‘Whole person’ caregiver assessment: Insights from heart failure and stroke caregivers

Abstract

Caregivers of persons with heart failure and stroke describe increased responsibilities after hospitalization. To better characterize the needs of caregivers who self-selected into the intervention, we conducted a convergent parallel mixed methods secondary analysis from our pilot study of The Heart Failure and Stroke Resilience Intervention for Caregivers (HEROIC) program. Using transcriptions of Visit 1, coders analyzed pre-defined domains of caregiving for features of perceived strength, neutrality, or challenge. The domains included a) Values (not rated); b) Caregiving Intensity; c) Dyadic Relationship; d) Social Support; e) Financial Resources; f) Home/Built Environment; g) Other Responsibilities of Caregiver; and h) Health/Health Care. Two coders identified salient quotes and rated the seven domains for each transcript. Discrepancies were resolved by group consensus. Data were then integrated with common demographic features and level of patient care need. Of the 20 caregivers, seven (35%) described low needs, 11 (55%) reported moderate needs, and two (10%) described high needs. The moderate/high needs group had worse self-reported health and caregiving-related outcomes when compared to the low needs group. Although a limited sample, this analysis highlights the spectrum of caregiving needs. While the HEROIC intervention was intended to capture caregivers with moderate/high needs, the self-selected inclusion of caregivers reporting low needs suggests that caregiving interventions broaden inclusion criteria. Quantitative measures of caregiving-related outcomes may miss caregivers who feel that they would benefit from targeted intervention. Overall, caregiving interventions that incorporate a whole-person approach have the potential to decrease caregiver burden and improve quality of life.

1. Introduction

Managing the consequences of a serious cardio- or cerebrovascular disease frequently requires family and friend caregiver support. Caregiving support may be defined as providing physical care, supporting activities of daily living, assisting with gaps in cognitive ability, and intervening to support and assume more responsibility for tasks that the patient can no longer perform independently (Buck et al., 2024; Zawawi et al., 2020). Due to the possible changes in physical and cognitive function and medication regimens, patients often need assistance from caregivers to maximize optimal outcomes.

Due to the diversity of caregivers themselves, experts have called for a greater understanding of the dynamic and contextual factors of family caregivers. This includes their access to other social support and community resources and their unmet needs. In this specific study, we focused particularly on caregivers of persons with heart failure and/or stroke. Unmet needs have been identified including increased stress, financial strain and social isolation, but interventions to address these needs in heart failure and stroke caregivers have had mixed results (Bakas et al., 2022; Evangelista et al., 2016; Kitko et al., 2020; Nicholas Dionne-Odom et al., 2017). Additionally, there are differences in lived experience within caregiving subgroups. These differences suggest a need for tailored support (Fabius et al., 2020). which must be based on a deeper understanding of the many contextual factors that contribute to the mental, physical, and social well-being of caregivers rather than a simplistic view.

To address the dearth of tailored caregiver interventions within serious vascular disease, we conducted a feasibility pilot of a caregiver support program: Heart Failure and Stroke Resilience Intervention for Caregivers (HEROIC). The HEROIC intervention consists of five nurse-led, remote sessions, with Visit 1 guided by a holistic assessment to understand the context of caregiving. This holistic assessment is whole-person-based and uses a semi-structured interview format to guide structured activities and goal setting in the intervention. The overarching goal of the HEROIC study was to assess feasibility and acceptability of the intervention.

For this analysis, we applied a convergent parallel mixed methods design to existing qualitative and quantitative data obtained from baseline data collection and Visit 1 of the HEROIC study. Although qualitative and quantitative data were collected simultaneously, we merged these data in a secondary analysis (Creswell & Plano, 2011). This merging allows for interrogation of data and findings and enriches insights that separate analyses could not offer. The purpose of this secondary analysis was to explore and characterize the existing challenges, strengths, and needs of heart failure and stroke caregivers, using recordings from the whole person assessment to 1) characterize the experiences of caregivers and their unmet needs, and 2) assess if our recruitment strategies were effective to reach caregivers with high needs.

2. Research context & background

More than 20% of Americans report caregiving (The National Alliance for Caregiving, & AARP, 2020). It is a ubiquitous human activity occurring in all generations, race/ethnic groups and across socioeconomic levels (The National Alliance for Caregiving, & AARP, 2020). Although literature tends to focus on one construct or factor, the intersection of identities and lived history is essential to understanding the caregiving experience (Jacobson et al., 2025). In 2016, the National Academy of Medicine highlighted the complex nature of caregiving and brought attention to the benefits and burdens of caregiving (Schulz & Eden, 2016). Because of this, it is challenging to isolate and silo the many factors that contribute to the caregiving experience.

2.1. Existing literature reports conflicting findings for caregiving health outcomes

Caregiver health is multi-dimensional and is commonly examined in terms of physical and mental health, caregiver burden, and benefit. Caregiving is frequently associated with worsening mental health outcomes including depression, anxiety and global measures of mental health (Pinquart & SÖrenson, 2003, pp. P112–P128; Roth et al., 2009, 2018). Seminal studies have demonstrated increased mortality and chronic disease risk for caregivers explained by consequences of heightened stress and stress hormones (Bidwell et al., 2021; Haley et al., 1996, 2019; Schulz & Beach, 1999). However, some studies have shown health benefit of caregiving including reduced risk of mortality (Ballew et al., 2025; Fredman et al., 2008). These contrasting findings have led some to test the ‘healthy caregiver hypothesis’ which posits that healthier adults take on caregiving roles and derive health benefits from increased cognitive and functional caregiving activities (Bertrand et al., 2012; Fredman et al., 2015).

2.2. Several factors, including identities, relationships, and lived experiences, influence caregiving health outcomes

2.2.1. Relationship to the care recipient and gender

Caregivers are predominantly spouses, adult children, or other members of the family (Schulz et al., 2020; Wolff et al., 2025). This type of relationship to the care recipient can be beneficial (i.e., closeness, comfortability, shared family history) and simultaneously nuanced (i.e., relationship shifts). Fenton and colleagues (2022) investigated the differences in caregiving by caregiver-care recipient relationship. Their findings demonstrated that when compared to spousal caregiving, adult children spent significantly fewer hours on caregiving but reported significantly higher emotional and social burden (Fenton et al., 2022). Because caregivers are typically women (i.e., spouse or adult daughters of the care recipient (The National Alliance for Caregiving, & AARP, 2020; Wolff et al., 2016), they also investigated factors of gender concordance and caregiver employment. In their study, gender concordance and caregiver employment furthered the difference in emotional and social burden between the two groups (Fenton et al., 2022).

2.2.2. Employment

Caregivers who are working face competing demands. The National Alliance on Caregiving 2025 report highlights that 70% of caregivers ages 18–64 are employed (The National Alliance on Caregiving & AARP, 2025). More than half of these caregivers work full-time; on average, employed caregivers are working almost 35 h per week in addition to their caregiving duties (The National Alliance on Caregiving & AARP, 2025). According to analysis from the nationally representative National Study on Caregiving, for the estimated 8.8 million working caregivers in 2015, nearly a quarter experienced caregiving-related work productivity loss within the past month (Keita Fakeye et al., 2023). The estimated annual loss of income was $5,600, which across the 8.8 million individuals represents an aggregated $49.1 billion. Work demands are also associated with caregiving interference and caregiver burden (Gordon et al., 2012). However, workplace support was associated with reduced caregiver burden, suggesting that improvements in workplace policies may improve overall caregiving experiences.

2.2.3. Caregiving across generations

Another consideration with family caregiving is that adult-child caregivers are possibly also caring for their own children. These caregivers (coined the “sandwich generation”) represent almost a quarter of caregivers in the United States (Lei et al., 2023). Although they have reported spending similar number of hours on caregiving, sandwich caregivers report increased financial, emotional challenges and caregiver overload when compared to their “non-sandwich” counterparts (Lei et al., 2023). A recent narrative review reported several themes surrounding the “problems” that sandwich caregivers face (Pashazade et al., 2024). These themes (categorized as dimensions) centered on physical health, psychological health, employment, finances, and their own marriages and relationships. Several key problems identified spoke to challenges across caregivers. However, challenges such as decreasing quality of marital relationships and couple burnout may be more unique to this subgroup of caregivers (Pashazade et al., 2024).

2.2.4. Type of care provided

The type of care provided by the caregiver can be highly varied. Caregiving may refer to assisting with household tasks, self-care, mobility, and supervision; providing emotional and social support; involvement in health and medical care; advocating and promoting care coordination; and decision making and surrogacy (Schulz & Eden, 2016). In a nationally representative profile of caregiving, authors estimated nearly 6.5 million U.S. caregivers provide substantial help, which included assisting with both care coordination and medication assistance (Wolff et al., 2016). Those who provide substantial help are more likely to live with the care recipient, provide more care hours per week and report longer duration of caregiving than those who provide less substantial help. Overall, given diversity in tasks, preparedness is critical yet lacking. Nationally representative data indicates that only 8.3% of caregivers of persons with dementia received caregiver training (Kim et al., 2025). Looking at the broader Medicare population, the rate of unmet needs that required training ranged from 8.2 to 16.0% by activity type. The lowest value (8.2%) represented training for household chores, while the highest value (16.0%) represented training for assisting with self-care tasks (Burgdorf et al., 2021). Despite the low rates of training, preparedness has shown significant association with increased resilience and decreased caregiver burden (Gutierrez-Baena & Romero-Grimaldi, 2022).

2.2.5. Culture and social constructs

The sociocultural implications of caregiving are vast, representing the impact of a variety of identities. When considering race, we are challenged by using a social construct to investigate subgroups. Because of this, it is important to consider the cultural identities, histories, and differences that these groups represent. For instance, while caregiver burden has been reported to be lower among Black caregivers, Black caregivers experience greater hours of caring compared to White caregivers (Fabius et al., 2020). In comparing Asian, White, and non-White Hispanic caregivers, Asian caregivers have reported greater worry about their own knowledge on how to provide care when compared to their counterparts. This same study reported that White and non-White Hispanic caregivers were more likely to offer emotional support compared to Asian caregivers (Tran et al., 2023). These findings are not to judge caregiving subgroups but instead highlight the underlying complexity of these experiences that race as a construct does not fully capture. Additionally, non-White Hispanic and Asian caregivers had less knowledge of caregiving-related government policies and available support programs and services (Tran et al., 2023).

2.3. Whole person care offers a theoretical shift in caring for the caregiver

The conflicting literature on caregiver health (i.e., ‘healthy caregiver hypothesis’) poses a question regarding the construct of caregiving burden versus caregiving experience. Many studies point to the benefits of caregiving: increased intimacy, emotional closeness and sense of meaning and purpose (Pendergrass et al., 2019; Ping et al., 2025). While benefits of caregiving are studied less commonly than burden, caregivers repeatedly affirm beneficial aspects of caring. Benefits persist throughout the caring trajectory and often decrease as the patient approaches end of life (Ping et al., 2025). The tension between burden and benefit represents the intricate intersections of caregivers’ identities, histories, roles, and lived experiences (Clair et al., 2023). This population is not a monolith and thus interventions and support must be tailored to the unique caregiving situation.

Considering the “whole person” can identify individual caregiver needs and inform specific caregiver interventions. Traditionally, whole person care describes how providers integrate biology and biography into the provider-patient relationship. This type of care emphasizes integration of biopsychosocial health over time within care relationships (Hutchinson, 2017; Kaslow et al., 2007; Thomas et al., 2023). Whole person care offers a rich understanding of the context and values of the person receiving care (Goldfarb et al., 2024). To date, many studies and guidelines have called for approaches that look beyond symptoms or acute stressors and instead consider the patient as a whole person. A narrative review by Jonas and Rosenbaum (2021) reported that, across 21 models of whole person care, this approach resulted in improved patient health outcomes, satisfaction, and cost effectiveness (Jonas & Rosenbaum, 2021). Although assessment of these models and their impact on outcomes is challenged by heterogeneity, the overarching theoretical construct appears to positively affect patients’ health and their care experiences. These promising findings indicate possible benefit in applying whole person care beyond the patient.

For caregivers, whole person care has been conceptualized as “caring for the caregiver.” Several frameworks have been developed to focus on whole person care for caregivers. Holliday and colleagues developed the CARE framework, which focused on caregiver well-being, advanced care planning, respite, and education (Holliday et al., 2022). This framework aimed to reduce caregiver burden with application in the clinical setting of the patient’s visit. While the caregiver well-being tenet involves investigating coping and needs, the suggested application is “Simply asking caregiver how things are going, and meaningfully engaging with their answer, is one quick yet effective intervention to gauge a caregiver’s coping.” (Holliday et al., 2022) Although an important question to begin the conversation, this framework lacks depth in implementation. The CARE framework was followed by the Interprofessional Family Caregiving Competences, which was developed by Harvath and colleagues to strengthen support for caregivers and enhance their capacity to provide care (Harvath et al., 2023). The four domains of the Competencies include 1) the nature of family caregiving; 2) family caregiving identification and assessment; 3) providing family-centered care; and 4) the context of family caregiving (Harvath et al., 2023). These domains do center on the caregiving experience, but similarly, this framework is intended for application in the patient’s visit and lacks sufficient guidance for implementation. These two frameworks represent a key first step and are critical for informing research and policy. They both highlight the importance of whole person care for the caregiver. However, the operationalization of these frameworks is limited. It raises questions of measurement, assessment, implementation, and action.

Whole person-centered assessment of caregivers is needed to target supports, yet there are few holistic assessments designed for caregivers. Specific to this study, many consensus statements and guidelines call for improving support to caregivers broadly and specifically within heart failure and stroke (Adeoye et al., 2019; Bakas et al., 2022; Heidenreich et al., 2022; Lindsay et al., 2014; Schulz & Eden, 2016).

3. Methods

3.1. Population of interest

The HEROIC study enrolled caregivers of persons 1) living with heart failure and evidence of decompensation and/or 2) recently hospitalized with stroke. Additional requirements for inclusion included the following: 1) the patient had experienced decompensated heart failure or stroke in the last 6 months (i.e., hospitalization, outpatient diuresis appointments, or failure to achieve guideline-directed medical therapy); 2) the caregiver lived with the patient or visited them to provide care more than 3 times per week; and 3) the caregiver was 18 years or older. Caregivers were excluded from the study if they themselves had a terminal diagnosis or were cognitively impaired (as determined by 6-item screener) (O’Sullivan et al., 2016). These caregivers were excluded due to a likely difference in their response to the intervention and data collection. Additionally, caregivers who consented to a previous piloted version of HEROIC (The Caregiver Support Study) and participated in at least one intervention visit were excluded from participation in HEROIC (Abshire Saylor, Pavlovic, et al., 2023).

3.2. Ethics approval and informed consent

The overall HEROIC study was approved on August 18, 2021, by the Johns Hopkins Medicine Institutional Review Board (IRB00277814). Prior to completing baseline data collection, participants reviewed the consent form with a data collector by phone. Participants then provided verbal informed consent (waiver of documentation of consent) to participate in the study.

3.3. Study and intervention design

This is a mixed methods secondary analysis using data from the HEROIC study; for this analysis we used a convergent parallel design (Creswell & Plano, 2011). In the HEROIC study, quantitative (baseline) and qualitative (Visit 1) data were collected separately. For this secondary analysis, these data were merged to address our research objectives. Merging involved initial qualitative analysis informing quantitative analyses, which then strengthened our interpretation of original qualitative findings. Our intention was to investigate the same phenomena using differing, yet complementary data (Creswell & Plano, 2011; Fetters et al., 2013; Guetterman et al., 2015).

The caregiver support program, HEROIC, consists of five nurse-led, virtual sessions over a 10-week period (Abshire Saylor et al., 2023). Following baseline data collection, caregivers were provided with intervention materials and randomized to either immediate intervention or waitlist control groups. The immediate intervention group received the HEROIC intervention between baseline (0 weeks) and 12 weeks, while the waitlist control group received the intervention between 12 weeks and 24 weeks. The HEROIC intervention includes the following activities: 1) Whole Person Assessment; 2) Life Purpose Statement; 3) Social Support Circles; 4) Instrumental Support Resource Identification; and 5) Goal setting.

The Whole Person Assessment is the main activity of the first visit, in which the nurse interventionist uses a semi-structured conversation guide to build rapport and understand how the context of caregiving has impacted domains related to well-being (Abshire Saylor et al., 2023; DeGroot et al., 2021; Nelson et al., 2022; Ryff, 2013). The Whole Person Assessment was developed as a clinical tool and in recognition that many available tools tend to focus on caregiving burden, rather than strengths of the caregiver and their context, limit the definition of caregiving, or ignore important contextual aspects of care (Clair et al., 2023; Nelson et al., 2022). We developed this assessment based on our previous qualitative and mixed methods studies of caregivers, which indicated that caregivers had varying past and present caregiving experiences, competing demands/interests and different levels of unmet needs (Abshire Saylor et al., 2023; DeGroot et al., 2021; Abshire et al., 2021; DeGroot et al., 2021; Nelson, K. E. et al., 2023). Domains include hobbies, education, employment, finances, health of the caregiver, culture and spirituality, caregiving experience, and social support, with 4–8 question prompts for each domain (Fig. 1). During the first intervention visit, the interventionist explains the activity:

Sometimes in the clinic or hospital it is easy for nurses and providers to miss important details because we don’t spend enough time learning from and listening to caregivers. So, today we are going to do the ‘Whole Person Activity’, which is designed to be a conversation for me to get to know you better and understand your role as a caregiver across various domains of your life. You are the expert, and this will simply help me understand a fuller picture of your life. Then we can talk about if there are any priorities we could focus on together. If there’s something you don’t see on the wheel, or you feel like I’ve missed anything, please feel let me know and we can talk about it.

Fig. 1. Domains of the Whole Person Assessmen.

Description: The figure displays the Whole Person Assessment activity from HEROIC Visit 1. The activity consists of eight domains, including caregiving experience, support, hobbies, education, employment, finance, health, and community, culture, and spirituality.

The caregiver is then encouraged to select a domain to start. A conversation guide with the prompts (Supplemental Material) is provided to support the flow of conversation and rapport building, but interventionists were trained to allow conversations to vary by individual.

To ensure consistency across nurse interventionists, we provided an intervention manual and robust training. The intervention manual and training were developed from the previous pilot of HEROIC, (The Caregiver Support Study), and informed by the experiences of those nurse interventionists. Training prework included reviewing the intervention manual and listening to four recordings of Visit 1 from The Caregiver Support Study. The HEROIC interventionists then met with Research Coordinator and Principal Investigator to review their assessment of these Visit 1 recordings, identifying (mis)alignment with the manual. For the first six months of the HEROIC study, interventionists met with the Research Coordinator and Principal Investigator biweekly to discuss their assessment of their own Visit 1 and subsequent visits. Any issues or concerns with fidelity were addressed in these meetings. Starting at the seventh month, meeting frequency was reduced to monthly.

3.4. Data collection

The baseline survey was collected via REDCap or over the phone, depending on the caregiver’s preference. In addition to demographic characteristics and information, the survey included several measures of quality of life and well-being. Quantitative measures are summarized in Table 1. Data collection for the overall HEROIC study ran from May 2022 to August 2024.

Table 1.

Theoretical constructs and associated quantitative measures included at HEROIC baseline assessment.

Theoretical construct	Instruments and interpretation	# of items	Cronbach’s alpha
Mental well-being	Mental health component of the 36-Item short form health survey (SF-36). Derived from 4 sub-scales. Scores range from 0 to 100 with higher scores indicating a more favorable health state	18	0.85 McHorney et al. (1994)
Physical well-being	Physical health component of the 36-Item short form health survey (SF-36). Derived from 4 sub-scales. Scores range from 0 to 100 with higher scores indicating a more favorable health state	18	0.85 McHorney et al. (1994)
Quality of Life	Short Form of the Positive Affect and Well Being (PAW) measure of the NeuroQOL. Scores range from 9 to 45 with higher scores indicating higher self-reported PAW	9	0.85 Cella et al. (2012)
Self-efficacy	13-Item Coping self-efficacy. Scores range from 0 to 130. Higher scores suggest a stronger belief in one’s ability to cope effectively with stressors and challenges.	13	0.91 Chesney et al. (2006)
Caregiver burden	Zarit burden interview (short version) Focuses on emotional aspects of caregiver burden. Summed scores range from 0 to 48 with higher scores indicating higher burden	12	0.86 Thornton & Travis (2003); Bédard et al. (2001)
	Oberst Caregiver Burden Scale Domains: Caregiving task time and difficulty. Both subscale scores range from 15 (minimal burden) to 75 (high burden).	30 (15 each sub-scale)	0.90(Bakas et al., 2004)
Depressive Symptoms	Patient Health Questionnaire. Scores range from 0 to 24 with higher scores indicating more severe depression.	8	0.86(Kroenke et al., 2009)
Perceived Stress	Perceived Stress Scale. Scores range from 0 to 56, with higher scores indicating higher levels of perceived stress.	10	0.82Cohen et al. (1983)
Social support	ENRICHD social support. Scores range from 8 to 34 with higher scores indicating greater social support.	7	0.89 Vaglio et al. (2004)

3.5. Data analysis

Whole person assessments (Visit 1) were recorded by the nurse interventionist and transcribed by a third-party qualitative transcription. We did not differentiate between immediate intervention and waitlist control groups in analysis. This decision was informed by our desire to capture the experiences of all participants, regardless of group assignment. Bereaved caregivers (i.e., bereaved after baseline data collection and prior to starting the HEROIC intervention) were excluded from coding for the Caregiving Intensity domain but included in the analysis.

Once the Whole Person Assessment was transcribed, two coders (CAC, NEW) independently coded the transcript using a deductive, a priori coding template. The template included the six domains of caregiving identified in our prior analysis of contextual factors for caregivers (Abshire Saylor et al., 2023; DeGroot et al., 2021). An additional domain was added to the template (“Health/Health care”). “Health/Health Care” was a domain in the Whole Person Assessment used in Visit 1, which supported its inclusion in the coding template. Table 2 highlights each coding domain and its definition. The coding template used in the analysis is available as Supplementary material.

Table 2.

Exemplar quotes by domain and coding.

Whole Person Assessment Domain		Exemplar Quotes
Values: Any expression of faith, what matters most to them, priorities	Not rated	“[Religion and spirituality] is very important. It’s been very important to me and my wife.” - Caregiver #1
Caregiving Intensity: Activities of caregiving (ADLs, IADLs and emotional support), degree of influence on caregiver’s lifestyle, physical health, mental health	Strength	“It’s been much more relaxing for me when it comes to stressing whether or not she’s doing all right by herself. So, yeah, it’s much more relaxing than worrying.” - Caregiver #2
		“Last year … I had to listen outside the shower to make sure that he was sitting in a chair and grabbing the bars. And, you know, it’s gone all the way from that kind of thing to almost back to normal.” - Caregiver #3
	Neutral	“I compartmentalize things probably too well in some cases … I don’t let things stress me. I try not to” - Caregiver #4
	Challenge	“[I am] basically his support. I do everything for him” - Caregiver #5
Dyadic Relationship: Who is the dyad? What is their relationship? Quality of relationship? Changes to relationship?	Strength	“[We] have a really good, pretty open relationship” - Caregiver #4
	Neutral	“He’s very stubborn. He does like being independent. I get that” - Caregiver #6
	Challenge	“My biggest problem of all is feeling excluded, not being allowed to be a part of it” - Caregiver #7
Social Support: People who are family, friends, neighbors, work colleagues, etc. (not the dyad) who provide emotional, instrumental (tasks), informational or appraisal/assessment supports, also those who negatively influence the perception of social support	Strength	“I can also [take] leave [from] the job, because my boss understands” - Caregiver #8
	Neutral	“I’m grateful for my best friend. She is such a beautiful person from the inside out. I love her very much, and I know that she will do anything for me … I would have to say that there’s not one person in my family or his family who is there for me and what has happened” - Caregiver #9
	Challenge	“I think everybody else is so used to me not taking any time that they make me feel guilty in addition to myself making me feel guilty. So like, then my kids will be like, “Oh, you got your nails done again, huh?” These just little snide, silly comments.” - Caregiver #10
Financial Resources: Resources available, strain caused by health events or caregiving in dyad, other financial strain, changes to employment, changes to insurance Home/Built Environment	Neutral	“We all have financial problems where we wish it could get better but I’m good with where I’m at” - Caregiver #5
	Challenge	“Finances are an issue.” - Caregiver #9
Home/Built Environment: How is the home environment supportive/problematic to performing caregiving tasks? Or patient mobility?	Strength	“I don’t worry as much because I know she’s in a safe environment.” - Caregiver #2
	Neutral	“Interviewer: So your son and daughter are out of the house. It’s just your husband. Do [the patients] live with you as well? Caregiver: No, they live 1.8 miles up the street.” - Caregiver #11
	Challenge	“My granddaughter got evicted from her home, COVID, she lost her job and she’s going through a thing of she’s got gastric issues and her and my great-granddaughter are with me, a four-year-old. And now I don’t- I think I like my peace and quiet better.” - Caregiver #12
Other Responsibilities of the Caregiver: Work, family roles such as grandparenting, church or hobby-related commitments	Strength	“That’s a stress reliever too. Just, you know, being around [my grandchildren], they’re fun.” - Caregiver #13
	Neutral	“Sure. It’s me, my mom, my dad, my three kids. We got a couple of animals <laughs> … we have two pit bulls. I also have a Yorkie. They have five water turtles, and they have a bearded dragon < laughs>… We all enjoy each other” - Caregiver #8
	Challenge	“And then the bank stuff is going to be the hardest stuff I know, because I’m actually still the executor on my mother’s estate. She passed away a year and a half ago, and I’ve been having all kinds of problems, because she was in [Country 1], and she has accounts in [Country 2], and I’m down here, and everybody wants something different and things like that.” - Caregiver #14
Health/Health Care: Patient and/or caregiver’s challenges and facilitators from healthcare encounters, pharmacists, etc. How connected is the caregiver to the patient’s providers? Are they taking care of their (the caregiver’s) own health needs?	Strength	“[Clinician] is a part of the cardiology team that takes care of my mom. I love [her]. [She] and I are- she’s just- she’s an amazing individual. She lifts me up when I need it.” - Caregiver #11
	Neutral	“I’m not still doing the [mental health] therapy, because life happens, and it was just a lot of things happened in the latter part of last year. I had cataract [surgery].” - Caregiver #15
	Challenge	“So for about three and a half years things have been hard health-wise” - Caregiver #9

After individual coding, both coders independently rated each domain as “Strength,” “Neutral,” or “Challenge,” representing integration of qualitative and quantitative data through data transformation (Fetters et al., 2013). Two domains had a different rating structure: Values (not rated) and Financial Resources (rating options limited to “Neutral” or “Challenge”). Values were not rated due to the personal nature of an individual’s priorities, values, and beliefs. Because it was not rated, Values was not included in the Overall Needs score calculation. Financial Resources rating did not include “Strength” based on a priori analytic decision. The study team determined that a “Strength” rating for someone’s financial position risked overinterpretation by the coders (Garg et al., 2024).

Ratings were then transferred to an analytic spreadsheet; an example of the spreadsheet is included in the Supplementary materials. In the case of disagreement with coder ratings, the two coders met with the Principal Investigator (MAS) to discuss and reach consensus. Disagreement, consensus process, and conflict resolution were documented in the spreadsheet. Throughout the process, we conducted constant comparison of our coded categories and sub-categories with the original text to ensure reliability and credibility (Flick, 2009). Overall, prior to consensus and conflict resolution, Financial Resources and Health Care had the fewest disagreements (n = 6), while Other Responsibilities of the Caregiver had the most (n = 10). Coding occurred over the course of five months; at the initial month, four interviews were discussed with 14 disagreements across the seven domains (percent agreement = 50.0%). At the final meeting, four interviews were discussed with 10 disagreements across the seven domains (percent agreement = 64.3%). This disagreement was primarily driven by Other Responsibilities of the Caregiver. This domain’s percent agreement at Month 1 (25%) was unchanged at Month 5.

Once consensus was reached, the domains coded as “Challenge” were summed for an Overall Need score per transcript. The Overall Need score was reported categorically: Low (0 challenges), Moderate (1–2 challenges), and High (3+ challenges). These categories were informed by prior research on the context of caregiving (Abshire Saylor et al., 2023; DeGroot et al., 2021). Overall, the individual coding templates and analytic spreadsheet enhanced trustworthiness of our analysis, as it allowed the team to maintain an audit-trail of decision-making (Flick, 2009).

For analyses of baseline demographic characteristics and quantitative measures, descriptive statistics are presented. Continuous data is presented as mean and standard deviation, and discrete data is presented as count and percentage.

To achieve mixing, we combined qualitative and quantitative data that were collected and analyzed independently, to confirm and enhance findings (Fetters et al., 2013; Guetterman et al., 2015). The qualitative data informed our categorical “needs” scoring, which then informed stratification for quantitative analyses. We collapsed the moderate and high need categories in the quantitative analysis for several reasons. Primarily, the goal of the overall HEROIC study was to capture caregivers with moderate and high needs. Given this goal, we separated low needs (0 challenges) from a combined category of moderate (1–2 challenges) and high (3+ challenges) needs. Secondary and tertiary reasons included the small number of high needs caregivers in our sample and very different qualitative descriptions of their roles compared to the low needs caregivers. Given the differences in qualitative description, we felt confident in our analytic decision to separate the low needs group from the moderate and high need groups. We present our quantitative findings for the entire sample and by Overall Need score.

3.6. Analytic team

CAC was the primary coder and at the time of analysis, was a doctoral candidate in social and behavioral sciences at the Johns Hopkins Bloomberg School of Public Health. She brought substantial experience in qualitative and mixed methods analyses focused on older adults and caregivers. The secondary coder (NEW) was an undergraduate student at Johns Hopkins University with no prior qualitative research experience. At time of analysis, the Principal Investigator (MAS) was an Assistant Professor at Johns Hopkins School of Nursing with significant research, clinical, and lived experience in caregiving.

4. Results

4.1. Sample characteristics

Of the 31 caregivers who completed baseline data collection, 24 completed the Whole Person Assessment visit and 20 had a useable audio recording (see CONSORT diagram in Fig. 2). Caregivers (N = 20) were on average 61 years old (SD: 9.3), 75% female and 40% Black (Table 3). Nine care recipients functionally dependent and caregivers supported 5.8 ± 1.9 instrumental activities of daily living (IADL) on average. Additionally, nearly half of caregivers reported financial strain (45%), which was defined as “Having just enough money” or “Having not enough money” to make ends meet by the end of the month. Of the two conditions (e.g., stroke, heart failure), most of the caregivers supported persons with heart failure (55%). When comparing the 31 caregivers to the 24 that completed the Whole Person assessment, there were no significant differences in demographic characteristics.

Fig. 2. CONSORT Diagram.

Description: This figure describes the flow of participants in the overall HEROIC study and specifically for this convergent parallel mixed methods secondary analysis.

Table 3.

Sociodemographic characteristics by level of need.

Variable	Overall N = 20	Qualitative Groupings
		Low Needs N = 7	Moderate/High Needs N = 13
Sociodemographic Characteristics
N (%) unless otherwise indicated
Age, Mean (SD)	60.9 (9.3)	62.4 (9.5)	60.1 (9.5)
Gender
Men	5 (25)	3 (43)	2 (15)
Women	15 (75)	4 (57)	11 (85)
Race
White	12 (60)	1 (14)	11 (85)
Black	8 (40)	6 (86)	2 (15)
Currently Employed
Yes	9 (45)	2 (29)	7 (54)
No	11 (55)	5 (71)	6 (46)
Relationship
Spouse or partner	11 (55)	3 (43)	8 (62)
Child	8 (40)	3 (43)	5 (38)
Other	1 (5)	1 (14)	0 (0)
Financially Strained	9 (45)	2 (29)	7 (54)
Care Recipient Condition
Heart Failure	11 (55)	3 (43)	8 (62)
Stroke	9 (45)	4 (57)	5 (38)
Functional Dependence
Heart Failure (n = 11): Assistance with ≥1 Activity of Daily Living	4 (36)	2 (67)	2 (25)
Stroke (n = 9): Barthel Index of ≤90	5 (56)	1 (25)	4 (80)
Instrumental Activities of Daily Living (IADL) Dependence
N (%) unless otherwise indicated
Total sample assisting with >2 IADLs	19 (95)	6 (86)	13 (100)
Mean ± SD across total sample	5.8 (1.9)	4.9 (2.1)	6.3 (1.7)

4.2. Findings

4.2.1. Overall ratings

Across the rated domains, the “Social Support” domain had the greatest number of coded strengths (n = 12). “Dyadic Relationship” had the greatest number of challenges (n = 5), and “Social Support,” “Financial Resources,” and “Home/Built Environment” had the least challenges (n = 2). See Fig. 3 for summary of quantitative ratings per domain.

Fig. 3. Needs Assessment Ratings by Domain.

Description: Across the caregivers in the study (n = 20), each was rated across seven domains: Challenge, Neutral, or Strength. This figure highlights the ratings per domain.

Table 2 outlines representative quotes for the “Strength”, “Neutral”, and “Challenge” ratings for each domain.

4.2.2. Low vs moderate/high needs caregiving

Of the 20 caregivers, we rated seven (35%) low needs, 11 (55%) moderate needs, and two (10%) high needs. We note that the two high needs caregivers were both in the heart failure diagnosis group. The two bereaved caregivers both reported moderate level of need at the time of Visit 1.

Table 4 compares quantitative measures by Overall Needs Score from our qualitative analysis. The moderate/high needs group had a higher proportion of women, White race, working, financially strained and spousal caregivers. For all quantitative measures, the low needs group had higher self-reported health and better caregiving-related quantitative outcomes when compared to the moderate/high needs group.

Table 4.

Health and caregiving-related outcomes by level of need.

Variable		Overall N = 20	Qualitative Groupings		Mean difference
			Low Needs N = 7	Moderate/High Needs N = 13
Mean ± SD
Physical Health		63.4 (22.1)	66.3 (14.3)	61.8 (25.7)	−4.5
Mental Health		64.1 (22.7)	76.1 (13.9)	57.6 (24.3)	−18.5
Quality of Life		35.3 (6.0)	39.4 (4.3)	33.0 (5.6)	−6.4
Coping Self-Efficacy		87.2 (26.1)	92.3 (32.2)	84.7 (23.7)	−7.6
Caregiver Burden	Emotional	12.2 (5.8)	7.7 (3.9)	14.6 (5.3)	6.9
	Task Time	48.8 (13.0)	46.7 (15.1)	49.9 (12.3)	3.2
	Task Difficulty	47.2 (16.8)	43.3 (19.1)	49.2 (15.8)	5.9
Depressive Symptoms		5.4 (4.3)	3.0 (2.7)	6.7 (4.5)	3.7
Perceived Stress		15.7 (9.1)	7.4 (4.6)	20.2 (7.6)	12.8
Social Support		18.5 (6.3)	20.7 (7.1)	17.2 (5.8)	−3.5

5. Discussion

This convergent parallel mixed methods study explored the challenges, strengths, and needs of caregivers enrolled in the HEROIC study with the goal of understanding those who participated in the program and their needs. Our findings suggest that caregivers who chose to participate in HEROIC have varying levels of needs. Although we aimed to include caregivers with moderate to high needs, we found that 35% of caregivers who participated were categorized as low needs. While HEROIC did reach the intended target population, there are other caregivers who self-select into programs for caregiving resources and support. This speaks to the need for tailoring caregiver interventions to the context and experience of the caregiver. Our findings enhance the caregiving science literature by characterizing caregivers who self-select into interventions and support programs. Although caregiver studies often characterize the entire sample of caregivers involved by reporting demographics, to our knowledge no studies have conducted a mixed methods within-group analysis using a whole person caregiver assessment to highlight the varying caregiver experience. Overall, we believe that the analysis presented here provides preliminary support for construct alignment of the Whole Person Assessment for caregivers. To promote this future exploratory research and aid with reproducibility, the Whole Person Assessment is provided in the Supplemental Materials.

Prior studies have analyzed quantitative measures to understand the balance of needs and resources among caregivers and identified similar variation in groups. One study by Brantner and colleagues used latent class analysis to examine differences on measures of well-being for caregivers, reporting that the largest class was low intensity, low strain (35.5%) and the smallest class was high intensity, high strain (12.7%) (Brantner et al., 2023). These findings align with our own, where low need caregivers represented 35% of our sample and high need caregivers represented 10%. Our previous work also supports this variability of needs and our findings regarding the quantitative measures; in our previous dyadic heart failure patient-caregiving study, which used semi-structured interviews to inform development of the Whole Person Assessment and HEROIC intervention, we found that the high needs group had the lowest average score on multiple caregiver measures, including caregiver burden, social support, and quality of life (mental well-being) (Abshire Saylor et al., 2023; DeGroot et al., 2021).

In this study, we found that our moderate/high need caregivers scored worse on all measures compared to their low need counterparts. This highlights the appropriateness of tailored caregiver interventions such as HEROIC. Others have called for categorizing caregivers by differing levels of care intensity and caregiving strain to help those at most risk for poor psychosocial outcomes, suggesting that it could aid in determining which caregivers might benefit from specific supports (Brantner et al., 2023). This call may be well-intentioned but limits which caregivers receive support. Although we aimed to enroll moderate to high needs caregivers, we found that caregivers of all need levels self-selected into the study. Future iterations of HEROIC and other caregiver support studies and programs could consider a dose-response intervention structure, where caregivers can be ‘dosed’ based on their needs (Bakas et al., 2022). In our study, we retrospectively categorized the caregivers into levels of needs; HEROIC did not have this information prior to a caregiver’s participation. Future iterations of this work may consider holistic assessment and dosing number of visits of programs like HEROIC accordingly. Additionally, our combination of whole person assessment and quantitative measures could be blended for caregiver support programs. For example, some quantitative measures could be utilized as pre-visit needs assessment to prioritize discussion that particularly in determining dose-response structure and tailoring resources.

The Whole Person Assessment is a wealth of information from the caregiver’s own perspective and has important clinical implications for care of the caregiver. Increasingly, health systems are recognizing the need to include caregivers, and some advocates recommend assigning medical record numbers to caregivers when patients with serious illness are admitted for care (Applebaum, A., 2021; Applebaum, A. J. et al., 2021). This provides a means to assess and deliver caregiver-focused care while also billing for the time. The Whole Person Assessment is a major component of HEROIC and takes approximately 1 h to complete. To effectively deliver caregiver-focused care programs, we must strike a balance of time and efficiency to adequately assess caregivers. Our study also builds on past approaches to caregiver assessment using quantitative measures and highlights that caregiver burden is inadequate as the sole assessment (Baird et al., 2023; Clair et al., 2023; Family Caregiver Alliance; Riffin et al., 2021). Clinics who are engaging in caregiver-inclusive practices must consider more than a single measure of caregiver burden and should not measure burden if adequate supports are not available.

Regarding future inquiry, we acknowledge the importance of research, clinical, and lived expertise with and as a caregiver to complete this type of work. Without these personal and professional experiences, it limits the ability to assess caregivers’ needs. Our analyses were challenged by the lack of agreement in coding certain domains (i.e., Other Responsibilities of Caregivers), particularly among study team members with less clinical and personal experiences of caregiving. Upon consideration of the disagreement, despite repeated training, we believe that the limited sample size hindered the ability for coders’ agreement to improve. A larger study, in combination with an improved training strategy, would provide opportunity for retraining to be implemented more successfully.

There are several limitations of this analysis. First, although 31 caregivers completed baseline data collections, only 24 caregivers completed Visit 1 and 20 had viable recordings for analysis. The small sample size for this analysis limits generalizability, which is further compounded by the self-selection of caregivers into the study. Additionally, for the caregivers who were lost to follow up (n = 7) and without a recording (n = 4), we do not know their level of needs. If caregivers were lost to follow-up due to moderate or high caregiving needs, this would inflate our proportions of low needs caregivers. Second, this was a deductive analysis, utilizing a structured coding template that was determined a priori. While this provided structure to our coders, this limited the ability for needs beyond our domains to emerge through the data. The predefined categories and analytic methodology could not capture the intersection of many covariates, including gender, caregiving duration, employment intensity, health status, marital dynamics, and cultural factors. This limits our understanding of the complexity that exists within a caregiving experience (Clair et al., 2023). Third, the “Financial Resources” domain was limited to ratings of “Challenge” and “Neutral.” Although determined a priori, this introduces structural bias into the categorization of this domain. Finally, collapsing our moderate and high needs categories for quantitative analyses introduces potential reductionist bias as described in literature on categorizing race and ethnicity (Read & Fairfax, 2025). This analytic decision may oversimplify caregiving experiences and mask heterogeneity existing within the collapsed category.

Despite these limitations, the strengths of this analysis lie in its mixing of qualitative and quantitative methods, which offers deeper insight compared to only utilizing one of these methodologies. This holistic, mixed methods approach to understanding the context of caregiving has provided an in-depth understanding of an intervention component in HEROIC. Our initial qualitative analysis informed our quantitative descriptive analyses, which then further enriched our interpretation of original qualitative findings. Our domains were expansive, and while each domain was weighted equally calculating the Needs score, this study represents the first step. Further exploration may deepen understanding if domain weighting is feasible and/or appropriate in this population. The study findings highlight the variety of caregiving situations and how caregivers perceive their own needs in terms of self-selecting into a support program. It also provides several opportunities for future identification of unmet needs through use of the HEROIC Whole Person Assessment or measures associated with needs categories.

6. Conclusions

Our mixed methods secondary analysis to explore the needs of caregivers in the HEROIC study underscores the importance of understanding the varying levels of needs across this population. Although quantitative measures are a critical method for identifying need, our findings highlight the complexity of needs and caregiving situations. The merging of qualitative with quantitative data enhanced our understanding that caregivers, particularly those with low needs, may elect to receive additional supports. Findings from this analysis will inform future research and implementation of tailored caregiver support interventions, especially for caregivers of persons living with heart failure and stroke. Future interventions and programs may consider the balance between identifying caregivers with higher levels of need and allowing caregivers to self-select into their programming. This research also supports the development of caregiving policies that center unique caregiving experiences and do not provide assistance solely based on quantitatively measured need and/or burden.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.ssaho.2026.103066.

Data availability statement

The dataset generated during and analyzed during the current study is not publicly available due the sensitive nature of data collected but is available from the corresponding author on reasonable request.