Comprehensive Older Adult and Caregiver Help (COACH): A person-centered caregiver intervention prevents elder mistreatment

STRUCTURED ABSTRACT

BACKGROUND:

Elder mistreatment (EM) harms individuals, families, communities, and society as a whole. Yet research on interventions is lagging, and no rigorous studies demonstrating effective prevention have been published. This pilot study examines whether a first-of-its-kind coaching intervention reduced the experience of EM among older adults with chronic health conditions, including dementia.

METHODS:

We used a double-blind, randomized controlled trial to test a strengths-based person-centered caregiver support intervention, developed from evidence-based approaches used in other types of family violence. Participants (n=80), family caregivers of older adults who were members of Kaiser Permanente, completed surveys at baseline, post-test, and 3-month follow-up. The primary outcome was caregiver-reported EM; additional proximal outcomes were caregiver burden, quality-of-life, anxiety, and depression. Nonparametric tests (Mann-Whitney U, Fisher’s Exact, Wilcoxon Signed Rank, and McNemar’s) were used to make comparisons between treatment and control groups and across time points.

RESULTS:

The treatment group had no EM after intervention completion (assessed at 3-month follow-up), a significantly lower rate than the control group (treatment=0%, control=23.1%, p=0.010).

CONCLUSIONS:

In this pilot study, we found that the COACH caregiver support intervention successfully reduced EM of persons living with chronic illness, including dementia. Next steps will include: 1) testing the intervention’s mechanism in a fully powered RCT and 2) scaling the intervention for testing in a variety of care delivery systems.

Elder mistreatment (EM) takes a devastating toll on the independence, health, and well-being of millions of older adults each year. Defined as “an intentional act or failure to act by a caregiver or another person in a relationship involving an expectation of trust that causes or creates a risk of harm to an older adult,”¹ EM includes emotional, physical, and sexual abuse, neglect, and financial exploitation. Over 10% of community-dwelling older Americans experience EM annually,^2,3 with rates over twice that high for individuals with physical or cognitive impairment.^4,5 EM may result in physical injuries,⁶ emotional and physical pain,⁷ eroded trust, financial devastation,⁸ exacerbated illness, functional decline,^3,9 and increased risk of institutionalization, hospitalization,^10,11,12 and death. Although EM harms individuals, families, communities, and society as a whole, research on interventions is limited,^{2,13,14,15,16} and no rigorous intervention studies demonstrating effective prevention have been published.^17,18

The well-being and health of older adults who need assistance due to physical, functional, or cognitive impairment depends, in part, on family caregivers, who provide more than 80% of their care.^19,20 Caregivers play a role in protecting vulnerable older adults from mistreatment, particularly in situations involving risk of financial exploitation.^21,22 Often, family caregivers undertake their caregiving responsibilities with little education or training to prepare them for or support them in carrying out an often challenging role. Although most caregivers do their best, needing care increases the risk of experiencing EM. For example, studies have found up to one quarter of caregivers to engage in physical abuse²³ or potentially harmful behavior, like withholding food from the care recipient.²⁴ Risk for EM is especially high among those caring for older adults living with dementia,^5,23 with rates as high as 47%.⁵ However, studies suggest several potential avenues for risk reduction and prevention of EM from supportive strategies, such as educating caregivers^15,25,26 and older adults,^15,26 addressing caregiver anxiety and depression,²⁷ and comprehensive care planning and support.^25,26

Evidence on risk factors and effective prevention from other fields of family violence—specifically child maltreatment and intimate partner violence (IPV)—offers potential lessons for EM prevention.²⁸ As an initial step in developing an EM prevention intervention, we identified risk factors that EM shares with child maltreatment and IPV. Risk factors for the victim and/or perpetrator include isolation, previous trauma, dependency, inability to cope with aggressive behaviors, lack of support, dysfunctional attitudes/lack of respect (e.g., ageism, sexism), and family stress, burden, depression, and anxiety.²⁹ We then identified characteristics of prevention interventions in child maltreatment and IPV that offered promise to prevent EM. These include supportive home visiting programs that provide a tailored, proactive, multicomponent approach with regular visits conducted for up to 12 sessions and strengths-based “check-up” tools. Among these tools are motivational interviewing (a communication skill that meets the individual’s perception of the problem and ambivalence about desired changes and solutions), regular self-evaluation, social support, modeled behavior, and attitude reframing.²⁹

Drawing on these learnings from the EM field and the diverse literature on evidence-informed prevention interventions in family violence, the “Comprehensive Older Adult and Caregiver Help” (COACH) intervention was developed for caregivers of older adults with care needs due to chronic illness, including dementia. COACH was developed by clinical and non-clinical academics, caregivers, and caregiver stakeholders and in collaboration with geriatricians and clinical staff (geriatrics, palliative, family, and internal medicine physicians, nurses, and social workers) at a large urban medical center. The intention of COACH is to improve caregivers’ ability to provide high-quality care and mitigate negative psychosocial impacts of caregiving that might lead them to engage in abusive behavior. Further, it aims to bolster caregivers’ role as “capable guardians”^21,22 who are vigilant against EM being perpetrated by other parties. Details of the intervention are provided in Supplemental Table S1, organized using the TIDieR framework.³⁰

This paper introduces the COACH intervention and reports results from a pilot study that rigorously evaluated this evidence-informed intervention, assessing the program’s impact on its primary outcome, preventing EM, and four proximal outcomes: reducing caregiver depression, anxiety, and burden, and increasing caregiver quality of life (QoL).

METHODS

Study Design

In this pilot study, we conducted a double-blind randomized controlled trial of the COACH intervention compared to an information-only control condition to determine the intervention’s preliminary efficacy. Patients and their caregivers were recruited through Kaiser Permanente Southern California’s Los Angeles Medical Center (Kaiser LAMC), a managed care system that houses a geriatrics program and specialized memory clinic and annually serves over 235,000 people aged 65 or older. Survey administration and intervention delivery were through the University of Southern California (USC). The study was approved by both institutions’ Institutional Review Boards.

The Intervention

COACH is a multicomponent, person-centered, strengths-based education and social support intervention for caregivers. It was administered by bachelor’s-level “care coaches” who received training in topics and approaches including motivational interviewing and active listening (full list in Supplemental Table S2). The training was designed to be administered during one 40-hour work week. Two of the four coaches were bilingual Spanish speakers; all intervention materials were developed in both English and Spanish.

Coaches met with caregivers weekly for up to 12 sessions to listen to their concerns and guide them through a personally tailored behavioral and educational intervention, providing caregiving tools and coping strategies drawn from a Caregiver Toolkit and corresponding Caregiver Workbook. Originally designed to be in-person, the program transitioned to telephone in response to the COVID-19 pandemic. During their first session, the coach and caregiver focused on strengths and potential areas for growth, drawing from the caregiver’s baseline survey. Based on this discussion, the coach and caregiver developed a personalized plan of topics to be covered during the 3–12 weekly sessions, which could be updated as new needs or priorities emerged. Details of the intervention, materials, and coaching topics are provided in Supplemental Tables S1-S5). The Caregiver Toolkit was made available to participants in both groups, allowing us to test the impact of the interpersonal care coaching compared with simply providing information to the caregiver.

Sampling and Screening

Inclusion criteria for family caregiving dyads in this study were English- or Spanish-speaking, care recipients aged 65 and older needing assistance with ADLs or IADLs because of a physical, mental, emotional, or memory problem, and the family caregiver most responsible over the past month for care (including hands-on help with tasks, supervising the care recipient to ensure safety, providing and/or assisting with an ADL or IADLs). Exclusion criteria were care recipients enrolled in hospice, living in custodial nursing care facilities, or experiencing homelessness, and caregivers residing more than 50 miles away from the care recipient.

Participants were referred by Kaiser LAMC’s primary care and palliative care physicians, hospital discharge planners, and providers in the memory clinic, and screened by Kaiser Permanente Southern California’s research division. Those screened into the study were referred to the USC study team for further screening and formal consent.

Randomization and Blinding

Randomization to treatment or control was done through REDCap after recruitment. We used a 1:1 allocation scheme stratified by primary language (English or Spanish). The allocation sequence was created using a random number generator and was not shared with staff involved in recruitment or participant surveys. Participants and surveyors were blinded to treatment assignment.

Data Collection

Participants were asked to complete 3 surveys: a baseline pre-test survey, a post-test survey at the end of the intervention (treatment group) or at a timeframe randomly selected to mirror the intervention duration (control group), and a follow-up survey 3 months after the post-test. Each survey took 60–90 minutes and was administered by graduate student surveyors who were blinded to treatment condition. Participants were given the option to complete surveys by phone with a surveyor or self-administered either online or using a hard copy delivered to their home.

Measures

Best-in-class measures were selected for brevity and strong psychometric properties. Caregiver-reported EM was operationalized dichotomously using the Geriatric Mistreatment Scale (GMS), a 22-item instrument that measures the presence of all five types of EM and has been demonstrated to have good internal, face, construct, and convergent validity.³¹ The look-back period for EM in the baseline survey was the past year; for the post-test and 3-month follow-up surveys, the look-back period was since the participant’s last survey. As such, the post-test survey detected EM experienced during the intervention period, and the 3-month follow-up survey detected EM experienced after the completion of the intervention. Positive responses to any GMS subtype yielded a positive EM score. This study focused on participants’ EM status measured at post-test and 3-month follow-up, regardless of their EM status at baseline.

Despite initial plans to survey all care recipients who had capacity to consent, only 18 completed surveys at one or more time points. EM was determined using caregiver report, after a comparison with care recipient responses revealed differences in only one control group case, where the caregiver reported EM at baseline but the care recipient reported none.

Four proximal outcomes were included. Caregiver QoL was measured on four domains (physical, social, psychological, and environmental) using a modified WHO QoL-BREF.³² Caregiver depression was measured with the PHQ-9 (categorized into minimal, mild, moderate, moderately severe, or severe),³³ caregiver anxiety with the GAD-7 (categorized into minimal, mild, moderate, or severe),³⁴ and caregiver burden with the Zarit Burden Interview - Short Version.³⁵

Other baseline measures included caregiver and care recipient demographics (age, gender, race/ethnicity, education), relationship to each other, and living arrangement; the care recipient having a dementia diagnosis, as reported by the caregiver; the care recipient’s need for constant monitoring, as reported by the caregiver; and the caregiver’s exposure to adverse childhood experiences (ACEs).³⁶

Data Analysis

Randomized treatment and control groups were compared on demographics and other measures at baseline and on key outcome measures at pre-test baseline, post-test, and 3-month follow-up. Because of the relatively small sample and the possibility of cell sizes smaller than 5, we used nonparametric tests for bivariate comparisons: Mann-Whitney U tests for continuous variables and Fisher’s Exact test for categorical variables. We assessed within-person differences, comparing individuals’ scores at pre-test baseline, post-test, and 3-month follow-up using Wilcoxon Signed Rank test for continuous variables and McNemar’s test for categorical variables. However, as discussed below, this was not possible for all categorical variable comparisons, as McNemar’s test cannot produce results when only one category is endorsed at one time point. Although this paper simultaneously tests multiple outcomes, no adjustments were made for multiple comparisons due to the study being an exploratory pilot test.^37,38 All analyses were performed in SAS 9.4.

RESULTS

Demographic Characteristics

Caregivers had a mean age of 61.3; most (76.3%) were female. The sample was racially/ethnically diverse, with 28.2% White, 37.2% Hispanic/Latine, 18.0% Black/African American, 14.1% Asian/Pacific Islander. Educational attainment included 37.5% with some college or an associate degree, 25.0% with a bachelor’s degree, 21.3% with a graduate degree. Care recipients had an average age of 81.9; half (50.6%) were female. They were racially/ethnically diverse (24.1% White; 41.8% Hispanic/Latine, 20.3% Black/African American, 13.9% Asian/Pacific Islander), with lower educational attainment than their caregivers (e.g., 24.3% less than high school, 24.3% high school graduates). Most (73.8%) caregivers lived with the care recipient and had a family relationship (33.8% spouses/partners, 51.3% children/grandchildren, 10.0% other family). Over half (54.7%) of care recipients had a dementia diagnosis reported by their caregiver. Almost all caregivers reported that the care recipient needed constant monitoring always (66.3%) or sometimes (27.5%).

No significant differences were found between treatment and control groups on caregiver or care recipient sociodemographic characteristics, relationship type, co-habitation, caregiver ACE count, care recipient dementia, or care recipient need for constant monitoring (Table 1). There were also no significant differences between those who left the study prior to the 3-month follow-up and those who completed the full study (Supplemental Table S7). Only one variable, categorical caregiver age, showed a statistically significant difference (p=0.006) between those who dropped out of the study prior to the post-test and those who completed the post-test. However, there were no differences in continuous caregiver age, and the significant difference for categorical age did not persist to 3-month follow-up (Supplemental Table S7). At baseline, no statistically significant differences were found between treatment and control groups on the presence of EM or the proximal outcome measures of caregiver burden, caregiver depression, caregiver anxiety, or caregiver QoL (Table 2).

Table 1.

Sample characteristics, by treatment status

	Full Sample (n=80)	Control (n=40)	Treatment (n=40)
	M(SD) or %	M(SD) or %	M(SD) or %	p
Caregiver Characteristics
Age	61.3 (14.0)	59.8 (14.4)	62.7 (13.7)	0.127
<55	28.8	37.5	20.0	0.392
55–64	32.5	30.0	35.0
65–74	21.3	17.5	25.0
75+	17.5	15.0	20.0
Female (Gender)	76.3	80.0	72.5	0.600
Race/Ethnicity				0.597
White	28.2	22.5	34.2
Hispanic/Latine	37.2	45.0	29.0
Black/African American	18.0	15.0	21.1
Asian/Pacific Islander	14.1	15.0	13.2
Other	2.6	2.5	2.6
Education				0.166
Less than High School	7.5	2.5	12.5
High School Graduate	8.8	7.5	10.0
Some College or AA	37.5	42.5	32.5
Bachelor’s Degree	25.0	32.5	17.5
Graduate Degree	21.3	15.0	27.5
Screened in Spanish	12.5	12.5	12.5	1.000
Relationship to CR				0.924
Spouse/Partner	33.8	30.0	37.5
Child/Grandchild	51.3	55.0	47.5
Other Family	10.0	10.0	10.0
Other	5.0	5.0	5.0
CG Lives with CR	73.8	75.0	72.5	1.000
ACEs				0.172
0 ACEs	40.0	50.0	30.0
1 ACE	30.0	20.0	40.0
2–3 ACEs	11.3	12.5	10.0
4+ ACEs	18.8	17.5	20.0
Care Recipient Characteristics
Age	81.9 (8.2)	81.8 (8.3)	82.1 (8.2)	0.450
65–74	23.1	23.1	23.1	1.000
75–84	39.7	38.5	41.0
85+	37.2	38.5	35.9
Female (Gender)	50.6	47.5	53.9	0.655
Race/Ethnicity				0.964
White	24.1	25.0	23.1
Hispanic/Latine	41.8	42.5	41.0
Black/African American	20.3	17.5	23.1
Asian/Pacific Islander	13.9	15.0	12.8
Other	0.0	0.0	0.0
Education				0.971
Less than High School	24.3	21.6	27.0
High School Graduate	24.3	27.0	21.6
Some College or AA	17.6	18.9	16.2
Bachelor’s Degree	24.3	24.3	24.3
Graduate Degree	9.5	8.1	10.8
Screened in Spanish	13.8	15.0	12.5	1.000
Dementia diagnosis	54.7	64.9	44.7	0.106
Needs Constant Monitoring				1.000
No	6.3	7.5	5.0
Sometimes	27.5	27.5	27.5
Yes	66.3	65.0	67.5

Note: CG=caregiver, CR=care recipient, AA=Associate of Arts degree, ACEs=adverse childhood experiences. Data are missing for CG race/ethnicity (treatment n=2), CR age (control n=1, treatment n=1), CR gender (treatment n=1), & CR race/ethnicity (treatment n=1).

Table 2.

Sample outcomes, by treatment status

	Baseline			Post-test			3-month Follow-up
	Control (n= 40) M(SD) or %	Treatment (n=40) M(SD) or %	p	Control (n= 32) M(SD) or %	Treatment (n=32) M(SD) or %	p	Control (n= 28) M(SD) or %	Treatment (n=27) M(SD) or %	p
Primary outcome
Elder mistreatment	15.4	22.5	0.568	15.6	16.7	1.000	23.1	0	0.010
Proximal outcomes
CG burden	15.6 (9.4)	14.7 (9.5)	0.380	16.7 (9.1)	14.2 (9.5)	0.152	14.1 (9.8)	13.2 (9.0)	0.417
CG depression			0.616			0.645			0.176
Minimal	59.0	65.0		65.6	58.1		60.7	69.2
Mild	33.3	20.0		28.1	25.8		32.1	11.5
Moderate	5.1	10.0		6.3	9.7		7.1	15.4
Moderately severe	2.6	2.5		0.0	6.5		0.0	3.9
Severe	0.0	2.5		0.0	0.0		0.0	0.0
CG anxiety			0.431			0.902			0.413
Minimal	61.5	60.0		71.9	74.2		71.4	73.1
Mild	25.6	30.0		21.9	16.1		14.3	23.1
Moderate	10.3	2.5		3.1	6.5		14.3	3.9
Severe	2.6	7.5		3.1	3.2		0.0	0.0
CG QoL - physical	15.4 (2.2)	15.2 (2.0)	0.289	15.5 (2.3)	15.0 (2.4)	0.226	15.6 (2.4)	15.3 (2.3)	0.374
CG QoL - psychological	14.7 (3.1)	14.9 (2.6)	0.410	15.0 (3.5)	14.5 (3.0)	0.316	14.8 (3.0)	14.3 (3.0)	0.256
CG QoL - social	15.1 (3.9)	15.9 (3.0)	0.247	14.1 (4.3)	16.7 (2.2)	0.006	15.5 (3.6)	16.4 (2.7)	0.218
CG QoL - environmental	14.3 (1.6)	13.9 (1.5)	0.070	14.3 (1.4)	13.9 (1.9)	0.189	14.0 (1.9)	14.2 (1.7)	0.453

Note: **p<0.01. CG=caregiver, QoL=quality of life.

Retention and Intervention Completion

As shown in the CONSORT Diagram (Figure 1), 110 of the 636 referred dyads met initial screening criteria and were forwarded to the study. Of these, 92 met inclusion criteria, and 80 (87.0%) agreed to participate and completed baseline surveys. Of these 80 (40 treatment, 40 control), 64 (80.0%) completed a post-test survey (32 treatment, 32 control). 55 caregivers (68.8% of baseline participants) completed a 3-month follow-up survey (27 treatment, 28 control). Reasons for non-completion are presented in Figure 1. Most (87.5%) caregivers in the treatment group participated in three or more intervention sessions (Supplemental Table S6). There were no significant differences in baseline EM or proximal outcomes between those who completed and those who did not complete the post-test or 3-month surveys (Supplemental Table S7).

Figure 1:

CONSORT Diagram. Note: USC=University of Southern California, CR=care recipient.

Primary Outcomes

As shown in Figure 2, treatment group participants experienced an overall reduction in EM over time (see individual question responses in Supplemental Table S8). While there was no significant difference in the control group’s rate of EM across the three time points (15.4% baseline, 15.6% post-test, 23.1% 3-month follow-up; Supplemental Table S9), the rate of EM in the treatment group dropped from 22.5% at baseline to 16.7% at post-test (EM that occurred during intervention delivery) and 0.0% at 3-month follow-up (EM that occurred after intervention completion). Among treatment group participants who provided EM responses at both baseline and post-test (n=31), change across these two time points was not significant (23.3% baseline, 16.7% post-test, p=0.727; Supplemental Table S9). Test statistics could not be calculated for the reduction from baseline or post-test to 3-month rates (n=27 and n=26, respectively), due to the inability of McNemar’s Exact test to operate when an event (i.e., EM) does not occur in any participants at one of the two time points. However, the reduction from EM rates of 22.2% at baseline to 0% at 3-month follow-up and from 11.5% at post-test to 0% at 3-month follow-up (Supplemental Table S9) indicates that COACH led to an elimination of EM in the treatment group following the completion of the intervention, while EM rates remained relatively stable in the control group.

Figure 2.

Elder mistreatment rates for treatment vs. control group over three time periods (n=80). Bivariate Fisher’s Exact tests were run for each period to compare the rates of elder mistreatment across the two groups. During intervention delivery n=64; after intervention completion n=55.

Secondary Outcomes

There was little impact of the intervention on the proximal outcome variables. Significantly higher rates of social QoL were detected in the treatment group compared with the control group at post-test (M=16.7, SD=2.2 treatment; M=14.1, SD=4.3 control; p=0.006), but these did not persist through 3-month follow-up (Table 2). No other QoL domains saw significant differences between treatment and control groups.

Although caregiver burden declined in the treatment group across the three time points, the decline was not statistically significant. No significant differences in burden, depression, or anxiety emerged between treatment and control groups at any time point.

DISCUSSION

In this pilot study, we found that the COACH caregiver intervention successfully reduced EM experienced by care recipients aged 65 and older. We used a randomized controlled trial to evaluate effects of the COACH intervention on self-reported EM and other indicators of caregiver well-being, including burden, anxiety, depression, and four aspects of QoL. Utilizing one-on-one care coaches to support family caregivers of older adults who needed care due to dementia, chronic health conditions, and/or functional impairment, the COACH intervention was linked to higher social QoL (satisfaction with personal relationships and social support) in the treatment group at the end of the intervention and to lower rates of EM in the treatment group following the completion of the intervention. The COACH program is the first intervention study that we are aware of to find evidence of EM prevention.

Although risk factors for and precursors of EM are still somewhat unclear, the COACH intervention was developed on the assumption that caregiver support offers a promising path to prevent EM against care recipients³⁹ by helping caregivers manage the stressful and emotionally taxing activities that the role often requires. To optimize this support, components of the COACH intervention were structured to be as comprehensive as possible, with a focus on addressing real world issues and providing tangible assistance in connecting participants to community services and resources to support the practical needs that arise in a caregiver’s day-to-day tasks. The intervention was individually tailored to meet each caregiver’s goals, needs, and circumstances.

Although this pilot study does not have the power to detect independent mechanisms of change, initial findings support the rationale for caregiver support as one potential approach to preventing EM.³⁹ They further suggest that caregiver social engagement may be an important factor, which comports with prior studies.^2,40 As such, elements of the COACH intervention that were included to enhance social support, such as honing communication skills, building a support system, and the supportive presence of the care coach, may be particularly important intervention components. Further research is needed to test the intervention in other settings and elucidate the mechanisms contributing to intervention efficacy, particularly given the minimal impact this study detected on the proximal outcomes. If these findings are supported in larger samples in diverse settings, the impact on the costly and tragic phenomenon of EM would be substantial.

Importantly, change was detected three months after the end of the COACH intervention, which measured EM following the completion of the intervention. If data collection had been stopped at the post-test, no treatment effect on EM would have been identified. The 3-month follow-up survey asked about EM that would have occurred once the intervention had ended and caregivers continued to practice, apply, and develop mastery over skills learned during the intervention.⁴¹

Limitations

The pandemic introduced two major limitations to this study. First, changes to data collection and intervention delivery were necessary due to stay-at-home orders, implemented in March 2020, and subsequent pandemic-related participant safety protocols. The in-home intervention, developed based on evidence from other forms of family violence, had to be abandoned and replaced by remote contact. While all intervention sessions were delivered by phone, surveys were administered via multiple remote modes but were not tracked to allow for comparison between modes. Second, the study had a small sample size, due in part to the Kaiser LAMC being unexpectedly taxed to respond to a surge of community and staff COVID-19 infections. The enrollment period was compressed due to the necessary addition of safety protocols, which halted recruitment. Kaiser research staff reassignments were necessary, and the sharp reduction of non-COVID-19 clinical interactions decreased the pool of patients who could feasibly be referred to a research study. We also believe that prospective participants were less likely to express interest due to fear and turmoil from the pandemic’s unprecedented circumstances. These circumstances may have, in turn, contributed to a selection bias, wherein caregivers who chose to participate may not have been representative (e.g., less distressed than non-participants).

This pilot study lacked the power to investigate primary versus secondary prevention or determine which of the intervention’s many components were responsible for the EM findings. Further studies should attempt to delineate the specific mechanisms of change, to guide intervention refinement and inform fidelity requirements. Further, despite attempting to collect data from care recipients, few were able to respond, requiring us to rely on caregiver self-reports of mistreatment. While this may result in under- or over-reporting, caregiver reports are the field standard for EM research with family caregivers⁵ and have been shown in child maltreatment research to reduce recall bias.⁴²

Implications

In spite of challenges related to the COVID-19 pandemic, the complete absence of EM in the COACH treatment group at 3-month follow-up provides compelling support for this intervention. The COACH program holds promise for improving care provision and reducing risk of EM among older adults with care needs. The strengths-based, person-centered approach provided caregivers with resources to understand and respond to existing and emerging caregiving issues. As health care systems respond to increasing numbers of older adults with chronic conditions, the onus will be on trusted others who will oversee patients’ medical, functional, and emotional care. Health care institutions have been chiefly patient-focused, which is insufficient for this population. There is a growing need for large-scale caregiver-focused interventions; the COACH program may provide a novel entrée to the paradigmatic transition to include patients’ primary caregivers in medical care and other interactions.

In educating and counseling caregiving dyads, clinicians can seek to employ similar strategies to enhance individuals’ capacity to engage in the caregiving role. This is salient in the current environment, where health care institutions are giving more attention to social factors impacting health outcomes. Participating in the COACH program is an opportunity for health care professionals to directly contribute to preventing the morbidity, mortality, hospitalizations, and institutionalization that results from EM. This may be a viable way for health care systems to invest proactively in reducing the enormous human costs and other consequences of EM.

The next step for this research includes a fully powered randomized controlled trial to confirm preliminary efficacy findings identified in the pilot study and detect the intervention’s mechanism (e.g., social support). The COACH intervention is well suited for effectiveness testing and scaling in practice settings, given its origination within a large health care organization and its ability to be delivered remotely, which may be even more desirable in an environment of limited resources and staffing. The COVID-19 pandemic exposed the importance and fragility of our health care workforce, and the COACH program’s utilization of trained non-clinical staff increases its reproducibility in lean health care settings.

KEY POINTS BOX:

Key Points

• Treatment group caregivers reported less mistreatment against their care recipient, which dropped from 22.5% at baseline to 0% following the completion of the intervention, compared with control group caregivers, whose reported rates did not change significantly (15.4% at baseline to 23.1%).

• Treatment group caregivers reported greater social quality of life than control group caregivers in the post-intervention survey, but this effect did not persist three months following intervention completion.

Why Does This Paper Matter?

Elder mistreatment affects 10% of older adults and up to 50% of persons living with dementia. Its victims experience lower quality of life, higher rates of morbidity, hospitalization, and nursing home placement, higher system costs, and early mortality. Preventing mistreatment is crucial to improving the quality of life and health outcomes for millions of older adults who are victimized annually. This is the first intervention study we are aware of to demonstrate preliminary efficacy of an elder mistreatment prevention intervention. As a caregiver support program that can be delivered remotely, the COACH intervention is designed to enable replication and scaling within health care systems or by providers of long-term services and supports, offering the potential to have a dramatic impact on the health and well-being of older adults.