Understanding the Impact of Patients’ Disease Types on Caregiving Time and Caregiver Burden: An Analysis of the Health Information National Trends Survey


 Background: Patients with different medical conditions often have distinctive caregiving needs that could result in varying levels of caregiver burden. However, despite empirical advances in this area, little is known about how patients’ disease types interact with caregiving time and caregiver burden. To bridge this gap, we examined the impact of patients’ disease types on caregiving time and burden. Methods: Data were analyzed from the 2018 Health Information National Trends Survey 5 Cycle 2. Only participants self-identified as caregivers were included in the final analysis. Data on patients’ disease types, caregiving time (i.e., caregiving duration and caregiving hours spent per week), and caregiver burden (i.e., caregivers’ self-rated health, body mass index, and psychological distress) were examined using logistic regression analysis.Results: Patients’ disease types impacted caregiving time and burden. Caregivers of patients with neurological disease spent the greatest amount of time. For caregiver burden, caregivers of patients with cancer or aging related disease experienced worst self-rated health, caregivers of patients with orthopedic disease have the greatest likelihood to be overweight or obese, while cancer caregivers experienced greatest levels of psychological distress. Conclusions: Patients’ disease types had highly varied effects on caregiving time and burden. This study underscores the need for healthcare researchers to adopt a nuanced approach in acknowledging and addressing the burden of care experienced by caregivers, such as tailoring interventions based on both patients and caregivers’ characteristics and preferences.


Introduction
Caregiving is a challenging experience that often results in physical and psychological burden on informal caregivers [1][2][3]. An informal caregiver is a person who offers unpaid or ill-compensated care to a family member, friend, or partner due to illness-related reasons or old age. Results indicate that informal caregivers often experience substantial physical and psychological health consequences as a result of caregiving, such as fatigue [4,5], loss of sleep [6,7], perceived stigma [8,9], anxiety symptoms [10,11], depression disorders [12,13], worsened subjective wellbeing [14,15], and compromised quality of life [16,17]. Evidence further shows that informal caregivers' selfrated health has been in decline for the past ve years, further widening the gap between caregivers and the general population' health status [18]. As more than one in ve Americans (~ 53 million) is an informal caregiver [18], research is needed to better understand factors that in uence caregiver burden so that more effective health solutions could be developed to alleviate resultant heath disparities.
Across disease contexts, factors that may impact caregiver burden have been examined, including caregivers' sociodemographic characteristics (e.g., gender) [19], caregiver self-care abilities (e.g., self-e cacy) [20], social support factors (e.g., caregiving unmet needs) [21], and patient-related factors (e.g., time spent on caregiving) [22]. While less studied, many thanks to the increasing recognition of the interactive nature of caregiving, research on patient-related factors' impact on caregiver burden is gaining momentum [23,24]. A growing body of evidence suggests that patients' disease types could have a considerable impact on caregiving time and caregiver burden [25][26][27][28][29]. Comparing caregivers of patients face renal transplantation and caregivers of patients undergo hemodialysis, researchers found that caregivers of patients receive hemodialysis had signi cantly higher levels of caregiver burden [26]. Findings further indicate that caregivers of frontotemporal lobar degeneration and caregivers of dementia with Lewy bodies patients experienced signi cantly more burden than caregivers of people with Alzheimer's [28].
Con icting ndings are also present in the literature [24,30]. In a comparative study, researchers found no signi cant difference in levels of burden reported by leukemia caregivers and cerebral palsy caregivers [24]. Comparing the impact of Alzheimer's disease and frontotemporal dementia on caregiver burden, study also indicates that there are no signi cant differences in levels of burden experienced in these two groups of caregivers [31]. Though evidence is needed to clarify mixed ndings in the literature, research is scarce. Overall, only a handful of studies have investigated the impact of patients' disease types on caregiving consequences such as caregiving time and caregiver burden [26,28,29,31,32]. Furthermore, considering that most of the available studies are either conducted in the cognitive impairment disease context (e.g., comparing different types of dementia conditions) [28,29] or only examined two to three types of distinctive medical conditions' in the analysis [26,31,32], these studies are limited in the diversity of insights they can offer. To bridge this gap, we aim to examine the impact of patient's disease types on caregiving time (i.e., caregiving duration and caregiving hours spent per week) and caregiver burden (i.e., caregivers' self-rated health, body mass index or BMI, and psychological distress), using data that examined a comprehensive list of patients' disease types (see Table 1).

Acute conditions
Health conditions that are severe and sudden in onset, such as a broken bone or an asthma attack.

Aging or related issues
Aging and related health issues not listed in the other categories above.

Multiple disease conditions
We coded this variable to represent patients with multiple health conditions mentioned above or different patients caregivers that have two or more distinctive conditions.

Other conditions
Medical conditions that are not listed above.

Unknown conditions
Conditions that are unknown to the caregiver and/or patients.

Study Design and Participants
This is a cross-sectional study using data from the Health Information National Trends Survey (HINTS, 2018) Version 5, Cycle 2, a nationally representative survey of U.S. adults aged ≥ 18 years (civilian and non-institutionalized) [33]. The survey is rst conducted by the National Cancer Institute (NCI) in 2003, aiming to gauge how Americans seek, share, and use cancer-related health information in their daily interactions [34]. The HINTS 5 (Cycle 2) survey used in this study was administered from January 26 to May 2, 2018. We utilized data from the 2018 HINTS survey because, by far, caregiver burden related variables were only included in this particular HINTS survey. Of all the 14,585 surveys mailed, 3,527 participants returned their questionnaires (response rate: 24.2%) [34]. As this study mainly focuses on the caregiver population, we used "Are you currently caring for or making health care decisions for someone with a medical, behavioral, disability, or other condition?" as the screening question to identify caregivers. A total of 458 (13.0%) participants who self-identi ed as caregivers (i.e., replied "yes" to the question) were included in the nal data analysis. Measures To adequately gauge the impact of patients' disease types on caregiving consequences, both direct and indirect consequences were evaluated in this study. Based on insights gained from the literature [35][36][37][38], direct consequence of caregiving is measured by caregiving time, in terms of caregiving duration and caregiving hours spent per week, while indirect consequences of caregiving are measured by caregiver burden in terms of caregivers' self-rated health, BMI, and psychological distress. Details of all measures used in this study are reported by the National Cancer Institute and available online [34].

Caregiving time
We adopted two measures to gauge caregivers' time spent on providing care to patients: caregiving duration and caregiving hours spent per week [34].
Caregiving duration. Participants were directed to "Think about the individual for whom you are currently providing the most care. About how long have you been providing care for this person?" and asked to choose one response from the 5 available choices (i.e., "Less than 30 days"=1, "1 to 6 months" = 2, "7 months to 2 years" = 3, "3 to 5 years" = 4, "More than 5 years"=5).
Caregiving hours spent per week. Participants were asked to "Think about the individual for whom you are currently providing the most care. About how many hours per week do you spend in an average week providing care?" and indicate hours they spent per week on offering care to the patients.
Psychological distress. Participants' psychological distress was measured with the four-item Patient Health Questionnaire (PHQ-4) [34,40]. Participants completed four items in response to "Over the past 2 weeks, how often have you been bothered by any of the following problems?": (a) Little Interest, (b) hopeless, (c) nervous, (d) worrying on scales from 0 (Not at all) to 3 (nearly every day). Based on available guidelines, the four items were subsequently summed [34,40]. The summed scores ranged from 0 to 12, with a higher score indicating a higher level of psychological distress [40].

Data Analyses
First, the characteristics of the participants were described using descriptive statistics (means accompanied with SDs or frequencies, as appropriate). Second, to explore the relationship between health outcome and independent variables, we used multivariate logistic regression method for analyzing the self-rated health variable and ordered logistic regressions for the BMI and psychological distress variables. In order to test the potential interaction, we generated several interaction terms by multiplying the categorical variables. Third, we calculated the mean scores of health indicators and caregiving burden of caregivers. In order to adjust for HINTS' multistage probability sampling design, a set of 50 jackknife replicate weights was applied to all analyses, estimating the model parameters for the U.S. population as a whole. Finally, to explore the association between caregiver burden, caregiver health outcomes, and independent variables, we used ordered logistic regression for analyzing self-rated health, BMI, psychological distress and applied linear regression analyses to examine caregiving duration and caregiving hours spent per week to answer the research question.
Odds ratios (ORs) or regression coe cients with CI were employed to depict the relationships between caregiver burden, caregiver health outcomes, and independent variables while controlling for covariates. Listwise deletion of each subjects were used for participants who provided invalid or missing responses for the dependent variables; sample sizes for each regression analysis are noted in Table 3 ranging from 3,256 to 3,264. All covariates with missing data were multiple imputed. All analyses were performed using Stata 14 [42] .

Results
The characteristics of participants who completed the HINTS questionnaire were illustrated in Table 3   95% con dence intervals in parentheses; *** p < 0.001, ** p < 0.01, *p < 0.05 Results on factors associated with self-rated health, BMI, and psychological distress were presented in Table 2. Both the interaction effect between caregiver and gender (OR = 0.63, 95% CI = 0.43-0.91) and the interaction effect between caregiver and employment status (OR = 0.61, 95% CI = 0.43-0.88) were statistically signi cant. That is, caregivers experienced greater psychological distress for men than women, and unemployed caregivers experienced greater psychological distress than employed caregivers.
As presented in   [38,43,44], and further highlights the need for intervention studies that could acknowledge and address caregiver burden. As approximately 21.3% of U.S. adults or 53.0 million Americans are informal caregivers who often do not have the adequate knowledge or skills needed to take care of the patients [18], whether or not successful interventions can deliver much-needed health solutions to these caregivers may impact not only the health of patients and caregivers, but also the health of the society [45,46].
Adding insights to the literature [38,[47][48][49], ndings of this study further suggest that caregivers who are female, older, had lower levels of education, physical inactivity, SES, unmarried, unemployed, not having health insurance experienced greater levels of caregiver burden. In other words, caregivers who belong to underserved communities are most likely to experience worse levels of caregiver burden. This worrying nding underscores the urgent need for timely and effective interventions that could alleviate these underserved populations' caregiver burden, such as intervention programs that are tailored to these caregivers' unique unmet needs [50][51][52]. As these disadvantaged caregivers are often di cult to reach via traditional intervention measures [53,54], researchers may consider using more agile and exible programs, such as local reach programs [55][56][57][58][59] or technology-based solutions [60,61], to deliver relevant and effective health solutions to these caregivers in a convenient and cost-effective fashion. Based on ndings of this study, researchers could also consider tailoring interventions based on patients and caregivers' needs and preferences [66][67][68][69]. Findings suggest though tailored and generic interventions both have potential to decrease caregivers' anxiety and depression symptoms, tailored interventions are more likely to induce sustained long-term health improvements in caregivers [66]. Evidence further indicates that dyad-based interventions can often take account of both patients' and caregivers' characteristics and preferences, and therefore have greater potential in improving patients and caregivers' coping skills and health outcomes than interventions designed solely for patients or caregivers [70][71][72]. Overall, ndings of this study ll considerable gaps in the literature and offer critical insights that could help further the research eld.

Limitations
First, the survey was cross-sectional in nature, which limits causal inference. We adopted a validated and reliable nationally representative survey, HINTS 5 Cycle 2, as our data source. While there are many bene ts in secondary data analysis of a nationally representative survey (e.g., cost-effective, time-e cient, and ability to generate ndings with greater replicability and comparability) [73][74][75], as the survey is not speci cally designed for this study, substantial research rigor could be lost due to the lack of purposeful survey design. To address these drawbacks, future research could develop research with tailored and robust survey design (e.g., longitudinal or mixed-methods) to increase research rigor.

Conclusion
This is the rst study that investigated the impact of patients' disease types on caregiving time and caregiver burden, using a comprehensive list of disease types surveyed in a nationally representative survey. Findings of this study show that underserved caregiver populations often subject to greater levels of caregiver burden in terms of self-rated health, BMI, and psychological distress. Results also indicate that great variations are found in the impact of patients' disease types on caregiving time and caregiver burden. These insights extend our understanding towards the relationship between patients' disease types and caregiving consequences. Furthermore, ndings of this study also underscore the need for healthcare researchers to adopt a nuanced approach in acknowledging and addressing caregiver burden, such as tailoring interventions based on patients and caregivers' needs and preferences. Overall, our study lls important gaps in the literature and offers great opportunities for better understanding the interaction between patients' disease characteristics and caregivers' burden and health outcomes.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.