Maternal Health Matters: Exploring the Links Between Parental Health and Illness Factors and Adolescent Functional Somatic Symptoms

Abstract

Objective

Functional somatic symptoms, which are physical symptoms that cannot be fully explained by any underlying medical condition, run in families. Besides genetic predispositions, parental model behaviors may contribute to the intergenerational transmission of proneness to somatic symptoms. However, studies examining the long-term influences of parental health and illness factors on somatic symptoms in offspring are scarce.

Method

Using data from the Lifelines Cohort Study, cross-sectional associations were investigated between parental somatic symptoms, (chronic) medical conditions, health care utilization, medication use, and health literacy—assessed through a questionnaire and professional involvement in health care—and somatic symptoms reported by their adolescent offspring (13-17 years old) using the Somatic Complaints subscale of the Youth Self-Report. Long-term associations were then examined with the same predictors and symptom reports at follow-up in early adulthood. The study included 3,735 mother–adolescent dyads and a subsample of 2,668 mother–father–adolescent triads containing maternal and paternal information. Negative binomial regression analysis was used to examine associations between parental health factors and offspring symptoms.

Results

Maternal (chronic) medical conditions and pain medication use were significantly associated with symptom reports of adolescents, particularly girls, in both mother–father–adolescent triads and mother–adolescent dyads. Additionally, in the dyads, mothers’ somatic symptoms predicted somatic symptoms in adolescents, especially daughters. Only maternal somatic symptom reports were significantly associated with somatic symptoms of young adult daughters 3 to 4 years later. No significant associations were found for parental health care utilization, health literacy, or fathers’ health and illness factors after multiple comparison correction.

Conclusion

This study examined various parental somatic health indicators, emphasizing the significant role of maternal health in somatic symptom burden of adolescent children. This influence diminishes as adolescents transition into adulthood. These findings highlight the importance of prioritizing maternal health in clinical interventions to reduce somatic symptoms in adolescents. By focusing on early identification and empowering mothers with health education, health care providers can promote healthier behaviors and potentially reduce the intergenerational transmission of somatic symptom tendencies.

Study registration information

Parental factors associated with the persistence of somatic symptoms from adolescence into adulthood; https://osf.io/5htsb/overview

Plain language summary

This study explored how parents’ health, including somatic or physical symptoms, medical conditions, and healthcare utilization, as well as healthcare literacy relate to physical symptoms in their teenagers. Researchers analyzed the Lifelines Cohort Study of Dutch families, which includes 3,735 mother-adolescent pairs, of which 2,688 also have paternal information. Maternal medical conditions and pain medication use were associated with their daughter’s somatic symptoms during adolescence and early adulthood. However, paternal health and healthcare utilization showed no clear associations.

Functional somatic symptoms are common physical symptoms, such as headaches, back pain, and gastrointestinal issues, that cannot be sufficiently explained by organic pathology. Persistent functional somatic symptoms are frequently observed in adolescents.^1,2 Key risk factors for youth developing functional somatic symptoms include a family history of somatic symptoms, such as recurrent parental headaches, and general parental ill health.2, 3, 4 For instance, a study linked somatic symptoms and pain levels in adolescents to parental chronic pain.⁵ After 1 year, adolescents both with and without a parent with chronic pain showed increased symptoms, with the highest levels reported by adolescents with parents who had chronic pain.⁵ Individual genetic predispositions may contribute to the development of somatic symptoms.⁶ Additionally, children of parents with chronic medical conditions may be at greater risk due to social learning mechanisms, such as parental modeling of pain behaviors and reinforcement of the child’s pain.7, 8, 9 The latter refers to parental behaviors such as providing contingent attention and concern, overaccommodating the child’s needs, which can inadvertently perpetuate the child’s experience of pain. How parents interpret, respond to, and communicate about somatic illness may be crucial in the intergenerational transmission of risk for developing functional somatic symptoms.

Health behaviors and outcomes are influenced by individual differences in educational attainment and health literacy, ie, the degree to which people have the capacity to obtain, process, and understand basic health information and services needed to make appropriate health decisions.10, 11, 12, 13 Health literacy levels vary, with people who work in health care professions, such as nurses and physicians, typically having higher levels.¹⁴ Thus, health literacy of parents may affect their illness-related attitudes and behaviors toward their own and their child’s symptoms, influencing the child’s health outcomes.

Previous studies suggest that medical conditions in mothers play a more significant role than medical conditions in fathers in transmitting symptom burden to their offspring.¹⁵ For example, somatic symptoms in 15-year-old adolescents were more strongly associated with poor health in mothers than in fathers, particularly among female participants.¹⁶ However, fathers are often underrepresented in health research, as only mothers, being considered primary caregivers, are usually invited to participate.¹⁷ Additionally, health and illness behaviors of fathers may affect health outcomes of offspring differently. The likelihood of offspring developing somatic symptoms has been found to be higher when both parents, rather than just one, experience recurrent pain.² Therefore, studies should include assessments of a broad range of somatic symptoms and health and illness behaviors of both mothers and fathers to allow for a balanced comparison between maternal and paternal influences and the identification of unique contributing factors and their potential interactions.

Although social learning mechanisms are recognized as key factors in the intergenerational transmission of functional somatic symptoms, empirical evidence explaining why and how symptoms cluster within families is limited and primarily focused on chronic pain conditions.^4,7,18 However, common symptoms may also include nonpain symptoms, such as fatigue or dizziness.

Few studies have explored whether parental health literacy affects children’s health, though 1 study found a negative association between parents’ education and back pain in offspring.¹⁹ Moreover, it remains unclear if parental health literacy, including knowledge of health and disease, associated behaviors, and profession, also impacts the somatic symptom burden of offspring. Furthermore, longitudinal evidence on the persistence of functional somatic symptoms from childhood to adulthood is scarce, with most studies focusing on short-term follow-ups within adolescence⁵ or limited to adulthood.¹⁸ However, parent–child influences may change over time, potentially weakening as offspring gain autonomy during adolescence to early adulthood, a period marked by increased independence in health care decisions and pain management. Research on social influences found that family effects on help-seeking behavior are strong in adolescence but decrease as individuals age, though they remain stronger than peer influences in early adulthood.²⁰ The question remains which aspects of parental health factors have lasting influence on offspring into adulthood and which diminish as adolescents become more autonomous in their health decisions.

This study aimed to fill existing knowledge gaps by using data from a large, prospective 3-generation cohort to link maternal and paternal health factors to the transmission of somatic symptom burden to their offspring during adolescence and early adulthood. Specifically, we assessed the contribution and cross-sectional associations of parental somatic symptoms, medical conditions, health behaviors, and health literacy with functional somatic symptoms in adolescents ages 13 to 17 years. We also investigated which parental factors predicted symptom levels into adulthood (≥18 years).

We hypothesized that parental health factors, such as chronic medical conditions, high somatic symptoms, increased medication use, or frequent health care visits, would be associated with higher rates of functional somatic symptoms in adolescents. We also expected lower parental health literacy to be linked to higher adolescent symptom rates, with these associations persisting into early adulthood.

Method

Participants

This study used data from the Lifelines Cohort Study, a multidisciplinary, population-based cohort examining health and behaviors across 3 generations in 167,729 people in Northern Netherlands. It assesses biomedical, sociodemographic, behavioral, physical, and psychological factors contributing to health and disease, with a focus on multimorbidity and complex genetics. In Lifelines, participants’ sex was derived from municipal registries (58% female), and a large majority of the cohort was born in the Netherlands (97%) and self-reported a White race (98%).²¹ Full cohort and recruitment details are available elsewhere.^22,23 We selected 3,950 participants ages 13 to 17 years with at least 1 parent also enrolled in Lifelines. For the second aim, we used data from their first follow-up assessment in adulthood (≥18 years old). The Lifelines study complies with the Declaration of Helsinki and was approved by the Medical Ethical Committee of the University Medical Center Groningen, the Netherlands. This study was preregistered on Open Science Framework before data analysis (https://doi.org/10.17605/OSF.IO/5HTSB). For details of a few methodological adaptations after registration, see Supplement 1, available online.

Measures

Somatic Symptoms

Somatic symptoms in adolescent participants were measured with the Somatic Complaints subscale of the Youth Self-Report (YSR).²⁴ The Somatic Complaints subscale contains items that refer to somatic symptoms without a known medical cause or without obvious reason. Two items, skin problems and eye problems, were excluded from the analyses, as prior studies reported on low factor loadings, suggesting that these items do not represent the underlying construct functional somatic symptoms well.^1,25 The remaining 7 items were aches/pains, headaches, nausea, stomachaches, vomiting, overtiredness, and dizziness, having 3 Likert-type response options: 0 = “not at all or never”; 1 = “sometimes or a little bit”; 2 = “often or very much.” The total score range was 0 to 14. Adolescents indicated whether and to what extent they had experienced any of these symptoms in the past 6 months.

Somatic symptoms in adulthood were assessed with the 12-item Symptom CheckList-90 Somatization subscale (SCL-90 SOM),²⁶ which has been recommended for large-scale studies.²⁷ Using a Likert-type scale ranging from 1 = “not at all” through 5 = “very much,” participants were asked to indicate the extent to which they were hampered by any of the symptoms listed in the past 7 days. Example items were pain in the lower back, dizziness, nausea, or upset stomach. Due to wrong translation to Dutch language for the baseline assessment, the item “feeling weak physically” was slightly different from the original SCL-90 item “feeling weak in parts of your body.” This mistake was corrected from the first follow-up assessment onward. Participants’ total scores were adjusted by subtracting 12, resulting in a possible range of 0 to 48.

Parental Health

The occurrence of common somatic symptoms in parents was assessed with the SCL-90 SOM (see assessment details under “Somatic Symptoms”). Self-reported lifetime prevalence of (chronic) medical conditions in parents was assessed at baseline using a broad list of prespecified conditions that were categorized into the following 8 disease clusters: respiratory conditions (asthma, chronic obstructive pulmonary disease or chronic bronchitis, sinusitis); neurological conditions (Parkinson’s disease, Alzheimer’s disease, dementia, dizziness, epilepsy, back or neck hernia, migraine, multiple sclerosis, spasticity, stroke, stenosis); musculoskeletal conditions (amputation, rheumatoid arthritis, osteoarthritis, rheumatoid osteoporosis, rheumatoid fibromyalgia, other rheumatism, hip fracture, repetitive strain injury); genitourinary conditions (incontinence, chronic cystitis, disturbed kidney function, kidney disease, kidney stones, renal failure); gastrointestinal conditions (Crohn’s disease, celiac disease, gallstones, hepatitis, irritable bowel syndrome, liver cirrhosis, stomach ulcer, ulcerative colitis); cardiometabolic conditions (aneurysm, angioplasty, arrhythmia, atherosclerosis, diabetes, heart attack, heart failure, heart valve problems, high cholesterol, hypertension, narrowing of a carotid artery, narrowing of a coronary artery); cancer (any type); and any other disease or illness not listed (maximum 5 response options).

When participants indicated having 1 or more medical conditions within a single cluster, this would result in a score of 1. When participants reported having conditions within 2 different clusters, this would result in a score of 2, and so on. Thus, we calculated a score for number of different clusters of medical conditions in parents that ranged from 0 to 8.

Parental Illness Behaviors

Health care utilization was defined as the number of different categories of health care providers that parents reported to have visited in the past 12 months, based on a list of health care providers included in the questionnaire. These health care providers were categorized into 5 broad categories: general practitioner (general practice center, general practitioner); mental health care providers (addiction care provider, mental health care provider, private practice psychiatrist, self-employed psychologist, general welfare work); paramedical health care providers (dietitian, occupational therapist, physical therapist, physiotherapist, speech therapist); medical specialist (other than psychiatrist); and complementary and alternative care providers. The score for parental number of different types of health care providers visited within the past 12 months ranged from 0 to 5 types of health care providers.

Medication use was defined as the self-reported number of 9 different types of over-the-counter (OTC) medication used in the past 12 months from the following list (total score range 0-9): medications for cough, cold, flu, sore throat, etc; medications for gastrointestinal symptoms; remedies that boost the immune system (eg, Echinacea purpurea [Echinaforce; A.Vogel AG, Roggwil, Switzerland]); laxatives; medications for pain and fever; sleeping pills and tranquilizers; supplements (eg, vitamins, minerals); homeopathic remedies; and other OTC medications not listed. When parents indicated that they have used pain medication (acetaminophen, ibuprofen, and/or acetylsalicylic acid), they were asked for each type the frequency used per year using the following 4 answer categories: <10, 10 to 50, 50 to 200, and >200. A sum score was calculated by adding categories 1 (none or <10 a year) to 4 (>200 a year) of each of the 3 types of pain medication, resulting in a total composite score of 3 to 12. Of note, we included OTC medication use, but not prescription medication use, as nonspecific somatic symptoms—assessed here with the SCL-90 SOM—are typically managed with OTC medication in the Dutch health care context.28, 29, 30 OTC medication is often the first and sometimes only step in self-care for such symptoms.^31,32

Parental Health Literacy

Parental health literacy was measured using a health literacy questionnaire consisting of 3 questions that were administered during the second follow-up assessment^11,33:

• 1.“How often do you have someone help you read hospital materials?”
• 2.“How confident are you filling out medical forms by yourself?”
• 3.“How often do you have problems learning about your medical condition because of difficulty understanding written information?”

Answers to these questions were rated on a 5-point Likert scale and summed (the first and third questions were reverse-coded) for a composite health literacy score that ranged from 3 to 15, with higher scores indicating a higher level of health literacy.

As an additional indicator of parental health literacy, we determined whether or not a parent’s profession (at baseline) was in health care according to the International Standard Classification of Occupations (ISCO-08).³⁴ We assumed that parents who had been working as health professionals would have more medical knowledge and thus health literacy than parents working in a profession not related to health care. All professions listed under the International Standard Classification of Occupations Sub-Major Group 22 Health Professionals were included. Parental profession in health care was defined as dichotomous variable; thus a parent either worked in health care or not (1 = “yes”; 0 = “no”).

Statistical Analyses

Data analyses were performed using IBM SPSS version 28 (IBM Corp, Armonk, NY) software package. To test the relation between several parental health and illness factors and symptom levels in their offspring, we used generalized estimating equations. We accounted for dependence among observations of participants from the same family (clustered data) by including family identification number as a within-subject variable. We assumed an exchangeable correlation structure, in which the dependence between observations within each family was assumed to be similar across all families. As the distribution of our dependent variables was comparable to a count variable (ie, potentially right-skewed non-negative integers ranging from 0 to 14 in case of the YSR and 0 to 48 in case of the SCL), a negative binomial log-link function was used.

We distinguished between mother–adolescent dyads and mother–father–adolescent (family) triads and constructed 2 models each; one tested associations between parental factors and participants’ symptom levels in adolescence (at baseline), and one tested associations between parental factors (assessed at baseline) and participants’ symptom levels in adulthood (at follow-up). Of note, the mother–adolescent dyads were also included in the family triads. The models included all parental variables, each of them separately for fathers and mothers: parental health (somatic symptoms, medical conditions), parental illness behaviors (health care utilization, OTC medication use, pain medication use), and parental health literacy (health literacy scores, parental profession). Thus, 7 independent variables were included in the models testing associations in the sample of mother–adolescent dyads, and 14 independent variables were included in the models examining associations in the family triads. Age of offspring participants was included as a covariate. Outcome measures were somatic symptom levels of offspring participants at the baseline assessment (in adolescence, measured by the YSR) and at the follow-up assessment (in early adulthood, measured by the SCL-90). In the model testing associations with participants’ symptom levels in adulthood, baseline symptom level during adolescence was included as a covariate, thereby accounting for within-subject correlations between somatic symptom reports at baseline and at follow-up.

To investigate whether parents’ health and illness factors influenced each other in their association with somatic symptoms of adolescents (ie, when both parents, rather than only 1 parent, indicated high rates of, for example, pain medication use), we examined whether one parent’s score moderated the effect of the other parent’s score for each independent variable. Findings were considered significant at p ≤ .002, after Bonferroni correction for multiple testing, based on 7 (in the family dyads) plus 14 (in the family triads) parental predictor variables (0.05/21 = 0.00238095 ≈ 0.002). Nonsignificant associations that passed the conventional, uncorrected α < .05 were reported as well (Supplement 1, available online).

Missing Data and Imputation

Because we included adolescents with at least 1 parent also participating in Lifelines, our sample consisted of both mother–adolescent dyads and father–mother–adolescent (family) triads. We tested our hypotheses in both subsamples because including the family dyads resulted in the largest sample size and thus statistical power, whereas the family triads enabled us to study influences of both parents and thus the full family context. As the subsample of father–adolescent dyads was relatively small, we did not include separate analyses of those father–adolescent dyads.

We used multiple imputation with fully conditional specification and predictive mean matching for cases in which parts of the parental health and illness variables were missing. Five imputed datasets were generated, including 14 imputed parent predictor variables, the dependent variable, and auxiliary variables: offspring age and sex, family identification, parents’ ages, and YSR and SCL somatic symptom scores of offspring. The pooled result was used for analysis. To ensure consistency and robustness, results from the pooled imputed dataset were compared with those from the original dataset.

Results

Descriptive Statistics

A total of 3,950 adolescents participated in the Lifelines youth baseline assessments. Five adolescents were excluded due to missing YSR scores, and 3 more were excluded for being 18 years old at the time of submission, leaving 3,942 adolescents with valid YSR data. We selected mother–adolescent dyads and a subsample for which both maternal and paternal information was available, forming mother–father–adolescent triads. As 181 adolescents had only paternal data, these cases were excluded from further analysis.

For the mother–adolescent dyads, 207 adolescents were excluded due to missing maternal health data, resulting in 3,735 adolescents (mean [SD] age = 14.8 [1.4] years; 54.7% female) with valid YSR scores and maternal health data. Of these, 1,017 (27.2%) completed the SCL-90 SOM questionnaire in the adult follow-up (mean [SD] follow-up period = 3.6 [1.1] years). Mean (SD) age of mothers at baseline was 43.8 (4.0) years.

For the mother–father–adolescent triads, 2,668 adolescents (mean [SD] age = 14.8 [1.4] years; 52.8% female) met the selection criteria, having valid YSR data and parental health data. Of these, 728 (27.3%) completed the SCL-90 SOM in the adult follow-up (mean [SD] follow-up period = 3.6 [1.1] years). Mean (SD) age of mothers was 43.9 (3.9) years, and mean (SD) age of fathers was 46.1 (4.3) years. Sample characteristics are presented in Table 1, and parental health variables are presented in Table 2.

Table 1.

Offspring Participant Demographics and Outcome Measures, Stratified Per Sample Type and Time of Assessment

Somatic symptoms outcome in offspring	n	Age, y		Somatic symptoms score
		Mean	(SD)	Median, mean	(SD)
	Mother–adolescent dyads (n = 3,735)
Girlsa at baseline	2,043	14.9	(1.4)	3, 3.1	(2.5)
Boysa at baseline	1,692	14.7	(1.4)	1, 1.9	(1.9)
Womenb at follow-up	652	19.2	(1.1)	16, 17.3	(5.0)
Menb at follow-up	365	19.1	(1.1)	14, 15.7	(4.0)
	Mother–father–adolescent triads (n = 2,668)
Girlsa at baseline	1,410	14.8	(1.4)	3, 3.0	(2.4)
Boysa at baseline	1,258	14.7	(1.4)	1, 1.8	(2.0)
Womenb at follow-up	459	19.2	(1.1)	16, 17.1	(4.9)
Menb at follow-up	269	19.1	(1.1)	14, 15.6	(4.0)

Note: ^aIn adolescent girls and boys, the Youth Self-Report was used.

^bIn women and men, the Symptom CheckList-90 was used.

Table 2.

Parental Health and Illness Factors

Factors	Mother–adolescent dyads (n = 3,735)		Mother–father–adolescent triads (n = 2,668)
	Maternal variables		Maternal variables		Paternal variables
	Mean	(SD)	Mean	(SD)	Mean	(SD)
Health
Somatic symptoms	16.2	(4.3)	16.1	(4.3)	14.9	(3.7)
Medical conditions	1.6	(1.5)	1.6	(1.5)	1.0	(1.2)
Health and illness behaviors
Health care utilization	1.9	(0.9)	1.9	(0.9)	1.7	(0.9)
Tyes of OTC medication used	2.2	(1.3)	2.2	(1.3)	1.5	(1.1)
Amount of pain medication used	4.1	(1.1)	4.1	(1.2)	3.7	(1.0)
Health literacy
Health literacy score	13.5	(1.4)	13.5	(1.4)	13.4	(1.4)
	n	(%)	n	(%)	n	(%)
Profession in health care	296	(7.9)	208	(7.8)	63	(2.3)

Note: OTC = over-the-counter.

Associations Between Parental Health and Illness Factors and Symptom Reports of Offspring

Results are presented as exponentiated coefficients (ExpB) and CIs (Tables 3 and 4), representing odds ratios (ORs) for each predictor to provide a clearer measure of effect size. These coefficients indicate the multiplicative change in the outcome variable for a 1-unit increase in the predictor. Original B values and CIs are provided in Tables S1A and S1B, available online.

Table 3.

Associations Between Maternal Health and Illness Factors and Adolescents’ Symptom Reports in Mother–Adolescent Dyads

Factors	Mother–adolescent dyads	SE	Exp(B)	95% Wald CI for Exp(B)		p
Somatic symptoms	Adolescent baseline (n = 3,735)	0.004	1.014	1.006	1.022	<.001∗∗∗
	Girls (n = 2,043)	0.005	1.019	1.009	1.028	<.001∗∗∗
	Boys (n = 1,692)	0.006	1.009	0.998	1.021	.117
	Adult follow-up (n = 1,017)	0.008	1.014	0.998	1.030	.084
	Women (n = 652)	0.009	1.026	1.008	1.044	.006∗
	Men (n = 365)	0.016	0.995	0.966	1.026	.763
Medical conditions	Adolescent baseline	0.012	1.042	1.017	1.067	<.001∗∗∗
	Girls	0.014	1.038	1.010	1.067	.008∗
	Boys	0.019	1.048	1.010	1.089	.013∗
	Adult follow-up	0.023	1.021	0.977	1.067	.361
	Women	0.027	1.022	0.969	1.079	.417
	Men	0.039	0.998	0.925	1.077	.957
Health care utilization	Adolescent baseline	0.019	0.984	0.948	1.021	.391
	Girls	0.023	0.999	0.953	1.047	.965
	Boys	0.032	0.957	0.899	1.019	.170
	Adult follow-up	0.039	1.007	0.933	1.088	.853
	Women	0.046	1.006	0.920	1.100	.903
	Men	0.072	1.034	0.898	1.191	.641
Types of OTC medication used	Adolescent baseline	0.012	1.016	0.993	1.041	.169
	Girls	0.014	1.002	0.975	1.029	.875
	Boys	0.020	1.043	1.002	1.085	.040∗
	Adult follow-up	0.027	1.022	0.969	1.077	.424
	Women	0.034	1.009	0.944	1.080	.786
	Men	0.042	1.062	0.978	1.154	.153
Amount of pain medication use	Adolescent baseline	0.014	1.051	1.022	1.081	<.001∗∗∗
	Girls	0.016	1.055	1.021	1.090	.001∗∗∗
	Boys	0.025	1.022	0.972	1.074	.386
	Adult follow-up	0.030	1.063	1.001	1.128	.046∗
	Women	0.033	1.044	0.978	1.114	.198
	Men	0.065	1.076	0.948	1.221	.257
Health literacy score	Adolescent baseline	0.013	0.969	0.944	0.996	.028∗
	Girls	0.014	0.968	0.940	0.995	.021∗
	Boys	0.021	0.982	0.940	1.025	.526
	Adult follow-up	0.025	1.016	0.967	1.067	.534
	Women	0.032	1.004	0.941	1.071	.906
	Men	0.044	1.059	0.971	1.154	.194
Profession in health care	Adolescent baseline	0.058	1.077	0.961	1.208	.204
	Girls	0.076	1.162	1.002	1.347	.048∗
	Boys	0.092	0.944	0.788	1.129	.526
	Adult follow-up	0.141	1.190	0.902	1.571	.218
	Women	0.173	1.038	0.740	1.458	.906
	Men	0.218	1.417	0.925	2.171	.109

Note: OCD = over-the-counter.

^∗Uncorrected p ≤ .05; ∗∗∗Bonferroni-corrected p ≤ .002.

Table 4.

Associations Between Parental Health and Illness Factors and Adolescents’ Symptom Reports in Mother–Father–Adolescent Triads

Factors	Mother–father–adolescent triads	SE	Exp(B)	95% Wald CI for Exp(B)		p
Somatic symptoms—mothers	Adolescent baseline (n = 2,668)	0.005	1.005	0.996	1.014	.240
	Girls (n = 1,410)	0.006	1.010	0.999	1.022	.072
	Boys (n = 1,258)	0.007	1.005	0.991	1.019	.455
	Adult follow-up (n = 728)	0.001	1.016	0.998	1.035	.089
	Women (n = 459)	0.010	1.035	1.014	1.056	.001∗∗∗
	Men (n = 269)	0.020	0.991	0.953	1.031	.663
Somatic symptoms—fathers	Adolescent baseline	0.005	1.010	1.000	1.021	.047∗
	Girls	0.006	1.012	1.000	1.023	.046∗
	Boys	0.009	1.010	0.991	1.028	.302
	Adult follow-up	0.011	0.998	0.976	1.021	.884
	Women	0.013	1.004	0.978	1.030	.775
	Men	0.020	0.991	0.953	1.031	.668
Medical conditions—mothers	Adolescent baseline	0.013	1.045	1.017	1.073	.001∗∗∗
	Girls	0.016	1.041	1.008	1.074	.013∗
	Boys	0.021	1.055	1.012	1.100	.011∗
	Adult follow-up	0.027	0.996	0.945	1.051	.894
	Women	0.031	0.990	0.931	1.054	.758
	Men	0.047	0.984	0.898	1.079	.730
Medical conditions—fathers	Adolescent baseline	0.017	0.999	0.967	1.033	.962
	Girls	0.021	0.993	0.953	1.035	.727
	Boys	0.026	1.009	0.960	1.062	.717
	Adult follow-up	0.034	1.016	0.950	1.086	.651
	Women	0.041	0.993	0.917	1.076	.865
	Men	0.062	1.074	0.950	1.215	.250
Health care utilization—mothers	Adolescent baseline	0.021	0.989	0.950	1.029	.601
	Girls	0.026	1.018	0.968	1.073	.481
	Boys	0.035	0.954	0.890	1.023	.183
	Adult follow-up	0.048	0.988	0.899	1.086	.806
	Women	0.058	0.984	0.878	1.104	.786
	Men	0.079	1.020	0.873	1.191	.803
Health care utilization—fathers	Adolescent baseline	0.024	1.019	0.970	1.069	.446
	Girls	0.029	1.017	0.962	1.077	.544
	Boys	0.037	1.010	0.939	1.088	.779
	Adult follow-up	0.047	1.026	0.935	1.126	.584
	Women	0.058	1.027	0.918	1.149	.643
	Men	0.083	1.006	0.855	1.184	.941
Types of OTC medication used—mothers	Adolescent baseline	0.015	1.007	0.978	1.037	.632
	Girls	0.018	0.986	0.952	1.020	.422
	Boys	0.024	1.039	0.991	1.088	.111
	Adult follow-up	0.034	1.015	0.951	1.085	.621
	Women	0.043	0.998	0.917	1.086	.960
	Men	0.052	1.035	0.934	1.147	.513
Types of OTC medication used—fathers	Adolescent baseline	0.017	1.016	0.982	1.051	.356
	Girls	0.022	1.031	0.989	1.076	.147
	Boys	0.027	1.005	0.952	1.060	.864
	Adult follow-up	0.039	1.018	0.942	1.100	.648
	Women	0.043	1.022	0.939	1.113	.608
	Men	0.079	1.031	0.883	1.205	.699
Amount of pain medication used—mothers	Adolescent baseline	0.018	1.058	1.021	1.094	.002∗∗∗
	Girls	0.020	1.071	1.029	1.115	<.001∗∗∗
	Boys	0.029	1.011	0.954	1.070	.714
	Adult follow-up	0.038	1.076	0.999	1.158	.053
	Women	0.039	1.056	0.978	1.140	.165
	Men	0.084	1.089	0.924	1.285	.308
Amount of pain medication used—fathers	Adolescent baseline	0.020	0.976	0.938	1.015	.224
	Girls	0.025	0.971	0.925	1.021	.256
	Boys	0.032	1.000	0.939	1.064	.989
	Adult follow-up	0.046	0.980	0.895	1.072	.654
	Women	0.049	1.016	0.922	1.119	.749
	Men	0.071	0.850	0.739	0.977	.023∗
Health literacy score—mothers	Adolescent baseline	0.015	0.978	0.950	1.008	.150
	Girls	0.017	0.978	0.946	1.011	.196
	Boys	0.025	0.988	0.941	1.038	.616
	Adult follow-up	0.033	1.026	0.959	1.097	.440
	Women	0.033	1.011	0.946	1.080	.752
	Men	0.070	1.049	0.908	1.212	.503
Health literacy score—fathers	Adolescent baseline	0.015	0.985	0.957	1.014	.309
	Girls	0.017	0.976	0.946	1.009	.157
	Boys	0.026	1.005	0.955	1.058	.845
	Adult follow-up	0.045	1.037	0.941	1.143	.429
	Women	0.045	1.053	0.958	1.157	.268
	Men	0.077	1.005	0.854	1.183	.947
Profession in health care—mothers	Adolescent baseline	0.068	1.055	0.924	1.206	.428
	Girls	0.087	1.113	0.938	1.320	.222
	Boys	0.108	0.977	0.790	1.208	.831
	Adult follow-up	0.175	1.215	0.862	1.714	.266
	Women	0.208	1.032	0.686	1.553	.880
	Men	0.262	1.639	0.981	2.739	.059
Profession in health care—fathers	Adolescent baseline	0.126	1.162	0.908	1.487	.234
	Girls	0.200	1.251	0.845	1.853	.263
	Boys	0.157	0.996	0.733	1.355	.982
	Adult follow-up	0.269	1.008	0.591	1.719	.976
	Women	0.363	1.144	0.552	2.371	.713
	Men	0.345	0.870	0.442	1.713	.688

Note: OCD = over-the-counter.

^∗Uncorrected p ≤ .05; ∗∗∗Bonferroni-corrected p ≤ .002.

Associations Between Maternal Factors and Adolescents’ Symptoms in Mother–Adolescent Dyads

Significant positive associations (p ≤ .002) were found between mothers’ somatic symptoms, number of medical conditions, and pain medication use with adolescents’ somatic symptom reports. Each additional maternal somatic symptom was associated with a 2% increase in adolescent symptoms. An additional maternal medical condition from a different illness cluster corresponded to a 4% increase, and a 1category increase in maternal pain medication use (ranging from <10 to >200 per year) was linked to a 5% rise in adolescents’ symptoms (Table 3). Stratification by sex showed that these associations were significant for girls, but not for boys. No other maternal predictors showed significant associations with adolescent symptoms.

In the subsample of participants ages 18 and older at follow-up (n = 1,017), none of the maternal health and illness factors were significantly associated with somatic symptoms in early adulthood. Associations meeting the uncorrected p ≤ .05 threshold are reported in Supplement 1, available online.

Associations of Maternal and Paternal Factors With Adolescents’ Symptoms in Family Triads

In the subsample in which both parents provided health data, mothers’ medical conditions and pain medication use were significantly associated with adolescents’ somatic symptoms. Stratification by sex showed maternal pain medication use was linked to a 7% increase in girls’ symptoms, but not in boys’ symptoms (Table 4). No other maternal or paternal factors showed significant associations. Associations passing the uncorrected p ≤ .05 threshold are reported in Supplement 1, available online.

Moderation analyses found no significant interactions between maternal and paternal health factors and adolescent symptoms. At the p ≤ .05 level, moderation effects of parental somatic symptoms and pain medication use were noted specifically in girls, in whom maternal effects on symptoms intensified with more paternal symptoms and pain medication use (Table S2, available online).

For a subsample (n = 728) with follow-up data in early adulthood, no parental health factors were significantly associated with somatic symptoms in young adults. However, stratification by sex revealed that each additional maternal somatic symptom at baseline was associated with a 3.5% increase in symptoms in daughters 3 to 4 years later. No significant moderation effects were found for parental factors on symptoms in young adults.

To cross-validate the effect of maternal health and illness factors across both the dyad and the triad samples, we conducted a post hoc sensitivity analysis, analyzing associations between mothers’ health and illness factors and adolescents’ symptom reports in the family triads, but omitting the fathers’ data. In both samples of mothers, we found comparable effects, ie, the strongest predictors of adolescent symptom reports were maternal number of medical conditions and maternal pain medication use (see Table S3, available online, and compare with Table 3). In addition, in the mother–adolescent dyads, but not in the family triads, maternal somatic symptom reports were significantly associated with adolescent symptoms.

Discussion

The current study aimed to examine associations between parental health and illness–related factors and somatic symptom burden in their adolescent and adult offspring. Using data from a large, prospective multigeneration population cohort, we found that maternal somatic symptoms and maternal medical conditions were positively associated with functional somatic symptoms in their adolescent offspring. Moreover, maternal pain medication use was positively associated with functional somatic symptoms in adolescent girls, but not in boys. Analyses in a subsample of family triads revealed similar associations of these maternal factors, whereas no significant associations were found for paternal health and illness factors. Thus, in line with our hypothesis, factors such as parents’ own health and their illness behaviors, albeit only in mothers, were significantly associated with somatic symptoms reported by adolescents. However, we found no indications for a relation with parental health care utilization or OTC medication use. Also, we did not find the expected protective influence of relatively higher parental health literacy. The association perpetuated into early adulthood only for maternal somatic symptom reports and only in girls.

A strength of the current study is the large sample sizes of mother–adolescent dyads and mother–father–adolescent triads, providing sufficient statistical power to test for potentially small-sized effects. In a cross-validation of the effects of maternal health and illness factors between the family triads (N=2,668) and the mother-adolescent dyads (N=3,735), we found very comparable effects (Table 3; Table S3, available online). One exception were the maternal somatic symptom reports, which were significantly associated with adolescent symptoms only in the dyads sample, but not in the family triads, which might be related to the larger sample size. Moreover, analyses in the subsample of family triads enabled us to test for the differential influences of both parents on symptom levels in adolescents. Another strength is that participants were recruited from the general population, and therefore the risk for selection bias is likely small. Furthermore, we included several parental health and illness factors, ranging from parents’ own somatic symptoms and medical conditions to illness behaviors. We also specifically investigated associations with parents’ health literacy, through both a questionnaire and by measuring parental medical knowledge through their profession. Most research thus far used the broader measure level of education. This allowed us to assess the unique contribution of specific parental factors to health outcomes of adolescents. Finally, a strong aspect of this study is its long-term perspective, assessing associations between parental health and illness factors and somatic symptoms in their offspring, first in adolescence and then again, several years later during early adulthood.

A limitation of our study is that 2 different instruments with differing response options were used for the assessment of somatic symptoms (YSR and SCL-90). Whereas the YSR Somatic Complaints subscale explicitly assesses functional somatic symptoms, the SCL-90 subscale assesses common, but not necessarily functional somatic symptoms. Although the type of symptoms largely overlaps between both instruments, and the symptoms in the SCL-90 typically remain unexplained,³⁵ use of 2 different instruments limits comparability of assessments between adolescence and adulthood. Another limitation is that we cannot be certain whether adolescents’ symptoms could have been explained by an underlying medical condition, as the data were self-reported (YSR). Additionally, due to the lack of available data on chronic illnesses and diseases in the adolescent participants, we were unable to adjust for these factors in our analyses. However, this might not be problematic, as current conceptualizations of persistent physical symptoms avoid distinguishing between medically explained symptoms (symptoms associated with a disease) and medically unexplained symptoms (functional symptoms).

Furthermore, parental health care utilization was not optimally assessed; parents were asked to select from a list of health care providers those they had visited in the past year. Therefore, it was not possible to determine the number of different medical specialists consulted or the frequency of visits. Consequently, our analyses were limited to counting the number of distinct types of health care providers visited across 5 broad categories. We could not differentiate between participants who sought care from multiple subspecialties within a category and participants who had a single consultation. This constraint limited the range and accuracy of our assessment of parental health care utilization in the current study. This limitation may have led to an underestimation of the true extent of health care engagement among parents. If higher utilization—particularly across multiple subspecialties or frequent visits—is associated with increased adolescent functional somatic symptom reports, our findings may underestimate the strength of associations between parental health care use and child outcomes.

Future studies should aim to incorporate more detailed measures of health care utilization to better capture its complexity and potential influence on study outcomes. In addition, although we had rich information about various parental health and illness variables, information on adolescents’ illness behaviors, such as medication or health care use, or other health measures, such as lifestyle behaviors, was not available. Future studies should therefore add other relevant health outcome measures, such as adolescent OTC medication use, in a sample that is well balanced for parental and adolescent sex and age. Another limitation is that prescription medication use was not included as a parental health and illness factor in the current study; prescription medication use might serve as a valid indicator of parental health and illness behavior and might have shown associations with offspring symptom reports. However, we reasoned that in the Dutch health care context OTC medication use may be a more relevant indicator of health-related behavior than prescription medication. Dutch general practitioners have a gatekeeping role, adhering to a cautious, wait-and-see approach for mild symptoms.³⁰ General practice guidelines and prescribing culture emphasize restraint in pharmacological intervention, particularly for mild or nonspecific symptoms.^28,29

Moreover, the health literacy questionnaire was the only health and illness factor that was not included in the baseline assessments, but rather at the first questionnaire follow-up measurement. This may explain the relatively large amount of missing information for this variable. In consequence, relatively more data had to be imputed, and findings may therefore be less reliable. Also, it should be noted that definitions of the concept of health literacy have changed throughout the years, with the latest definition by the World Health Organization circumscribing health literacy on both a personal and an organizational level.³⁶ We acknowledge that our health literacy questionnaire did not fully encompass this broader definition of health literacy and suggest the use of a more comprehensive measure in future studies. Furthermore, given the relatively high number of independent variables included, we chose to apply Bonferroni correction based on 21 parental health and illness factors (7 maternal factors in the family dyads, 14 maternal and paternal factors in the family triads). However, we did not include additional tests, such as interactions between parental factors and associations with follow-up data. Consequently, whereas the risk of type 1 error in our findings was substantially reduced, it was not as low as it would be in a univariate analysis.

Our findings are in line with a recent study in a general population sample that compared somatic symptom burden in 15-year-old adolescents with healthy parents (no illnesses) with adolescents having an ill parent.¹⁶ Similar to the current study, poorer general maternal health and having an ill mother predicted higher rates of somatic symptoms, particularly in adolescent girls, whereas illness in fathers was less influential. Similarly, other studies found that self-reported poorer general health and more somatic symptoms, such as chronic pain, in fathers were only moderately associated with offspring health outcomes, whereas general health and pain status in mothers were strong predictors.^15,37 Although we found that the effects of mothers’ somatic symptoms and pain medication use on symptom reports of their daughters were stronger when fathers reported more somatic symptoms and increased pain medication use, those findings were nonsignificant. A prior study similarly found that compared with offspring whose parents were free of chronic pain, offspring with both parents experiencing 1 or 2 chronic pain sites showed an increased risk (OR = 1.33) of developing additional chronic pain sites. This risk was further heightened if both parents had ≥3 chronic pain sites (OR = 1.46).¹⁸

We found that specifically maternal pain medication use, but not other OTC medication use, was associated with adolescent functional somatic symptom reports. Testing whether this association might be specific to pain-related symptoms, we conducted a post hoc analysis distinguishing between adolescent pain-related and non–pain-related functional somatic symptoms. Indeed, pain medication use by mothers was more strongly associated with their adolescents reporting pain symptoms than nonpain somatic symptoms, although effects differed only slightly (Table S4, available online).

Using follow-up data, we found that almost none of the parental health and illness factors predicted somatic symptoms of offspring in adulthood. An exception was somatic symptoms of mothers, which were significantly associated with their daughters reporting somatic symptoms in early adulthood. Longitudinal studies that examined parental influences across a wide age range and into adulthood are scarce. In a Norwegian population cohort sample, parental chronic pain was found to be a risk factor for pain in their adult offspring.³⁸ In addition, follow-up assessments suggested that specifically maternal, but not paternal, chronic pain predicted development of additional chronic pain sites in their adult offspring.¹⁸ Our study explored how associations between parental health and illness factors with somatic symptom reports might change during different developmental periods, investigating links during both adolescence and early adulthood. Adolescence is regarded as a period of increasing independence and autonomy, and therefore one might expect parental influences to decrease with age. This is corroborated by our findings of fewer associations found in adulthood compared with adolescence. Regarding help-seeking behaviors in mental health care, others have found that during adolescence, influences from family members were dominant, whereas in late adolescence and early adulthood, a shift occurred favoring self-influence, though with still considerable effects of family influences.²⁰ It may thus depend on both the predicting parental health and illness factor and the offspring outcome measure, in how strong the parental influence is during adolescence, and, in addition, to what degree it endures.

Together, these findings suggest that though paternal influences might play a role as well, a mother’s health has a stronger and longer-term influence on a child’s health, specifically a daughter’s health, than a father’s health. Indeed, prior studies have pointed to mothers as important sources for teaching their adolescent children about pain recognition, pain prevention, and pain management.³⁹ In addition to mothers’ health, their attitudes toward somatic symptoms, their health knowledge, and their pain management practices, as well as modeling and reinforcing illness behaviors, were identified as highly influential on symptom experiences and pain management strategies of adolescents.^37,40,41 Although our study investigated more broadly parental health and illness effects on somatic symptoms of offspring, we also found that maternal pain medication use was a significant predictor of adolescent somatic symptoms. However, as we did not assess adolescents’ own medication use, we could not directly relate maternal and offspring pain management behaviors. In contrast, we did not find significant associations between parental health literacy and adolescent symptom burden, which in part may be explained by the relatively high nonresponse rate to the health literacy questionnaire and the relatively small number of parent participants in the sample who had a profession in health care.

It has been suggested that the parent–child transmission of somatic symptoms may be sex specific, with daughters being sensitive specifically to their mothers’ health status and fathers’ health and illness behaviors being salient models for their sons.⁴² However, our data did not reveal clear sex-specific parent–adolescent transmissions of symptoms. Although maternal health and illness factors were more influential, and associations were often specifically found in girls, the paternal health and illness factors seemed less influential overall. A potential explanation for this finding may be that mothers more often take on the task as primary caregiver, spending more time with their children, and thereby may have more opportunity to model maladaptive health and illness behaviors than fathers.¹⁷ Future studies may further explore whether, for example, time spent with children vs time spent at work moderates the associations between parental health and illness factors and child health outcomes.

Research suggests a genetic basis for the predisposition to develop somatic symptoms.⁶ Interestingly, genetic effects have been found to be more pronounced in males than in females.⁶ In contrast, our findings suggest stronger maternal health influences—particularly on daughters—thereby pointing to the important role of interpersonal processes. Furthermore, on one hand, the observed reduction in associations with greater independence in young adulthood may support the notion that psychosocial factors contribute to the intergenerational transmission of somatic symptom proneness. On the other hand, fewer associations found between parental health factors and offspring symptom reports may be due to the smaller sample size at follow-up.⁶ Our findings have several clinical implications; for example, given the significant impact of maternal health on the somatic symptom burden of adolescents, health care providers should prioritize early identification and intervention strategies that focus on maternal health and well-being. Furthermore, health education initiatives could focus on empowering parents, particularly mothers, with knowledge about and awareness of their own illness behaviors (eg, pain medication use) to promote healthier lifestyles, which may, in turn, reduce the intergenerational transmission of somatic symptom tendencies. Our findings align with broader evidence indicating that sex and gender contribute independently to health symptoms and diagnostic pathways.^35,43,44 In primary care, sex inequalities influence diagnostic trajectories for patients with somatic symptoms, leading to disparities in health outcomes between women and men. The associations found between maternal health factors and symptom reports of their offspring highlight intergenerational influences on health perceptions and symptom reporting. This may reflect social and cultural factors—including gendered expectations surrounding health, care-seeking behaviors, and symptom attribution—that warrant further investigation within the context of health equity.

In conclusion, our study investigated the impact of various parental somatic health indicators, including a balanced comparison between maternal and paternal influences, and highlights the significant role maternal health plays in the functional somatic symptom burden of adolescents. As these adolescents transition into adulthood, the influence of parental health and illness diminishes.

CRediT authorship contribution statement

Sarah M. Burke: Writing – review & editing, Writing – original draft, Formal analysis, Data curation. Elske Hogendoorn: Writing – review & editing. Arnout C. Smit: Writing – review & editing, Validation, Methodology. Judith G.M. Rosmalen: Writing – review & editing, Supervision, Funding acquisition, Conceptualization.