Background
Attention-deficit/hyperactivity disorder (ADHD) is a neurodevelopmental disorder affecting both children and adults [
1‐
3]. Inattention, hyperactivity, and impulsivity constitute core symptoms of ADHD [
4]. Children and adolescents with ADHD are at increased risk of conduct disorder [
5], learning disabilities [
5], substance use disorder [
6], criminality [
7], and injuries [
8], which might add to the overall burden on caregivers.
A multi-country online survey conducted among 2326 caregivers of children/adolescents with ADHD showed that the caregivers experienced considerable burden in terms of work, social activity, family life, and parental stress, despite pharmacotherapy received by all children/adolescents during the previous 6 months [
9]. In addition, the survey found that ADHD medication adherence was associated with reduced caregiver burden related to work and social activity [
9]. One limitation of the survey study was the lack of an optimal proxy for off-medication time.
High caregiver burden has been linked to depression among caregivers of patients with different health problems [
10‐
12]. Previous research found that parents of offspring with ADHD tend to show more depressive symptoms and parental stress than parents of offspring without ADHD [
13‐
15]. While pharmacotherapy is effective in reducing core symptoms of ADHD [
16,
17], and associated with decreased risks of adverse health outcomes [
7,
8,
18], to date, it remains unclear whether pharmacotherapy for ADHD in offspring may benefit the parents, for example, by mitigating parental depression.
One substantial challenge to such investigation results from the genetic overlap between ADHD and depression. The most up-to-date meta-analysis of genome-wide association studies (GWASs) of ADHD reported a moderate genetic correlation (about 0.4) between ADHD and depression [
19]. Because of the genetic sharing between offspring and their parents, offspring who received pharmacotherapy for ADHD might be more likely to have parents with depression than offspring without ADHD, even in the absence of any causal relationship. This is an issue similar to confounding by indication in pharmacoepidemiological studies [
20]. Nonetheless, the Swedish national health registers provide medical records of hospital discharge and drug dispensation, enabling us to compare the rate for depression-related specialty care visits by parents during the time periods when their ADHD-affected offspring are on medication with the rate during all other time periods. When comparisons are made within the same parents across different time periods, all time-constant factors, such as genetic makeup and disease severity at baseline, are implicitly cancelled out.
Using data from the Swedish national registers, we conducted a cohort study to investigate the concurrent association between pharmacotherapy for ADHD in offspring and rate of depression-related specialty care visits by the parents with a history of depression.
Discussion
In this large cohort study, we followed up 5605 parents (3872 mothers and 1733 fathers) with a history of depression and an offspring diagnosed with ADHD for 8 years to investigate the concurrent association between pharmacotherapy for ADHD in offspring and rate of depression-related visits in parents.
Between-individual analyses yielded no difference in the hazard rate for depression-related visits in parents during exposed periods when their offspring were on medication for treatment of ADHD as compared to unexposed periods. Nonetheless, the analyses could not take into consideration individual specific baseline severity of depression, and other unmeasured time-constant or time-varying factors. In within-individual analyses, we compared the same parent across time periods when his or her offspring was on and off medication for treatment of ADHD. Only those parents who had at least one outcome event during follow-up and whose offspring had both on-and off-medication periods contributed to the analyses. Within-individual comparisons helped rule out bias arising from time-constant factors, such as genetic makeup and disease severity at baseline, and thus provided more valid estimates than between-individual comparisons. Based on within-individual comparisons, we detected an association between pharmacotherapy for ADHD in offspring and a decreased hazard rate for depression-related visits in fathers. No statistically significant association was observed in mothers. The pattern of results remained in all sensitivity analyses, suggesting the findings were not due to our definitions of study cohort, exposure, or outcome event.
In the current study, we speculate that the change in hazard rate for depression-related visits in parents might to some extent reflect the change in severity of depression during exposed and unexposed periods. Our findings then could be explained by proper management of ADHD in offspring leading to decrease in severity of depression in the fathers. The process might be mediated by father-perceived improvements in children’s ADHD and subsequent stress-relief, as proposed by two prior clinical studies reporting parent-perceived improvements in children’s ADHD-related behavioral symptoms, parenting stress, and depressive symptoms in parents following methylphenidate treatment for ADHD in the children [
33,
34]. Alternatively, our findings could be attributed to that while offspring were on medication for treatment of ADHD, their parents were likely to receive interventions aiming at training parents themselves to appropriately handle children’s ADHD-related behaviors, reducing parental stress, and increasing parental confidence [
35], which might also lead to improvement in depression. In the current study, we did not observe statistically significant association between pharmacotherapy for ADHD in offspring and depression-related visits by mothers. Nonetheless, this does not necessarily mean that there was no decrease in severity of maternal depression associated with pharmacotherapy for ADHD in the offspring, given the lack of direct measurements for depressive symptoms.
To examine whether the observed associations for fathers were due to reverse causation (i.e., offspring with ADHD might be more likely to be on medication for treatment of ADHD when their parents were less depressed), we plotted the rates of depression-related visits by parents per every 8 weeks during 40 weeks before and 40 weeks after treatment initiation in offspring. We observed no obvious drop in the rate shortly before treatment initiation either in mothers or in fathers, suggesting that treatment initiation in offspring was unlikely triggered by parents being less depressed. Nonetheless, we cannot rule out the possibility that decrease in severity of depression in fathers might help improve their offspring’s compliance to treatment of ADHD. More research is needed to unravel the role of the complex and dynamic parent-offspring interactions in the associations under study.
The main strengths of our study include the large sample and prolonged follow-up, both enhancing the statistical power for detecting small effects. The Multi-Generation Register allowed unambiguous identification of parent-offspring relations. The prospectively collected data on medication dispensations and depression-related specialty care visits precluded recall bias. Furthermore, we employed an innovative within-individual comparison design to control for unmeasured time-constant factors that often threaten the validity of observational studies. In addition, since the exposure and outcome events were separately defined in offspring and their parents, our study was not subject to bias due to confounding by contraindication (i.e., in an observational study, a medication might seem to have a protective effect on an outcome when the outcome is a contraindication for the medication) [
36].
We also acknowledge some limitations of the study. First, we are not aware of any validation study on ADHD diagnosis in the National Patient Register. Nevertheless, a prior study of 19,150 twins in Sweden found that 70% twins with an ADHD diagnosis also screened positive for ADHD by their parents [
37]. In the current study, over 77% of the children with ADHD were also dispensed ADHD medications, indicating that they were likely to be true ADHD cases. More importantly, it was mainly the parents of these 77% children that contributed information to the within-individual analyses. Second, on-medication periods were defined empirically using medication dispensations. Although used in many studies, the definition might not precisely reflect the actual consumption of ADHD medications by the offspring. The potential misclassification in exposure, however, would tend to bias the estimates towards null and therefore would not be responsible for the observed significant associations in fathers. Indeed, when on-medication period was redefined using two alternative approaches, the results remained similar to the main results. Third, as in many other studies using within-individual comparisons, unmeasured or even unknown time-varying factors could not be handled by the analyses. Although our results restricted to the three sub-cohorts were less likely affected by time-varying factors associated with change in ADHD status of parents themselves or ADHD/depression status of other family members, other comorbidities or life events occurring to the offspring or their parents during follow-up might still bias the estimated associations. Fourth, considering that pharmacotherapy for ADHD is reserved to relatively severe cases in Sweden [
38], our findings apply primarily to parents of offspring with ADHD that was severe enough to warrant pharmacotherapy. Consequently, the generalizability of the findings to other populations is unclear but worth further investigation. Fifth, we were not able to take into account the influences of parent-offspring interactions and non-pharmacological treatments received by either offspring or parents on the associations of interest due to the lack of relevant data.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.