The role of medical/nursing skills training in caregiver confidence and burden: A CanCORS study

Abstract

Background

Informal cancer caregivers provide essential support to cancer patients, including performing direct medical/nursing tasks, assistance with activities of daily living, and offering social support. We examined associations between receipt of medical/nursing skills training with caregiver burden, and the mediation of caregiving confidence on this relationship in a sample of caregivers of lung and colorectal cancer patients.

Methods

Caregivers identified by cancer patients in the Cancer Care Outcomes Research and Surveillance (CanCORS) Consortium completed a questionnaire assessing care provided, type of medical/nursing skills training received, burden (measured by the short-form Zarit Burden Interview), and confidence caring for the patient’s physical needs. Regression models adjusting for sociodemographic, caregiver, and care recipient characteristics assessed the relationship between training received and burden, and mediation analysis assessed the role of confidence in this relationship.

Results

641 caregivers performed some type of medical/nursing task, with 59% (n=377) reporting that they did not receive training for all care provided. Caregivers reported moderate levels of burden (mean summary score = 32.07, SD = 12.66, possible range 14–70), and lack of receipt of training was associated with greater levels of burden (b = 2.60; SE 0.98; p-=0.01). Confidence partially mediated the relationship between training and burden (Sobel’s t =1.90; p= 0.03).

Conclusions

As the number of cancer patients and caregivers increase, understanding how best to reduce caregiver burden is necessary. Skills training is a potential area for interventions, but research on how best to provide training for caregivers (i.e., content, mode of delivery, timing) is needed.

Introduction

Approximately 39.8 million U.S. adults are currently estimated to be serving as an informal caregiver for an adult patient with a chronic health condition, with 2.8 million estimated to be caring for adult cancer patients.¹ With improvements in diagnostics and therapy, many patients diagnosed with cancer are living longer and requiring substantial care outside of traditional healthcare settings. Informal caregivers, including family members and close friends, provide essential support to cancer patients at all phases of the illness trajectory,² and are often managing many complex patient needs. This support includes assistance with activities of daily living (ADLs), providing psychosocial support, and performing direct medical or nursing tasks (i.e. managing/attending appointments, administering treatments, managing symptoms). Caring for an individual with cancer can be burdensome and demanding for the caregiver.^3–6 In addition, caregiver well-being has implications for the patient’s perceived quality of care as well as quality of life.⁷

A growing number of interventions have been developed to address the burden placed on caregivers, and these interventions typically focus on providing psychosocial support to caregivers.^{3, 8} While research indicates that interventions aimed at cancer caregivers can reduce burden,³ several reviews have indicated a lack of interventions that provide training on the practical skills that caregivers need to assist their care recipients with ADLs.^{8, 9} A review of randomized controlled trials of caregiver interventions on cancer patient outcomes revealed mixed evidence of efficacy, with one conclusion indicating that building caregivers’ practical skills may be more effective for improving patient symptoms than general psychosocial support or educational information for caregivers.^{7, 10, 11} Van Houtven and colleagues created a framework for informal caregiver interventions based on a review of 121 studies, postulating that such interventions improve caregiver knowledge, confidence, and self-efficacy, and in turn benefit both caregivers and care recipients.¹² In order to further understand the role of skills training, the purpose of this study was to examine (1) the association between receipt of medical/nursing task skills training and level of caregiver burden, and (2) whether caregiver confidence in their ability to care for their care recipient’s physical needs mediates the relationship between training and caregiver burden. We hypothesized that more training would be associated with lower level of caregiver burden and that this effect would be mediated by caregiver confidence.

Methods

Data Source

This study used data from the Share Thoughts on Care Caregiver Study conducted by the Cancer Care Outcomes Research and Surveillance (CanCORS) consortium as an ancillary study to data collection on lung and colorectal cancer patients. Detailed information about the CanCORS study protocols is available elsewhere (https://cancors.org/public).^13–15 To summarize, the CanCORS consortium consisted of 7 study sites ascertaining patients from either cancer registries (5 sites) or health care systems (2 sites); the resulting sample was demographically representative within the CanCORS regions.¹⁶ Cancer patients (herein referred to as care recipients) in the core CanCORS survey nominated an informal/family caregiver to participate in the study; that caregiver was then invited to participate via mail with information about the study, a self-administered questionnaire, postage-paid return envelope, and a $20 incentive. Caregivers were identified either shortly after the baseline (n=825) or follow-up interviews (n=805) with the patient, and completed the questionnaire on average 7.3 (baseline) or 15.6 (follow-up) months after the care recipient’s diagnosis. Study procedures were approved by institutional review boards of all participating institutions.

Measures

Outcome Variable: Subjective Burden

Caregiver burden was assessed with a modified, 14-item version of the short-form Zarit Burden Interview (ZBI),^{17, 18} designed to reflect stresses experienced by caregivers. The original version was most consistently used in dementia caregiving research¹⁷ but has been used in cancer caregiving research as well.¹⁹ Example questions from the ZBI include, “Do you feel that, because of the time you spend with your relative, you don’t have enough time for yourself?” and, “Do you feel stressed between caring for your relative and trying to meet other responsibilities for your family or work?” Respondents were asked to respond to items with a 1–5 Likert scale, where 1 is never, and 5 is nearly always. Each item is then summed for a total burden score (possible range 14–70), with higher scores reflecting greater burden.¹⁸

Predictor Variable: Receipt of Training

Receipt of needed training was assessed using two survey questions. First, respondents were asked whether they performed tasks in each of the following areas: medical tasks, medication administration, and management of symptoms, with response options of “Yes”, “No”, and “Not Needed.” A follow-up question then asked if the caregiver received training in each of the areas, with the same response options. After matching the type of care provided (medical tasks, medication administration, and management of symptoms) with the specific training received or not received, we then identified 4 groups: (1) Did not perform medical/nursing tasks; (2) Training received for all medical/nursing tasks performed; (3) Training not received for all medical/nursing tasks performed; and (4) Performed tasks but reported not needing training.

Mediating Variable: Confidence

Confidence in the caregiver’s ability to care for the care recipient’s physical needs was assessed with the following item, taken from the Caregiver Preparedness Scale²⁰: How confident are you that you can take care of your Care Recipient’s physical needs? Response options ranged from “not at all confident,” “a little confident,” “somewhat confident,” “very confident,” to “extremely confident.” This variable was positively skewed, resulting in small sample sizes in the five categories. We therefore collapsed this variable into very or extremely confident versus all other responses.

Caregiver characteristics

Caregivers reported their current age, sex, race/ethnicity, relationship to care recipient, marital status, hours per day providing care, income, number of persons in household supported by income, and employment status (works for pay vs. does not work for pay). We created a 4-level summary variable crossing caregiver gender with employment status to account for gender differences in employment status. We also created a poverty level variable that categorized respondents as being at/above versus below poverty level, based on income and number of persons supported by income.

Care recipient characteristics

Care recipient sociodemographic variables, including age, sex, race/ethnicity, type of cancer, and Surveillance, Epidemiology, and End Results (SEER) summary stage at diagnosis were obtained through the CanCORS core data.

Final Sample Selection

Out of the 1630 caregivers in the dataset, 338 were excluded due to missing data on covariates (Figure 1), resulting in 1292 caregivers with complete data. We then included caregivers who reported performing some type of medical/nursing task (medical tasks, medication administration, and management of symptoms) and excluded those who did not (Group 1 shown on Figure 1). To ascertain training received for those who performed medical/nursing tasks, we excluded those who indicated that they performed tasks but did not need training in any area (group 4 shown on Figure 1). Our final analytic sample included groups 2 and 3, comprised of 641 caregivers.

Figure 1.

Sample selection of cancer caregivers from the CanCORS Caregivers Study

Statistical analyses

Bivariate and multivariable regression

Utilizing available case analysis, we examined caregiver and care recipient characteristics with descriptive statistics and assessed bivariate associations between training, confidence, and caregiver/care recipient characteristics and burden. We conducted multivariable regression analyses between 1) training and burden, 2) confidence and burden, and 3) training and confidence. Models controlled for caregiver age, race/ethnicity, timing of survey (baseline or follow-up), sex, relationship to care recipient, distance to care recipient, poverty level, hours of care provided per day, and confidence; and care recipient race/ethnicity, age, sex, cancer stage, and cancer type. Only characteristics significantly associated with outcome variables in bivariate regressions and those used in similar analyses with these data7 were included as control variables. We then tested for the potential mediation of confidence on the relationship between training received and burden using Sobel’s mediation test,²¹ controlling for the same covariates. Univariate and multivariable logistic regression analyses were performed using SAS 9.4 (SAS Institute, Cary NC), and statistical significance declared at p<0.05.

Results

Respondent and corresponding care recipient characteristics

Table 1 depicts the characteristics of cancer caregivers and care recipients (n=641), including caregivers identified at baseline (n=365) and follow-up (n=276). The majority of caregivers were between 51–65 years of age at survey (41.5%), received some college or greater (66.5%), female (80.5%), and Non-Hispanic White (72.5%). Most caregivers were the spouse of the care recipient (62.6%), provided care between 1–6 hours per day (56.5%), and slightly more than half worked for pay aside from their caregiver role (51.2%). Our sample of caregivers provided care for approximately equal numbers of lung and colorectal cancer patients, who were mostly older (52.4%), male (66.5%), and more than half had stage 3 or 4 cancer (59%).

Table 1.

Patient and Informal Caregiver Characteristics

Informal Caregiver Characteristics (N=641)	N	%
Age
20–50	189	29.49
51–65	266	41.50
Over 65	186	29.02
Education
High School or Less	213	33.49
College or Greater	423	66.51
Gender
Female	516	80.50
Male	125	19.50
Race/ethnicity
White	465	72.54
African American	95	14.82
Other	81	12.64
Marital Status
Married	500	78.00
Not Married	141	22.00
Poverty
Household on or above poverty threshold	558	87.05
Household below poverty threshold	83	12.95
Relationship to Care Recipient
Spouse	401	62.56
Child	90	14.04
Other	150	23.40
Hours per day CG provides care
Less than 1 hour a day	126	19.66
1–2 hrs a day	166	25.90
3–4 hrs a day	113	17.63
5–6 hrs a day	83	12.95
7–8 hrs a day	30	4.68
More than 8 hrs a day	123	19.19
Employment Status
Works for pay	328	51.17
Does not work for pay	313	48.83
Corresponding Care Recipient Characteristics	N	%
Age
0–64 years	305	47.58
65+	336	52.42
Gender
Female	215	33.54
Male	426	66.46
Race/ethnicity
White	479	73.79
African American	88	13.73
Other	74	11.54
Marital Status
Married	473	73.79
Not Married	168	26.21
Cancer Site
Lung	322	50.23
Colorectal	319	49.77
SEER Summary Stage
Stage 0 or Local	150	23.40
Stage 2	113	17.63
Stage 3 or Regional	212	33.07
Stage 4	166	25.90

The majority of caregivers reported that they managed care recipient symptoms (n=504; 81.82% of 666 caregivers performing medical/nursing tasks), while more than half administer medications (n=389; 62.8%) and far fewer changed bandages (n=173; 28.1%). Most of the caregivers in this study reported that they did not receive training for all of the care provided (n=377; 56.6% of caregivers performing medical/nursing tasks). There were no significant in receipt of training based on differences in sociodemographic characteristics or timing of survey (baseline or follow-up). The majority of respondents (n=449; 70.1%), however, reported being very or extremely confident in their ability to care for the care recipient’s physical needs. In addition, caregivers reported modest levels of burden (mean summary score=32.1; standard deviation= 12.7; range 14–70).

In bivariate analyses, not receiving needed training was significantly associated with less confidence (unstandardized b =−7.1, p<0.0001) and greater reported burden (b=2.59, p=0.01) (Table 2). Caregivers who reported greater levels of confidence reported significantly less burden (b=−7.1, p<0.0001). Those who provided care for a greater number of hours per day also reported greater burden (b=0.94, p=0.001). In addition, as compared to white caregivers, African American caregivers (b=−4.05, p=0.01), as well as those who provided care for African American care recipients (b=−3.13, p=0.03) reported less burden than those who cared for white care recipients.

Table 2.

Associations with Caregiver Burden (N=641)

	Unadjusted		Adjusted
	b	p value	b	SE	p value
Not all needed training received	2.59	0.01	2.60*	0.98	0.01
Caregiver Confidence	−7.1	<0.0001	−6.42**	1.07	<0.0001

^*Adjusted for covariates: Caregiver characteristics: age, relationship to recipient, distance from recipient, gender-employment, poverty level, race/ethnicity, hours of care provided per day, survey time (baseline or follow-up), confidence; Care recipient characteristics: cancer type, gender, age, race/ethnicity, cancer stage

^**Adjusted for covariates: Caregiver characteristics: age, relationship to recipient, distance from recipient, gender-employment, poverty level, race/ethnicity, hours of care provided per day, survey time (baseline or follow-up); Care recipient characteristics: cancer type, gender, age, race/ethnicity, cancer stage

After adjusting for covariates, those who did not receive all needed training also reported less confidence (b=−0.08, SE=0.04, p=0.04), and greater burden (b=3.09, SE=1.01, p<0.01). The association between training and burden remained even after controlling for confidence, but the effect was attenuated (b=2.60, SE=0.98, p=0.01). Table 2 displays unadjusted and adjusted associations between training and burden and confidence and burden. The adjusted mean burden was 33.13 for those not receiving all training, compared to those who received all training (mean= 30.55) (Figure 2). Increased caregiver confidence was significantly associated with less burden, after adjusting for previously stated covariates and training (b=−6.42, SE=1.07, p<0.0001).

Figure 2.

Adjusted burden summary scores with standard errors

We hypothesized that the association of caregiver training on burden might be at least partially mediated through caregiver confidence. After controlling for covariates, confidence did partially mediate the relationship between training and reported burden (Figure 3). The figure presents the coefficients and associated standard errors for the mediational model. There was a significant effect of training on burden mediated through confidence (Sobel test statistic=1.90, p=0.03).

Figure 3.

Coefficients and associated standard errors for caregiver burden are presented. In the model, confidence partially mediates the relationship between caregiver training and burden*

*Adjusted for covariates: Caregiver characteristics: age, relationship to recipient, distance from recipient, gender-employment, poverty level, race/ethnicity, hours of care provided per day, survey time (baseline or follow-up), confidence; Care recipient characteristics: cancer type, gender, age, race/ethnicity, cancer stage

Conclusions

This study assessed the relationships between training, confidence, and burden among caregivers of patients with lung and colorectal cancer in a large, multisite sample. This study extends the work of van Ryn and colleagues,²² which detailed the care tasks and needs of cancer caregivers within this sample. Caregivers reported moderate levels of burden, consistent with previous studies.^{3, 8} We also found that African American caregivers had lower levels of burden than White caregivers. These results support previous work demonstrating racial/ethnic minority differences in the impact of caregiving.²³ Researchers hypothesize that African American caregivers may experience less burden in part because of higher levels of intrinsic motivation based on familial norms, higher levels of religious commitment, or greater availability of informal support.²³

More than 58% of our sample did not receive training for all of the tasks they performed. Furthermore, lack of needed training was associated with higher reported levels of burden, and confidence partially mediated this relationship. It is possible that not receiving needed training to perform medical/nursing tasks may impact caregiver’s confidence, thereby increasing burden. While confidence in ability to perform physical tasks is important, our results indicate that the training also has a direct impact on perception of burden. Indeed, while training alone does not guarantee competence, it may improve actual skills as well as perceived confidence to handle caregiving responsibilities.

Our findings confirm earlier work showing that receipt of training is associated with caregiver mental health, although previous studies have been mostly focused on caring for dementia patients and those with neurocognitive disorders.^24–26 Studies specifically exploring medical/nursing skills training for cancer caregivers and the effect on subsequent burden are lacking, however. A meta-analysis,³ as well as a more recent systematic review²⁷ indicate that skills training interventions are less prevalent than psychoeducational interventions, and skills interventions focused mostly on coping and problem-solving skills. A study of cancer caregivers within the Veteran’s Administration system did find that medical/nursing skills training improved caregiver self-efficacy, although this was limited to colorectal cancer care recipients.²⁸ Our study includes caregivers of lung and colorectal cancer patients, which represent the second and third leading sites of new cancer cases.²⁹ In addition, care recipients represented all cancer stages, and it is important to note that cancer stage could impact level of care required by caregivers. Our results point to the need for empirical studies that examine whether medical/nursing skills training for cancer caregivers can have an impact on caregiver outcomes over time.

The results of this study are timely and important, especially considering the increasing attention to the value of the range of support provided by caregivers to their care recipients. Recent data on caregiving generated by the National Alliance for Caregiving¹ and a study on family caregiving for older adults published by the National Academies of Sciences, Engineering, and Medicine (NASEM)³⁰ highlight not only the ubiquity of informal caregivers but also the expectations and burden placed on them. The importance of the caregiver role is reflected in new state-based legislation under the Caregiver, Advise, Record, and Enable (CARE) Act. The CARE Act requires hospitals to ascertain/record the identity of a family caregiver in the medical record of the recipient, as well as alert and involve caregivers in discharge planning.³¹ Taken together, these reports and legislation underscore the potential role that appropriate training can play in reducing caregiver burden.

While results of the current study indicate that skills training is a potentially high impact area in which to intervene to reduce caregiver burden, the delivery of such interventions warrants further research. Previous work focuses on training in areas of coping, communication, and problem solving,³ and there is little research on the appropriate delivery of training for medical/nursing tasks. Such delivery includes specific components (e.g., administering medications, changing dressings, monitoring for side effects), mode of delivery (e.g., in-person, virtual/online, telephone), timing (eg., at end of treatment/transition to home, hospitalization discharge, or other), and frequency/interval of skills training. While the CARE Act requires involvement of the caregiver in discharge planning, which could include such skills training, not all cancer patients are treated in inpatient settings.³² In fact, the delivery of the majority of cancer therapies have moved to outpatient and community settings, making the need for similar policies important here as well. In addition, ²⁵ caregivers (3.9%) of our overall sample reported that they provided care, but did not need training, signaling that because of varying health literacy levels and experience, training may not be necessary for all. Future research should focus on flexible interventions that are tailored to the needs of the specific cancer caregiver. Such research may benefit from preliminary mixed methods work prior to randomized trials.

The findings of this study should be considered in light of some limitations, including the fact that this study was cross-sectional and that receipt of training was self-reported and does not reflect quality of training. Follow-up research should include longitudinal studies including medical record audits to confirm training received and measures of quality. In addition, the measure of confidence was a single item originally intended to be part of a larger scale, the Caregiver Preparedness Scale.²⁰ The CanCORS survey tools in total, however, were pilot tested and showed results consistent with previous population-based studies of cancer patients.¹⁵ Strengths of the current study include the large, multi-site data, as well as the specific matching between care provided by the caregiver and training received. In addition, these results point to important implications for caregivers of lung and colorectal cancer patients with varied stages of disease, a research area that has previously been limited.

In summary, we found that approximately 40% of the eligible caregivers in the dataset performed some type of medical/nursing task, however more than half of these caregivers did not receive training for all tasks they performed. Our findings indicate that caregiver medical/nursing skills training is associated with reduced burden, a relationship which is partially mediated by confidence in ability to provide physical care for the care recipient. Finding ways to efficiently and effectively integrate caregiving medical/nursing skills training into diverse delivery systems and within the constraints of competing staff demands is an imperative. Researchers and clinicians should work together in concert with patients and caregivers to tailor task training interventions to facilitate caring for patients with complex demands of a cancer diagnosis, treatment, and long-term effects.