Unmet Needs of Family Cancer Caregivers Predict Quality of Life in Long-term Cancer Survivorship

Abstract

Purpose.

To identify demographic and caregiving characteristics associated with caregivers’ unmet needs and to examine associations of caregivers’ unmet needs with their quality of life at 2 and 5 years since their patients’ initial diagnosis.

Methods.

Family cancer caregivers completed a prospective longitudinal surveys at 2 (T1) and 5 years (T2) post-diagnosis. Demographic and caregiving characteristics were measured at T1. Unmet needs and quality of life were measured at T1 and T2.

Results.

Younger and spousal caregivers reported greater unmet needs (B > 2.03, p < .05). Independent of demographic characteristics, caregivers’ perception that providing care to their relative with cancer was overwhelming was consistently associated with unfulfillment of their needs in various domains, concurrently and prospectively (B > 2.50, p < .05), across the long-term survivorship phases.

Conclusions.

Findings highlight the contribution of earlier subjective caregiving stress to family caregivers’ needs not being met both currently and years later, which in turn related to poorer quality of life across different family caregivership trajectories. Findings suggest identifying at-risk subgroups of family caregivers based on demographics and assessing caregiving stress as a priority in psycho-oncology research and clinical practices.

Implications for Cancer Survivors.

Family caregivers’ quality of life is affected by cancer survivors’ illness trajectory many years after the initial cancer diagnosis, so are the caregivers’ needs. Cancer survivorship care plan should take careful consideration of the nuanced long-term contributions of caregivers’ unmet needs to their specific aspects of quality of life.

Introduction

Improved 5-year survival during the past four decades and accumulating evidence on long-term mental and physical sequelae of cancer treatment have informed care plans for cancer survivors for years after their completion of cancer treatment [1–3]. Studies have also found psychological distress and poor quality of life among family caregivers who are likely to outnumber the survivors [4–7]. This calls for adequate care plans for family caregivers of cancer survivors years after the initial diagnosis and treatment. To date, however, studies of cancer caregivers focus mainly on the time of initial treatment and end-of-life phases and most are cross-sectional. Efforts to address these problems, which are essential to inform care plans for caregivers, are largely lagging. An important initial step is better understanding how the cancer patients’ trajectory may affect their caregivers’ quality of life according to the degree to which caregivers’ needs are being fulfilled [8].

By the end of first five years, approximately two thirds of cancer patients are alive: some in remission, others in cancer care; one-third have passed away [2,3,9]: Family members experience the same passages. By five years after their relative’s initial cancer diagnosis, family caregivers who were once actively involved in cancer care diverge to different paths: some become former caregivers whose patients no longer need their care, others continue or resume cancer care to the patient, and others become bereaved [10,11]. Family caregivers’ needs presumably vary at that time depending on their patients’ prognostic status. Unmet needs of these three groups of caregivers may be critical determinants of their quality of life. This study aimed to identify demographic and early caregiving characteristics that are associated with greater caregivers’ unmet needs and to examine the extent to which caregivers’ unmet needs are associated with their quality of life at 2 and 5 years since their patients’ initial cancer diagnosis, concurrently and longitudinally.

Method

Participants and Procedure

The National Quality of Life Survey for Caregivers was designed to longitudinally assess the impact of cancer on the quality of life of family members and close friends who were caring for cancer survivors [10]. Caregivers were nominated by the cancer patients for whom they provided consistent help during a patient’s cancer experience. Eligibility criteria also included being (a) 18 years and older, (b) able to speak/read English or Spanish, and (c) residing in the United States.

At caregivers’ initial assessment (T1: n = 896), patients were on average 2.2 years (SD = 0.6 years) post-diagnosis [10]. Follow-up data collection at five years (T2)¹¹ included a survey designed for caregivers whose index patient had passed away (Bereaved Caregivers), caregivers who were actively engaging in caregiving for the index case (Current Caregivers), and caregivers who no longer provided cancer care for the index cases, who were alive and did not need cancer care from the family member (Former Caregivers, Remission) by T2 [11]. Compared with caregivers who provided complete data at both T1 and T2 (n = 470), those who did not provide complete data at T2 did not differ in study variables at T1 (p > .10), with five exceptions. Females, caregivers with higher household income, and spouses of the patients were more likely to provide complete data at T2. Completers reported better mental functioning and less caregiving stress at T1 and had patients whose cancer was less fatal (p < .02). Characteristics of caregivers with complete information are reported in Table 1.

Table 1.

Sample Characteristics and Descriptives of Study Variables

	2-year (n = 470)	5-Year (n = 470)
	m (SD) or %	FCR (n = 340) m (SD) or %	CC (n = 62) m (SD) or %	BC (n = 68) m (SD) or %
Sample Characteristics at T1
Age	55.80 (12.56)	54.94 (12.76)	57.74 (10.72)	58.34 (12.67)
Gender (Female)	66.0%	61.5%	71.0%	83.8%
Ethnicity (White)	93.8%	93.8%	91.9%	95.6%
Lower Education	23.6%	21.2%	29.0%	30.9%
Lower Income	18.1%	14.7%	24.2%	29.4%
Spouse of Patient	70.4%	73.5%	69.4%	55.9%
Caregiving Stress	1.59 (0.58)	1.51 (0.52)	1.73 (0.64)	1.83 (0.69)
Patient’s Cancer Severity	0.17 (0.25)	0.11 (0.18)	0.19 (0.24)	0.48 (0.29)
Unmet Needs
Psychosocial Unmet Needs	1.56 (2.19)	0.64 (1.35)	1.64 (1.79)	-
Medical Unmet Needs	1.18 (1.81)	0.36 (0.91)	1.39 (1.50)	-
Financial Unmet Needs	1.02 (2.30)	0.70 (2.09)	1.16 (2.13)	-
Daily Activity Unmet Needs	0.83 (1.28)	0.37 (1.10)	1.05 (1.33)	-
Managing the Loss Unmet Needs	-	-	-	1.61 (1.65)
Reintegration Unmet Needs	-	-	-	0.92 (1.30)
Quality of Life
Mental Functioning	50.27 ( 9.81) (n=456)	50.88 (9.57) (n=324)	46.88 (11.52) (n=61)	45.19(10.90 (n=66)
Physical Functioning	50.35 (10.54) (n=456)	49.75 (9.43) (n=324)	48.29 (11.87) (n=61)	52.92 ( 8.70) (n=66)

Note. FCR = Former Caregivers - Remission; CC = Current Caregivers: BC = Bereaved Caregivers; Gender = 1 for female, 0 for male; Ethnicity = 1 for non-Hispanic White, 0 for other ethnicity; Lower Education = 1 for high school and lower education, 0 for college and higher education; Lower Income = 1 for household annual income < $39,999, 0 household annual income ≥ $40,000 and cases preferred not to answer Spouse of Patient = 1 for spousal caregivers of the survivor, 0 for non-spousal caregiver.

This study was approved by the Emory University Institutional Review Board. Returning the completed survey served as evidence of informed consent to participate. Two cycles of mailing and telephone follow-up calls were made during an eight-week data collection period corresponding to each time point.

Measures

Demographic and caregiving variables.

Caregivers’ self-reported age, gender, ethnicity, education, income, spousal status were obtained at T1. Two indicators of early caregiving characteristics were assessed at T1. First, subjective caregiving stress (the extent to which caregivers felt overwhelmed by care tasks and responsibilities) was measured by the 4-item stress overload subscale of the Pearlin Stress Scale (e.g., “I work hard as a caregiver but never seem to make any progress”), with responses ranging from 1=not at all to 4=completely [12]. Higher mean scores reflect a greater perceived level of caregiving stress. The subscale had good internal consistency (α=.81).

Second, a cancer severity index was created based on mortality rates calculated by cancer type (10 cancers studied) and stage (localized, regional, distant), and the time since diagnosis at T1 [13], information that was obtained from the cancer registry. This index ranges from 0 to 1, with higher scores reflecting more severe cancer diagnosis and therefore greater objective caregiving strain.

Unmet needs.

Family caregivers’ unmet needs were assessed using the 27-item Needs Assessment of Family Caregivers-Cancer (NAFC-C) scale [14] at both T1 and T2, except for bereaved caregivers at T2. Caregivers rated each of the 27 need items on (a) how important the need has been for them during the past four weeks, and (b) how satisfied they have been with the fulfillment of the need during the past four weeks, each using a 5-point Likert-style scale (0 = not at all, 4 = extremely). For each need item, the satisfaction rating was reverse coded; the item then was scored by multiplying the importance rating by the (reversed) satisfaction rating. Thus, scores for each need item can range from 0 to 16, with the score being an index of unfulfillment. The NAFC-C has four subscales: psychosocial (9 items), medical (7 items), financial (3 items), and daily activity (8 items) unmet needs. The subscales had acceptable internal consistency at T1 (.69< α <.87) and at T2 (.70< α <.80).

For assessing needs of bereaved caregivers at the 5-year mark, the 20-item Needs Assessment of Family Caregivers-Bereavement to Cancer (NAFC-BvC) [15] was used. Bereaved caregivers at T2 rated each of the 20 need items on importance and satisfaction, each using a 5-point Likert-style scale (0=not at all, 4=extremely). The bereavement unmet need scores for each need item can thus also range from 0 to 16, with the score being an index of unfulfillment. The NAFC-BvC has two subscales: unmet needs for managing the loss (12 items) and unmet needs for reintegration (8 items). The subscales had good internal consistency at T2 (α= .83 and .80, respectively).

Quality of life.

Caregivers’ mental and physical functioning levels were measured using the Medical Outcomes Study 12-Item Short Form Health Survey (MOS SF-12) [16] at T1 and T2. The mental/social functioning score was a composite of weighted vitality, social functioning, role-emotional, and mental health subscale scores. The physical functioning score was a composite of weighted physical functioning, role-physical, bodily pain, and general health subscale scores. Higher composite scores reflect better mental/social and physical functioning.

Statistical Analyses

Means and standard deviations or percentages of frequency of study variables are reported in Table 1. The extent to which demographic and caregiving factors correlate with caregivers’ unmet needs was examined first cross-sectionally. General linear modeling (GLM) examined the four types of unmet needs simultaneously at T1 and T2 for the Former Caregivers-Remission and Current Caregivers, and two types of unmet needs simultaneously at T2 for the Bereaved Caregivers (Table 2). GLM was also employed to predict quality of life by demographic, caregiving, and unmet needs factors cross-sectionally (Table 3) and longitudinally (Table 4).

Table 2.

Demographic and Caregiving Correlates of Caregivers’ Unmet Needs

Predictors	Psychosocial		Medical		Financial		Daily Activity
	B	t	B	t	B	t	B	t
At 2-year (n = 463)
Age	−0.018	−2.1983*	−0.004	−0.490	−0.024	−2.633**	−0.005	−0.966
Gender	0.395	1.752†	−0.187	−0.969	0.128	0.530	−0.144	−1.130
Ethnicity	0.232	0.564	−0.022	−0.063	0.195	0.443	−0.022	−0.095
Education	−0.021	−0.090	0.159	0.781	−0.266	−1.047	−0.151	−1.130
Income	0.455	1.730†	0.236	1.047	1.001	3.556***	0.182	1.231
Spouse	0.912	3.819***	0.185	0.904	0.673	2.631**	0.047	0.351
Caregiving Stress	0.939	5.231***	0.595	3.863***	0.442	2.296*	0.834	8.242***
Cancer Severity	−0.180	−0.444	−0.083	−0.238	0.780	1.795†	0.458	2.006*
R2	.040***		.017***		.028***		.069***
At 5-year, Former Caregivers - Remission (n = 333)
Age	−0.002	−0.381	0.000	0.048	−0.013	−1.296	−0.004	−0.826
Gender	0.013	0.080	0.053	0.468	−0.142	−0.553	0.082	0.613
Ethnicity	−0.322	1.057	−0.118	0.564	0.296	0.623	−0.236	−0.955
Education	−0.329	−1.777†	−0.156	−1.227	−0.112	−0.387	−0.279	−1.858†
Income	0.338	1.601	0.186	1.284	0.637	1.935*	0.349	2.032*
Spouse	0.091	0.490	0.206	1.613	0.589	2.031*	0.151	1.002
Caregiving Stress	0.526	3.557***	0.112	1.100	0.523	2.268*	0.277	2.305*
Cancer Severity	1.059	2.637**	0.828	3.009**	1.639	2.619**	1.114	3.416***
R2	.020***		.009		.023***		.017**
At 5-year, Current Caregivers (n = 62)
Age	0.047	1.931†	−0.002	−0.076	−0.029	−1.168	0.028	1.623
Gender	0.015	0.027	0.403	0.850	0.352	0.630	−0.346	−0.896
Ethnicity	−0.719	−0.842	−0.039	−0.052	0.269	0.306	−1.254	−2.066*
Education	0.003	0.006	0.099	0.217	1.170	2.170*	−0.088	−0.236
Income	−0.778	−1.347	0.321	0.638	0.597	1.005	−0.590	−1.436
Spouse	0.363	0.681	0.263	0.566	1.347	2.459*	−0.346	−0.914
Caregiving Stress	0.564	1.533	0.252	0.786	1.163	3.075**	0.548	2.097*
Cancer Severity	0.357	0.364	0.661	0.775	1.311	1.303	1.076	1.547
R2	.166**		.109		.185**		.106

	At 5-year, Bereaved Caregivers (n = 66)
	Managing the Loss		Reintegration
	B	t	B	t
Age	−0.270	−2.894**	0.208	1.822†
Gender	−1.555	−0.514	−1.114	−0.309
Ethnicity	0.348	0.069	−17.276	−2.818**
Education	−0.539	−0.239	−4.273	−1.547
Income	−3.099	−1.309	−5.137	−1.775†
Spouse	0.292	0.132	−1.969	−0.728
Caregiving Stress	1.344	0.863	−5.009	−2.630**
Cancer Severity	4.115	1.206	−1.818	−0.436
R2	.052		.175***

^†p < .10

^*p < .05

^**p < .01

^***p < .001

Note. B = Unstandardized coefficients; Gender = 1 for female, 0 for male; Caucasian = 1 for Caucasian, 0 for non-Caucasian; Gender = 1 for female, 0 for male; Ethnicity = 1 for non-Hispanic White, 0 for other ethnicity; Education = 1 for high school and lower education, 0 for college and higher education; Income = 1 for household annual income < $39,999, 0 household annual income ≥ $40,000 and cases preferred not to answer Spouse of Patient = 1 for spousal caregivers of the survivor, 0 for non-spousal caregiver.

Table 3.

Predicting Quality of Life – Cross-sectional Analysis

	QOL at 2-Year (n = 450)				QOL at 5-Year: FCR (n = 317)				QOL at 5-Year: CC (n = 61)				QOL at 5-Year: BC (n = 66)
	Mental		Physical		Mental		Physical		Mental		Physical		Mental		Physical
	B	t	B	t	B	t	B	t	B	t	B	t	B	t	B	t
Demographic and Caregiving Predictors
Age	0.153	4.185***	−0.302	−7.440***	0.089	2.009*	−0.244	−5.672***	0.001	0.009	−0.319	−2.200*	0.208	1.822†	−0.270	−2.894**
Gender	−0.159	−0.166	0.734	0.687	1.050	0.894	−0.494	−0.433	−2.659	−0.819	−0.080	−0.025	−1.144	−0.309	−1.555	−0.514
Ethnicity	−1.761	−0.985	1.762	0.885	0.500	0.230	3.795	1.801†	5.256	1.031	2.280	0.450	−17.276	−2.818**	0.348	0.069
Education	0.388	0.383	−2.661	−2.362*	−0.041	0.031	−2.583	−2.019*	6.708	2.141*	−7.471	−2.401*	−4.273	−1.547	−0.539	−0.239
Income	−3.743	−3.309**	−1.230	−0.977	−2.734	−1.803†	−0.373	−0.254	0.042	0.012	0.921	0.269	−5.137	−1.775†	−3.099	−1.309
Spouse	−0.244	0.237	2.613	2.284*	0.278	0.211	2.225	1.742†	−9.351	−2.897**	8.129	2.535*	−1.969	−0.728	0.292	0.132
Caregiving Stress	−4.549	−5.917***	−1.643	−1.921†	−2.800	−2.681**	−2.672	−2.639***	−3.353	−1.521	−7.085	−3.237**	−5.009	−2.630**	1.344	0.863
Cancer Severity	−6.527	−3.735***	3.997	2.056*	−9.696	−3.451***	3.114	1.143	−8.443	−1.442	6.724	1.156	−1.818	−0.436	4.115	1.206
R2	.187***		.161***		.109***		.119***		.184**		.289***		.182***		.124*
Unmet Need Predictors
Psychosocial	−1.385	−5.888***	0.154	0.545	−1.084	−2.246*	−1.199	−2.310*	−2.164	−2.157*	1.633	1.510
Medical	0.174	0.638	0.090	0.275	−0.399	−0.617	0.939	1.351	−1.957	−1.938†	−0.261	−0.240
Financial	−0.477	−2.592**	−0.321	−1.452	−0.902	−3.343**	−0.341	−1.174	0.383	0.475	−1.619	−1.860†
Daily Activity	−0.469	−1.208	−0.842	−1.809†	−1.318	−2.233*	0.376	0.593	0.283	0.202	−1.350	−0.895
Managing the Loss													−2.022	−2.232*	−1.437	−1.772†
Reintegration													−1.316	−1.237	1.063	1.115
R2	.273***		.169***		.246***		.130***		.332**		.350***		.407***		.153*

^†p < .10

^*p < .05

^**p < .01

^***p < .001

Note. B = Unstandardized coefficients; FCR = Former Caregivers - Remission; CC = Current Caregivers: BC = Bereaved Caregivers; Gender = 1 for female, 0 for male; Ethnicity = 1 for non-Hispanic White, 0 for other ethnicity; Education = 1 for high school and lower education, 0 for college and higher education; Income = 1 for household annual income < $39,999, 0 household annual income ≥ $40,000 and cases preferred not to answer; Spouse of Patient = 1 for spousal caregivers of the survivor, 0 for non-spousal caregiver.

Table 4.

Predicting Quality of Life at 5 Years – Longitudinal Analysis

	QOL at 5-Year: FCR (n = 300)				QOL at 5-Year: CC (n = 60)				QOL at 5-Year: BC (n = 63)
	Mental		Physical		Mental		Physical		Mental		Physical
	B	t	B	t	B	t	B	t	B	t	B	t
2-year Demographic, Caregiving, and QOL Predictors
Age	0.042	1.008	−0.100	−2.680**	−0.048	−0.398	0.019	0.163	0.172	1.740†	−0.155	−1.642
Gender	0.633	0.625	−0.839	−0.929	0.036	0.015	−0.836	−0.352	−1.913	−0.646	−0.647	−0.229
Ethnicity	2.653	1.367	3.896	2.249*	5.413	1.438	2.181	0.596	−9.965	−1.977*	−0.552	−0.115
Education	0.180	0.158	−1.664	−1.637	4.470	1.827†	−3.068	−1.288	−5.074	−2.283	0.053	0.025
Income	−0.782	−0.585	1.911	1.602	5.030	1.853†	−1.810	−0.686	−1.982	−0.815	−1.149	−0.495
Spouse	−0.723	−0.630	0.114	0.111	−6.038	−2.402*	4.740	1.939†	−1.758	−0.801	0.188	0.090
Caregiving Stress	−0.303	−0.323	−2.097	−2.501*	−0.886	−0.483	−1.513	−0.847	0.197	0.111	3.065	1.814†
Cancer Severity	−5.123	−2.070*	3.604	1.632	−9.562	−2.173*	2.266	0.529	−0.804	−0.236	2.328	0.716
Mental Functioning	0.544	9.889***	0.120	2.440*	0.682	6.039***	0.196	1.780†	0.689	5.936***	0.118	1.061
Physical Functioning	0.154	3.119**	0.609	13.802***	0.051	0.502	0.679	6.821***	0.208	1.625	0.449	3.688***
R2	.333		.464		.574		.653		.597		.424
2-year Unmet Need Predictors
Psychosocial	−0.309	−1.055	−0.204	−0.775	−1.452	−2.678**	0.461	0.822	0.292	0.297	−0.012	−0.013
Medical	−0.293	−0.805	−0.169	−0.519	0.138	0.148	−0.069	−0.071	−0.260	−0.464	0.323	0.592
Financial	−0.259	−1.059	0.311	1.420	−0.025	−0.039	−0.807	−1.236	−0.319	−0.896	−0.182	−0.525
Daily Activity	−0.032	−0.063	−0.133	−0.286	1.447	1.217	−0.904	−0.734	−0.891	−1.001	−0.278	−0.321
R2	.348		.470		.652		.679		.619		.434
5-year Unmet Need Predictors
Psychosocial	−0.330	−0.692	−0.745	−1.678†	−1.348	−1.650	1.193	1.320
Medical	−0.259	−0.414	1.281	2.205*	−1.637	−2.074*	−0.338	−0.387
Financial	−0.503	−1.751†	−0.050	−0.187	0.606	0.957	0.067	0.095
Daily Activity	−1.496	−2.551**	−0.299	−0.549	0.328	0.285	0.838	0.659
Managing the Loss									−1.511	−1.834†	−0.625	−0.712
Reintegration									−1.747	−1.853†	1.502	1.495
R2	.407		.483		.733		.719		.698		.461

^†p < .10

^*p < .05

^**p < .01

^***p < .001

Note. B = Unstandardized coefficients; FCR = Former Caregivers – Remission; CC = Current Caregivers: BC = Bereaved Caregivers; Gender = 1 for female, 0 for male; Ethnicity = 1 for non-Hispanic White, 0 for other ethnicity; Education = 1 for high school and lower education, 0 for college and higher education; Income = 1 for household annual income < $39,999, 0 household annual income ≥ $40,000 and cases preferred not to answer; Spouse of Patient = 1 for spousal caregivers of the survivor, 0 for non-spousal caregiver.

To examine the degree to which unmet needs predict QOL cross-sectionally, a set of demographic and caregiving factors was entered into the equation (Table 2), followed by a set of unmet needs factors in the second block (Table 3). For longitudinal analysis (Table 4), quality of life at T1 was entered to the equation in the first block along with demographic and early caregiving factors. Unmet needs at T1 were added to the equation in the second block to test their unique contribution to predicting the quality of life at T2. Then, unmet needs at T2 were added in the third block, in order to test the unique contribution of unmet needs at T2 predicting the quality of life at T2, beyond what the demographic, caregiving, quality of life, and unmet needs factors at T1 accounted for.

Results

As shown in Table 1, caregivers were primarily middle-aged, females, non-Hispanic White, relatively educated and affluent, and spouses of the index patient. Caregivers reported overall mild to moderate levels of caregiving stress at T1, but those who either were actively engaging in cancer caregiving or became bereaved in the next three years (at T2) reported significantly greater caregiving stress than those whose patient did not need their care at T2, F=10.661, p<.001. Similarly, although the patients overall had had relatively treatable kinds of cancer, those who had passed away by T2 (for bereaved caregivers) had had the most fatal types of cancer, and those who were alive and did not need further cancer care at T2 had had the least fatal kind of cancer, F=88.853, p<.001.

Caregivers’ needs were mostly fulfilled at both T1 and T2. Caregivers at T1 and caregivers in the former caregiver (FCR) group at T2 reported mental and physical functioning that are comparable to the US population norms, whereas caregivers in the current caregiver (CC) group at T2 reported mental and physical functioning at the 25^th and 50^th percentile of the US population norms, respectively; caregivers in the bereaved group (BC) at T2 reported mental functioning at 25^th percentile and physical functioning at normative levels of the US population norms [16].

Demographic and Caregiving Correlates of Caregivers’ Unmet Needs

As shown in Table 2 (under “At 2-year” rows), younger caregivers and spouses of the patients reported greater psychosocial and financial unmet needs at the 2-year mark. Caregivers with lower household income reported greater financial unmet needs. Caregivers who perceived providing care to the patients as overwhelming reported their psychosocial, medical, financial, and daily activity needs as less fulfilled. Caregivers whose patients had more fatal kinds of cancer reported their daily activity needs as less met.

At the 5-year mark, different patterns of associations of demographics and early caregiving characteristics with unmet needs emerged, depending on the patients’ status. Among former caregivers whose patients no longer need cancer care at T2 (Table 2, At 5-year, Former Caregivers - Remission), caregivers with lower household income reported greater financial and daily activity unmet needs. Spousal caregivers also reported greater unmet financial needs. Furthermore, the more stressed they had been while providing care earlier, the more likely they were to report their psychosocial, financial, and daily activity needs not being met at T2. Caregivers whose patients had had more fatal kinds of cancer at T1 yet did not need cancer care at T2 reported greater unmet needs of all four types at T2.

Among caregivers actively involved in cancer care at T2 (Table 2, At 5-year, Current Caregivers), ethnic minorities reported greater daily activity unmet needs; less educated and spousal caregivers reported greater financial unmet needs. Greater earlier subjective caregiving stress was also associated with greater financial and daily activity unmet needs at T2.

Finally, among bereaved caregivers at T2 (Table 2, At 5-year, Bereaved Caregivers), younger caregivers reported their needs for managing the loss more unfulfilled. Ethnic minority caregivers and those who had perceived providing care as overwhelming while the patients were alive reported greater unmet needs for reintegrating their lives after the loss.

Predicting Quality of Life: Cross-sectional Analyses

As shown in Table 3, at the initial 2-year mark, poorer mental functioning was related to younger age, lower household income, greater subjective caregiving stress, and providing care to patients whose cancer was more fatal. Above and beyond these demographic and early caregiving effects, caregivers whose psychosocial and financial needs were unfulfilled reported poorer mental functioning. On the other hand, poorer physical functioning was related to older age, lower education, being non-spouse of the patient, providing care to the patients whose cancer was less fatal and (marginally) with greater caregiving stress. Above and beyond these effects, greater daily activity unmet needs were marginally related to poorer physical functioning.

Similar patterns were found predicting mental and physical functioning of Former Caregivers – Remission at T2 (Table 3, QOL at 5-year: FCR columns), with some exceptions. Compared with the 2-year mark, at the 5-year mark, when caregivers were no longer involved in cancer care for the index patient, caregivers with lower household income now reported only marginally poorer mental functioning and non-spousal caregivers now reported only marginally poorer physical functioning, whereas earlier stressed caregivers now reported significantly poorer physical functioning.

On the other hand, other associations that were either non-significant or significant at T1 became either significant or non-significant at T2. For example, caregivers’ greater daily activity unmet needs became a significant predictor of caregivers’ poorer mental functioning. Non-Hispanic White caregivers now reported marginally poorer physical functioning and greater psychosocial unmet needs now significantly predicted poorer physical functioning. Patients’ cancer severity became a non-significant predictor of physical functioning.

Compared with former caregivers whose patients were in remission, for Current Caregivers at T2 (Table 3, QOL at 5-year: CC columns), psychosocial unmet needs remained a significant predictor of poorer mental functioning; older age, lower education, and greater subjective caregiving stress remained predictor of poorer physical functioning. However, other factors no longer significantly predicting mental or physical functioning. In addition, higher education, spousal caregiver status, and (marginally) greater medical unmet needs became unique predictors of poorer mental functioning. Non-spousal caregiver status and greater financial unmet needs became stronger (to be marginally significant) predictors of poorer physical functioning.

For Bereaved Caregivers at T2 (Table 3, QOL at 5-year: BC columns), compared with Former Caregivers-Remission, younger age (marginally), lower income, and greater subjective caregiving stress remained predictors of poorer mental functioning and older age remained a predictor of poorer physical functioning. However, other variables no longer significantly predicted mental or physical functioning. Ethnic minority caregivers uniquely reported poorer mental functioning. Beyond these contributions, Bereaved Caregivers’ unmet needs for managing the loss, but not for reintegration, significantly related to poorer mental functioning and marginally to poorer physical functioning.

Predicting Quality of Life: Prospective Longitudinal Analyses

We also examined the degree to which unmet needs at T2 predict quality of life at T2, beyond prediction from caregivers’ demographics, earlier caregiving, quality of life, and unmet needs at T1. As shown in Table 4, among Former Caregivers - Remission (see QOL at 5-year: FCR columns), more fatal kind of cancer and poorer mental and physical functioning at T1 predicted poorer mental functioning at T2. Beyond these associations, greater daily activity unmet needs were concurrently related to poorer mental functioning. Regarding physical functioning, older age, ethnic minority, greater earlier caregiving stress, poorer mental and physical functioning at T1 significantly predicted poorer physical functioning at T2. Above and beyond, less medical unmet needs were concurrently related to poorer physical functioning.

Among Current Caregivers at T2 (see QOL at 5-year: CC columns), spousal status to the patient, patients having more fatal kind of cancer, poorer mental functioning at T1, and greater psychosocial unmet needs at T1 significantly predicted poorer mental functioning at T2. Beyond these contributions, concurrent medical unmet needs were related to poorer mental functioning. On the other hand, poorer physical functioning at T1 was the only significant predictor of poorer physical functioning at T2.

Finally, among Bereaved Caregivers by T2 (QOL at 5-year: BC columns), being non-Hispanic White and having poorer mental functioning at T1 were significant predictors of poorer mental functioning at T2. Beyond these effects, greater unmet needs for managing the loss and for reintegration were both marginally associated with concurrent poorer mental functioning. Poorer physical functioning at T1 was the only predictor of poorer physical functioning at T2.

Discussion

The findings illustrate that in the beginning of long-term survivorship, younger and spousal caregivers reported greater unmet needs. The same demographic factors were associated with unmet needs as the long-term survivorship proceeds. Independent of demographic characteristics, caregiving factors were substantially related to caregivers’ unmet needs. Caregivers’ perception of providing care to their relatives with cancer as overwhelming was consistently associated with their needs in various domains being unfulfilled, concurrently and prospectively, across the long-term survivorship phases. On the other hand, having provided care to relative with more fatal kind of cancer was associated with various types of unmet needs only among caregivers whose patients were alive and no longer need cancer care at the 5-year mark.

Findings suggest investigations to identify pathways by which demographic and early caregiving characteristics are linked to caregivers’ needs being unfulfilled. Becoming a caregiver to an adult with cancer may not be expected to younger adults. Spouses are more likely to be primary or sole caregivers who must navigate the complex cancer care system. As unprepared [17,18] and primary [19–21] caregivers have reported greater burden from caregiving, lack of psychological and tangible resources required for carrying out the new caregiver role may be a pathway for certain demographic factors linking to greater unmet needs in caregiving. The importance of considering the caregivers’ place in their own life course should also be considered, as the severity of unmet needs and distress varied by the caregivers’ age and familial relationship to the patient [22]. Findings also suggest that programs designed to address cancer caregivers’ needs should target younger and middle-aged adults who are in the transition to longterm survivorship or to bereavement by the 5-year mark.

Furthermore, findings suggest that stress management interventions for cancer caregivers [23,24] should be cognizant of the complex dynamic of family caregivers’ needs being associated with their mental and physical functioning, which vary by their patients’ prognostic status. Psychosocial unmet needs impair caregivers’ mental functioning regardless of their patient’s prognostic status. On the other hand, at the 5-year mark, financial and daily activity unmet needs are impactful only for caregivers whose index patients no longer need cancer care due to remission; medical unmet needs are impactful only for caregivers actively involved in cancer care for the index patient, and unmet needs for managing the loss are impactful only for bereaved caregivers. Furthermore, newly emerged unmet needs relating to daily activity and medical issues, besides those needs the caregivers had years earlier, are related to their current quality of life. Psychosocial unmet needs during providing cancer care but not current had lasting effects on mental quality of life only among caregivers who were actively involved in cancer care at the 5-year mark.

The nuanced long-term contributions of caregivers’ unmet needs to their specific aspects of quality of life should also be investigated, as the majority of existing studies focus on psychological aspects and around the time of diagnosis. For example, owing to the worldwide trend of aging populations, investigating how unmet needs and burdens of cancer caregiving may contribute to physical health in particular, such as higher rates of morbidity and mortality [25–27], particularly among older caregivers, will be essential. In addition, prospective information that begins from the cancer diagnosis of the care recipient and follows through for years is valuable in precisely documenting the cancer trajectory, not only from the care recipients’ perspective during the survivorship phases, but also in bridging the survivorship to bereavement phases from caregivers’ perspective.

These are concerns that must be incorporated in future studies as the first step to develop evidence-based guidance for programs and policies aimed at reducing family caregivers’ unmet needs and improving their quality of life at key transition points across diverse illness trajectories that may span many years. Addressing each of these concerns in future studies will help not only in advancing the science but also in bettering public health by improving family caregivers’ health.

Conclusion

Findings highlight the contribution of earlier subjective caregiving stress to predicting family caregivers’ needs not being met currently and years later. Unmet psychosocial needs in particular is a critical concern, as they link to poorer mental and physical functioning across different trajectories of family caregivership. Research priority should be given to identifying at-risk subgroups of family caregivers based on their age and spousal status to the patient, as well as to assess their stress evaluation of the new cancer caregiver role and appraisal of the patients’ cancer severity earlier in the cancer survivorship phases. Psycho-oncology practitioners also should ensure that survivorship care plans include plans for family caregivers, and be equipped to deliver support to family caregivers who are at different places in the cancer journey many years after their patients’ initial cancer diagnosis, because needs differ at different places on that journey.