Nearly 6% of children in Sweden grow up with a parent with mental illness (Hjern and Manhica
2013), but international figures point to as many as one in five children (Reupert et al.
2012). International as well as studies from Scandinavia estimate that a substantial minority of patients in adult psychiatric care have children under 18 years old (Lauritzen
2014; Maybery and Reupert
2018; Ruud et al.
2019). The increased risk of developing mental health problems is well established in children growing up with a parent with mental illness (LeFrancois
2012; Leijdesdorff et al.
2017; Ramchandani and Psychogiou
2009).
Parenting can be both greatly rewarding and a daunting task for anyone, but the complex task of parenting is more challenging when the parent with severe mental illness (Campbell et al.
2012; Wan and Green
2009). Fathers and mothers with severe mental illness experience all of the challenges of other adults struggling to balance their roles as workers, spouses, and parents. The symptoms of mental illness, however, may cause disturbances in thought and behaviors that result in difficulties handling the parenting role (Blegen et al.
2010,
2012; Krumm et al.
2013). The relationship between mental illness in parents and its impact on parenting and child outcomes is complex and remains to be fully elucidated. First, there can be a direct causal relationship based on genetic inheritance of liability from parents to offspring (Ripke et al.
2014), the effect of maternal mental illness in pregnancy through fetal programming (Stein et al.
2014), and the effect of maternal and paternal severe mental illness on parenting behavior, the parent–child relationship, and general family functioning (Smith
2004).
Studies have reported impaired parenting capacity in parents with psychosis (Diaz-Caneja and Johnson
2004; Engur
2017, Evenson et al.
2008). Many studies of mothers with psychosis reveal that they struggled with problems in everyday parenting, including ensuring and asserting discipline and maintaining boundaries, problems in their parental role due to symptoms of the illness or side effects of medication, and feelings of guilt, shame, and fear of the possible effects of their illness on their children (Dolman et al.
2013). However, a large study of parents with psychosis (
n = 234) reports that most had good-enough or adequate parenting defined as participation, interest, and competence in childcare over the last 12 months as rated by trained clinicians (Campbell et al.
2018).
Parenting is multiply determined and influenced by characteristics of the parent, the child, and the family’s social context (Belsky
1984). Recent research on parenting has attempted to develop domain-specific models to investigate discrete dimensions of parenting (Grusec and Davidov
2010). A domain-specific model recognize that caregiving is multi-dimensional and involves a variety of specific practices or actions in a variety of specific contexts. The need of investigating specific parenting practices used in specific contexts or domains of child activities, behaviours, or needs has been put forward by socialization researchers (Bornstein et al.
2008; Grusec and Davidov
2010; Turiel
2010). General and global characterizations of parenting can fail to recognize that parenting practices may be adequate and positive in one context but inadequate or harmful in another. Grusec and Davidov (
2010) used a behavioral systems approach and described parenting in the context of the different needs and childrearing goals that are activated in different situations. Parenting is depicted as using different practices flexibly depending on their childrearing goals, their children’s needs, and the types of child behaviors toward which their parenting is directed. The model proposes five main domains of socialization: protection; reciprocity; control; guided learning; and group participation. The different domains require different parenting competences and skills and encourage children to acquire different social, cognitive, and emotional competences. (Each domain is briefly described in the Results section).
Method
Participants
The study included 15 parents (10 mothers and 5 fathers) with an average age of 42 years (range 36–56, SD = 5.70). Two of the parents worked half-time or more, one was on parental leave, and the other 12 were on sick leave. All but two of the parents had 12 or more years of education. One participant had been in contact with psychiatric services for two years, three participants for five to six years, four for eight to ten years, and six for more than ten years. The participants’ primary diagnoses were schizoaffective disorder (8), schizophrenia (2), psychotic disorder (3) and major depressive disorder with psychotic episodes (2).
Eight parents were married or co-habiting and seven were divorced and lived as single parents. Ten parents lived with their children, three had alternating living arrangements with the other parent, and two had a right of access to their children, i.e., no legal custody but regulated allowance to see their child. The 15 parents had 17 children (11 boys and 6 girls) between 3 and 16 years of age, with an average age of 10 years (SD = 4.09).
Procedure
The data comes from a project examining the implementation of two child-focused interventions for parents with mental illness: Let’s Talk About Children (Solantaus and Toikka
2006) and Beardslee’s family intervention (Beardslee et al.
2003). The parents were recruited from outpatient clinics specializing in psychosis in 2015–2017 in the second largest city in Sweden.
Eight open-care psychosis service units were contacted via management and agreed to participate. Seven units were included, with one excluded because it aimed to treat only elderly people. Inclusion criteria were being admitted to a psychiatric outpatient clinic specializing in psychosis and being a parent of at least one minor child. An unknown number of parents who were considered unable to participate in research by their care provider due to intellectual or psychiatric problems were not invited to participate.
Adult patients who had participated in one of the child-focused interventions received verbal and written information about the research project via care providers and were later given further information by the research staff. If the parent agreed to participate, a meeting was scheduled at a place preferred by the parent. Data collection took 2 to 3 h and included a questionnaire and a qualitative interview. Each participant was given a 10-euro voucher for participating. The study was approved by the Regional Ethics Review Board, University of Gothenburg (ref. nr. 599-15).
Measures
The interview guide was semi-structured and included questions on parenting and on parents’ perceptions of their child, and experiences of participating in family interventions. For this particular study, questions related to the intervention were excluded.
The interview questions aimed to gain an understanding of the informant’s perceptions of family life, their parenting, how their parenting was affected by their illness, and to get an image of how they perceived their child, and their child’s wellbeing and social situation. Examples of interview questions are; “Could you describe an ordinary day in your family?”, “How do you handle conflicts with your child?”, “How do you think your psychosis affects you as a parent?”, and “How would you describe your child?” All questions were open-ended, and follow-up questions were used to gain a deeper understanding of the participants’ experiences and perceptions and to clarify potential misunderstandings. The interviews lasted from 45 min to 1.5 h.
Because parenting can be difficult to discuss with people whose parenting ability might be questioned, the interviews were conducted in an informal style and the interviewer occasionally talked about own experiences of being a parent. For some, this approach seemed to have a calming effect and they became more talkative.
Data Analysis
Data was analyzed according to Braun and Clarke’s thematic analysis (
2006). After an initial reading of the interview transcripts, all data related to the research aim were analyzed deductively using Grusec and Davidov’s (
2010) model of parenting. The data within each domain were coded inductively to form the subthemes. During thematic analysis, the authors approached the data from a critical realist perspective, applying a semantic, low-level interpretation.
First, all meanings units in which parents themselves made a link between their illness and their parenting were selected for analysis. The aim was not to give an overall view of parenting, but to focus specifically on links between illness and parenting. However, if participants spoke about being on sick leave because of their psychosis and discussed how this affected their presence as parents, the meaning unit was marked. The authors then discussed and explored the essence of the five theoretical domains of protection, reciprocity, control, guided learning, and group participation that constitute Grusec and Davidov’s (
2010) model of parenting. The essence of each domain was summarized, and concepts that were significant were noted as codes. During the third step, the first author categorized all relevant data deductively using these codes. Fourth, the second author read through all meaning units to make sure they were placed in the most suitable domain. When disagreements arose about the fit of a quotation, the quotations were discussed and recoded when appropriate. Details of how each participant contributed to each domain are presented in the supplementary online material. During the fifth step of the analysis, the first author recoded the meaning units in each domain inductively (on a semantic level). In the sixth step, the semantic codes were conceptualized at a higher level of interpretation. All authors discussed and reorganized the inductive codes and quotations into subthemes within each domain.
Reflexivity and Ethics
It is acknowledged that the background and standpoint of the researcher will interact with the research process and influence the conclusions drawn (Elliott et al.
1999). One solution applied to avoid preconceptions about how psychosis affects parenting was parallel coding, an efficient means of ensuring rigor in the analysis (Yardley
2000). An equally important aspect of doing research on sensitive issues is ethics (Yardley
2000). Besides following ethical guidelines, we applied ethical considerations to every step of the study. For example, the parent’s well-being was always a higher priority than the research question. Likewise, a child’s well-being was always a higher priority than their parent’s; when there was concern for a child, the patient’s care provider was informed and, when necessary, social services were alerted.
Discussion
The present study aimed to explore how parents with psychosis experience that their illness affects parenting. All domains in Grusec’s and Davidov’s model of parenting (
2010) were described as negatively affected by symptoms. Periods of depression, lack of energy, and fatigue hindered many of the parents to create and maintain family routines. Furthermore, hallucinations and delusions seemed to affect aspects of parenting such as providing protection and support and taking part in reciprocal interaction. To compensate for some of these insufficiencies, parents described withdrawing from the child and/or seeking support from their social and professional networks to protect and meet the needs of their children.
In the domain of protection, appropriate parenting involves responding to distress in a way that the child perceives as comforting and helpful (Grusec and Davidov
2010). According to our findings, depression and distracting voices obstructed the parents’ capacity to protect and comfort their children. Some parents also expressed that their children’s need for comfort made them distressed and anxious. These results are consistent with previous studies indicating that symptoms such as delusions, hallucinations, depression, and fatigue affect parents’ attention and ability to meet the child’s emotional needs (Healy et al.
2016; Kahl and Jungbauer
2014).
Grusec and Davidov (
2010) also stress that unneeded protection and comfort are inappropriate responses (Grusec and Davidov
2010). Our findings suggests that some parents, because of hearing voices or paranoia, protect their children from their own purely internal threats. For children, such seemingly unmotivated protective behaviors can be unpredictable, confusing, and restrictive of their necessary explorations. Although some parents, in periods of severe symptoms, withdrew from their parental role in order to protect the children, it is likely that the child was left alone to cope and understand what was happening.
An important outcome of appropriate parental protection is that children develop a greater ability to manage stress and to self-regulate distress (Grusec and Davidov
2010). There is robust evidence that children of parents with schizophrenia are at heightened risk of attention deficit hyperactivity disorder (ADHD) (Sanchez-Gistau et al.
2015), a disorder characterized by deficits in self-regulation. These parents’ descriptions of unpredictable comfort and difficulties in regulating the child’s distress may be important to consider in exploring the association between parental psychosis and ADHD in children.
In the domain of reciprocity, parenting involves accommodating the child’s reasonable requests of for interaction (Grusec and Davidov
2010). In this study, fatigue following a psychotic episode or medical treatment, as well as difficulties in staying focused because of voices, interfered with parents’ capacities to communicate and play with the child. According to a review of mother-infant interaction in parents with schizophrenia, there is a lack of studies exploring interaction beyond 36 months (Davidsen et al.
2015). However, three studies showed that mothers with schizophrenia, compared to mothers with affective disorders, had less mutually satisfying, engaged, and smooth interaction at 4 months (ibid). Studies exploring parent-child interaction from the child’s perspective showed that children experience their parent with psychosis as mentally absent, emotionally unavailable, and as more focused on their own personal problems and symptoms than on the interaction with their child (Duncan and Browning
2009; O’Connell
2008). Our findings indicate, unsurprisingly, that hearing voices has negative effects on parents’ ability to focus when interacting with their child. However, the fact that some parents related the difficulties to side-effects of medication may be important for health care staff to notice. Furthermore, how medical treatment with for example neuroleptics affects parenting abilities is a blind spot in research, and requires further exploration considering the amount of patients who receive medical treatment and have minor aged children. Another important finding was that the children were described as reacting with anger and frustration or becoming calm and withdrawn. A focus in future studies could be to explore the relationship between externalizing and internalizing behavioral problems (Donatelli et al.
2010) and parents’ capacity to comply with their children’s wishes for mutual interaction.
Appropriate caregiving in the control domain involves applying an appropriate amount of authority in order modify the child’s behavior (Grusec and Davidov
2010). In general, control and discipline was an absent topic in the interviews and the parents attributed their lack of control to their lack of energy. Some parents also expressed being afraid of upsetting the child, dreading both the child’s and their own emotional reactions. There is a dearth of studies investigating parental discipline or control in this specific sample. However, parental self-efficacy, warmth, and positive discipline practices have been found inversely related to depressive symptoms (Jones and Prinz
2005). Explanations of these results seem to be in line with what the parents in the present study described; that avoiding disputes by letting the child have its’ own way, requires less energy than setting limits.
Effective parenting in the guided learning domain involves suitable structure, information, strategies, and feedback that children need to learn and improve important cognitive, emotional, and social skills (Grusec and Davidov
2010). The major issue in these interviews was the parents’ concern about how to talk to and inform their children about their mental illness. The prominence of this theme may have been related to the parents’ previous participation in interventions aimed to facilitate communication about parental illness. The results, however, indicate that despite participating in these interventions, parents’ provided their children either unrestrained or insufficient guidance on how to comprehend and cope with their parent’s illness. Another important finding was that parents expressed a wish for support in their parenting role to enable them to facilitate their child’s development. Sensitive scaffolding in the area of guided learning requires awareness of the child’s developmental stages and an ability to take the child’s perspective (Grusec and Davidov
2010). Difficulties in taking the child’s perspective may relate to a preoccupation with one’s own problems; leaving less energy to focus on the child’s emotional needs. The scarcity of examples concerning guided learning could also be related to the results in the domain “Group participation”, that mental illness, unemployment, and limited social networks restricts the parent’s capacity to support the child’s abilities in a domain focusing on cognitive and social scaffolding. Furthermore, guidance in this domain requires parents to be goal oriented, structured, and strategic (Grusec and Davidov
2010), abilities that may be undermined by low socio-economic status and mental health problems.
In the domain of group participation, parents are seen as models for how things are done in wider social contexts and groups (Grusec and Davidov
2010). The most prominent obstacle to the parents’ promotion of their child’s participation in in-groups was financial difficulty, which hindered the children from having the same material standards as their peers. Financial difficulties as an obstacle for parental functioning and child well-being is a problem that appears in studies from both parents’ and children’s perspective (Dolman et al.
2013; Kahl and Jungbauer
2014). Being unable to dress as other youths do and having an identity influenced by the parent’s mental illness could place the child’s chances of joining an in-group on shaky ground. Parents’ ambivalence toward themselves as role models, along with their descriptions of exposing their children to periods of odd beliefs and behaviors, could also explain unusual social and behavioral expressions in children of parents with psychosis.
Finally, the new/additional theme of “Unpredictable absences” crosses all domains in Grusec and Davidov’s model (2010) as parental psychosis involves mental and/or physical absence from the child. Mental withdrawal due to psychoses, and longer periods of hospitalization pose a constant threat to the fundamentally important continuity in the relationship between parent and child. This theme may be important to consider when examining the heightened risk for children of mentally ill parents to be insecurely attached (e.g., Davidsen et al.
2015).
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.