Caregiver Burden in Caregivers of Children With Special Healthcare Needs and Association With Chronic Pain

Abstract

Background/Objectives:

Chronic non-cancer pain (CNCP) affects millions of individuals in the US but evidence of its prevalence among caregivers of children with special health care needs is sparse. We sought to estimate the prevalence of CNCP and its association with caregiver burden, in a nationally representative sample.

Methods:

Retrospective cross-sectional study using pooled Medical Expenditure Panel Survey data for 2010–2015. Within interviewed households, family groups consisting of at least one parent and one child (0–17 y) were identified. CNCP was identified by one or more ICD-9-CM codes utilizing previously published approaches. Level of caregiver burden was defined using a validated screener questionnaire identifying children with high burden of care (i.e. special health care needs), e.g., high or low burden. We estimated prevalence of CNCP as a function of caregiver burden, as well as the association of risk factors with CNCP, including parent sociodemographic features, clinical diagnoses, and family level characteristics.

Results:

We identified 46,525 caregivers of whom 3.6% reported experiencing high caregiving burden. The prevalence of chronic non-cancer pain (CNCP) was 25.5% and 14.0% among parents with high compared to low caregiving burden, respectively. Odds of CNCP were higher among parents with high compared to those with lower caregiver burden (OR 1.29 95% CI 1.06 – 1.55). Being obese, experiencing disability, and having a mental health diagnosis were associated with higher odds of CNCP.

Conclusions:

Chronic pain is more common among caregivers with high caregiver burden. Our findings highlight the need to further explore the nature and impact of risk factors on caregiver health and disability.

Introduction

Approximately 14.1 million adult individuals in the United States care for a child with special health care needs (CSHCN). CSHCN include children with chronic physical, developmental, behavioral, or emotional conditions; with different levels of disability or functional impairment.^1,2 CSHCN vary in terms of medical needs and complexity of care; greater functional limitations in children are associated with more intensive home-care requirements and greater family burden.^3,4 The burden and stress of caring for a child with special health care needs may contribute to worse parental overall health and quality of life.^5–7

There is some evidence to suggest that pain or pain-associated conditions, such as low back pain or migraines, are more likely to be present among parents experiencing higher caregiver burden than among their counterparts.⁷ Caregiving may increase risk of chronic pain directly (e.g., accidental musculoskeletal injury caring for a child with behavioral outbursts) or indirectly by reducing opportunities for self-care (e.g., social isolation, poor preventative health care, sedentary lifestyle).⁵ Caregivers may also be more likely to be exposed to individual or environmental factors that increase the risk for chronic pain (e.g., depression or limited access to resources).⁶ There is limited evidence on the occurrence of chronic pain and associated individual, family, and contextual factors in the US population of CHSCN caregivers.

Although chronic pain is considered a public health issue, evidence is still limited on its prevalence and risk factors among vulnerable populations, such as caregivers of young children with special health care needs. Using the caregiver burden framework proposed by Raina et al, we explore the individual, family, and contextual characteristics associated with caregiving burden and chronic pain.⁸ Our goal was to estimate the prevalence of chronic pain among caregivers comparing those with high versus low caregiver burden, accounting for parental and family characteristics. As a secondary goal, we examined factors at the individual and family level associated with caregiver chronic pain.

Methods

We conducted a retrospective cross-sectional study using pooled Medical Expenditure Panel Survey (MEPS) Household Component data from 2010 to 2015.⁹ MEPS collects detailed information on the relationship between individuals in each family unit, demographic characteristics, health status, health insurance coverage, medical conditions, income, and employment. Medical conditions are classified through International Classification of Diseases, Ninth Revision (ICD-9) codes truncated at three digits to preserve subject confidentiality. Truncated ICD-9 codes limit the capacity to identify chronic pain per se, so Clinical Classification Codes were added to further specify conditions associated with chronic pain.¹⁰

Study Cohort

From every household in the respective full year consolidated data files, we identified all families that were comprised of: (i) civilian, noninstitutionalized members; (ii) at least 1 parent, aged 18 years or older; and (iii) at least 1 child, aged 17 years or younger. MEPS defines a family as two or more persons living together in the same dwelling unit (household) who are related by blood, marriage, or adoption, as well as foster children. Unmarried individuals living together in the same household who consider themselves a family, are included in this definition.

Our study population was restricted to the parents in these families. We identified parents using a variable constructed by concatenating the dwelling unit identifier (DUID) and either the mother identifier (MOPID42X) or the father identifier (DAPID42X). These parent-identifiers include biological, adopted and stepparents but not foster parents or non-parent guardians.

Variables

Dependent Variable:

Parents were classified as having chronic non-cancer pain (CNCP) if, during the second interview round of each survey year, they communicated one or more conditions associated with chronic pain identified through ICD-9 and Clinical Classification Codes. The selection of these codes was informed by clinical expertise as well as by other studies that have identified chronic pain associated conditions in MEPS.^10–12 (see Supplemental Digital Content A for list of ICD-9 codes).

Independent Variable:

Level of caregiving burden was derived from parents’ responses to the CSHCN Screener developed by the Child and Adolescent Health Measurement Initiative, and included as part of MEPS (see Supplemental Digital Content B)¹³. We classified parents as being exposed to higher caregiver burden if they had at least one child for whom the parent responded “yes” to all 5 stem questions and sub questions on the CSHCN Screener, otherwise they were deemed as exposed to lower caregiving burden. We excluded parents who were ineligible to answer the CSHCN Screener or who did not know, could not ascertain, or refused to answer any of the five questions in the tool.

Covariates:

We assessed parental sociodemographic and clinical characteristics potentially associated with chronic pain including: age, sex, race, ethnicity, insurance coverage, highest educational degree attained, employment status, mental health comorbidities, obesity and parental self report of disability (see Supplemental Digital Content C). At the family level we assessed characteristics such as single parenting or co-parenting, family income category based on income as a percentage of the poverty line. Because parents of children with behavioral disorders experience additional care burden we also measured whether families had at least one child with a mental health disorder.¹⁴ All parental characteristics were assessed in the second interview round of each survey year.

Statistical Analyses

We pooled annual 2010–2015 MEPS Household Component files and examined the differences between levels of caregiver burden by individual and family level characteristics. The six-year, snapshot provides average estimates over the pooled years.

Prevalence of CNCP was estimated by calculating the proportion of the parents who reported any of the included pain-associated chronic medical conditions, over the total number of parents in the sample. We compared the distribution of sociodemographic and clinical characteristics between a) parents with high or low caregiver burden, b) presence or not of chronic pain. Rao Scott Chi square tests and Student’s t-tests were used to compare the parent and family-level characteristics.

Multivariable logistic regression was used to estimate the association (Odds Ratio [OR] and 95% confidence interval [95% CI]) between caregiving burden and CNCP, adjusting for covariates. Robust variance estimation techniques were used to account for possible correlation between repeated measurement of parents’ characteristics, as parents in the pooled six-year cross section were assessed twice by MEPS panel design. Lastly, we applied a domain analysis approach, and person-level survey weighting to obtain appropriate estimates and standard errors for the national prevalence of CNCP and the association between caregiver burden and CNCP.

Statistical analyses were performed using SAS 9.4 (SAS Institute, Inc., Cary, NC). We considered p<0.05 to be statistically significant for all multivariable analyses. The institutional review board at University of Maryland Baltimore authorized an exemption from review for this study.

Results

Across pooled years (2010–2015) we identified 46,525 parents of whom, 3.6% were classified as experiencing high caregiving burden. Cohort characteristics are described in Table 1. Being female and a single parent were more common among parents with high caregiving burden (60.8%, and 26.7% respectively), compared to parents with lower caregiving burden (56.5% and 18.5%). About forty-nine percent (48.7%) of parents with high caregiving burden had one or more children with a mental disorder, compared to 7.3% of parents with lower caregiving burden. A higher proportion of parents in the high burden group had disability, self-perceived fair/poor health status, at least one mental health condition, lower median family income, were on public insurance and unemployed compared to those with lower caregiver burden (Table 1). The prevalence of CNCP was 25.5% among parents with high caregiving burden, compared to 14.0% among those with lower caregiving burden.

Table 1.

Characteristics of US Caregivers by Caregiver Burden, Medical Expenditure Panel Survey 2010 – 2015

Caregiver Characteristics	Total	Caregivers with no children with severe functional limitations	Caregivers with at least 1 child with severe functional limitations	p value
Sample, N	46,525	44,856	1,669
Pooled weighted, N %	66,181,521	63,778,284 (96.4)	2,403,237 (3.6)
Sociodemographic Characteristics
Age years, Mean 95%CI *	38.6 (38.4 – 38.9)	38.6 (38.4 – 38.8)	39.5 (38.7 – 40.2)	0.03
Age category, years, %				0.02
18 – 25	7.3	7.5	3.9
26 – 35	30.8	30.8	30.5
36 – 45	38.4	38.3	41.0
46 – 55	20.6	20.6	22.0
56 and older	2.8	2.8	2.6
Sex, %				<.001
Females	56.6	56.5	60.8
Males	43.4	43.5	39.2
Race/Ethnicity, %				<.001
Non-Hispanic White	59.0	58.8	66.2
Non-Hispanic African/American	11.0	10.9	13.4
Hispanic	21.2	21.4	15.2
Non-Hispanic Asian/Other	8.8	8.9	5.2
Region, %				0.005
Northeast	16.5	16.4	19.0
Midwest	21.1	20.9	26.3
South	37.4	37.5	36.9
West	25.0	25.2	17.8
Parent Type, %				<.001
Single Parent	18.8	18.5	26.7
Co-parent	81.2	81.5	73.3
Level of Education, %*				0.44
High School Diploma/GED or Lower	40.5	40.4	41.5
Beyond High School Diploma/GED	59.1	59.1	58.3
Unknown	0.4	04	0.2
Employment Status, %				<.001
Employed in Survey Round	77.3	77.7	65.0
Unemployed in Survey Round	22.5	22.1	34.8
Unknown	0.2	0.2	0.2
Median Total Family Income, $ 95% CI	$80,079 ($77,591 - $82,568)	$80,425 ($77,914 - $82,935)	$70,912 ($64,432 - $77,391)	0.01
Family Income as % of Poverty Line, %				<.001
Poor/Negative	15.3	15.1	21.7
Near Poor	4.6	4.5	7.5
Low Income	14.8	14.7	16.0
Middle Income	31.5	31.6	29.7
High Income	33.8	34.1	25.1
Insurance Type, %				<.001
Any Private	70.6	70.9	64.1
Public Only	14.0	13.6	23.3
Uninsured	15.4	15.5	12.6
Clinical Characteristics
Caregiver Obesity, %				<.001
Yes	31.2	30.9	39.8
No	66.3	66.6	58.2
Unknown/Refused	2.5	2.5	2.1
Caregiver Mental Disorder Status, %				<.001
Yes	11.4	10.8	25.7
No	88.6	89.2	74.3
Caregiver Disability Status, %				<.001
Some/Serious	1.5	1.3	6.8
None	98.1	98.3	92.3
Unknown	0.4	0.4	0.9
Caregiver has at least 1 Child with a Mental Disorder, %				<.001
Yes	8.8	7.3	48.7
No	91.2	92.7	51.3
Perceived Health Status, %				<.001
Excellent	25.9	26.2	16.1
Very Good	36.7	37.0	26.7
Good	28.2	28.0	35.8
Fair	7.6	7.2	16.4
Poor	1.6	1.5	4.9
Unknown	0.0	0.0	0.2
Has Chronic Pain, %				<.001
Yes	14.4	14.0	25.5
No	85.6	86.0	74.5

^*-Chi-square statistic is not statistically significant at alpha = 0.05; There was not a statistically significant difference between parents with or without a child with severe functional limitations, in terms of parents’ age in years and level of education. All other risk factors assessed showed a statistically significant association.

The proportion of parents experiencing high caregiver burden among those with CNCP was 6.4% (Table 2). Also, a larger proportion of parents with CNCP had one or more children with a mental disorder. Parents with CNCP were overall more likely to be older, single, with higher prevalence of current mental health disorders, limitations in activities of daily living, and obesity.

Table 2.

Characteristics of US Caregivers, by Chronic Pain Diagnostic Status, Medical Expenditure Panel Survey 2010–2015

Caregiver Characteristics	Total	Caregivers without Chronic Pain	Caregivers with Chronic Pain	p value
Sample, N	46,525	40,680	5,845
Pooled weighted, N %	66,181,521	56,634,762 (85.6)	9,546,760 (14.4)
Sociodemographic Characteristics
Age years, Mean (95%CI)	38.6 (38.4 – 38.9)	38.1 (37.9 – 38.4)	40.9 (40.5 – 41.3)	<.001
Age category, years, %				<.001
18 – 25	7.3	7.9	4.0
26 – 35	30.8	32.2	22.7
36 – 45	38.4	37.9	41.5
46 – 55	20.6	19.4	27.8
56 and older	2.8	2.6	4.0
Sex, %				0.60
Females	56.6	56.6	57.1
Males	43.4	44.3	39.2
Race/Ethnicity, %				<.001
Non-Hispanic White	59.0	57.6	67.4
Non-Hispanic African/American	11.0	11.2	9.7
Hispanic	21.2	22.2	15.1
Non-Hispanic Asian/Other	8.8	8.9	7.9
Region, %				<.001
Northeast	16.5	16.1	18.3
Midwest	21.1	20.8	23.3
South	37.4	38.2	32.7
West	24.8	24.9	25.7
Parent Type, %				0.001
Single Parent	18.8	18.4	21.4
Co-parent	81.2	81.6	78.6
Level of Education, %				0.007
High School Diploma/GED or Lower	40.5	40.8	38.3
Beyond High School Diploma/GED	59.1	58.7	61.4
Unknown	0.4	0.4	0.3
Employment Status, %				<.001
Employed in Survey Round	77.3	78.2	71.7
Unemployed in Survey Round	22.5	21.6	28.1
Unknown	0.2	0.2	0.1
Median Total Family Income, $ (95% CI)*	$80,079 ($77,591 - $82,568)	$80,152 ($77,703 - $82,602)	$79,753 ($75,925 - $83,581)	0.80
Family Income as % of Poverty Line, % *				0.56
Poor/Negative	15.3	15.2	16.1
Near Poor	4.6	4.6	4.8
Low Income	14.8	14.9	14.3
Middle Income	31.5	31.7	30.6
High Income	33.8	33.7	34.2
Insurance Type, %				<.001
Any Private	70.6	70.3	72.3
Public Only	14.0	13.2	18.7
Uninsured	15.4	16.5	9.1
Clinical Characteristics
Caregiver Obesity, %				<.001
Yes	31.2	30.0	38.5
No	66.3	67.4	59.8
Unknown/Refused	2.5	2.6	1.7
Caregiver Mental Disorder Status, %				<.001
Yes	11.4	9.0	25.4
No	88.6	91.0	74.6
Caregiver Disability Status, %				<.001
Some/Serious	1.5	0.8	5.7
None	98.1	98.9	93.6
Unknown	0.4	0.3	0.6
Caregiver has at least 1 Child with a Mental Disorder, %				<.001
Yes	8.8	7.7	15.4
No	91.2	92.3	84.6
Perceived Health Status, %				<.001
Excellent	25.9	28.3	11.1
Very Good	36.7	37.4	32.3
Good	28.2	27.3	33.9
Fair	7.6	6.0	16.8
Poor	1.6	0.9	5.7
Unknown	0.0	0.0	0.1
Caregiver has at least 1 Child with Severe Functional Limitations, %				<.001
Yes	3.6	3.2	6.4
No	96.4	96.8	93.6

^*-Chi-square statistic is not statistically significant at alpha = 0.05; There was not a statistically significant difference between parents with or without chronic pain diagnoses, in terms of caregiver’s family income as a percentage of the poverty line. All other assessed risk factors for chronic pain showed a statistically significant association.

Parents with higher caregiving burden had an increased odds of having CNCP, compared to those with lower burden (AOR 1.29 95% CI 1.06 – 1.55) (Table 3.) Odds were also higher among those with a child with a mental health diagnosis (AOR 1.39 95% CI 1.22 – 1.58). Being unemployed, experiencing limitations in daily living, and having a mental disorder diagnosis were associated with higher odds of CNCP.

Table 3.

Association Between Level of Caregiving Burden and Caregiver Chronic Pain, Medical Expenditure Panel Survey 2010–2015.

Characteristic	Adjusted odds ratio Caregiver Chronic Pain, (AOR, [95% CI])
Exposure
Caregivers with at least 1 child with severe functional limitations	1.29 (1.06 −1.55)*
Caregivers with no children with severe functional limitations	Reference
Sociodemographic
Age category, years
56 and older	2.72 (1.98 – 3.73)*
46 – 55	2.68 (2.12 – 3.39)*
36 – 45	2.12 (1.67 – 2.70)*
26 – 35	1.39 (1.10 – 1.77)*
18 – 25	Reference
Sex
Females	0.93 (0.85 – 1.02)
Males	Reference
Race/Ethnicity
Hispanic	0.73 (0.65 – 0.82)*
Non-Hispanic African American	0.81 (0.72 – 0.92)*
Non-Hispanic Asian/Other Race	0.85 (0.72 – 1.00)
Non-Hispanic White	Reference
Region
Northeast	0.93 (0.78 – 1.11)
Midwest	0.92 (0.81 – 1.04)
South	0.81 (0.71 – 0.91)*
West	Reference
Parent Type
Single Parent	1.08 (0.96 – 1.23)
Co-parent	Reference
Level of Education
High School Diploma/GED or Below	0.96 (0.87 – 1.06)
Above High School Diploma/GED	Reference
Employment Status
Unemployed in Survey Round	1.22 (1.10 −1.35)*
Employed in Survey Round	Reference
Family Income as % of Poverty Line
Poor/Negative	1.11 (0.95 – 1.30)
Near Poor	1.21 (0.98 – 1.50)
Low Income	1.14 (0.98 – 1.33)
Middle Income	1.04 (0.92 – 1.17)
High Income	Reference
Insurance Type
Public only	1.16 (0.99 – 1.35)
Uninsured	0.64 (0.55 – 0.75)*
Any Private	Reference
Caregiver Obesity
Yes	1.33 (1.21 – 1.45)*
No	Reference
Caregiver Mental Disorder Status
Yes	2.52 (2.24 – 2.84)*
No	Reference
Caregiver Disability
Some/Serious	3.35 (2.58 – 4.33)*
None	Reference
Caregiver has at least 1 Child with a Mental Disorder
Yes	1.39 (1.21 – 1.58)*
No	Reference

^*-Statistically significant association with alpha level set to 0.05

Discussion

This study highlights the high prevalence of chronic pain among caregivers of CSHCN experiencing higher caregiver burden, among a nationally representative sample. In our sample, 21.4% of caregivers with high burden reported having fair or poor health. Recently, the BFRSS found that more than 19% of caregivers across the US report being in fair or poor health with variation across states.¹⁵ Caregivers who were obese, with current disability, and mental health conditions were more likely to have a clinical diagnosis associated with chronic pain. Despite the evidence on the psychological and emotional impact of caregiving, very little is known about the direct consequences on physical health. Poor caregiver health can have an effect on children’s development and health, especially for children with special health care needs. Parents who have poor health may have challenges attending frequent child appointments and implementing behavioral and physical health interventions at home to support the child’s development. To date few studies have explored the physical health impact of caregiving, especially among those providing care for a child with complex health care needs.

Caregiving is an important public health issue that impacts the health and quality of life across the lifespan of caregivers, care-recipients, and their families.¹⁶ A significant challenge in assessing the impact of caregiving is the limited availability of population-based data measuring caregiving burden. MEPS provides a unique opportunity to overcome this gap as it identifies family units, and collects detailed information on medical conditions and care among family members. This in turn allows estimation of chronic pain among a nationally representative sample, providing valuable information on the impact of caregiving at a national level.

This study has some limitations. First, the definition used for CNCP was based on truncated ICD-9 codes consistent with previously published work.^10–12,17 To reduce overestimation of CNCP we used Clinical Classification Codes when possible as they are based on five digit ICD-9 codes.^18–20 Second, our results are limited to caregivers who completed the CSHCN Screener, however these represent the majority of individuals in the sample. Third, for single or dual parent households we cannot identify which parent carries the highest burden as it is possible that other family provide support when caring for the child. Our assumption is that parents are likely to have the majority of responsibilities and as such were the only family members included in the analyses. The strengths of this study include identification of a large sample from a nationally representative dataset, ability to identify households and family groups, and inclusion of covariates typically absent from secondary data such as caregiver body mass index and disability.

Having estimated the prevalence of chronic pain and associated risk factors among caregivers of children, this work raises new questions. For example, given the high prevalence of chronic pain among caregivers we need to understand access and use of health care, and pain management in this vulnerable population. Assessing the causality between burden of care and development of chronic pain is also critical to identify potential early intervention targets. Longitudinal studies that capture changes in caregiver burden as children develop could help identify primary and secondary chronic pain prevention targets. Further research on co-located pediatric and adult services, could provide unique opportunities to deliver adult chronic pain prevention and interventions. Moreover the role of community level factors in lessening burden or improving caregiver health remains to be explored.

Through this work we provide evidence of the high prevalence of chronic pain among caregivers of children with special health care needs who require high levels of care. Our findings suggest that chronic pain co-occurs with higher care burden among adults who are at the prime of their productive years and considered a high risk group for substance misuse or abuse. This work highlights the need to further explore risk factors for chronic pain and their impact on caregivers’ health and disability.