Pain, physical and psychosocial functioning in adolescents at-risk for developing chronic pain: A longitudinal case-control study

Abstract

This longitudinal case-control study aims to 1) compare symptoms and functioning in otherwise healthy adolescents with versus without a parent with chronic pain (Parent CP+/Parent CP−) 2) test adolescent sex as a moderator of the relation between parent CP group and child functioning, and 3) determine changes in adolescent pain over one year. Adolescents (n=140; ages 11–15) completed tests of pain responsivity and physical function, as well as self-report measures assessing pain characteristics, somatic symptoms, and physical and psychosocial functioning. Self-reported pain and somatic symptoms were re-assessed one year later. Adolescents in the Parent CP+ group reported greater pain, somatic symptoms, and worse physical health than Parent CP− youth. Parent CP+ youth performed worse on all tests of physical function. Some observed effects were stronger for girls than boys. There were no differences between groups on pain responsivity. Both groups reported increased pain and somatic symptoms from baseline to one-year follow-up, with the Parent CP+ group reporting the highest level of symptoms at both time points. This study highlights the potential impact of parental pain status on children, particularly daughters, and is the first to document objective physical functioning differences in youth at risk for developing chronic pain.

Introduction

The burden experienced by adults with chronic pain extends to their children, who are at increased risk for developing chronic pain compared to children of parents without pain ³². Research with clinical, community, and epidemiological samples consistently shows increased risk of chronic pain and disability among offspring of adults who have chronic pain ^11,28,53,59, and a large (n>5000) family linkage study demonstrated that maternal and paternal chronic pain increased the odds of chronic pain in adolescents ³³. These findings and others suggest that in addition to genetics, family environment and parent behavior play an important role in conferring chronic pain risk⁶². Beyond pain, youth are at risk for additional negative outcomes, including higher rates of general somatic complaints and internalizing problems ^{29,42, 58}. A recent metaanalysis of studies of offspring of parents with chronic pain found higher pain, and increased internalizing and externalizing problems in these children compared to controls ³²

Psychosocial models of the intergenerational transmission of chronic pain and somatic symptoms have been examined in mothers with a variety of pain conditions ^{40, 60, 61, 65}. Results show that parent pain and behavior influence child pain, functional impairment, somatic symptoms, and healthcare utilization ^{41, 42, 54, 61}. Some studies have found sex-dependent effects with familial pain history having a stronger relation to females’ pain than to males’ pain ^{18, 21}. In healthy children, girls who perceived that their mothers had recurrent pain or observed their mothers exaggerate responses to laboratory pain reported higher levels of their own pain intensity in laboratory pain tasks compared to boys ^7,19. Thus, girls may be more vulnerable to the potential negative impact of observing parental pain behaviors. However, previous work generally has not examined sex or gender differences in outcomes among the offspring of parents with chronic pain, or has controlled for child sex but not examined it as a primary variable e.g.,⁴². Although research on clinical and high risk samples has been encouraged, to our knowledge no prior research has examined daily pain experiences and response to pain under controlled laboratory conditions among the offspring of a clinical sample of parents with chronic pain⁷⁷.

To fill this gap the current study examined laboratory pain responses, daily pain, psychosocial (i.e., depression, pain catastrophizing, fear avoidance^{24, 37, 67}), and health-related functioning in adolescents who may be at increased risk for developing chronic pain due to having a parent with chronic pain. The study utilized a clinical sample of parents with chronic pain conditions, as well as a pain-free parent control group and was designed to capture youth during the vulnerable transition to adolescence when pain prevalence increases, particularly for females⁵⁹, and often persists into adulthood^{8, 36} Aims are threefold The first is to describe differences in pain, somatic symptoms, psychosocial functioning and physical functioning between otherwise healthy adolescents with and without a parent with chronic pain (Parent CP+/Parent CP−). It was hypothesized that youth in the Parent CP+ group would report higher pain frequency and intensity, and higher somatic symptoms. It was also hypothesized that Parent CP+ youth would evidence greater pain sensitivity in the laboratory, and demonstrate poorer physical and psychosocial function. A second aim was to test adolescent sex as a moderator of the relation between parent CP group and child functioning. It was hypothesized that adolescent females of parents with chronic pain would demonstrate the poorest functioning across domains. The third aim was to examine pain and somatic symptoms over time with longitudinal data collected at one year. It was hypothesized that the CP+ group would have greater increases in pain and somatic symptoms over one year compared to the CP− group.

Method

Participants

Participants were parent-child dyads (n = 140, children aged 11–15 years) participating in a longitudinal study evaluating risk for chronic pain in adolescents. All participants in the study were included in the current analyses. Recruitment, enrollment, and baseline data collection took place between August 2011 and November 2013. Power analyses were conducted using G*Power version 3.1²⁰. The study was originally powered to detect baseline group differences with a planned sample size of n=70 per group, which was selected to provide over .80 power with alpha of .05 to detect a 1 point group difference in usual pain intensity on a 0–10 NRS scale and a 1 point group difference on a 0–6 pain frequency scale. Previous manuscripts from the baseline dataset have evaluated the impact of parent chronic pain and mental health on parenting and responses to child pain ^26,76 and neighborhood characteristics of these dyads ⁵⁶ This is the first manuscript to report on group differences in laboratory pain and physical function tasks, and to examine longitudinal findings.

Parents were from two groups: 1) a clinical sample of adults receiving medical treatment for a chronic pain condition (n=62; Parent CP+ group), and 2) a community sample of adults who did not have chronic pain (n=78; Parent CP− group). Specific inclusion and exclusion criteria for parents in the chronic pain group were: 1) presence of chronic pain defined by the International Association for the Study of Pain (IASP) guidelines as pain that persisted for >3 months, 2) currently receiving specialty care medical treatment for their pain problem, and 3) reported during screening that pain occurred weekly or more frequently. Parents with chronic pain were excluded if pain was due to a serious disease or chronic illness (e.g., cancer, arthritis). Potentially eligible parents in the healthy group were excluded if they: 1) reported having a past or current pain problem that persisted for >3 months or 2) had a spouse or partner living in the home who had a pain problem lasting for >3 months. Parents in both groups were excluded if they had a serious chronic medical problem (e.g., cancer, inflammatory bowel disease, diabetes) or were not fluent in English. All information about parent pain and health characteristics was obtained via parent self-report.

General study inclusionary criteria for children were: 1) between ages 11–15 years at study enrollment, 2) no presence of serious or chronic illness, 3) biological child of the participating parent, 4) English speaking, and 5) no cognitive impairment defined as being diagnosed with developmental disability or intellectual disability. Participating youth of parents with chronic pain are referred to as “Parent CP+ youth” and participating youth of parents without chronic pain are referred to as “Parent CP− youth”.

Only one target child was selected to participate per family. In families where two children in the age range were willing to participate, the research assistant did a coin flip to select which child would participate.

Procedure

This study was approved by the institutional review board at the participating academic medical center. Potentially eligible adults with chronic pain were identified through the university’s Comprehensive pain center medical records and sent a letter and brochure in the mail inviting them and their child to participate in the study. Flyers about the study were also posted in the Pain Center’s clinical offices, and were provided to chronic pain support and treatment groups in the community. Flyers were also sent via electronic mail to a list of adults with fibromyalgia who previously agreed to be contacted about research. Pain-free parents and their children were recruited from announcements in the community and postings in public places, including the medical center’s research recruitment website. Flyers instructed participants to call if they were interested in participating in a parent and teen health study, or to enter their contact information on a study website.

The study coordinator screened and enrolled potentially interested participants from both groups via phone; consent/assent was obtained verbally and participants were scheduled for a baseline study visit. Five potential participants who were screened did not meet inclusion criteria (in 3 families the child in the target age range was not a biological child; 2 families screened for the healthy group were excluded because a biological parent had a history of chronic pain). No participants declined participation during the phone screen, but one child became ill during the study visit and the family declined to complete study procedures (active decline). Two other families did not attend scheduled study visits and thus were not enrolled (passive decline).

This prospective study consisted of two time points, a baseline visit (in-person) and a 12-month follow-up. At the initial visit written consent/assent was obtained from parents and children. Children completed laboratory pain testing, tests of physical function, and had height and weight collected. Participants completed survey questionnaires at baseline and 12 months via Research Electronic Data Capture (REDCap), a secure web-based computerized survey system ³⁰. At baseline, the REDCap survey was completed during the lab visit. At 12 month follow-up parent and youth were sent links to the surveys via email and were instructed to complete surveys independently. Parent and child participants were compensated with gift cards to local stores ($30.00 for parents and $40.00 for children for their participation in the baseline assessment, and $20.00 each for completion of follow-up surveys).

Measures

Participant Characteristics

Parent and family sociodemographics

Parents completed a Family Information Form to document demographic and socioeconomic information. Measures of sociodemographic status included child ethnicity, family income, parental education level, and parental marital status. Child and parent age at the initial study visit were calculated from dates of birth.

Adolescent body mass index (BMI)

Adolescent height and weight were taken at the initial study visit. The CDC’s pediatric BMI calculator ¹⁴ was used to calculate BMI percentile based on the adolescent’s age at the day of visit.

Adolescent Pain and Somatic Symptoms

Adolescents reported on their pain characteristics using a retrospective Pain Questionnaire. This brief instrument asks children to report on pain characteristics in the past 3 months, including location, frequency, and usual pain intensity. Presence of pain in nine standardized body regions was assessed via a body map, coded as 0 = No, 1 = Yes ³⁹. Pain frequency items were based on questionnaire items used in the World Health Organization’s survey on health behaviors in school-age children ¹⁶, which have been used previously to assess recurrent pain frequency in this age group ⁵⁹. Closed-ended response options include 0 (Not at all), 1 (Less than 1 time per month), 2 (1–3 times/month), 3 (About 1 day/week), 4 (2–3 days/week), 5 (4–6 days/week), and 6 (Daily). Pain intensity was measured using an 11-point Numeric Rating Scale (NRS; 0–10). The Pain Questionnaire has been used in previous studies with this age group and psychometric data on these pain perception variables are available ^{48, 51}.

Adolescent somatic symptoms were assessed using child-self report on the Children’s Somatization Inventory-24 (CSI-24). The CSI-24 assesses the presence of somatic symptoms including aches, pains and other physical complaints such as nausea. The CSI-24 has demonstrated excellent psychometric properties ⁷², and internal consistency in the current sample was excellent, Cronbach’s α = 0.91.

Laboratory pain responses were assessed at a study visit. Adolescents were separated from parents and underwent a pain testing protocol that included pressure pain threshold assessment using a computerized algometer (AlgoMed; Medoc Ltd; 2010). Pressure was applied to the flexor carpi radialis muscle on the anterior forearm of the non-dominant hand, and participants were instructed to push a response button when the pressure became painful.

Amount of pressure (kPa) that is being administered at the time the response button was pushed is recorded in the AlgoMed program. A pressure pain threshold score was obtained by averaging pressure pain thresholds across administration of three stimuli to the flexor carpi radialis. Similar protocols are commonly used to assess pressure pain threshold in adults with and without chronic pain⁴⁹. Adolescents also completed a cold pressor task (CPT) to assess cold pain tolerance time and cold pain rating. The CPT is a commonly used lab pain task, has established guidelines for use ⁷⁰, and is considered ethical for use in pediatric samples ⁶ A laboratory grade CPT apparatus was utilized to maintain water at 10° C via a dip cooling unit and water circulating/temperature control unit (Techne, Inc.; 2008). In the current study, the CPT was administered with an uninformed four minute ceiling, with time from hand immersion until withdrawal being the measure of cold pain tolerance. Participants rated pain intensity of the CPT immediately after their hand was withdrawn, using a 0–10 NRS. The order of pain task administration was not counterbalanced (i.e., pressure pain threshold testing always preceded CPT).

Adolescent Physical Functioning

Adolescents underwent two standardized laboratory tests of physical function. The sit-to-stand task assesses ability to move rapidly from a seated to standing position during a 1 minute period. This score is the number of complete movements from seated to standing that the adolescent completed during the task time limit, with higher scores indicating better functioning. A timed 10 meter walk was used to assess walking speed. Similar tasks have been used previously in the assessment of physical functioning in healthy youth and youth with chronic pain ^{1,17, 50}. While normative references are not available for the exact timed motor tests used in this study, the psychometric properties of highly similar measures (including the Five Time Sit-to-Stand test and timed gait speed tasks of various distances) have been evaluated in healthy samples of children and adolescents. Such simple motor tasks have been shown to be normally distributed, reliable, and stable in youth over the age of 6.

Adolescents also reported on their own physical health-related quality of life using the 5-item Physical Health subscale of the Pediatric Quality of Life Inventory, Short Form (PedsQL), a reliable and well-validated measure ⁶⁶. Respondents rate how much of a problem they have had with physical activities in the past month on a 5-point scale ranging from (0) Never to (4) Almost Always. Responses are reverse scored, transformed, and averaged to result in a scaled score ranging from 0–100. Higher scores indicate better physical health-related quality of life. In the present study this scale demonstrated acceptable internal consistency, with a Cronbach’s α = 0.77.

Adolescent General and Pain-specific Psychosocial Functioning

Adolescent depressive symptoms

The 10-item major depressive disorder subscale of the Revised Child Anxiety and Depression Scale (RCADS child-report) was used to assess child depressive symptoms. Good internal consistency has been demonstrated. The measure shows good one-week test-retest reliability, and validity has been shown through correlations with other depression symptom measures ^{12, 13}. This scale demonstrated good internal consistency in the current sample, with a Cronbach’s α = 0.87.

Adolescents also reported on their psychosocial health-related quality of life using the Psychosocial Health score of the PedsQL, a well-validated measure that has been used in a variety of pediatric populations ⁶⁶ Respondents rate how much of a problem they have had with emotions, school, and social activities in the past month on a 5-point scale ranging from (0) Never to (4) Almost Always. Responses are reverse scored, transformed, and averaged to result in a scaled score ranging from 0–100. Higher scores indicate better psychosocial health-related quality of life. In the present study this scale demonstrated good internal consistency, with a Cronbach’s α = 0.84.

Adolescent pain catastrophizing in response to daily pain experiences was assessed using adolescent report on the 13-item Pain Catastrophizing Scale for Children (PCS-C). Response options are on a 5-point scale (0–4) ranging from 0 (not at all) to 4 (extremely). This scale shows good internal consistency and reliability and has been validated for use with 8 to 16 year old children ¹⁵. Internal consistency was excellent in the current sample, Cronbach’s α = 0.92.

Fear-avoidance beliefs about pain and physical activity were assessed via adolescent report on the 5-item Fear and Avoidance Beliefs Questionnaire – Physical Activity (FABQ-PA) subscale ⁷¹. This valid and reliable measure assesses the degree to which a respondent fears or believes that physical activity will cause harm and worsen pain (e.g., “Physical activity makes my pain worse”). Response items range from 0 (completely disagree) to 6 (completely agree), with higher scores indicating higher fear-avoidance. This measure has been used in previous studies of clinical samples of adolescents with chronic pain ⁷⁸. Internal consistency in the present sample was acceptable, Cronbach’s α = 0.78.

Statistical Analysis

Data analysis was conducted with SPSS v.24. Data analytic procedures included calculation of descriptive statistics, including ranges, skewness, means and standard deviations on all study variables in the total sample and in the Parent CP+ and Parent CP− groups. Bivariate associations were conducted between potential parental and family covariates and study outcome variables. T-tests and chi-squared tests were used to test group differences on covariates. Parental and family covariates that were correlated with outcomes and differed significantly in the Parent CP+ and Parent CP− groups were entered as covariates in subsequent analyses. Adolescent age and sex were selected a priori as potential covariates for analyses of all outcome domains, and adolescent sex was examined as a potential moderator of study group in all MANCOVAs. Further, preliminary analyses indicated significant associations between parent age, parent education, and children’s performance on laboratory tasks, and were thus included in MANCOVAs as covariates. In addition, BMI percentile was selected a priori as a covariate of physical functioning, and was included in the MANCOVA predicting physical functioning outcomes. Chi-square analysis was used to test differences in rates of weekly or more frequent pain in the two groups.

Multivariate analysis of covariance (MANCOVAs) were conducted via the general linear model command in SPSS to test Parent CP+ vs. Parent CP− group differences, and group by sex interactions across the four domains of interest: pain characteristics and somatic symptoms, laboratory pain responses, physical functioning, and psychosocial functioning. Variables were grouped into these domains a priori. In instances in which the omnibus model testing group differences was significant, univariate tests were used to ascertain the significance of each variable in the model. Given prior research that has tested depressive symptoms as the primary psychological outcome in samples of offspring of parents with chronic pain ³², we a priori elected to examine the univariate test for group differences in depressive symptoms.

Repeated measures analysis of variance (ANOVAs) were conducted via the general linear model command in SPSS to test changes in primary pain outcomes (pain frequency, pain intensity, and somatic symptoms) in both groups over time. A group by time interaction was conducted to test whether the Parent CP+ group increased significantly more than the Parent CP− group. All statistical tests used an alpha level of 0.05

Results

Participant characteristics and missing data

Of the n = 140 parent-child dyads, data from 4 children was dropped or missing. Thus, the final sample size in the current analyses was n = 136, representing n = 76 in the Parent CP− group (87 .3% mothers) and n = 60 in the Parent CP+ (90.2% mothers) group. The clinical sample of adults with chronic pain included the following primary pain problems: fibromyalgia. (40.3%); back pain (19.3%); other musculoskeletal pain (e.g., neck, shoulder, 24.1%); headache (8.1%); abdominal pain/IBS (6.5%); other chronic pain problem (1.7%). Participating youth were 60.6% female and 11–15 years of age (M = 13.39 years, SD = 1.47). participating parents were 88.6% female and 27–59 years of age (M = 43.82, SD = 6.71). Adolescent and parent sex and adolescent age were evenly distributed across the Parent CP− and Parent CP+ groups. Child ethnicity and racial background were similar in the two study groups. Parent marital status was also similar in the two groups (see Table 1). Parent age, education, and family income were significantly different in the two study groups, with all being lower in the Parent CP+ group (see Table 1). Follow up data were available for n = 114 youth (83.9% of the original sample). The CP− and CP+ groups did not differ significantly on follow up retention, which was 88.2% in the CP− group and 78.7% in the CP+ group. Youth who did and did not complete follow up data collection were not significantly different on any of the baseline variables of interest or covariates examined in the current study.

Table 1.

Sample sociodemographic characteristics by group

	Parent CP− n = 76 M (SD) / n (%)	Parent CP+ n = 60 M (SD) / n (%)	Total Sample n = 136 M (SD) / n (%)
Child characteristics:
Age in years (range 11–15)	13.36 (1.48)	13.62 (1.48)	13.47 (1.48)
Sex (% female)	49 (62.0%)	36 (59.0%)	85 (60.7%)
Ethnicity
Hispanic or Latino	4 (5.1%)	6 (9.8%)	10 (7.1%)
Not Hispanic or Latino	75 (94.9%)	55 (90.2%)	130 (92.9%)
Racial background
African American	2 (2.5%)	3 (4.9%)	5 (3.6%)
American Indian/Alaska Native	0 (0.0%)	1 (1.6%)	1 (0.7%)
Asian	2 −2.5%)	1 (1.6%)	3 (2.1%)
Caucasian	63 (79.7%)	51 (83.6%)	114 (81.4%)
Mixed Ethnicity	12 (15.2%)	5 (8.2%)	17 (12.1%)
Family and parent characteristics:
Parent age in years (range 2 7– 59)*	44.81 (6.46)	42.46 (13.62)	43.82 (6.71)
Parent sex (% female)	69 (87.3%)	55 (90.2%)	124 (88.6%)
Family income*
Up to $25,000	4 (5.1%)	15 (24.6%)	19 (13.6%)
$25,000 – 49,999	16 (20.3%)	18 (29.5%)	34 (24.3%)
$50,000 – 74,999	16 (20.3%)	9 (14.8%)	25 (17.9%)
$75,000 – 99,999	20 (25.3%)	9 (14.8%)	29 (20.7%)
$100,000 – 149,999	10 (12.7%)	6 (9.8%)	16 (11.4%)
$150,000 – 199,999	7 (8.9%)	3 (4.9%)	10 (7.1%)
$200,000 and above	6 (7.6%)	1 (1.6%)	7 (5.0%)
Parent education level*
Some high school	0 (0.0%)	1 (1.6%)	1 (0.7%)
High school or GED	6 (7.6%)	9 (14.8%)	15 (10.7%)
Technical school	1 (1.3%)	5 (8.2%)	6 (4.3%)
Some college/Associate’s	15 (19.0%)	24 (39.3%)	39 (27.9%)
Bachelor’s degree	24 (30.4%)	12 (19.7%)	36 (25.7%)
Master’s or Doctoral	33 (76.7%)	10 (16.4%)	43 (30.7%)
Parent marital status
Single (never married)	6 (7.6%)	6 (9.8%)	12 (8.6%)
Married	51 (64.6%)	30 (49.2%)	81 (57.9%)
Separated or divorced	13 (16.5%)	14 (23.0%)	27 (19.3%)
Widowed	0 (0.0%)	1 (1.6%)	1 (0.7%)
Remarried	4 (5.1%)	6 (9.8%)	10 (7.1%)
Living with partner	5 (6.3%)	3 (4.9%)	8 (5.7%)
Other	0 (0.0%)	1(1.6%)	1 (0.7%)

Note:

^*denotes variables significantly different between Parent CP+ and Parent CP− groups, p < .05

Group Differences

Four separate MANCOVA analyses were conducted to compare adolescents in the Parent CP+ and Parent CP− groups on pain characteristics and somatic symptoms, laboratory pain responses, physical functioning, and psychosocial functioning variables. A priori, all MANCOVAs included parent age, parent education, and child age as covariates. The MANCOVA testing physical functioning outcomes also included child BMI percentile.

Pain characteristics and somatic symptoms

In the first analysis evaluating group differences on adolescent pain characteristics (i.e., pain frequency, pain intensity) and somatic symptoms, the multivariate model was significant for group as hypothesized, Wilk’s Λ =.90, F(3,127) = 4.70, p=.004, η²=.10. All outcome variables were moderately correlated as expected with r values ranging from .58 to .61 (p values < .001). Follow-up univariate tests revealed the youth in the Parent CP+ group, compared to youth in the Parent CP− group, reported significantly greater pain frequency, pain intensity, and somatic symptoms at baseline (see Table 2). Specifically, the rate of weekly or more frequent pain, defined as a response of 3 (about 1 day per week) or higher on the pain frequency item, was significantly higher in youth in the CP+ group (46.7%) compared to the CP− group (18.4%; Chi-square = 12.53, p < .001). Adolescent sex significantly moderated the relation between group and somatic symptoms. Specifically, girls in the Parent CP+ group reported significantly more somatic symptoms than boys in the Parent CP+ group (see Figure 1), but girls and boys reported similar levels of somatic symptoms in the Parent CP− group.

Table 2.

Pain and somatic symptoms: Descriptive statistics and MANCOVA results

Descriptive Statistics
	Parent CP+ M (SD)	Parent CP− M (SD)
Pain Frequency	3.73 (1.74)	2.82 (1.08)
Pain Intensity (0–10 NRS)	3.58 (1.86)	2.54 (1.46)
Somatic Symptoms	15.43 (14.23)	9.95 (8.21)
MANCOVA Results: Univariate tests
	F	p	η2
Pain Frequency
Group	11.78	.001	.084
Group x Sex	.944	.333	.007
Pain Intensity
Group	10.56	.001	.076
Group x Sex	.138	.711	.001
Somatic Symptoms
Group	4.91	.028	.037
Group x Sex	3.92	.050	.029

Notes: Covariates in model include adolescent age, adolescent sex, parent age, and parent education; Significant covariates: none

Figure 1.

Somatic symptoms by study group and child sex

Note: Parent CP+ Females significantly higher than Parent CP+ males, p = .05

Physical functioning

Groups significantly differed on physical functioning based on results from the multivariate analysis, Wilk’s Λ = .90, F(3,123) = 4.31, p = .006, η² = .09. Physical functioning variables were all significantly correlated as expected with r values ranging from .25 to .51 (p values < .01). Follow-up univariate tests revealed differences in the hypothesized direction between groups on all physical functioning variables at baseline (sit to stand task, timed walk, and self-report of physical function on the PedsQL; see Table 3). Specifically, adolescents in the Parent CP+ group compared to adolescents in the Parent CP− group completed fewer movements during the sit to stand task, took more time to complete the 10 meter walk, and reported poorer physical function on the PedsQL. Analyses also indicated a group x sex interaction for the timed walk, with girls in the Parent CP+ group having significantly slower timed walks than boys in the Parent CP+ group (see Figure 2). Girls and boys in the Parent CP− group did not significantly differ on the timed walk.

Table 3.

Physical functioning: Descriptive statistics and MANCOVA results

Descriptive Statistics
	Parent CP+ M (SD)	Parent CP− M (SD)
Sit to Stand	28.69 (7.33)	33.20 (8.74)
Timed Walk	5.61 (1.16)	5.04 (.80)
Physical HRQOL	81.44 (18.22)	89.80 (10.80)
MANCOVA Results: Univariate Tests
	F	p	η2
Sit to Stand
Group	6.06	.015	.046
Group x Sex	.823	.441	.013
Timed Walk
Group	5.68	.019	.043
Group x Sex	5.92	.004	.086
Physical HRQOL
Group	7.66	.007	.058
Group x Sex	1.24	.293	.019

Notes: Covariates in model include adolescent age, adolescent sex, adolescent BMI, parent age, and parent education; Significant covariates: child sex on sit to stand (p = .027)

Figure 2.

Timed 10 meter walk by study group and child sex

Note: Parent CP+ Females significantly higher than Parent CP+ males, p = .004

Psychosocial functioning

Psychosocial functioning variables were significantly correlated with each other, with r values ranging from .26 to .73 (p values < .01). Contrary to hypothesized results, the analysis evaluating group differences on psychosocial functioning showed the overall multivariate model was not significant by group, Wilk’s Λ = .96, F(4,126) = 1.40, p = .239, η² = .04. Planned univariate tests did reveal group differences on depressive symptoms, F(1,129) = 4.64, p = .04, with youth in the Parent CP+ group showing more depressive symptoms than youth in the Parent CP− group (CP+ M = 7.32, SD = 5.94; CP− M = 5.09, SD = 3.46). Both groups had mean scores in the normative range. Youth in the CP+ and CP− groups were similar on other psychosocial variables, including psychosocial HRQOL (CP+ M = 74.92, SD = 17.41; CP− M = 80.76, SD = 13.06), pain catastrophizing (CP+ M = 11.20, SD = 9.47; CP− M =8.09, SD = 7.49), and fear avoidance (CP+ M = 9.77, SD = 8.21; CP− M = 8.08, SD = 6.02).

Laboratory pain responses

Laboratory pain responses were all significantly correlated as expected with r values ranging from .22 to .51 (p values < .01). Results from the multivariate model examining group differences on laboratory pain tasks were not significant, Wilk’s Λ =.97, F(4,125) = .869, p = .485, η² = .03, nor were there any significant group x sex interactions when including the covariates of parent age, parent education, and child age in the model. Adolescent age was a significant covariate predictor of pressure pain threshold (p = .012). Youth in the CP+ and CP− groups were similar on lab pain responses, including CPT tolerance time in minutes (CP+ M = 1.81, SD = 1.43; CP− M = 2.44, SD = 1.43), CPT pain intensity ratings on 0–10 NRS (CP+ M = 6.23, SD = 2.32; CP− M =5.58, SD = 2.17), and pressure pain threshold in kPa (CP+ M = 231.28, SD = 106.53; CP− M = 253.18, SD = 113.26).

Post hoc, all MANCOVA analyses were conducted with the addition of family income as a covariate. The overall pattern of significance of omnibus and univariate results was the same, with the exception of depressive symptoms, which was no longer significantly different by group (p = .07).

Longitudinal pain outcomes

To evaluate children’s longitudinal pain outcomes by parental chronic pain group, repeated measures ANOVAs were conducted to examine child reports of pain frequency, pain intensity, and somatic symptoms at baseline and at one year by group. There was a significant within-subject effect of time from baseline to one year follow-up for both groups across pain outcomes. Specifically, adolescents reported increases in pain frequency, Wilk’s Λ =.96, F(1,112) = 4.88, p = .029, η² = .04, pain intensity, Wilk’s Λ =.92, F(1,112) = 9.79, p = .002, η² = .08, and somatic symptoms, Wilk’s Λ =.91, F(1,112) = 10.65, p = .001, η² = .09. Further, significant effects of parental chronic pain group were observed for all pain outcomes with adolescents in the Parent CP+ group compared to adolescents in the Parent CP− group reporting greater pain frequency, F(1, 112) = 14.49, p < .001, η² = .12, pain intensity, F(1, 112) = 18.44, p < .001, η² = .14, and somatic symptoms, F(1, 112) = 10.42, p = .002, η² = .09. The time by group interaction was non-significant for all pain outcomes: pain frequency, Wilk’s Λ =1.00, F(1,112) = .01, p = .915, η² = .00, pain intensity, Wilk’s Λ =.97, F(1,112) = 3.22, p = .075, η² = .03, and somatic symptoms, Wilk’s Λ=.98, F(1,112) = 2.07, p = .15, η² = .02. Overall, adolescents reported increased pain frequency, pain intensity, and somatic symptoms from baseline to one year follow-up with adolescents in the Parent CP+ group reporting the highest levels of symptoms at both time points (see Table 4 for mean scores by group and percent increases over the one year time period by group).

Table 4.

Descriptive statistics of changes in adolescent pain characteristics from baseline to 1 year follow-up by group

	Parent CP+				Parent CP−
	Baseline M (SD)	1 year M (SD)	Change	Percent increase	Baselin M (SD)	1 year M (SD)	Change	Percent increase
Pain Frequency	3.77 (1.64)	4.06 (1.67)	0.29	7.7%	2.82 (1.07)	3.15 (1.58)	0.33	11.7%
Pain Intensity	3.51 (1.80)	4.45 (2.06)	0.94	26.8%	2.60 (1.51)	2.85 (1.97)	0.25	9.6%
Somatic Symptoms	13.94 (11.07)	18.36 (12.68)	4.42	3 1.7%	9.66 (8.33)	11.37 (10.08)	1.71	17.7%

Discussion

Results highlight the impact of parental chronic pain on adolescent pain and functioning across a number of domains over one year. Similar to prior research ³², this study demonstrated higher pain and somatic symptoms in adolescents who have a parent with chronic pain even after controlling for key covariates. Findings highlight additional domains of physical functioning that differ in adolescents of parents with and without pain. Both self-report and objective performance tests of physical function were lower in the Parent CP+ group, even after controlling for child BMI. Given previous research showing reduced physical function in pediatric chronic pain samples ⁷⁹, one possible interpretation is that lower physical functioning puts adolescents at increased risk for developing chronic pain. Although the correlational data does not allow for determinations regarding whether lower physical functioning is best considered a risk factor, correlate, or consequence of pain, physical function differences are observable in this sample of high risk adolescents. Youth in the Parent CP− group may be less vulnerable to chronic pain because of higher physical function. Future research should continue to examine physical function as a potential risk or resiliency factor.

Youth in the Parent CP+ group also reported higher levels of depressive symptoms, consistent with research showing offspring of parents with chronic pain experience higher internalizing symptoms compared to controls³². This finding should be interpreted with caution; the omnibus test for group differences in psychosocial function was not significant, and when additional covariates (specifically family income) were included in the model in post hoc analyses the groups were no longer significantly different on depressive symptoms (p = .07). Additionally, mean levels of depressive symptoms in both groups were within the non-clinical range. More work in this area is needed as prior pediatric research has found depressive symptoms are associated with reduced physical function ^{2, 9,47, 75} and pain persistence³⁴. Early adolescence is a prime developmental period to intervene on physical functioning, as perceptions of ability and activity levels become increasingly fixed ^{31, 44,57}.

Contrary to hypotheses, no differences in lab-based pain responsivity were observed. This may reflect fundamental differences between transient, predictable and controllable pain administered in a laboratory and the less predictable nature of real life pain experiences. While several studies have demonstrated differences in acute pain responsivity between high vs. low risk or pain vs. pain-free groups in the adult literature e.g., ^{3,23, 43} the overall pediatric experimental pain literature is equivocal⁷⁷. Inconsistency of findings may be due to heterogeneity of samples (e.g., variation in pain frequency, location, presence of chronic pain), different pain modalities tested, variations in experimental conditions (e.g., order effects), as well as social-contextual variables that may influence response to pain testing ^{63, 64, 77} Promising findings in the adult literature show the predictive value of quantitative sensory testing in pain populations. Thus future research might determine best practices for pain testing in pediatric samples who may be at risk for chronic pain.

Our secondary aim, to explore adolescent sex as a moderator of the relation between parent CP and child functioning, was partially supported. Daughters in the CP+ group reported more somatic symptoms and had slower walk times compared to Parent CP+ male children. Daughters of parents with chronic pain may have increased susceptibility to poorer outcomes relative to their male counterparts. Although these sex differences were observed across all domains of functioning, they were not consistently observed across all variables within each domain, which introduces challenges to interpretation. Understanding how female sex confers additional risk within an already high risk group is an important area for future exploration.

Finally, longitudinal analyses indicate youth in both groups reported increases in pain and somatic symptoms over one year. Thus, differences in these groups are observable in early adolescence and appear to remain stable over a year. This is consistent with epidemiological data on the incidence of chronic pain increasing between early and mid-adolescence³⁸, which has been demonstrated in a number of large samples ^{27, 35, 46, 55}, suggesting that the physiological changes, role transitions and stressors that characterize this developmental period may increase vulnerability to pain. Although there was no group-by-time interaction effect, the increase in pain and somatic symptoms within the high risk group may reflect clinically important change ^{68, 69} For example, the small increase in pain frequency represents a shift from pain experienced monthly, on average, to pain experienced weekly; likewise, the nearly 30% increase in pain intensity represents nearly a full point 0–10 NRS increase, which is clinically significant⁵. Results suggest that offspring of parents with chronic pain may be on a negative trajectory of greater pain frequency, intensity, and multiple somatic complaints, which may continue across development.

Strengths and Limitations

This longitudinal case control study has a number of strengths. To our knowledge, this is the first study to obtain objective physical function and lab pain data from children of parents with clinically significant chronic pain. The multi-method assessment of physical function was a particular strength, highlighting lower perception of physical function and lower performance of physical tasks in youth in the Parent CP+ group. Our multi-method assessment of pain was also a strength; discrepant findings between pain self-reports and lab pain responses is hypothesis generating. Including other aspects of pain responsivity may prove more valuable (e.g., conditioned pain modulation, temporal summation). The focus on early adolescents, inclusion of a low-risk control, and re-assessing the sample at 12 months, are additional strengths of our study.

There are also limitations to consider. First, our sample was predominantly mothers. The higher percentage of mothers is representative of gender ratios in clinical chronic pain populations ²² and is commonly observed in developmental literature⁵². More work is needed to determine how paternal pain impacts children, and whether this impact is also moderated by child sex. Some of the observed gender differences may be a reflection of parent sex; mothers may be more potent models for daughters than sons, particularly for gendered behavior like pain ^4,10. Parental factors not included in the present analyses (e.g., type of parental pain, pain duration, catastrophizing, physical activity participation, pain interference) likely contribute to adolescents’ pain experiences⁶². Qualitative work with the current sample indicates that CP+ parents report difficulty engaging in physical activities with their child, and impacts on general parenting (e.g., more inconsistency, impatience)⁷⁶. Future research should examine how these factors impact risk in offspring. Additionally, the impact of parental pain may be greater if children experience pain more frequently themselves; this might provide more opportunities for maladaptive parent-child patterns of co-rumination and catastrophizing^{45, 73}. Finally, because the sample was largely Caucasian and non-Latino results may not be generalizable to diverse groups. Cultural beliefs, socio-economic status and family income may affect the intergenerational transmission of chronic pain in complex ways (e.g., stressful family environment, mistrust of medical providers, pain acceptance) that were not captured in the current analyses.

Results identify key child domains that are impacted by parental chronic pain and suggest intervention targets. Family-level interventions are needed for comprehensive pain prevention and treatment. Programs might be modeled on existing prevention programs for youth with family history of depression⁷⁴, or on treatments for youth with chronic pain and their parents, which are actively being studied and might be adapted for prevention²⁵.

Highlights/Perspective.

• Having a parent with chronic pain increases adolescent risk for experiencing pain.

• Physical function is lower in youth of parents with pain compared to those without.

• Group differences in pain and somatic symptoms persist over one year.

• Family based interventions may aid in comprehensive pain prevention and treatment.

Perspective.

Adolescents who have a parent with chronic pain demonstrate higher pain and lower physical function than adolescents who have a parent without chronic pain. Group differences in pain and somatic symptoms persist over one year. Family based interventions are needed for comprehensive pain prevention and treatment.