical (n = 14)	10.19 (7.45)	1.901	.172	3.19 (1.75)	15.327	<.001	1.48 (1.17)	7.777	.007
Non-identical (n = 61)	7.69 (4.18)			1.97 (1.77)			0.76 (0.83)
Type of communication
Task-related (n = 40)	6.96 (3.93)	3.877	.053	1.52 (1.19)	10.354	.002	0.76 (0.82)	0.329	.568
Social informal (n = 35)	9.82 (5.73)			2.97 (2.11)			1.04 (1.05)

Mean: relative frequency, total frequency per minute divided by total length of video; used ranked ANCOVA test controlling dyad.

Abbreviations: CR: care recipient; CG: caregiver.

^aN = 73 due to the missing value.

Group Differences in Communication Patterns

Male caregivers exhibited more facilitative communication than female caregivers (F = 5.090, p = .027). Caregivers of female care recipients exhibited more facilitative communication than those of male care recipients (F = 5.531, p = .021). Caregivers of care recipients with other types of dementia exhibited more facilitative communication than those of care recipients with Alzheimer’s disease (F = 7.617, p = .001). Caregivers with lower depression scores exhibited more facilitative communication than those with higher depression score (F = 4.124, p = .046).

Caregiver competence level, burden level, care recipient education level, dementia diagnosis length, gender difference, and type of communication were also associated with communication patterns. Caregivers with higher competence scores exhibited more disabling communication (F = 10.254, p = .002) and their care recipients exhibited more challenging (F = 9.518, p = .003) and neutral (F = 5.061, p = .028) communication than the lower competence score group. Caregivers with lower burden scores exhibited more disabling (F = 5.040, p =.028) and neutral (F = 6.395, p = .014) communication than those with higher burden scores. Caregivers of care recipients with lower education levels exhibited more disabling communication (F = 5.484, p = .022). Caregivers of care recipients with shorter dementia diagnosis length exhibited more disabling communication (F = 20.658, p < .001) and their care recipients exhibited more engaging (F = 4.596, p = .035), challenging (F = 6.073, p = .016), and neutral (F = 7.551, p = .008) communication than the longer dementia diagnosis length group. Caregivers in same-gender dyads exhibited more disabling communication (F = 7.335, p = .008) and their care recipients exhibited more challenging (F = 15.327, p < .001) and neutral (F = 7.777, p = .007) communication than those in different-gender dyads. Finally, caregivers exhibited more disabling communication (F = 7.497, p = .008) and care recipients exhibited more challenging communication (F = 10.354, p = .002) during social communication than task-related communication.

Discussion

Understanding caregiver-care recipient communication patterns and the relationships with individual characteristics are critical in developing individualized interventions for family caregivers and persons living with dementia. This exploratory study examined communication patterns in dementia care interactions by individual and dyadic characteristics.

Dementia Diagnosis Length

The dementia diagnosis length was associated with communication patterns of care recipients and caregivers. Caregivers of care recipients diagnosed with dementia for less than 5 years used disabling communication more frequently, and their care recipients used engaging, challenging, and neutral communication more frequently. There are 2 possible reasons for these dyadic communication patterns. First, during the early years of a dementia diagnosis, family caregivers would also be in their early years of caregiving for care recipients and had limited knowledge and skills to accommodate the changed communication abilities of care recipients, such as slight difficulty with object naming, conversational coherence, and complex language comprehension due to limited attention and executive function. ³ Therefore, caregivers may be more likely to exhibit disabling communication unintentionally, which may negatively affect the engagement of the care recipient in communication. ¹⁸ In this situation, care recipients may signal the need for re-engaging in communication, ¹⁸ which is indicated through more use of all types of communication.

Another probable reason is the association between caregiver appraisal of effective communication strategies and challenging communication of persons living with dementia. For example, Savundranayagam and Orange ³ found that persons living with dementia exhibit more challenging communication when caregivers incorrectly appraised the effectiveness of communication strategies on engagement in communication. Caregivers may be unaware of their communication patterns in their early years of caregiving, which can affect the subsequent behaviors of persons living with dementia.

The findings provide healthcare providers and researchers insights in delivering communication interventions for family caregivers. Interventions need to be initiated at the earliest time after dementia diagnosis to increase effective communication by caregivers to accommodate their communication to the needs of the persons living with dementia. Caregivers may need to improve their awareness and accommodate their communication, including capturing a specific signal from persons living with dementia to maintain adequate and engaging communication.

Gender and Dyadic Relationship

We found that caregivers exhibited more disabling and care recipients exhibited more challenging and neutral communication in the same-gender dyads than different-gender dyads. In addition, male caregivers and caregivers of female care recipients used more facilitative communication than female caregivers and caregivers of male care recipients. All caregivers in the same-gender dyads were adult children, among which 80% were mother-daughter and daughter-in-law dyads in this study.

Parent-child communication can be different from spousal communication. A prior study reported that adult-children caregivers experience more negative effects on their health, family life, and work than spouse caregivers. ¹⁹ Specifically, for adult children, the obligation and duty to care for both their parents and their minor children may be more prevalent motives associated with more stress in care.^20,21

On the contrary, spouse caregivers experience more intrinsic motivation to care, such as love or empathy for their spouse living with dementia. ²¹ In addition, prior studies revealed that daughter caregivers put substantial effort into rebuilding a relationship with their mothers living with dementia ²² and use various coping strategies to protect themselves from caregiving distress. ²¹ These motivations and negative experiences in care are positively associated with the caregiver’s disabling communication ² and the care recipient’s challenging and neutral communication. ⁶

We also found that caregivers of female care recipients exhibited more facilitative communication and marginally more neutral communication than caregivers of the male care recipient. There are inconsistent findings regarding caregiver patterns of positive and negative communication toward male and female care recipients. ⁶ Female spouse caregivers exhibited less positive and more negative communication in one study ⁶ but less positive and less negative communication in another study. ⁷ These findings suggest the need for a clear conceptualization and categorization of communication patterns considering the context of communication, such as characteristics of antecedent communication behaviors, ²³ relationships with persons living with dementia, or care recipients’ communication abilities and status. ²⁴ Neutral communication was defined as either facilitative or disabling depending on the context of the communication in this study. Further research is needed to categorize neutral communication and examine its relationship with gender.

Types of Communication

During social communication, caregivers and care recipients in this study used more disabling and challenging communication than task-related communication. In general, current evidence has shown that persons living with dementia respond negatively to negative caregiver communication behaviors (eg, harsh communication) and positively to positive communication behaviors (eg, soft tone of voice). ²⁵ Gallagher-Thompson et al ⁷ found that family caregivers and persons living with dementia used more facilitative communication during planning tasks than at mealtime. However, prior research is limited to investigating either type of communication, task-related or social communication only (eg, Braun et al ⁶ ), and the relationships between the types of communication and communication patterns have not been extensively examined. In this study, there was no significant relationship between types of communication and characteristics of caregivers and care recipients. The possible explanation for this finding may be related to different characteristics of social and task-related communication. Task-related communication included the interactions during ADL-related activities (eg, mealtime, cleaning, dressing) and activities and games in this study, which may have specific procedures or steps that can be a cue for caregivers to facilitate communication. In addition, task-related communication patterns are relatively structured, such as the use of a series of directives by caregivers and the low frequency of care recipient verbal communication. ²⁶ Although it was not significant, both caregivers and care recipients in this study exhibited more frequent communication behaviors in social communication than in task-related communication. These findings may indicate that dyads were engaged in social communication without specific cues or structured patterns, increasing the opportunity to exhibit more disabling and challenging communication. Since dementia family dyads communicate in many different contexts, such as social informal/task-related or supportive/discouraging communication, investigating diverse situations is critical to understanding dementia family communication.

Caregiver Depression

In this study, caregivers with low depression were more likely to exhibit facilitative communication than those with high depression. This finding is consistent with previous reports that depression is associated with more negative and neutral communication and less use of facilitative communication^2,6 among family caregiver-care recipient dyads. Since prior research found that communication of persons living with dementia is affected by how supportive the caregiver is, by more use of facilitative communication, ²⁷ future research focusing on the causal relationship between depression and facilitative communication on care recipient communication patterns is warranted.

Caregiving Burden and Competence in Care of Burden

Caregivers with high self-reported competence in dealing with burden were more likely to exhibit disabling communication, and their care recipients exhibited more challenging and neutral communication than those caregivers with low competence. In addition, caregivers with low burden used more disabling and neutral communication than those with high burden. These 2 findings are inconsistent with previous reports that a greater burden is associated with more frequent caregiver negative communication ² and problematic communication by persons living with dementia, ²⁸ and that caregivers with higher competence in dealing with burden exhibit less caregiving burden. ²⁹

A possible reason for these inconsistent findings may be related to different operational definitions of burden and communication patterns between previous studies and this study. In this study, caregiver negative (ie, disabling) communication included both verbal and nonverbal communication, such as no eye contact, walking away from the care recipient, checking memory, and correcting. In contrast, one previous study included only verbal communication (ie, verbal mistreatment by caregivers), such as screaming, yelling, criticizing, scolding, and threatening, which are harsher types of communication. ² In addition, a self-report questionnaire was used to assess use of harsh communication, and this study used video observations that may allow for more in-depths and accurate assessment of dyadic communication.

Caregiver burden is often defined as objective and subjective burden. From the caregivers’ perspective, communication problems can be perceived as care recipients’ intentional disruptive behaviors that significantly affect relationship quality between caregivers and care recipients. ²⁸ This negative perception by caregivers is related to higher subjective burden, such as emotional and social burden. ²⁸ On the contrary, objective burden (eg, physical burden) is not directly related to communication problems, rather overall problematic behaviors, such as aggression. ²⁸ The MZBS assesses both objective and subjective burden, therefore, the findings may not be consistent with previous studies.

Another possible explanation could be the nature of the self-reported assessments used to assess competence in dealing with burden. Caregivers may overestimate their competence in dealing with burden due to factors, such as a desire to present themselves in a positive light and appear capable and competent in their caregiving role. This can lead to an overestimated perception of their abilities, resulting in a mismatch between their perceived competence and their actual skill level. Consequently, it can potentially lead to use of more negative communication patterns.

It is important to note that self-reported competence may not always align with actual communication behaviors. To address this, caregiver education and support programs should emphasize the importance of continuous training and provide caregivers with evidence-based strategies for effective communication in dementia care. Continuous training and support for caregivers also can help caregivers to have a more accurate understanding of their communication abilities and address any gaps in their communication skills.

Limitations

This study has several limitations. Videos were collected from the intervention study to support family caregivers of persons living with dementia with various social informal and task-related communication situations. Thus, the generalization of results may be limited to dementia dyads in home care settings and the same contexts of the communication rather than to other diverse dyad populations in other care settings. In addition, the videos were recorded in a way that caregivers were aware of the recording, thus the recorded observation may be less natural than purely naturally occurring interactions; caregivers may have altered their communication during the recording. The use of non-normally distributed data was another limitation that might affect the interpretation of data. To prevent this, we used ranked ANCOVA with log-transformed data over other non-parametric methods (eg, Mann-Whitney test), which allows non-normally distributed data toward a normal distribution and the estimate unbiased. ¹⁷ Finally, because this study used secondary data, some variables that were associated with communication patterns (eg, relationship quality) based on prior research were not collected in the parent study and were not available for analysis. Although we screened the possible confounding variables, including communication-related intervention, there is a possibility that caregiver communication was affected by the intervention procedure.

Implications and Future Directions

Understanding communication patterns and related factors are essential to promoting a meaningful and satisfactory relationship between caregivers and care recipients in dementia care. The findings from this study can have important implications for care and treatment in dementia. Caregivers can benefit from understanding communication patterns and related factors to improve their interactions with persons living with dementia. Based on these findings, caregivers should consider incorporating effective communication strategies into their caregiving approach. Educational programs specifically designed for caregivers can be developed, providing them with necessary knowledge and skills to enhance communication with persons living with dementia. Policies can be formulated in primary care, long-term care, and community settings to support caregiver training programs focused on effective communication strategies. By integrating these programs into existing care systems, caregivers can receive ongoing support and guidance in improving their communication skills.

Future research can further investigate interaction effects of individual characteristics, such as relationship quality, on communication patterns in dementia care. Exploring these factors in more depth will provide a better understanding of how these factors influence communication dynamics. To enhance generalizability, conducting research with larger and more diverse samples is crucial, allowing for conclusions applicable to a broader population and providing stronger evidence for care and treatment practices. In addition, sequential analysis can be employed to understand reciprocal relationships between caregiver and care recipient communication behaviors, offering valuable insights into the dynamics of interactions. This knowledge can contribute to the development of targeted interventions to address the specific needs and challenges of caregivers and persons living with dementia.