Perceived parental alcohol problems and psychosomatic complaints among adolescents in Sweden

Highlights • Knowledge is lacking on the health of youth reporting parental problem drinking.• Data from a Swedish nationally representative sample of 9,032 teens were used.• Perceived parental alcohol problems were associated with psychosomatic complaints.• Attention and support to this group of adolescents is important.


Introduction
Adverse childhood experiences (ACEs) have a profound impact on individuals' lives both in the short and in the long run (Boullier & Blair, 2018;Petruccelli et al., 2019). One particular type of adversity relates to alcohol problems in the family of origin (Felitti et al., 1998). Children of problem-drinking parents have increased risks of short-and long-term adverse outcomes in terms of, e.g., higher levels of drinking and more problematic drinking themselves (Haugland et al., 2013;Karlsson et al., 2016;Olsson et al., 2019Olsson et al., , 2021Pisinger et al., 2017;Thor et al., 2022), higher levels of self-reported health problems (Pisinger et al., 2016;Wahlström et al., 2023a), increased risks of mental and behavioural disorders (Johnson & Leff, 1999;Landberg et al., 2019;Raitasalo et al., 2019), as well as poorer educational outcomes (Berg et al., 2016;Pisinger et al., 2023).
Living with a parent with alcohol problems can be regarded as a chronic stressor with potentially negative implications for health. Children of problem-drinking parents might have a challenging family situation with a poor sense of belonging, poor control over the situation, and lack of parental support (Windle, 1996;Lee & Williams, 2013). Indeed, adolescents with perceived parental alcohol problems have reported poorer parent-child relationships compared with adolescents without such problems (Pisinger et al., 2016;Wahlström et al., 2023b). The family situation may also involve feelings of shame, stigma and guilt, which could lead to social isolation (Hutchinson et al., 2014;Tamutienė & Jogaitė, 2019). Furthermore, adolescents with perceived parental alcohol problems have increased risks of being subjected to bullying (Ramstedt et al., 2022).
Psychosomatic complaints have been shown to correlate with perceived stress in adolescents, and can hence be regarded as an indicator of stress-related ill-health in this age group (Corell et al., 2022). To date, the association between parental alcohol problems and adolescents' self-reported psychosomatic complaints has been scarcely investigated, although there are some recent studies based on Swedish data which report links between parental alcohol problems and children's psychosomatic complaints , 2023bRamstedt et al., 2022;Wahlström et al., 2023a). Overall, however, there is limited knowledge and a relative lack of empirical evidence on this topic.
The aim of the current study was to examine perceived parental alcohol problems and the links with psychosomatic complaints among Swedish adolescents aged 15-18 years. The research questions are: 1) How common are perceived parental alcohol problems among Swedish adolescents, and does the reporting of such problems differ by sociodemographic characteristics? 2) Are perceived parental alcohol problems associated with a higher likelihood of reporting psychosomatic complaints, whilst adjusting for sociodemographic characteristics?

Material and methods
The data was drawn from the 2021 national school survey performed by the Swedish Council for Information on Alcohol and Other Drugs (CAN). The survey has been conducted yearly since 1971 in grade 9 (ages 15-16 years) and since 2004 also in grade 11 (ages 17-18 years) (CAN, 2021). The aim of the survey is to collect information about the consumption of alcohol, narcotics, tobacco, and doping use as well as about gambling experiences among adolescents. Since 2019, there has been a transition from pen and paper to a digital form of data collection, and in 2021, a complete digital data collection was performed for the first time.
Every year, CAN draws a random sample of 350 schools in each grade. In 2021, the number of students in the classes participating in the survey was 6,609 in grade 9 and 5,427 in grade 11. The number of absent students reported by teachers was 1,253 in grade 9 (19 %) and 1,041 in grade 11 (19 %). The number of students who refrained from participating in the survey was 11 in grade 9 (0.2 %), and 9 in grade 11 (0.2 %). The number of students sorted out from the survey due to inconsistent, exaggerated, or unrealistic answers was 91 in grade 9 (1.4 %), and 28 in grade 11 (0.5 %) (CAN, 2021). Accordingly, our data included 9,603 students (n = 5,254 in grade 9; n = 4,349 in grade 11). Out of these, 571 cases with missing information were excluded, rendering an analytical sample of 9,032 students (n = 4,892 in grade 9 and n = 4,140 in grade 11).

Ethics
CAN instructed all schools to perform the questionnaire in the classroom under examination-like conditions. Students used a one-time password to log in to the survey. The password was linked to a class but not to an individual, and the students did not include their names or any other personal information in the survey. The students received information that participation was voluntary. Parental consent was not obtained since according to the Swedish Ethical Review Act (SFS, 2003:460), this is not required for adolescents aged 15 or older if they realise what the research means for them.

Measures
Psychosomatic complaints were captured by the question: "During the past six months, how often have you…" and the items a) "had a headache"; b) "had a stomach ache"; c) "felt depressed or down"; d) "had difficulties to fall asleep"; and e) "slept poorly at night". The response categories were "Nearly every day"; "Several times a week"; "Once a week"; "Some time/a few times a month"; and "Rarely or never". In line with commonly used measures of psychosomatic complaints such as the Health Behaviour in School-aged Children (HBSC) symptom checklist (HBSC-SCL) (e.g., Agency, 2019; Cosma et al., 2020) and the Psychosomatic Problems (PSP) scale (Corell et al., 2022;Hagquist, 2009) the variable covers different types of complaints (i.e., psychological as well as somatic ones). The five items were weakly to moderately correlated (Pearson's r ranging between 0.34 and 0.50) except for the two items about sleep which were highly correlated (r = 0.75). Internal consistency was acceptable (Cronbach's α = 0.71). Each item was dichotomised to distinguish between students who reported the complaint more than weekly (i.e. nearly every day or several times a week) vs. less often. In addition, we constructed a composite measure that distinguished between students who reported two or more psychosomatic complaints more than weekly vs. others. Dichotomization makes the interpretation of the results easier to understand, and odds ratios are a meaningful measure of association (Farrington & Loeber, 2000).
Perceived parental alcohol problems were measured by the Children of Alcoholics Screening Test (CAST-6) scale (Hodgins & Shimp, 1995). The CAST-6 scale includes six yes/no statements instead of the original scale of 30 items. This short version has proven to be as effective as the original version (Hodgins & Shimp, 1995). These six questions cover four areas: perceptions of parental alcohol problems; attempts to control parental drinking; perceptions of marital discord; and efforts to escape adverse consequences. Internal consistency was high (Cronbach's α = 0.87). We created a binary variable distinguishing between students who answered affirmatively to at least three of the six statements vs. those who answered affirmatively to less than three of the statements (Hodgins & Shimp, 1995;Ramstedt et al., 2022).
Sociodemographic background characteristics included adolescents' gender (male; female; and other), grade (9 and 11), parental education (at least one parent with a university/college education level; no parent has a university/college-level education; and do not know), and parental country of birth (at least one parent born in Sweden; two parents born abroad; and do not know).

Statistical method
The prevalence of perceived parental alcohol problems by sociodemographic characteristics was evaluated by chi-square tests. The associations between perceived parental alcohol problems and psychosomatic complaints were examined by means of cross-tabulations and chi-square tests as well as by binary logistic regression analyses, adjusting for sociodemographic background characteristics. To take the clustered nature of the data into account, with students nested in classes, robust standard errors were estimated in the logistic regression analyses. The analyses were performed with Stata, version 17 (StataCorp, 2021). Table 1 shows descriptive statistics of the study sample. Overall, 34.6 % of adolescents experienced two or more psychosomatic symptoms weekly. Regarding the CAST-6 questions about parental alcohol problems, 12.9 % of adolescents responded affirmatively to three or more of these. For the specific items, the proportions who answered affirmatively varied between 20.6 % ("Parent drank too much") and 7.7 % ("Wanted to hide/pour parent's alcohol bottle"). Table 2 displays crosstabulations with chi-square tests of perceived parental alcohol problems (CAST-6) by sociodemographic characteristics. Perceived parental alcohol problems were more common among girls and among those with another gender identity compared with boys (p < 0.001), among students in grade 11 compared with those in grade 9 (p < 0.01), among adolescents with at least one parent born in Sweden compared with those with two parents born abroad (p < 0.001), and among those with no university-educated parent and those who responded "don't know" to the questions about parental education, compared with those who had at least one parent with a university education (p < 0.001).

Results
Adolescents' psychosomatic complaints (≥2 more than weekly) by perceived parental alcohol problems (CAST-6) are presented in Table 3. Psychosomatic complaints were more common among adolescents who had responded affirmatively to the CAST-6 items, compared with those who had not. This was true for each individual CAST-6 item as well as for the composite CAST-6 measure. All differences between groups were statistically significant (p < 0.001). Table 4 presents results from binary logistic regression analyses, with each individual psychosomatic complaint and the composite measure (≥2 complaints more than weekly) regressed on perceived parental alcohol problems, adjusted for sociodemographic characteristics. Perceived parental alcohol problems were positively and significantly associated with each psychosomatic complaint ( . Additional analyses with a less restrictive cutoff for CAST 6 (≥2 affirmative answers on individual items) showed slightly weaker but overall similar associations (see Appendix, Table A1). We also performed a set of linear regression analyses of the psychosomatic complaints operationalised as continuous measures (see Appendix, Table A2). In these analyses, each complaint was coded 1-5, with higher values indicating more frequent complaints. In addition, we created a continuous measure adding the values of all five complaints to a measure of psychosomatic complaints ranging 5-25. As seen in Table A2, the results from the linear regression analyses largely reflect those from the binary logistic regression analyses.
Subsequently, we included the interactions terms between perceived parental alcohol problems and gender and grade, one at a time, in the fully adjusted models. Through Wald tests, we found that gender moderated the associations between perceived parental alcohol problems and having a headache (p < 0.01) and feeling depressed and down (p < 0.05), while grade moderated the associations between parental alcohol problems and difficulties falling asleep (p < 0.05) and the composite measure of psychosomatic complaints (p < 0.01). Further analyses stratified by gender and grade revealed that the associations in question were stronger for boys compared with girls, and for students in grade 9 compared with students in grade 11, respectively (see Appendix, Tables A3-A7).

Discussion and conclusion
This study aimed to examine perceived parental alcohol problems and the links with psychosomatic complaints among Swedish adolescents aged 15-18 years. The results demonstrated that adolescents with perceived parental alcohol problems had about twice as high odds of reporting psychosomatic complaints compared with adolescents without perceived parental alcohol problems, even when adjusting for
Parental alcohol drinking problems CAST-6 sum at cutoff 3 Less than three At least three
Psychosomatic complaints (≥2 more than weekly) ( %) sociodemographic background characteristics. This finding reflects those of recent studies based on other Swedish data materials using different measures of parental alcohol problems (Ramstedt et al., 2022(Ramstedt et al., , 2023Wahlström et al., 2023aWahlström et al., , 2023b. While prior studies on CAST-6 and adolescent psychosomatic complaints were based on data collected among grade 9 students (aged 15-16) (Ramstedt et al., 2022(Ramstedt et al., , 2023, the present study utilised data from students in both grades 9 and 11, thus expanding the scope of generalisability. One key mechanism in the association between perceived parental alcohol problems and psychosomatic complaints is likely to be stress. Not only may parental alcohol problems be stressful in themselves, but they may also imply that the children have fewer resources that can help to cope with stress. Indeed, children who experience parental alcohol problems are at a risk of having poor family relationships, poor control over the situation in the life, and lack of parental support, possibly leading to a shortage of protective buffering factors and poor development of coping and resilience skills against stress in life (Windle, 1996;Lee & Williams, 2013). Furthermore, children may feel shame, stigma, and guilt because of their parents' drinking problems (Hutchinson et al., 2014;Tamutienė & Jogaitė, 2019). They may also be at an increased risk of bullying victimisation (Ramstedt et al., 2022). A child might try to avoid inviting friends home to reduce the risk for embarrassing situations and prevent peers from learning about the parents' alcohol problems, which could potentially lead to social isolation and/or exclusion. Parental alcohol problems can also be assumed to fall under the broader concept of adverse childhood experiences (ACEs) (Felitti et al., 1998). ACEs have consistently been linked to later adverse health outcomes and the proposed pathways highlight that the exposure to stressful events can lead to several negative biological responses followed by social, emotional, and cognitive impairment as well as the adoption of different health-risk behaviours (Boullier & Blair, 2018;Petruccelli et al., 2019). These kinds of consequences may, in turn, be manifested as psychosomatic complaints in the affected adolescents.
The current study also found that perceived parental alcohol problems was more commonly reported by girls, non-binary adolescents, students in grade 11, those with at least one parent born in Sweden, and respondents without university-educated parents. While previous research has found differences in perceived parental alcohol problems by gender (Haugland & Elgan, 2021;Pisinger et al., 2016;Ramstedt et al., 2022), the explanations behind this are not clear. Possible mechanisms could include that girls are more likely to disclose family-related behaviour or exposure to harmful experiences (Lev--Wiesel & First, 2018;van der Ploeg et al., 2022) and are more sensitive to parental drinking problems (Homel & Warren, 2019). Similarly, older adolescents could also be more likely to recognise and reveal parental drinking problems and the conflicts associated with it compared with younger children. In the case of adolescents with parents born abroad, they are a diverse group with differing background and cultural traditions. Nevertheless, alcohol abuse in Sweden tend to be less prevalent among migrants compared to the native population (Harris et al., 2019), which could explain our finding. Regarding parental education, prior research in the Swedish context has demonstrated that while individuals with a higher education drink more alcohol in total, they tend to drink less intensely compared with individuals with a lower education (Heckley et al., 2017). Higher educational attainment is also associated with a lower risk of alcohol use disorder in Sweden (Calling et al., 2019). Thus, adolescents with parents who have not studied at university may indeed experience more problems related to their parents' drinking.
Additionally, we found that the association between perceived parental alcohol problems and (some) psychosomatic complaints was stronger for boys and for students in grade 9. Since boys and younger students were less likely to report perceived parental alcohol problems, it is possible that more severe cases of parental drinking problems were captured for these adolescents. Some other possible explanations include that boys do not experience the same level of stress from, e.g., school, as girls and might therefore be more susceptible to other forms of stressors (Östberg et al., 2015). Younger students tend to be more dependent on their parents and may thus be more strongly affected by parental problem drinking compared with older adolescents who are probably more independent and have stronger bonds with peers.
To extend current knowledge, longitudinal, prospective studies that are based on children's and adolescents' reports in childhood, and that Table 4 Adolescents' psychosomatic complaints (more than weekly) regressed on perceived parental alcohol problems (CAST-6, cutoff ≥ 3). Odds ratios (OR) and 95 % confidence intervals (95 % CI) from binary logistic regression models with robust standard errors. (n = 9,032). follow the offspring until adulthood would be beneficial. Qualitative research might also improve the contextual understanding of the mechanisms between parental alcohol problems and children's selfreported health. Finally, research should investigate how the association between perceived parental alcohol problems and offspring health can be alleviated. It seems highly relevant to map how social support from, e.g., teachers, other school staff, and leaders in extra-curricular activities may buffer against problematic conditions in the family, and to propose how such knowledge may be translated into policy and practice.

Strengths and limitations
One strength of the current study is that the measure of parental alcohol problems was based on adolescents' survey responses. It is not uncommon to use data on treatment for alcohol-related disorders to measure parental alcohol problems. Since a majority of individuals with mild to moderate alcohol dependence do not seem to seek help and treatment (Wallhed Finn et al., 2014), such measures are however biased to the most severe (clinical) cases and findings may therefore have limited generalisability to the population at large. At the same time, it should be acknowledged that self-reported data may be conflated by bias. For instance, the present study's finding that girls (and participants who marked the "other" gender category) were more likely than boys to report perceived parental alcohol problems may possibly be due to gender differences in perceptions and/or reporting rather than in actual situations (for a discussion, see Wahlström et al., 2023b).
The large-scale nationally representative data, with a high response rate at the student level, can also be regarded as a strength. Notwithstanding, it should be acknowledged that the non-response was most likely systematic. First, it is reasonable to assume that students with a problematic situation at home and/or psychosomatic complaints were probably more likely to be absent on the day of the survey, implying a systematic bias in the non-response. Second, with regards to the students in grade 11 in our sample, there may be an additional bias as upper secondary school is not mandatory in Sweden. Although the majority of adolescents enrol in upper secondary school, it is likely that students with problematic familial alcohol use are more inclined to not continue to or to drop out from upper secondary school (Pisinger et al., 2023). Relatedly, this group of adolescents probably also have more psychosomatic complaints. However, these possible biases are more likely to imply underestimations of the associations, rather than the other way around.
Due to the cross-sectional study design, we were not able to establish a temporal sequence or draw conclusions about causality with support in the data. Yet, it seems implausible that adolescents' poor psychosomatic health causes parental alcohol problems. Nonetheless, we cannot rule out the possibility that there are omitted variables that may impact the reported associations, for instance, parents' mental health problems or (other aspects of) family psychosocial adversity.
Another limitation is that CAST-6 is a life time measurement scale with no referral to a specific time frame, whereas the questions about psychosomatic complaints refer to last six months. Hence, mapping a life time measure scale against six months confined psychosomatic complaints can have temporal validity issues.
Furthermore, it should be acknowledged that the data used in the present study was collected during the covid-19 pandemic, which may have affected the generalisability of the results. Even though no significant change in the alcohol consumption among the general adult population in Sweden during the pandemic has been detected (Guttormsson, 2022;Trolldal, 2022), it is possible that young people's lives were affected by covid-19 in various ways. For instance, due to the social distance recommendations (and to some extent, also distance learning at school), people in general including adolescents were more bound to their homes during this time than usual. It is possible to assume that social support from friends but also from teachers and other school staff was less accessible. Hence the potentially stress-buffering role of social support on the association between parental drinking and adolescent psychosomatic complaints may have been less prominent than at other times. Notwithstanding, the findings regarding both the prevalence of perceived parental alcohol problems and the associations with psychosomatic complaints in the present study are very similar those reported by Ramstedt et al. (2022Ramstedt et al. ( , 2023 which were based on survey data collected prior to the pandemic.

Conclusions
The findings of the present study highlight that adolescents with perceived parental alcohol problems have an increased risk of reporting psychosomatic complaints, and hence this group should be given adequate attention and support. The school health services could play an important role in identifying students who are living with problemdrinking parents and school-based interventions may be effective in mitigating against the adverse effects of parental drinking (cf. Låftman et al., 2021). Other ways for policy makers to address this issue is to target harmful parental drinking in general and to provide children in need with sufficient opportunities to seek and receive help.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability
The data are not publicly available. Access to the data can be applied for at the Swedish Council for Information on Alcohol and Other Drugs (CAN), under the condition that the Swedish Ethical Review Authority has granted permission.

Funding
Open access funding was provided by Stockholm University, Sweden. The study was financed by Systembolagets Alkoholforskningsråd (2020-0047). The funder had no role in the study design, the collection, analysis or interpretation of the data, nor in the writing of the manuscript.

Ethics approval and consent to participate
The paper builds on a voluntary, anonymous previously collected data material. The provided data, for the current paper, has been cleaned from other variables, other than presented, and thus does not fall under the Swedish Ethics Review Act.