Abstract
Purpose
Financial strain and stressful life events can constrain open communication within families. A cancer diagnosis can bring heightened emotional stress and financial strain for most cancer patients and their families. We evaluated how level of comfort and willingness to discuss important but sensitive economic topics affected longitudinal assessments of family relationships, exploring both within-person and between partner effects over two years after a cancer diagnosis.
Methods
A case series of hematological cancer patient-caregiver dyads (n=171) were recruited from oncology clinics in Virginia and Pennsylvania and followed for 2 years. Multi-level models were developed to examine the associations between comfort discussing economic aspects of cancer care and family functioning.
Results
Broadly, caregivers and patients who were comfortable discussing economic topics reported higher family cohesion and lower family conflict. Dyads’ assessments of family functioning were influenced both by their own and their partners level of communication comfort. Overtime caregiver but not patients perceived a significant decrease in family cohesion.
Conclusions
Efforts to address financial toxicity in cancer care should include examination of how patients and families communicate as unaddressed difficulties can have detrimental effects on family functioning in the long term. Future studies should also examine whether the prominence of specific economic topics, such as employment status, varies depending on where the patient is in their cancer journey.
Implications for Cancer Survivors:
In this sample, cancer patients did not perceive the decline in family cohesion that was reported by their family caregiver. This is an important finding for future work that aims to identify the timing and nature to best intervene with caregiver supports to mitigate caregiver burden that may negatively impact long term patient care and QoL.
Keywords: cancer caregivers, cohesion, conflict, employment, insurance, treatment costs, cross-partner effects
INTRODUCTION
Open communication among families can become strained by stressful life events and financial difficulties (1,2). The financial models of the U.S. health insurance and cancer care delivery systems means a cancer diagnosis brings both heightened emotional stress and financial strain for most cancer patients and their families. Research has shown that cancer patient and family caregivers mental and physical health outcomes are interconnected (3–6); there is much work that has characterized vulnerabilities and effects of financial burden among patients and their families (2,7–9). What is unclear is whether willingness vs. reluctance to be communicatively open about the financial aspects of cancer care, including the impact of the disease and treatment on the employment status of caregivers and patients, is associated with changes in patient and caregiver perceptions of their family relationships.
Financial burden (or toxicity) in the cancer survivorship literature is defined as “the combined impact on a family of objective material burden and subjective financial distress” (10). Factors contributing to financial burden can be direct such as out-of-pocket costs associated with cancer treatment or indirect such as reduced or lost income, limited or lost access to benefits and health insurance, or other changes to employment status (11,12). Families with lower income and unmarried caregivers report greater financial strain (13). Characteristics of cancer patients who are most at-risk of experiencing serious employment disruption are those who are younger, female, uninsured and work in hourly positions (7,9). Family caregivers report difficulty managing work and caregiving responsibilities (2), with caregiving strain correlating to conflict experienced at work (8). Changes in employment can impact financial stability for patients and their families making discussions about health insurance, medical bills and cancer treatment potentially fraught. It is not surprising that financial burden is associated with decreased QoL, depression and anxiety among patients and family caregivers (2,13).
The present analysis was undertaken to understand how level of comfort and willingness to discuss important but sensitive economic topics (employment, health insurance and treatment costs) affects longitudinal assessments of family relationships by the caregiver and the patient, exploring both within-person and cross partner effects over two years after a cancer diagnosis. Costs of cancer care and financial toxicity represent additional stressors that can further undermine cancer patients and caregivers social and emotional resources leading to poorer QOL outcomes. Understanding whether the ways in which patients and caregivers approach these difficult conversations results in changes and challenges to future family functioning is needed. The hypothesis was that discomfort discussing economic topics would be associated with poorer perceived family functioning, and that each partner’s comfort would have an independent effect on the other such that each individuals’ own comfort communicating would be associated with the dyad’s perception of family functioning.
METHODS
Participants
A total of 171 patients with cancer were recruited to this longitudinal study. Study participants were adults ( > 18 years) who were diagnosed with 1 of 3 hematological cancers (multiple myeloma [MM], Non-Hodgkin’s Lymphoma [NHL], or chronic lymphocytic leukemia [CLL]) at any stage of disease, who were employed immediately before their diagnosis, and had begun initial systemic treatment within the previous six months. Hematological cancers were chosen because less research has focused on the socioeconomic factors and patterns of family functioning among patients with these cancers as compared to solid tumors, and to minimize the heterogeneity in treatment trajectories (e.g., relapsing and remitting) that differentially impact employment, which was the primary focus of the study. Patients also needed to identify a primary informal (unpaid) caregiver defined as the person responsible for providing the majority of care. Both patient and caregiver needed to consent to the study for inclusion into the sample. A complete case series of patients were identified who fit the study parameters within the study recruitment phase. A telephone screen was conducted to confirm employment and caregiver parameters for final eligibility. Patients were excluded if they had not been employed immediately prior to diagnosis, did not have an informal caregiver, had cognitive impairments, were receiving hospice care, or only spoke a language other than English. Caregivers had to be adults (>18 years), cognitively able to participate and not being paid to provide patient care. Approval from the institutional review board for human subjects research was obtained from the IRB at each participating institution (IRB # HM13074).
Recruitment
Patients with cancer were recruited from 8 oncology clinics (both academic and community clinics) located in Pennsylvania and Virginia. Patients were identified using electronic medical records. Preliminary eligibility was confirmed by clinic staff (cancer diagnosis, no cognitive impairment or hospice care) and these patients were mailed invitation letters followed by a telephone call from research staff. Interested patients provided contact information for an informal caregiver and baseline interviews were scheduled once eligibility was determined. Of the 818 individuals contacted, approximately 21% declined participation, 57% were ineligible (primarily due to unemployment or the absence of a caregiver), and 21% provided informed consent. The final sample size was 171 patient-caregiver dyads (n= 342 individuals).
Interviews.
Patients and caregivers were separately interviewed every 6 months for 2 years. Between 2013 and 2018, semi-structured interviews were conducted either in participants’ homes or in a private room at one of the cancer centers. A small number of follow-up interviews were completed over the phone, for example for dyads who moved during the study period. Standardized scales were used to collect demographic information and cancer communication concordance. We extracted cancer diagnosis, stage and treatment dates from the patients’ medical records. Baseline interviews were approximately 90 minutes and follow-up interviews lasted 30-60 minutes. Interviews were audio recorded and performed by trained graduate-level research staff, but the participants self-administered standardized scales. Participants were provided a $50.00 honorarium per interview. The current analysis is based on baseline, wave 2 and wave 4.
Measures
Demographics included patient and caregiver age, sex, race, education, annual household income, marital status and caregiver relationship (spouse, parent, offspring/sibling, friend). Cancer stage, cancer type and date of first systemic treatment were extracted from the patient’s medical record.
Family Environment Scale (FES)(14).
The FES is composed of 90 true-false items measuring social and environmental characteristics of families. The Relationship Dimension consists of 27 items that assess three domains, two of which were collected from our participants: Family Cohesion or the degree of commitment, help, and support family members provide for each other (e.g., “There is a feeling of togetherness in our family.”) and Conflict, or the amount of openly expressed anger, criticism, and fighting among family members (e.g., “We fight a lot in our family.”). Raw scores range from 0 to 9 on each subscale; higher scores reflect greater cohesion/conflict. This instrument has moderate to high internal consistency and acceptable test-retest reliabilities (14). FES was collected at baseline, wave 2 and wave 4.
Comfort Discussing Financial Topics.
Caregivers and patients were asked to rate their own comfort discussing a variety of employment and finance topics related to the patient’s cancer using an item created for this study. A Likert scale was used ranging from 1 (not at all comfortable) to 5 (extremely comfortable) and used as a continuous variable in analyses. Topics included comfort discussing patient’s employment, caregiver’s employment, cost of treatment and health insurance.
Statistical Analysis
Descriptive information on demographic characteristics is presented for patients and caregivers on age, sex, race, education, income and marital status. Both within-person and within-dyad (i.e., cross-partner) bivariate correlations are presented for patient and caregiver comfort discussing employment, cost of treatment and health insurance with cohesion and conflict at each of the three waves.
The focus of the analysis was to understand how communication about potentially emotionally charged topics affected assessments of family cohesion and family conflict by the two major partners experiencing the cancer journey: the caregiver and patient dyad. To accomplish this, the data were analyzed using the multivariate two-level model for longitudinal data (15–17). This model enabled the simultaneous estimation of the unique effects for each dyad member as well as any cross-partner effects while controlling for interdependencies in the data. The inter-dyad effects, if not adequately modeled, may lead to misleading or incomplete understanding of the presence of cohesion and conflict in these dyads (18). To incorporate all these elements, the analysis combined a longitudinal model for individuals with a cross-sectional model for matched pairs.
Separate sets of models were tested for each of the FES subscales with both patient and caregiver independently rating on family conflict and cohesion. Separate models were run for patient and caregiver estimates of conflict and cohesion to facilitate examination of within-person and cross-partner communication effects. Each of these dependent variables were first estimated alone. For both the cohesion and conflict models, patient and caregiver scores were estimated capturing initial status and change over time (linear). Next, demographic variables were added including education, sex, and age. Using these as the base model for each dependent variable, a series of three additional models were run to explore the effect of caregiver and patient reported comfort discussing financial topics as follows: patient employment (model 2), caregiver employment (model 3) treatment costs (model 4) and health insurance (model 5). All models contained both caregiver and patient reported communication comfort with these topics to evaluate the participant’s own cohesion and conflict scores (within-person effects) and their partner’s cohesion and conflict scores (cross-partner effects). The models were estimated using full information maximum likelihood (FIML) via HLM 6.04(19) using all available data from all caregivers and patients.
Results
Table 1 presents descriptive statistics for the sample. Both patients and caregivers were 54 years on average, and the majority were married (75% and 83% respectively) and white (73%). Patients were primarily male (63.2%) and caregivers female (74.3%). Most participants had a high school diploma or some college and were in a higher income bracket. The most common diagnosis was NHL (63.7%) followed by MM (24.6%) and CLL (11.7%). Twenty-one (12.9%) patients died over the course of this study. Bivariate associations among participant comfort discussing financial topics associated with cancer care and family functioning at three waves (baseline, 12 and 24 months) are presented in Table 2. In general, greater comfort discussing each topic was associated with greater family cohesion and less family conflict, but there was variation in which of these relationships were significant at each timepoint.
Table 1.
Demographics: Total Sample
| Total Sample N=171 |
|
|---|---|
| Patient Age (M (sd)) | 54.2 (12.1) |
| Caregiver Age (M (sd)) | 54.2 (12.9) |
| Sex | N (%) |
| Patient Male | 108 (63.2) |
| Caregiver Male | 44 (25.7) |
| Race | |
| Patient Race | |
| African American | 37 (21.6) |
| Caucasian | 123 (72.0) |
| Other | 11 (6.4) |
| Caregiver Race | |
| African American | 39 (22.9) |
| Caucasian | 125 (73.5) |
| Other | 6 (3.5) |
| Education | |
| Patient Education | |
| < HS /HS/GED | 35 (20.5) |
| Some college | 59 (34.5) |
| Bachelors degree or higher | 77 (45.0) |
| Caregiver Education | |
| < HS /HS/GED | 41 (29.8) |
| Some college | 52 (30.4) |
| Bachelors degree or higher | 68 (39.1) |
| Income | |
| Patient Income | |
| <15-29,999 | 23 (13.4) |
| 30-59,999 | 20 (11.7) |
| 60-100,000+ | 118 (69.0) |
| Missing | 10 (5.8) |
| CG Income | |
| <15-29,999 | 20 (11.7) |
| 30-59,999 | 21 (12.3) |
| 60-100,000+ | 119 (69.6) |
| Missing | 11 (6.4) |
| Marital Status | |
| Patient Married | 129 (75.4) |
| Caregiver Married | 142 (83.0) |
| Relationship | |
| Spouse/partner | 126 (73.7) |
| Parent | 15 (8.8) |
| Adult child/Sibling | 22 (12.9) |
| Friend/Other | 8 (4.7) |
| Diagnosis | |
| NHL | 109 (63.7) |
| NHL Stage | |
| 0 | 9 |
| 1-2 | 33 |
| 3-4 | 50 |
| Unknown/Unclassified | 17 |
| CLL | 20 (11.7) |
| CLL Stage | |
| 0 | 3 |
| 1-2 | 6 |
| 3-4 | 5 |
| Unknown/Unclassified | 6 |
| MM | 42 (24.6) |
| MM Stage | |
| 0 | 2 |
| 1-2 | 10 |
| 3-4 | 19 |
| Unknown/Unclassified | 11 |
Table 2.
Bivariate Correlations Cohesion and Comfort for PT and CG with Topics Over Time
| CG | PT | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| PT Employment | CG Employment | Tx Costs | Health Insurance | PT Employment | CG Employment | Tx Costs | Health Insurance | ||
| Baseline | |||||||||
| PT | Cohesion | −0.05 | −0.04 | −0.11 | 0.10 | 0.23 | 0.21 | 0.35 | 0.16 |
| Conflict | −0.08 | −0.16 | −0.14 | −0.26 | −0.15 | −0.04 | −0.15 | 0.01 | |
| CG | Cohesion | 0.09 | 0.20 | 0.10 | 0.08 | 0.18 | 0.09 | 0.10 | 0.10 |
| Conflict | −0.16 | −0.07 | 0.05 | −0.06 | −0.11 | −0.10 | 0.04 | −0.03 | |
| T2 | |||||||||
| PT | Cohesion | 0.25 | 0.24 | 0.10 | 0.06 | 0.19 | 0.18 | 0.24 | 0.32 |
| Conflict | −0.22 | −0.16 | −0.18 | −0.07 | −0.19 | −0.22 | −0.08 | −0.05 | |
| CG | Cohesion | 0.28 | 0.34 | 0.15 | 0.18 | 0.23 | 0.05 | −0.01 | 0.08 |
| Conflict | −0.25 | 0.001 | −0.10 | −0.09 | −0.28 | −0.13 | 0.08 | 0.08 | |
| T4 | |||||||||
| PT | Cohesion | 0.04 | −0.03 | 0.05 | −0.11 | 0.20 | 0.21 | 0.25 | 0.33 |
| Conflict | −0.10 | −0.05 | −0.43 | −0.18 | −0.17 | −0.03 | 0.22 | −0.05 | |
| CG | Cohesion | 0.21 | 0.37 | 0.06 | 0.20 | 0.05 | −0.02 | 0.18 | 0.24 |
| Conflict | −0.18 | −0.17 | −0.25 | −0.34 | −0.22 | −0.08 | 0.12 | −0.10 | |
Multilevel Models for Patient and Caregiver Cohesion and Conflict
Significant relationships between comfort discussing financial topics and individual ratings of family cohesion and family conflict are depicted in Figure 1. The within-person significant effects are represented using solid line arrows; the cross-partner significant effects are represented using dashed line arrows. All supporting data from the models is presented in Tables 3 and 4.
Figure 1. Significant Associations of Patient and Caregiver Cohesion and Conflict with Comfort Discussing Topics.
Note: Solid lines are within-person effects; dashed lines are cross-partner effects
Table 3.
Multi-level Models for Patient Cohesion and Caregiver Cohesion
| Patient Perceived Family Cohesion | ||||||
|---|---|---|---|---|---|---|
| Model 1 | Model 1a | Model 2 | Model 3 | Model 4 | Model 5 | |
| β (se) | β (se) | β (se) | β (se) | β (se) | β (se) | |
| Intercept | 7.62 (0.12) | 8.08 (0.45) | 5.75 (0.68) | 6.06 (.88) | 5.46 (.81) | 6.15 (1.05) |
| Time | 0.04 (0.03) | 0.037 (0.03) | 0.032 (0.04) | 0.025 (0.05) | 0.12 (0.05)* | 0.04 (0.05) |
| Education | −0.025 (0.06) | 0.014 (0.06) | 0.08 (.06) | 0.043 (0.07) | 0.035 (0.07) | |
| Sex (Male) | 0.0021 (0.18) | −0.003 (0.16) | 0.06 (0.20) | 0.08 (0.21) | −0.13 (0.22) | |
| Age | −0.006 (.007) | 0.004 (.007) | −0.001 (.0008) | 0.003 (0.009) | −0.001 (0.009) | |
| CG Comfort Discussing PT Employment | 0.10 (0.09) | |||||
| PT Comfort Discussing PT Employment | 0.30 (0.09)** | |||||
| CG Comfort Discussing CG Employment | 0.027 (0.12) | |||||
| PT Comfort Discussing CG Employment | 0.28 (0.11)** | |||||
| CG Comfort Discussing Cost | −0.04 (0.10) | |||||
| PT Comfort Discussing Cost | 0.42 (0.10)*** | |||||
| CG Comfort Discussing Health Ins | 0.051 (0.14) | |||||
| PT Comfort Discussing Health Ins | 0.30 (0.13)** | |||||
| Chi-Sq | 66.63 | 65.33 | 13.03 | 8.54 | 15.3 | 18.9 |
| −2 Log Likelihood | 1463.6 | 1462.4 | 847.4 | 761.4 | 638.4 | 686.4 |
| AIC (Smaller is Better) | 1471.6 | 1476.4 | 865.4 | 779.4 | 656.4 | 704.4 |
| AICC (Smaller is Better) | 1471.7 | 1476.7 | 866.1 | 780.2 | 657.4 | 705.3 |
| BIC (Smaller is Better) | 1484.2 | 1498.4 | 892.9 | 805.7 | 682.7 | 730.7 |
| Caregiver Perceived Family Cohesion | ||||||
| Model 1 | Model 1a | Model 2 | Model 3 | Model 4 | Model 5 | |
| β (se) | β (se) | β (se) | β (se) | β (se) | β (se) | |
| Intercept | 7.89 (.14) | 8.62 (0.52) | 5.65 (1.29) | 4.57 (1.45) | 8.23 (1.68) | 3.67 (2.03) |
| Time | −0.075 (0.03)* | −0.077 (0.03)* | −0.23 (.10)* | −0.22 (.10)* | −0.21 (0.15) | −0.13 (0.14) |
| Education | 0.028 (0.07) | −0.010 (.10) | 0.00 (.09) | −0.07 (0.14) | −0.07 (0.12) | |
| Sex (Male) | −0.43 (0.21) | −0.42 (.30) | −0.51 (0.30) | −0.22 (0.42) | −0.43 (0.39) | |
| Age | −0.010 (.009) | 0.003 (.012) | 0.0085 (0.012) | 0.0025 (0.02) | 0.006 (0.02) | |
| CG Comfort Discussing PT Employment | 0.42 (.15)* | |||||
| PT Comfort Discussing PT Employment | 0.30 (.14)* | |||||
| CG Comfort Discussing CG Employment | 0.76 (.20)*** | |||||
| PT Comfort Discussing CG Employment | 0.04 (0.18) | |||||
| CG Comfort Discussing Cost | 0.25 (0.20) | |||||
| PT Comfort Discussing Cost | −0.13 (0.20) | |||||
| CG Comfort Discussing Health Ins | 0.56 (0.25)* | |||||
| PT Comfort Discussing Health Ins | 0.46 (0.21)* | |||||
| Chi-Sq | 100.92 | 93.91 | 14.83 | 6.04 | 7.04 | 7.57 |
| −2 Log Likelihood | 1528.8 | 1522.1 | 508.9 | 437.3 | 312.2 | 415.6 |
| AIC (Smaller is Better) | 1536.8 | 1536.1 | 526.9 | 455.3 | 330.2 | 433.6 |
| AICC (Smaller is Better) | 1536.9 | 1536.4 | 528.3 | 456.9 | 332.7 | 435.5 |
| BIC (Smaller is Better) | 1549.4 | 1558.1 | 552.5 | 479.8 | 354.0 | 457.9 |
Table 4.
Multi-level Models for Patient Conflict and Caregiver Conflict
| Patient Perceived Family Conflict | ||||||
|---|---|---|---|---|---|---|
| Model 1 | Model 1a | Model 2 | Model 3 | Model 4 | Model 5 | |
| β (se) | β (se) | β (se) | β (se) | β (se) | β (se) | |
| Intercept | 1.92 (0.14) | 1.61 (.61) | 3.27 (0.93) | 2.49 (1.14) | 3.83 (0.95) | 3.97 (1.31) |
| Time | −0.008 (0.03) | −0.007 (0.03) | −0.010 (0.04) | 0.033 (0.05) | 0.001 (0.05) | −0.042 (0.05) |
| Education | −0.009 (0.08) | 0.047 (0.08) | −0.037 (0.09) | −0.04 (0.09) | −0.027 (0.09) | |
| Sex (Male) | 0.26 (0.25) | 0.34 (0.25) | 0.058 (0.30) | 0.23 (0.27) | 0.39 (0.29) | |
| Age | 0.003 (0.01) | −0.001 (0.01) | 0.006 (0.01) | 0.002 (0.01) | −0.017 (0.01) | |
| CG Comfort Discussing PT Employment | −0.25 (0.11)* | |||||
| PT Comfort Discussing PT Employment | −0.12 (0.11) | |||||
| CG Comfort Discussing CG Employment | −0.16 (0.15) | |||||
| PT Comfort Discussing CG Employment | −0.06 (0.14) | |||||
| CG Comfort Discussing Cost | −0.43 (0.11)** | |||||
| PT Comfort Discussing Cost | −0.08 (0.11) | |||||
| CG Comfort Discussing Health Ins | −0.31 (0.15)* | |||||
| PT Comfort Discussing Health Ins | 0.027 (0.15) | |||||
| Chi-Sq | 163.51 | 162.87 | 51.3 | 51.1 | 42.4 | 34.1 |
| −2 Log Likelihood | 1566.9 | 1565.5 | 990.9 | 867.4 | 678.1 | 759.4 |
| AIC (Smaller is Better) | 1574.9 | 1579.5 | 1008.9 | 885.4 | 696.1 | 777.4 |
| AICC (Smaller is Better) | 1575.0 | 1579.8 | 1009.6 | 886.2 | 697.1 | 778.3 |
| BIC (Smaller is Better) | 1587.5 | 1601.5 | 1036.4 | 911.7 | 722.4 | 803.7 |
| Caregiver Perceived Family Conflict | ||||||
| Model 1 | Model 1a | Model 2 | Model 3 | Model 4 | Model 5 | |
| β (se) | β (se) | β (se) | β (se) | β (se) | β (se) | |
| Intercept | 1.82 (0.14) | 1.35 (0.58) | 3.21 (1.29) | 2.63 (1.43) | 2.62 (1.67) | 0.48 (1.8) |
| Time | −0.03 (0.03) | −0.031 (0.03) | −0.084 (0.10) | −0.05 (0.11) | 0.037 (0.14) | −0.040 (0.11) |
| Education | 0.024 (0.08) | 0.073 (0.10) | 0.065 (0.10) | −0.07 (0.14) | 0.078 (0.13) | |
| Sex (Male) | 0.39 (0.24) | 0.43 (0.30) | 0.39 (0.31) | 0.13 (0.43) | −0.16 (0.42) | |
| Age | 0.0025 (0.010) | 0.016 (0.01) | 0.008 (0.01) | −0.007 (0.12) | 0.007 (0.02) | |
| CG Comfort Discussing PT Employment | −0.23 (0.10)* | |||||
| PT Comfort Discussing PT Employment | −0.36 (0.16)* | |||||
| CG Comfort Discussing CG Employment | −0.31 (0.19) | |||||
| PT Comfort Discussing CG Employment | −0.10 (0.18) | |||||
| CG Comfort Discussing Cost | −0.37 (0.15)* | |||||
| PT Comfort Discussing Cost | 0.31 (0.20) | |||||
| CG Comfort Discussing Health Ins | −0.10 (0.24) | |||||
| PT Comfort Discussing Health Ins | 0.30 (0.21) | |||||
| Chi-Sq | 160.79 | 157.20 | 15.5 | 8.66 | 8.57 | 17.0 |
| −2 Log Likelihood | 1531.5 | 1528.4 | 514.8 | 433.8 | 304.8 | 396.1 |
| AIC (Smaller is Better) | 1539.5 | 1542.4 | 532.8 | 451.8 | 322.8 | 414.1 |
| AICC (Smaller is Better) | 1539.5 | 1542.6 | 534.2 | 453.4 | 325.4 | 416.0 |
| BIC (Smaller is Better) | 1552.0 | 1564.4 | 558.3 | 476.3 | 346.6 | 438.4 |
Patient Perceptions of Family Conflict and Cohesion: Within-Person and Cross-Partner Effects
Patient multi-level models for cohesion (β = .04, se = .03, p > .05) and conflict (β = −.008, se = .03, p > .05) provide no evidence of linear change in scores over time and no covariates including education, sex, race and age were significant for either variable. When patients feel comfortable discussing their own employment status (both patient [β = .30, se = .09, p < .01], caregiver employment [β = .28, se = .11, p < .0][), treatment costs (β = .42, se = .10, p < .001) or health insurance (β = .30, se = .13, p < .01) they report greater family cohesion. Patient multi-level models for cohesion showed no cross-partner effects associated with the caregiver’s level of comfort discussing sensitive topics.
Patient reports of family conflict are not associated with any patient communication topics but were positively related to caregiver discomfort discussing the patient’s employment status (β = −0.25, se = .11, p < .05) , treatment costs (β = −0 .43, se = .11, p < .05) or health insurance (β = −0.31, se = .15, p < .05).
Caregiver Perceptions of Family Conflict and Cohesion: Within-Person and Cross-Partner Effects
Caregiver multi-level models for cohesion evidenced a significant linear decline (β = −0.075, se = .03, p < .05) over time. Multi-level models of conflict provide no evidence of linear change (β = .008, se = .03, p > .05) in caregiver conflict scores over time. No covariates including education, sex, race and age were significant for either model. When caregivers feel comfortable discussing their own employment (β = .76, se = .20, p < .001), the patients’ employment (β = .42, se = .15, p < .05)) or health insurance (β = .56, se = .25, p < .05) they report greater cohesion. Caregivers also report higher cohesion when patients report feeling comfortable discussing their own (patient) employment (β = 0.30, se = .14, p < .05) or health insurance (β = 0.46 se = .21, p < .05).
Caregivers who report higher levels of family conflict report greater discomfort talking about patient employment status (β = −0 .23, se = .10, p < .05) and cost of treatment (β = −0.37, se = .15, p < .05). Caregivers also report greater family conflict when patients are uncomfortable discussing their own (patient) employment status (β = −.36 se = .16, p < .05).
DISCUSSION
In this longitudinal cohort analysis, we found support for our hypotheses that overall, patient and caregiver reports of positive family functioning were associated with comfort discussing the financial aspects of cancer treatment. Moreover, assessments of family functioning were influenced by their own and their partners’ levels of communication comfort. Interestingly, the longitudinal design also allowed us to examine differential patterns in caregiver versus patient perceptions of family functioning over time. This is the first study, to our knowledge, that has identified a significant decline in caregiver ratings of family cohesion over the first two years following a cancer diagnosis. By contrast, patient assessments of family functioning did not change. The implications of these findings are discussed in detail below.
Increased comfort discussing financial aspects of cancer treatment was generally associated with greater family cohesion and decreased family conflict however, these relationships varied over time. The strongest associations were found for caregivers’ comfort discussing employment status. Caregivers who felt greater comfort discussing either their own or the patient’s employment status expressed a greater sense of cohesion and less family conflict. This highlights the salience of employment as a source of stress for caregivers. Patient decisions to continue to work throughout a cancer diagnosis and treatment, the ability and willingness to take time off, retire or quit are all personal decisions influenced by the type of job held (fulltime, part time, hourly, salary), characteristics of the employer, and availability of benefits such as paid time off and health insurance (11,20,21). Employment is tied to security and there may be worries about lost savings and future security of the family, especially if health insurance is tied to the cancer patient. Financial toxicity is a known stressor for cancer caregivers (22,23); caregivers may fear introducing additional anxiety or worry to the patient (6) and there is little guidance available to support caregivers to engage in discussion of potential difficult financial topics. In fact, there is limited support for clinicians to engage patients and their families in discussions of financial costs and hardships associated with cancer care. Maintaining employment can also be important psychologically to the patient to maintain identity and independence. Nonetheless, we found that dyads avoid discussing employment status with important consequences on family functioning.
Over the two-year study period caregiver ratings of family cohesion significantly decreased but patient ratings did not. We found that patients’ ratings of family cohesion were only associated with their own levels of communication comfort. However, caregiver ratings of family cohesion were associated by both their own and their partners’ level of communication comfort. This is in line with other studies (3,24,25) that demonstrate that patients are relatively insensitive to, or shielded from, problems by family members but that caregivers take on the role of monitoring and responding to the patient’s emotional needs. Caregivers, as a result, report more stress and difficulties with communications (26), especially issues related to the costs and burdens of the disease. This protective feature of caregiving; a desire not to overburden patients, including withholding information that may be upsetting, in an attempt to keep them focused on treatment and maintain patient QOL, seems to come at the expense of caregiver well-being. (24). Our results suggest that caregivers who suffer in silence begin to lose the sense of shared commitment, help, and support from family over time. Notably, the loss of these feelings is not associated with changes in perceived family conflict but results in a loss of family cohesion.
Limitations—
One limitation is that we could not measure family functioning prior to the cancer diagnosis therefore it is unclear how prior relationship characteristics may have affected strain. Relationship strain refers to the continuation or exacerbation of pre-existing relationship problems and has been found to be a contributing factor to increased burden among caregivers of patients with forms of dementia (27,28). Given three waves of data, only linear associations could be examined. It is possible with additional points of measurement non-linear patterns of association over time could be detected. Moreover, we only examined outcomes of patients and their primary caregivers. A small number of studies have looked at secondary caregivers that indicates that a source of conflict is within the family caregiving unit (3). Lastly, there may be unknown cohort specific factors that affect generalizability of results.
Conclusion.
This study of hematological cancer patients and caregivers presents data on a cohort of patients diagnosed with cancers for which financial toxicity and family functioning features remain understudied. The issue of financial toxicity associated with cancer diagnosis and treatment has recently received much attention, particularly in terms of how patients and families receive support from health care providers and the institutions where they receive their care. Our work highlights that communication within families about financial aspects of cancer care is also important as unaddressed communication difficulties can have detrimental effects on family functioning in the long term. This work can help to characterize which caregivers and families are most likely to be vulnerable to these negative effects and can help pinpoint the timing for optimal supportive interventions. Future studies should also examine whether the prominence of specific economic topics, such as employment status, varies depending on where the patient is in their cancer journey. Such insights could help to refine the types of interventions developed and the timing of specific topics or audience (i.e., the caregiver, the patient, the dyad).
Funding:
NIH/NCI R01 CA168647
Footnotes
Competing Interests: None of the authors have competing interests
Animal Research (Ethics)-N/A
Consent to Participate and Publish (Ethics): Informed consent and consent to publish was obtained from institutional review board, IRB # HM13074
Data Availability:
dataset available upon request
References
- 1.Williamson HC, Karney BR, Bradbury TN. Financial strain and stressful events predict newlyweds’ negative communication independent of relationship satisfaction. Journal of Family Psychology. 2013. Feb;27(1):65–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hastert TA, Ruterbusch JJ, Nair M, Noor MI, Beebe-Dimmer JL, Schwartz K, et al. Employment Outcomes, Financial Burden, Anxiety, and Depression Among Caregivers of African American Cancer Survivors. JCO Oncology Practice. 2020. Mar;16(3):e221–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Siminoff LA, Wilson-Genderson M, Baker SJ. Depressive symptoms in lung cancer patients and their family caregivers and the influence of family environment. Psychooncology. 2010. Dec;19(12):1285–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Li C, Yuan J, Huang X, Zhang S, Hong Y, Zhong J. Correlation between depression and intimacy in lung cancer patients and their family caregivers. BMC Palliat Care. 2022. Jun 3;21(1):99. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Wang H, Yue H, Ren M, Feng D. Dyadic effects of family-functioning and resilience on quality of life in advanced lung cancer patients and caregivers: An actor-partner interdependence mediation model. Eur J Oncol Nurs. 2021. Jun;52:101963. [DOI] [PubMed] [Google Scholar]
- 6.Lin Y, Hu C, Xu Y, Zhao J, Li Q. The mutual impact and moderating factors of quality of life between advanced cancer patients and their family caregivers. Support Care Cancer. 2020. Nov;28(11):5251–62. [DOI] [PubMed] [Google Scholar]
- 7.Spencer JC, Rotter JS, Eberth JM, Zahnd WE, Vanderpool RC, Ko LK, et al. Employment Changes Following Breast Cancer Diagnosis: The Effects of Race and Place. J Natl Cancer Inst. 2020. Jun 1;112(6):647–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Kayaalp A, Page KJ, Rospenda KM. Caregiver Burden, Work-Family Conflict, Family-Work Conflict, and Mental Health of Caregivers: A Mediational Longitudinal Study. Work Stress. 2021;35(3):217–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.de Moor JS, Mollica M, Sampson A, Adjei B, Weaver SJ, Geiger AM, et al. Delivery of Financial Navigation Services Within National Cancer Institute–Designated Cancer Centers. JNCI Cancer Spectr. 2021. Apr 9;5(3):pkab033. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Carrera PM, Kantarjian HM, Blinder VS. The financial burden and distress of patients with cancer: Understanding and stepping-up action on the financial toxicity of cancer treatment. CA: A Cancer Journal for Clinicians. 2018;68(2):153–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Thomson MD, Siminoff LA. Managing work and cancer treatment: Experiences among survivors of hematological cancer. Cancer. 2018. Jul 1;124(13):2824–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Yabroff KR, Kim Y. Time costs associated with informal caregiving for cancer survivors. Cancer. 2009. Sep 15;115(18 Suppl):4362–73. [DOI] [PubMed] [Google Scholar]
- 13.Galtieri LR, Fladeboe KM, King K, Friedman D, Compas B, Breiger D, et al. Caregiver perceived financial strain during pediatric cancer treatment: Longitudinal predictors and outcomes. Health Psychology. 2022. Jan;41(1):43–52. [DOI] [PubMed] [Google Scholar]
- 14.Moos R, Moos B. Family Environment Scale manual and sampler set: Development, applications and research (4th ed.). Palo Alto, CA: Mind Garden; 2009. [Google Scholar]
- 15.Lyons KS, Sayer AG. Using multilevel modeling in caregiving research. Aging Ment Health. 2005. May;9(3):189–95. [DOI] [PubMed] [Google Scholar]
- 16.Lyons KS, Sayer AG, Archbold PG, Hornbrook MC, Stewart BJ. The enduring and contextual effects of physical health and depression on care-dyad mutuality. Res Nurs Health. 2007. Feb;30(1):84–98. [DOI] [PubMed] [Google Scholar]
- 17.Raudenbush SW, Brennan RT, Barnett RC. A multivariate hierarchical model for studying psychological change within married couples. Journal of Family Psychology. 1995;9(2):161–74. [Google Scholar]
- 18.Davey A, Fincham FD, Beach SR, Brody GH. Attributions in marriage: examining the entailment model in dyadic context. J Fam Psychol. 2001. Dec;15(4):721–34. [DOI] [PubMed] [Google Scholar]
- 19.Raudenbush SW, Bryk S, Cheong YF, Congdon R. HLM6: Hierarchical linear and nonlinear modeling. Chicago: Scientific Software International.; 2004. [Google Scholar]
- 20.Fitch MI, Nicoll I. Returning to work after cancer: Survivors’, caregivers’, and employers’ perspectives. Psychooncology. 2019. Apr;28(4):792–8. [DOI] [PubMed] [Google Scholar]
- 21.Greidanus MA, de Boer AGEM, de Rijk AE, Tiedtke CM, Dierckx de Casterlé B, Frings-Dresen MHW, et al. Perceived employer-related barriers and facilitators for work participation of cancer survivors: A systematic review of employers’ and survivors’ perspectives. Psychooncology. 2018. Mar;27(3):725–33. [DOI] [PubMed] [Google Scholar]
- 22.Sadigh G, Switchenko J, Weaver KE, Elchoufi D, Meisel J, Bilen MA, et al. Correlates of financial toxicity in adult cancer patients and their informal caregivers. Support Care Cancer. 2022. Jan;30(1):217–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Inguva S, Priyadarshini M, Shah R, Bhattacharya K. Financial toxicity and its impact on health outcomes and caregiver burden among adult cancer survivors in the USA. Future Oncol. 2022. Apr;18(13):1569–81. [DOI] [PubMed] [Google Scholar]
- 24.Langer SL, Rudd ME, Syrjala KL. Protective buffering and emotional desynchrony among spousal caregivers of cancer patients. Health Psychol. 2007. Sep;26(5):635–43. [DOI] [PubMed] [Google Scholar]
- 25.Langer SL, Brown JD, Syrjala KL. Intrapersonal and interpersonal consequences of protective buffering among cancer patients and caregivers. Cancer. 2009. Sep 15;115(18 Suppl):4311–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Li Q, Loke AY. A literature review on the mutual impact of the spousal caregiver-cancer patients dyads: “communication”, “reciprocal influence”, and “caregiver-patient congruence”. Eur J Oncol Nurs. 2014. Feb;18(1):58–65. [DOI] [PubMed] [Google Scholar]
- 27.Bannon SM, Reichman M, Popok P, Grunberg VA, Traeger L, Gates MV, et al. Psychosocial Stressors and Adaptive Coping Strategies in Couples After a Diagnosis of Young-Onset Dementia. Gerontologist. 2022. Feb 9;62(2):262–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Holdsworth K, McCabe M. The Impact of Dementia on Relationships, Intimacy, and Sexuality in Later Life Couples: An Integrative Qualitative Analysis of Existing Literature. Clin Gerontol. 2018. Feb;41(1):3–19. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
dataset available upon request