Health service and medication use of parents of childhood cancer survivors: a controlled comparison study

Abstract

Background

Parents of childhood cancer survivors may be vulnerable to experiencing poor health outcomes, but little is known about how these parents use healthcare. This study investigated the nature and extent of survivors’ parents’ healthcare and medication use relative to a comparison group. We also examined whether demographic or cancer-related factors were related to healthcare use and whether healthcare use was associated with parents’ general functioning.

Methods

We conducted a cross-sectional study involving 55 parents of cancer survivors recruited through eight Australian hospitals, and 135 parents of children without a cancer diagnosis, through an online recruitment platform. Participants responded to a questionnaire assessing their health service usage, regular medications, general functioning (engagement activities including work/study) and anxiety and depression symptoms (using PROMIS short forms). We performed regression analysis to determine factors related to healthcare and medication use in parents of survivors.

Results

More parents of survivors reported accessing mental health services than comparison parents (56% vs. 33%, p=.003), mainly due to their use of social workers. Fewer parents of survivors reported accessing other community health services, particularly general practitioners (51% vs. 78%, p<.001). Having a child survivor who was male was associated with greater use of community health services (B= −0.67, p=.008). No other demographic or cancer-related variables were associated with health service use. Health service use was not associated with general functioning, but greater medication use was associated with higher anxiety scores (B = 1.41, p=.008).

Conclusion

Parents of childhood cancer survivors showed different patterns of health service use relative to comparison parents, but the extent of their use was not significantly linked with demographic or cancer-related variables. Comprehensive assessment of parents’ needs in clinical encounters remains vital to identify and appropriately match support needs with available services.

Keywords:

• Childhood cancer
• health care utilisation
• medication use
• mental health
• oncology
• cancer survivors

Background

Medical treatment for childhood cancers has improved considerably, translating to more favourable prognoses and an overall survival rate reaching 85% in developed countries [1–4]. A burgeoning group of patients is now reaching the milestone of survivorship; however this period introduces unique stressors and healthcare needs not only to children but also their parents or caregivers [5].

Cancer-related distress is an important health concern for parents of cancer survivors [6]. While most parents’ distress levels reduce overtime after their child enters remission [5,6], a subset of parents experiences persistent distress, including anxiety and depression, up to five years post-treatment. This can include clinically significant post-traumatic stress symptoms [7–11].

Potential stressors can also emerge during survivorship and impact parents’ health outcomes [12,13]. Following cancer treatment completion, supporting their child’s reintegration with peers and participation in daily activities is a prominent concern [5,8,12,14]. This is especially true for survivors who experience late effects impacting their behavioural or physical functioning – such as neurocognitive effects in brain cancer. Such additional parenting considerations can negatively impact health-related quality of life, and parents of survivors commonly report sleep disturbances, headaches and chronic fatigue [10,15]. In these parents, chronic stress and poor sleep can mediate the relationship between caregiving burden and poor mental health outcomes, and impair their physical immune responses [13,16].

There is currently little research on healthcare use in parents of cancer survivors. However, existing literature suggests that caregivers of adult cancer patients often underutilise health services for themselves, with psychological support being a particular unmet need [17,18]. Evidence from distressed parents of children without cancer indicates that engagement with formal psychological services is relatively low [19]. Across studies of both distressed parents in the general population, and caregivers seeking healthcare, several common barriers to accessing healthcare emerged. Barriers included navigating the expenses of care [20,21], coordinating around patient or child needs which are often prioritised above their own [22], and a tendency to manage emotional concerns with informal support [19].

These barriers to healthcare-seeking are likely to be relevant to parents of childhood cancer survivors [19–22]. Ongoing caregiving demands for the child survivor can complicate the balance between work and family life for parents, who may be re-entering the workforce after their child’s treatment [8,23]. Many families continue to bear financial burdens, both from follow-up medical care and any reduced working hours [24–26]. At the same time, their geographical distance from the hospital can make receiving support from hospital-based healthcare professionals difficult [8,27].

There is also a lack of data concerning medication use among parents of cancer survivors. Salem et al. [28] noted that parents of cancer patients were at increased risk of receiving a first prescription of psychotropic medication in the two years after diagnosis, relative to control parents. However, the ongoing patterns of psychotropic medication use among parents of longer-term survivors (rates of continuing use, or cessation) remain unclear. Other types of medication use within this population have also been little studied to date.

Given their significant role in child and family functioning, understanding parents’ wellbeing throughout survivorship is important. Developing and evaluating interventions for this target population requires a thorough understanding of their needs and characteristics. This study was an exploratory investigation into the nature and extent of parents’ healthcare use and how these relate to their health outcomes. We aimed to address the following questions:

1. What are the rates of health and mental health service use in parents of cancer survivors relative to a comparison group?

2. What are the rates of medication use in parents of cancer survivors relative to a comparison group?

3. What sociodemographic and cancer-related factors are associated with the extent of health service and medication use in parents of survivors?

4. Are the rates of health service and medication use in parents of survivors associated with their psychosocial outcomes including anxiety, depression and general functioning?

Methods

This study was approved by the Hunter/New England Human Research Ethics Committee (HREC/14/HNE/44).

Study design

This cross-sectional study analysed retrospective data from two groups. We used baseline data collected from a randomised trial of an online psychological intervention for parents of childhood cancer survivors (‘Cascade’) [29], as well as from a general population comparison group.

Participants

Parents of childhood cancer survivors

Participants were eligible if they were parents of children ≤18 years-old who had completed cancer treatment with curative intent and were now in remission. This included patients who had previously relapsed, provided they had also finished relapse treatment. Participants were excluded if (i) their child was in active treatment or palliative care, (ii) they had insufficient English skills for the questionnaire or (iii) they had extremely severe anxiety or depression based on the Depression, Anxiety and Stress Scale short form (DASS21). Recruitment was carried out at eight Australian children’s hospitals [29].

Comparison group parents

Comparison participants were recruited via PureProfile (https://www.pureprofile.com/), an online recruitment service with a panel of 300,000 Australian households. Individuals were eligible if (i) they were parents of children 18 years old or younger, and (ii) their child had never been diagnosed with cancer. The comparison group was matched to a larger cancer parent cohort that included the subset involved in this study, based on parent sex (70% female) and their child’s age (50%: 0–4 years; 25%: 5–11 years; 25%: 6–18 years). Eligible participants were emailed an invitation with links to the participant information sheet, consent form, and online survey. Participants have reimbursed a nominal sum (<$10).

Measures

Both groups completed all measures in a questionnaire.

Demographics

Participants reported their age, sex, the highest level of education and current employment status (full-time, part-time, casual work, retired, student, unemployed). They also reported their child’s demographic information, including for the cancer group their child’s diagnosis, age at diagnosis, treatment type and date of treatment completion.

Healthcare use

Several purposely designed measures were developed for the questionnaire by a multidisciplinary team of psychologists, oncologists, nurses and health economists to assess parents’ health service use, medication use and general functioning. Parents were asked to indicate whether they had accessed different health services for support within the past six months, including medical and mental health professionals (general practitioner, nurse, psychologist, psychiatrist, social worker, and counsellor). For the cancer group, cancer support organisations were included, while comparison parents were asked about other community organisations (with examples of Australian community mental-health organisations, Lifeline and Beyond Blue, provided). Participants also reported any emergency department presentations and hospital admissions, including the number of visits/admissions in the last six months.

To compare healthcare use between the groups, we considered total healthcare use as well as several subcategories: (i) hospital-based services (emergency presentations and hospital admissions), (ii) community-based health services (general practitioners (GPs) and nurses), and (iii) mental health services (psychiatrists, psychologists, counsellors, social workers and community support organisations).

Medication use

Participants listed their regular medications and supplements within the past six months, including the reason for their use wherever possible. Responses were categorised by therapeutic class and purpose in consultation with a senior paediatric oncologist, in line with MIMS Online database classes (see Table 1).

Table 1. Characteristics of cancer and comparison group parents and their children.

	Cancer^a (n = 55)	Comparison (n = 135)	p-value
Parent characteristics
Parent age, mean (SD)	40.35 (6.03)	38.98 (8.47)	.28
Range	20–55	23–70
Parent gender, n (%)			.07
Male	7 (12.7)	33 (24.4)
Female	48 (87.3)	102 (75.6)
Highest level of education, n (%)			.89
Year 10 or below	1 (1.8)	4 (3.0)
Year 12	7 (12.7)	19 (14.1)
Apprenticeship	–	1 (0.7)
TAFE certificate/diploma/college	13 (23.6)	31 (23.0)
University degree	22 (40.0)	59 (43.7)
Postgraduate qualification	12 (21.8)	21 (15.6)
Current marital status, n (%)			.39
Married/defacto	48 (87.3)	111 (82.2)
Not married/defacto	7 (12.7)	24 (17.8)
Current employment status, n (%)
Full-time work	15 (27.3)	54 (40)	.10
Part-time work	37 (67.3)	29 (21.5)	.10
Casual work^b	5 (9.1)	7 (5.2)	.32
Unemployed	18 (32.7)	39 (28.9)	.60
Retired	–	2 (1.5)	–
Student	1 (1.8)	5 (3.7)	–
Child characteristics
Child age, mean (SD)	7.90 (4.43)	7.48 (4.78)	.58
Range	2–17	0–18
Child gender, n (%)			.48
Male	36 (65.5)	81 (60.0)
Female	19 (34.5)	54 (40.0)
Other children in family, n (%)			.004**
Yes	48 (87.3)	90 (66.7)
No	7 (12.7)	45 (33.3)
Number of other children, mean (SD)	1.70 (0.95)	1.44 (0.69)	.092
Age at diagnosis, mean (SD)	5.03 (4.59)	–	–
Type of diagnosis, n (%)			–
Leukaemias	14 (25.5)	–
Lymphomas	7 (12.7)	–
CNS tumours	11 (20.0)	–
Neuroblastoma	6 (10.9)	–
Sarcomas	8 (14.5)	–
Wilm’s tumour	4 (7.3)	–
Other	5 (9.1)	–
Treatment received^c, n (%)
Surgery	30 (54.5)	–	–
Chemotherapy	50 (90.9)	–	–
Radiotherapy	17 (30.9)	–	–
Bone marrow/stem cell transplant	7 (12.7)	–	–
Years since treatment completion, mean (SD)	1.89 (1.79)	–	–
Range	0–10

TAFE: technical and further education; CNS: central nervous system.

**Significant at p<.01 level.

^aNB: Our cancer group being recruited across multiple hospitals sites, we did not have ethical approval to collect data on parents who did not opt-in to do the study; as such it was not possible for us to calculate an opt-in response rate.

^bCasual work refers to work without guaranteed, fixed employment hours or entitlements to paid leave.

^cParticipants were able to select multiple options in this category.

General functioning

Participants indicated the number of days within the last four weeks that they had taken part in a range of productive activities (paid work, study/learning, exercise, hobbies and socialising). They also reported the number of days they were absent from paid work or study in the same period, and the main reasons for doing so (sickness, tiredness or low energy, low motivation or ‘feeling flat’, medical appointments, unable to keep up with the workload, and difficulty with friends/colleagues).

Depression and anxiety

Parents indicated their level of depression and anxiety in the past seven days using the eight-item Patient-Reported Outcomes Measurement Information System (PROMIS) Depression Short Form and seven-item PROMIS Anxiety Short Form. For each item, participants rated their experience on a five-point scale from 1 (never) to 5 (always). Higher total scores indicated greater symptom severity [30]. The PROMIS short forms are widely used validated measures with strong psychometric properties [31,32].

Data analysis

Participants’ data were collated in IBM SPSS Statistics (IBM, New York) and descriptive statistics calculated for cancer and comparison groups. We performed comparisons between groups using t-tests for continuous variables and chi-squared tests for categorical variables. For each of the healthcare subcategories (described above), we considered two aspects: whether participants had accessed any healthcare in the category (a binary outcome), and their extent of use, indicated by the number of different services accessed. The same two aspects were analysed for participants’ total medication use.

We then performed multivariable linear regression for the cancer group to determine if sociodemographic or cancer-related factors accounted for variance in the extent of healthcare or medication use. We chose independent variables based on literature that established links to parents’ socioeconomic situation and caregiving burden. Age and sex of both parent and child were included due to their influence on caregiving demands and the parent-child relationship [33]. Parent employment status was included due to its potential influence on financial or logistical barriers to accessing healthcare [20,21]. Cancer-related factors included were the child’s diagnosis and time since completion of their treatment. These factors may relate to the child’s health needs, which influence the parent’s caregiving burden [12,27]. Treatment received was excluded due to the lack of specific detail in our data about treatment types.

The potential covariates were then analysed using variance inflation factor (VIF) values for problematic levels of shared variance. We considered VIF ≥5 to be problematic [34,35]. High collinearity emerged between the child’s current age and age at diagnosis; hence, we used only the child’s current age in our final multivariable regression analyses.

Rates of health service and medication use were then analysed for associations with general functioning, depression and anxiety levels using multivariable linear regression.

Results

Participant characteristics

Among the total of 190 participants, 55 were parents of cancer survivors and 135 were comparison parents. Our cancer and comparison samples were well-matched in terms of parents’ age, their child’s age and child’s sex (see Table 1) Across groups, most participants were female (87% of cancer participants, 76% of comparisons). One significant difference emerged between groups: more parents in the comparison group had just one child compared to the cancer group (33% vs. 13%, p=.004).

Research question 1: what are the rates of health and mental health service use in parents of cancer survivors relative to a comparison group?

The level of overall health service use did not differ significantly between parents in cancer and comparison groups. However, parents of survivors were more likely to report having accessed mental health services, especially social workers, in the past six months compared to comparison parents (56% vs. 33%, p=.003; Table 2). Despite this, they appeared somewhat less likely to report visiting a psychiatrist (0%, compared to 20% of comparison parents; Table 2). They were also less likely to report having accessed support from other community-based health services (51% vs. 78%, p<.001; Table 2), particularly GPs, and reported lower rates of hospital-based healthcare use (p=.009; Table 2). A minority of parents in both groups reported accessing support organisations (22% of cancer participants, 20% of comparisons).

Table 2. Comparison of healthcare use (HCU) and medication use across different categories in the last six months, among parents in cancer and comparison groups.

	Cancer (n = 55)	Comparison (n = 135)	p-value
Any HCU^a, n (%)	38 (69.1)	108 (80.0)	.15
Total use, mean (SD)	2.04 (1.89)	2.46 (2.85)	.24
Hospital-based HCU^b, n (%)	12 (21.8)	40 (29.6)	.30
Total use, mean (SD)	0.22 (0.42)	0.46 (0.81)	.009**
Community-based HCU^c, n (%)	28 (50.9)	105 (77.8)	<.001**
General practitioner	26 (47.3)	104 (77.0)
Nurse	2 (3.6)	27 (20.0)
Total use, mean (SD)	0.74 (0.85)	0.97 (0.65)	.08
Mental HCU^c, n (%)	31 (56.4)	45 (33.3)	.003**
Psychiatrist	0 (0.0)	28 (20.7)
Psychologist	15 (27.3)	35 (25.9)
Social worker	18 (32.7)	27 (20.0)
Counsellor	13 (23.6)	29 (21.5)
Cancer/community support organisation	12 (21.8)	27 (20.0)
Total use, mean (SD)	1.07 (1.20)	1.08 (1.87)	.97
Medication use^c, n (%)	25 (45.5)	72 (53.3)	.32
Supplements^d	15 (27.3)	36 (26.7)
Antidepressants/anxiolytics/mood stabilisers	10 (18.2)	18 (13.3)
Other CNS	–	4 (3.0)
Analgesics (Opioids/GABA/benzodiazepines)	1 (1.8)	4 (3.0)
(Other)	3 (5.5)	10 (7.4)
Anti-inflammatory	–	7 (5.2)
Sleep	2 (3.6)	5 (3.7)
Endocrine	5 (9.1)	9 (6.7)
Cardiovascular	–	6 (4.4)
Gastrointestinal	2 (3.6)	8 (5.9)
Antibiotics	–	2 (1.5)
Total use, mean (SD)	1.04 (1.50)	1.17 (1.77)	.62

CNS: central nervous system; GABA: gamma-aminobutyric acid.

**Significant at p<.01 level.

^aAny HCU encompasses the subcategories of hospital-based HCU, community-based HCU and mental HCU.

^bHospital-based HCU included emergency presentations, hospital admissions and hospital-based nurses.

^cParticipants were able to select multiple options in this category.

^dSupplements included vitamins and minerals, fish oil, probiotics, herbal products, and homeopathic remedies.

Research question 2: what are the rates of medication use in parents of cancer survivors relative to a comparison group?

The likelihood of parents using at least one medication, and the number of medications used, did not differ significantly between groups. Across groups, the most frequently used category was supplemented (27.3% of cancer group, 26.7% of comparisons) followed by psychotropic medications (antidepressants, anxiolytics and mood stabilisers; reported by 18.2% and 13.3% of cancer and comparison groups respectively; Table 2). Supplements consisted mostly of vitamins and minerals, and in the cancer group, some parents also reported herbal medicines and homeopathic remedies. Medication use in other categories was similar across groups.

Research question 3: what sociodemographic and cancer-related factors are associated with the extent of health service and medication use in parents of survivors?

In our model examining sociodemographic factors, having a child survivor who was male was associated with higher self-reported rates of accessing community-based health services (B= −0.67, p=.008; Table 3). In our model considering cancer-related factors, there was no evidence that any independent variables were associated with parents’ reported healthcare use (Table 3).

Table 3. Multivariable linear regression models for predictors of the number of healthcare services used in the last six months by parents of cancer survivors.

(a)	B (95% CI)	p-value	F	R^2	Adjusted R^2
Total HCU^a		.66	0.65	.064	–.034
Parent age	0.05 (–0.6, 0.2)	.36
Parent gender (male vs.. female)	0.84 (–1.4, 3.0)	.45
Full-time employment	0.51 (–1.2, 2.2)	.54
Child age	0.01 (–0.1, 0.2)	.90
Child gender (male vs.. female)	–0.91 (–2.1, 0.3)	.13
Total hospital HCU^b		.75	0.54	.053	–.046
Parent age	–0.01 (–0.4, 0.0)	.28
Parent gender (male vs.. female)	–0.14 (–0.6, 0.4)	.57
Full-time employment	–0.05 (–0.4, 0.3)	.79
Child age	0.02 (–0.01, 0.1)	.26
Child gender (male vs.. female)	–0.04 (–0.3, 0.2)	.78
Total community-based HCU^c		.04*	2.57	.211	.129
Parent age	0.04 (–0.1, 0.1)	.09
Parent gender (male vs.. female)	0.29 (–0.6, 1.2)	.53
Full-time employment	0.21 (–0.5, 0.9)	.54
Child age	0.03 (–0.1, 0.1)	.40
Child gender (male vs.. female)	–0.67 (–1.2, −1.8)	.008**
Total mental HCU^d		.89	0.33	.033	–.068
Parent age	0.03 (–0.1, 0.1)	.49
Parent gender (male vs.. female)	0.69 (–0.7, 2.1)	.33
Full-time employment	0.35 (–0.7, 1.4)	.52
Child age	–0.04 (–0.1, 0.1)	.45
Child gender (male vs.. female)	–0.20 (–1.0, 0.6)	.59
Total medications used		.55	0.81	.076	–.018
Parent age	0.01 (–0.1, 0.1)	.84
Parent gender (male vs.. female)	0.44 (–1.2, 2.1)	.60
Full-time employment	–0.19 (–1.5, 1.1)	.78
Child age	–0.02 (–0.1, 0.1)	.73
Child gender (male vs.. female)	–0.70 (–1.6, 0.2)	.14
(b)
Total HCU		.70	0.35	.014	–.025
Diagnosis (blood vs.. other)	0.39 (–0.7, 1.5)	.47
Time since treatment	–0.08 (–0.4, 0.2)	.61
Total hospital HCU		.09	2.48	.090	.054
Diagnosis (blood vs.. other)	0.23 (–0.1, 0.5)	.05
Time since treatment	0.02 (–0.1, 0.1)	.49
Total community-based HCU		.58	0.55	.022	–.018
Diagnosis (blood vs.. other)	0.10 (–0.4, 0.6)	.68
Time since treatment	0.06 (–0.1, 0.2)	.38
Total mental HCU		.24	1.47	.055	.018
Diagnosis (blood vs.. other)	0.06 (–0.6, 0.7)	.87
Time since treatment	–0.16 (–0.3, 0.1)	.09
Total medications used		.43	0.87	.034	–.005
Diagnosis (blood vs.. other)	0.57 (–0.3, 1.5)	.20
Time since treatment	–0.04 (–0.3, 0.2)	.71

(a) Model for sociodemographic predictors. (b) Model for cancer-related predictors.

HCU: healthcare use.

*Significant at p<.05 level. **Significant at p<.01 level.

^aTotal HCU is the number of services totalled across the other subcategories: hospital HCU, community-based HCU and mental HCU.

^bHospital HCU included emergency presentations, hospital admissions and hospital-based nurses.

^cCommunity-based HCU included general practitioners and nurses.

^dMental HCU included psychiatrists, psychologists, social workers, counsellors and cancer support organisations.

Research question 4: are the rates of health service and medication use in parents of survivors associated with their psychosocial outcomes and general functioning?

Overall, the two groups of parents did not differ significantly in perceived health status, although a higher proportion of cancer survivors’ parents reported a good/very good/excellent health status (91% vs. 79% of cancer and comparison groups respectively; Table 4). Average depression scores appeared similar across groups, however, cancer group parents reported higher anxiety scores than comparison parents (p=.04; Table 4).

Table 4. General functioning, depression and anxiety levels in parents by group.

	Cancer (n = 55)	Comparison (n = 135)	p-value
Days off work^a, mean (SD)	1.73 (3.25)	0.96 (3.47)	.18
Reason for days absent^b, n (%)
Sickness	4 (7.3)	15 (11.1)	.47
Tiredness or low energy	2 (3.6)	7 (5.2)	.68
Low motivation or ‘feeling flat’	3 (5.5)	4 (3.0)	.38
Medical appointments	2 (3.6)	3 (2.2)	.55
Medical appointments for child/ren	13 (23.6)	5 (3.7)	<.001**
Unable to keep up with workload	1 (1.8)	3 (2.2)	.89
Difficulty with friends/colleagues	–	–	–
Days of activity engagement^a, mean (SD)
Paid work	7.8 (7.70)	10.6 (11.06)	.06
Study	0.7 (2.35)	1.2 (3.97)	.38
Exercise/sport	6.3 (7.36)	6.4 (7.42)	.92
Personal hobbies	5.7 (7.7)	7.6 (11.96)	.28
Socialising with friends	8.6 (7.79)	6.1 (7.81)	.05
Socialising with other parents	1.2 (1.97)	4.8 (7.32)	<.001**
Perceived health status, n (%)			.06
Poor/fair	5 (9.1)	28 (20.7)
Good/very good/excellent	50 (90.9)	107 (79.3)
Anxiety score^c, mean (SD)	16.06 (5.76)	13.92 (6.59)	.04*
Anxiety level, n (%)			–
None/slight	26 (47.3)	83 (61.5)
Mild	12 (21.8)	20 (14.8)
Moderate	13 (23.6)	28 (20.7)
Severe	2 (3.6)	4 (3.0)
Depression score^d, mean (SD)	15.75 (6.32)	15.4 (7.26)	.76
Depression level, n (%)			–
None/slight	26 (47.3)	82 (60.7)
Mild	20 (36.4)	25 (18.5)
Moderate	5 (9.1)	27 (20.0)
Severe	2 (3.6)	1 (0.7)

PROMIS: patient-reported outcomes measurement information system.

*Significant at p<.05 level.

**Significant at p<.01 level.

^aIn the last 4 weeks.

^bParticipants were able to select multiple options in this category.

^cAnxiety was measured using the PROMIS anxiety short form.

^dDepression was measured using the PROMIS depression short form.

There was no significant between-group difference in the number of days that parents had been absent from work in the last four weeks. However, when they did report absenteeism, survivors’ parents were more likely to have been absent due to medical appointments for their children (24% vs. 4%, p<.001; Table 4).

Across groups, parents’ reported engagement in paid work, study and personal hobbies did not differ significantly. However, parents of survivors estimated that they spent fewer days socialising with other parents in the last four weeks, relative to the comparison group (p<.001; Table 4).

Regression analyses for associations between healthcare use and general functioning in survivors’ parents are summarised in Table 5. Parents’ extent of healthcare use was not associated with their general functioning, depression, or anxiety. However, parents who reported using a greater number of medications were more likely to have a higher anxiety score (B = 1.41, p=.008; Table 5), and reported spending more time socialising (B = 1.70, p=.04; Table 5). Parents with moderate-severe anxiety (27%) comprised 70% of those who reported using antidepressant/anxiolytic medication use in the group.

Table 5. Multivariable linear regression for associations between number of specific health services used and general functioning in parents of survivors.

	B (95% CI)	p-value	F	R^2	Adjusted R^2
Anxiety score^a		.003	4.68	.281	.221
Total hospital HCU^d	2.13 (–1.6, 5.8)	.25
Total community-based HCU^e	0.75 (–1.3, 2.8)	.47
Total mental HCU^f	0.40 (–0.9, 1.7)	.55
Total medications used	1.41 (0.4, 2.4)	.008**
Depression score^b		.11	2.00	.143	.071
Total hospital HCU	2.69 (–1.7, 7.1)	.22
Total community-based HCU	2.17 (–0.3, 4.6)	.09
Total mental HCU	–0.72 (–2.3, 0.9)	.37
Total medications used	0.19 (–1.0, 1.4)	.75
Days off work^c		.06	2.50	.189	.113
Total hospital HCU	2.06 (–0.3, 4.4)	.08
Total community-based HCU	0.87 (–0.4, 2.1)	.17
Total mental HCU	0.18 (–0.7, 1.0)	.68
Total medications used	–0.08 (–0.7, 0.5)	.80
Days working/studying^c		.39	1.05	.081	.004
Total hospital HCU	–1.54 (–7.6, 4.5)	.61
Total community-based HCU	–1.72 (–5.1, 1.7)	.32
Total mental HCU	–1.11 (–3.3, 1.1)	.32
Total medications used	0.66 (–1.0, 2.4)	.44
Days socialising^c		.19	1.59	.117	.043
Total hospital HCU	–3.43 (–9.2, 2.3)	.24
Total community-based HCU	0.76 (–2.5, 4.0)	.64
Total mental HCU	0.11 (–2.0, 2.2)	.92
Total medications used	1.70 (0.1, 3.3)	.04*

HCU: healthcare use; PROMIS: patient-reported outcomes measurement information system.

*Significant at p<.05 level.

**Significant at p<.01 level.

^aAnxiety was measured using the PROMIS anxiety short form.

^bDepression was measured using the PROMIS depression short form.

^cIn the last 4 weeks.

^dHospital HCU included emergency presentations, hospital admissions and hospital-based nurses.

^eCommunity-based HCU included general practitioners and nurses.

^fMental HCU included psychiatrists, psychologists, social workers, counsellors and cancer support organisations.

Discussion

This study examined healthcare use among parents of childhood cancer survivors, whether it was related to their general functioning, and also whether sociodemographic factors influenced it. Our findings address questions surrounding how parents use health services beyond their child’s cancer treatment, and how vulnerable subgroups might be identified.

We found similar overall healthcare use among our cancer and comparison parent groups. However, parents of cancer survivors had a different pattern of engagement with health services compared to other parents. They were more likely to have accessed at least one mental health service in the last six months but less likely to have sought support from other community health services such as GPs or nurses. In particular, more parents of survivors accessed support from social workers relative to comparison parents, and fewer reported visiting a GP. Social workers are often involved in supporting parents and families of cancer patients, and such support can extend past the end of treatment [36,37]. Although our study did not assess parents’ support needs, our results may reflect a tendency for parents to seek support from support systems established during their child’s treatment over other community healthcare professionals. Supporting families to transition away from hospital-based support services is an important aspect of their reintegration into their local communities post-treatment. The existence of appropriate systems that can support parents to move from on-treatment support structures to other community services may therefore be important to consider.

Across groups, only a minority of parents (<25%) reported accessing a community-based support organisation. Since cancer-related support organisations often provide relevant services for patients and families during survivorship, there may be room for improvement in the levels of access among parents of survivors. Some parents who may benefit from support may be unaware of relevant services provided by these organisations; interactions between parents and healthcare professionals throughout their child’s treatment may therefore provide valuable opportunities for clinicians to offer such information before families leave the hospital system.

Demographically, our comparison parents were more likely to have just one child compared to the cancer group. This group difference may have had some effect on our findings, for example by impacting parents’ available time or resources to access healthcare. Further research may explore if and how the number of children impacts parents’ healthcare use.

Our analysis of medication use in parents of survivors also highlights an important, and to date understudied, facet of their wellbeing. Supplements were the most commonly reported medications used, usually for vitamin/mineral deficiencies. Several parents in our cancer group also reported using other complementary or alternative medicines (CAM), including herbal medicines and homeopathic remedies. Research on CAM use among parents of survivors is limited, but many paediatric cancer patients and survivors use or express interest in CAM [38–40]. More research is required to establish how and why parents might use CAM. Previous studies have reflected on the value of healthcare professionals discussing the relative risks and benefits of incorporating CAM use for cancer survivors; there may be value in extending this to parents, given our data may also point to their CAM use [38,41].

Few sociodemographic and cancer-related characteristics included in this study were associated with the extent of parents’ healthcare use. The only sociodemographic factor that predicted healthcare use was the child’s sex. Parents of male survivors on average reported accessing more community health services. Previous research has suggested that the health-related quality of life and overall adjustment reported by childhood cancer survivors can differ by gender [42–45]. It is possible that differences in the child’s adjustment are related to the amount of support that parents need or feel able to access for themselves, however, this remains speculative. Future research could use longitudinal data to investigate the association if it is replicated. In contrast with research identifying differences in healthcare use among childhood cancer survivors based on demographic and clinical factors [46], no other clear predictors of parental healthcare use emerged. This highlights the ongoing value of individualised assessment of parents’ support needs to identify vulnerable sub-groups.

Overall, parents of survivors reported functioning well. Despite parents in the cancer group reporting higher anxiety levels than the comparison group, almost all perceived their general health as good or excellent. They had similar levels of engagement with work and study relative to comparison parents, although reported fewer days socialising with other parents. This could be a result of parents feeling more isolated and less able to interact socially, or prioritising productive activities over leisure [47]. Parents taking more medications reported higher anxiety levels, which may reflect appropriate healthcare use given that pharmacological treatments are more likely to be offered when mental health symptoms are more severe [48]. Their extent of healthcare use did not otherwise relate to their level of distress or functioning. Future research in this area will benefit from developing a conceptual model to link parental health, mental health, and healthcare use; currently, no theoretically grounded or evidence-based models are available to guide analyses on this topic.

Strengths and limitations

Our study provides one of the first investigations into the healthcare use of parents of cancer survivors relative to a comparison group. Several limitations warrant consideration. Firstly, the cross-sectional study design was unable to examine changes in healthcare use or general functioning over time, or the directionality of any relationships found. Our sample size was also relatively small and inadequately powered to examine associations with individual subcategories of medication use. Future studies with larger samples, longitudinal designs and using other medication-tracking methods more reliable than retrospective self-reporting will benefit the field.

As parents in our cancer group were participants in an earlier intervention trial, they may not be entirely representative of the broader population. Participating parents may either have been more distressed and seeking support, or conversely, less distressed and therefore able to engage in a multi-session study with peers. We were unable to compare our sample’s demographic or psychosocial functioning to parents who opted out; further research is required to confirm the generalisability of our findings to the broader population of childhood cancer survivors’ families.

There was a non-significant trend for fewer fathers in our cancer group relative to the comparison group. Our findings may therefore reflect mothers’ experiences more than fathers'. Lower participation rates of fathers in paediatric oncology research are unfortunately common; ongoing efforts to improve participation are warranted [49].

This study did not assess parents’ current physical/mental health conditions or desire for greater support (unmet needs). Consequently, our data cannot determine the extent to which healthcare use was appropriately matched with their level of need. Finally, the extent of healthcare use by parents in this study was determined by the number of different services accessed. As such, participants accessing the same service on multiple occasions within the study period would not be reflected in the data.

Conclusion

This study fills a gap in current knowledge about parents’ wellbeing in the context of their child’s cancer survivorship. Parents of survivors reported greater reliance on mental health services relative to comparison parents, and could benefit from support to increase engagement with other community supports such as GPs and cancer support organisations. Overall, data from our cohort indicate that the factors influencing parents’ extent of healthcare use are likely to be nuanced, as no sociodemographic or cancer-related factors strongly accounted for parents’ healthcare use. Individualised survivorship care may therefore be valuable in identifying and meeting different parents’ support needs.

Acknowledgments

The authors would like to thank all the parents who participated in this study. We would like to acknowledge the following people for their contributions towards the Cascade study: Antoinette Anazodo, Felicity Anicich, Kate Awford, Julia Bänziger, Belinda Barton, Lauren Carlson, Amanda Carter, Richard De Abreu Lourenco, Luciano Dalla-Pozza, Emma Doolan, Peter Downie, Gadiel Dumlao, Holly Evans, Afaf Girgis, Martha Grootenhuis, Janelle Jones, Madeleine King, Stephanie Konings, Maggie Leung, Cherie Lowe, Maria McCarthy, Kate Marshall, Sanaa Mathur, Gordon Miles, Michael Osborn, Pandora Patterson, Eden Robertson, Nicole Schneider, Akshay Sharma, Emily Spencer, Julie Ta, Kate Turpin, Janine Vetsch, Rosalie Viney, Nicole Weavers, and Helen Wilson.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported in part by a project grant for the Cascade study, funded by Cancer Australia, Priority-Driven Collaborative Cancer Research Scheme [APP1065428]. Ursula Sansom-Daly is supported by Early Career Fellowships from the CINSW [ID: 14/ECF/1-11] and the National Health and Medical Research Council [ID: APP1111800]. Claire Wakefield is supported by a Career Development Fellowship from the National Health and Medical Research Council of Australia [APP1143767]. The Behavioural Sciences Unit is proudly supported by the Kids with Cancer Foundation and by the Kids Cancer Alliance as well as a Cancer Council New South Wales Program [Grant PG16-02] with the support of the Estate of the Late Harry McPaul.