Familial Risk Factors in Relation to Recurrent Depression Among Former Adolescent Psychiatric Inpatients

The family environment plays an essential role in the development of depression among children and adolescents, since children spend most of their time with their parents and other family members. Childhood family instability is commonly linked to offspring’s later mental health problems [7]. Parental mental health problems, especially those of mothers, are also associated with an increased risk of adolescent depression [8, 9]. Other predisposing factors for offspring’s depression are shown to include: single-parenthood, parental low education or unemployment and childhood sexual abuse [10‐15]. Researchers have described four possible mechanisms for a parent’s depression contributing to the onset of depression in their offspring: (1) genetic heritability, (2) innate dysfunctional neuro-regulatory mechanisms on the child, (3) child exposure to negative parenting patterns, and (4) stressful circumstances in a child’s life [16].

Exposure to Adverse Childhood Experiences (ACEs), such as having a depressed family member, has shown to have a negative effect on adolescent mental health [17]. Moreover, a depressed family member, especially the mother, can lead to other adverse experiences such as child neglect and abuse or parental substance abuse [8]. According to Shonkoff and workgroup [18], these experiences can generate toxic stress for a child, which, when accumulated, can increase the risk of childhood depression. The researchers of that study suggested that ACEs can disrupt the neurochemical circuitry in the brain, and other metabolic and/or organ functions, during sensitive developmental periods, leading to impairment of memory and learning functions. Further, ACEs can also increase the levels of inflammatory markers in children, which, in turn, can have long-lasting effects to health. This includes increasing the risk of depression, but also of somatic illnesses, such as cardiovascular diseases and asthma.

Long-lasting harmful parenting patterns can negatively affect a child’s developing brain architecture and cause long-lasting effects on a child’s behaviour and overall stress response system. Therefore, treatment of a mother’s depression does not necessarily translate to improving the mental health of their child [19]. One example of this is when depressed mothers have more difficulties in recognizing and reading their child´s happy faces and gestures, which might, in turn, lead to a decrease in responses to these gestures by the mothers. Thus, a child’s ability to learn of how to communicate their feelings may be hindered and skewed [20].

As summarised in the research paper by Pettit and colleagues [21], recurrent depression is defined by a person having subsequent depressive episodes after the initial episode and a symptomless remission period of approximately eight weeks. However, this definition leaves open what “symptomless” period and its length mean. They noted that the proposed mechanisms of recurrence in depression generally fall in to two categories, namely the “scar model” and the “liability model”. In the scar model, the process of having a depressive episode somehow changes the architecture of the brain in a way that increases the probability of a new episode. In the liability model, the patient has, even before their first episode of depression, one or more underlying risk factor which directly increase their chances of having a depressive episode, providing the underlying factor is present [21].

In previous studies, many patients are reported to fail to achieve complete recovery between successive episodes of depression [22]. The studies have also suggested that sub-acute inter-episodic symptoms associate with shorter asymptomatic periods, more rapid and increased rates of relapse and recurrence, and a greater number of, and more chronic, major depressive disorder (MDD) episodes [23], as well as with significant social and occupational impairment [24]. This partial remission is often characterized by the presence of symptoms such as depressed mood, anxiety, somatic symptoms and sexual dysfunction, without meeting the full criteria for MDD [24]. In recurrent depression, the second episode usually occurs within five years of the first episode, regardless of the age group (children, adolescents and adults) of patients [22]. Individuals with recurrent MDD have shown to experience an average of five separate depressive episodes during their lifetime [25].

Pettit and colleagues [21] analysed multiple possible factors relating to recurrence in depression and showed that adolescent’s own history of depression, as well as parental recurrent depression, were the most notable predictors for recurrent depression in adolescents. The review by Burcusa [26] and Iacono reported that additional risk factors for recurrent depression are the age at onset of first depressive episode, the number of previous episodes, the severity of the first episode and other psychopathological comorbidity. Of these risk factors, the latter two were only associated with adult recurrent depression [26].

Using the data of the Minnesota Twin Family Study, Wilson and workgroup [22] examined the effects of several premorbid risk factors of MDD of offspring aged 11 years. They reported that parental depression and antisocial behaviour, in association to the offspring’s negative emotionality and disconstraints, externalizing symptoms, and childhood related physical abuse, assault and maltreatment increased the offspring’s risk for early-onset depression (compared to late-onset depression) in both recurrent and non-recurrent subtypes of depression. In their study, notable risk factors for recurrent depression in offspring, compared to offspring with single-episode depression, included: lower positive emotionality, higher trait anxiety and childhood sexual abuse/assault, but not parental psychiatric disorders. The researchers summarised that risk of offspring recurrent depression was largely a function of its earlier onset of the disorder. In an adolescent study, previous history of MDD, dysphoric mood and dysfunctional thinking patterns were stronger predictors of recurrent episodes than to first onset episodes and, conversely, major stressful life events were more likely associated with first onset episodes than recurrent episodes [27].

The purpose of this study was to investigate the association of several familial risk factors with recurrence of depression, among former adolescent psychiatric inpatients diagnosed with depression at ages 13 to 17 years. By recognizing early enough the adolescents at high-risk of developing recurrent depression, we could allocate clinical resources more efficiently to this patient group and, thus, decrease risk for recurrence in depression of this vulnerable group of patients. Based on findings in the earlier literature, we hypothesised that poorer family conditions and exposure to stressful family environments in childhood make it more likely that children and adolescents will experience a recurrent path of depression.

This study is a part of an ongoing clinical follow-up project, which investigates psychosocial risk factors and long-term outcomes of psychiatric inpatients hospitalized between the ages of 13–17 years old. The original study population includes 508 adolescents (208 boys, 300 girls) admitted to acute psychiatric inpatient care in Oulu University Hospital in Finland, between April 2001 and May 2006 (later referred as index hospitalization). The catchment area of the hospital covers the whole of Northern Finland, accounting for 43% of Finland’s total geographical area. The study population was homogenous, with Caucasian ethnicity accounting for 98% of the population.

Both the adolescents and at least one parent’s (or guardian’s) signed informed consent was required before participation in this study was allowed. Subjects who were aged over 18 years at admission, mentally retarded, had organic brain disorders or did not provide written informed consent for participation were excluded from the study. Participation rate in the study was high, with 83.7% of the eligible adolescents (n = 637) included. The study protocol was approved by the Ethics Committee of University of Oulu, Finland.

The information on treatment episodes for depression of the study subjects was obtained from the National Finnish Care Register for Health Care (CRHC) until to the end of 2016. Information on inpatient treatments covered the whole life time of study participants, while data on outpatient visits to specialised level of health care was available from 1998 onwards. Depending on the year of data entry, the follow-up time after the index hospitalisation (during years 2001–2006) to the end of year 2016 varied between 10 and 15 years.

The definition of depression was based on the International Statistical Classification of Diseases and Related Health Problems, 10^th revision. According to the ICD-10, a depressive episode is classified as a minimum of two week period of almost constantly having two or more of the following symptoms: decreased mood, increased fatigue or decreased interest in previously enjoyable activities. Moreover, a patient has to have one or more of the following symptoms so that the total number of symptoms is four or more: excessive guilt, lowered self-confidence, recurrent thoughts of death, difficulties concentrating, agitation or retardation, changes in eating or in weight, difficulties sleeping, suicidal behaviour or attempted suicide. Depression is defined as being recurrent if the patient has two or more separate depressive episodes separated by at least two symptom free months [26].

For the purpose of our study, 272 study subjects with hospital-diagnosed depression (ICD-10: F3-codes) were identified from CRHC by the end of 2016. Those participants who had been diagnosed with bipolar disorder (ICD-10: F30, F31, F34.0) (n = 30) or schizophrenia spectrum disorder (ICD-10: F20, F21, F25) (n = 5) were excluded from the current study sample.

The final sample of the current study sample consisted of 237 patients with a diagnosis of depression over their lifetime. Of these, 84 (35.4%) patients had developed recurrent depression and 153 (64.6%) had only one depressive episode. The mean (SD) age at the onset of first depressive episode was 16.4 (3.1) years in recurrent and 15.8 (2.4) years in the non-recurrent depression group.

Statistical significance of group differences in categorical variables was assessed with the chi-square or Fisher’s Exact test, and in continuous variables, with Student’s t test or Mann–Whiney U test. A logistic regression analysis was used to examine the association of family-related factors (Model 1: family type, level of maternal/paternal education, employment status of mother/father, death of a parent, large family size, sibling status), perceived characteristics of the family members (Model 2: distant relationship with mother/father/siblings, psychiatric problems of mother/father/siblings, substance use problems of mother/father/sibling) and adolescent psychiatric disorders (Model 3: affective disorders, substance use disorders, conduct disorders, anxiety disorders, psychotic disorders) to likelihood of recurrence of depression by young adulthood. Model 4 includes all variables presented in Tables 1, 2, 3 as potential predictors for recurrence of depression. In all models, the variables were entered to the stepwise logistic regression model using forward selection criteria and Table 4 reports the results of final models showing statistically significant association with outcome variable (recurrence of depression). All statistical tests were two-tailed and the limit for statistical significance was set at p ≤ 0.05. The statistical software used in analyses was IBM SPSS, version 25.

Family-related factors, assessed during index hospitalization in relation to recurrence of depression by young adulthood of the study subjects are presented in Table 1. Generally, family adversities were emphasized in those suffering from recurrent depression. However, the only statistically significant difference between study-groups was observed in the family sizes of female participants. A greater proportion of females with recurrent depression came from large families (five or more children) compared to those with non-recurrent depression (23% vs. 7%) (χ² = 8.68, df = 1, p = 0.003).

Study subject’s perceptions of the closeness of their relationship with their family members (mother, father, siblings), and their perceived psychiatric and substance use related problems of family members in association to recurrence in depression are outlined in Table 3. The results showed that males reporting having a distant relationship with their mother were more likely to have diagnosis of recurrent depression by young adulthood, compared those in the non-recurrent depression group (45% vs. 22%) (χ² = 4.07, df = 1, p = 0.044). In females, psychiatric problems of fathers (21% vs. 9%) (χ² = 5.01, df = 1, p = 0.025) and siblings (18% vs. 5%) (χ² = 7.341, df = 1, p = 0.007) were related with recurrent depression.

Table 3 shows psychiatric disorders diagnosed during adolescent psychiatric inpatient care. The most common disorders during adolescence were affective (77%), substance use (37%) and conduct (36%) disorders. The prevalence of these adolescent psychiatric disorders did not statistically significantly differ between the recurrent and non-recurrent depression groups, either in male or female study subjects.

Three logistic regression models were performed to investigate the deeper the role of familial risk factors in the recurrence of depression in the study subjects (Table 4). In the final model 3, when all familial risk factors were simultaneously entered into the stepwise logistic regression analysis, an increased likelihood for recurrence in depression was associated to a distant relationship with the mother both in both male (OR = 3.9, p = 0.022) and female (OR = 2.2, p = 0.042) subjects. Further, in females, psychiatric problems in fathers (OR = 3.4, p = 0.013) and siblings (OR = 3.7, p = 0.032) as well as large family size (OR = 4.3, p = 0.005) were all associated to an increased likelihood to develop recurrent depression by young adulthood. None of the adolescent psychiatric disorders (see Table 3) was associated to the recurrence of depression of the study participants.

A strength of our study is that psychiatric diagnoses at the index period during adolescence were based on the semi-structured K-SADS-PL interview, showing high concurrent validity and inter-rater agreement (93 to 100%) [28, 48]. Further, the quality of the data from the Finnish National Care Register for Health Care, provided by the Finnish of Health and Welfare, has been shown to be high and provides access to comprehensive information about the use of national inpatient or outpatient health care resources at a national level [49]. In addition, by examining parental problems, as experienced by the study subjects during their adolescence, it was possible to obtain a more comprehensive and inclusive insight into each patient’s family life than we perhaps would have obtained had we just used information on parental diagnoses of psychiatric or substance use disorders. It also better represents the child’s feelings within their family and asks whether they find that their parents’ behaviour to be problematic or not [50]. We considered the possibility of patients under reporting family mental health problems, given that mental health and substance use problems are, in many cases, still taboo in our society. An important strength in our study is also its longitudinal nature, with a long follow-up period of 9 to 14 years after discharge from adolescent psychiatric inpatient care. It is noteworthy that this period mainly covers adolescence and early adulthood of the study subjects, the period of time when the onset of depression is most likely to occur. This gives us a unique insight into the developmental process of recurrent depression in young adults. Finally, the study population was very homogenous and selection bias can be assumed to have been small, given the high participation rate (83.7%).

The following limitations of our study are acknowledged. Because our study population consisted of former adolescent psychiatric inpatients, these findings are not directly generalizable to the general Finnish population of the same age. Another limitation is that, from our register-based data, it was impossible to determine whether or not the patient achieved complete recovery before a new treatment period for depression. For the purposes of this study, we considered an 8 week minimum period between subsequent treatments for depression, to indicate a ‘symptomless period’ as is also suggested by the ICD-10 criteria. In addition, we have no information on the quality and length of outpatient treatment offered to the adolescents after their first episode of depression, and whether they had access to individual psychotherapy or family therapeutic approaches. Our statistical findings do not allow making conclusions which of the adolescent-related predictors had major impact on depression recurrence by the young adulthood. It is assumable that some potential risk factors for depression recurrence will relate to later lifer of the study participants, and in this respect, our database is limited.

Since our data was collected in the early 2000s, it may be that the field of adolescent psychiatry care has somewhat changed over the time and, therefore, is remains unknown whether this change could have an impact on our results. Our current society and adolescent lifestyle is now much more interconnected across the globe, and the rise of social media may have created a new environment that can potentially affect adolescent mental health adversely [51]. A previous study, published using the same population as our study sample, reported that an interest in computers and video games did not increase the risk of any specific psychiatric disorder among participants but, conversely, a decreased likelihood of a substance-related disorder was observed in boys with computers as a hobby [52]. The authors concluded that social contacts and peers play an important role in preventing adolescent depression, a statement we can also whole-heartedly endorse.