Abstract
Studies on teenage pregnancy have usually focused on the mothers. There is good evidence that the levels of psychological symptomatology in pregnant teenagers and teenage mothers are higher than those observed in older pregnant women and mothers [1], [2]. There is less data on fathers in the setting of teenage pregnancy [3–12]. There is no data on the levels of psychological symptomatology experienced by teenage fathers. Indeed, there is limited data on the psychological impact of childbirth on fathers. The limited Australian data suggests that pregnancy may be the most stressful time for first-time fathers with the highest levels of anxiety reported during pregnancy [13].
Previous research has shown that compared to partners of older women, male partners involved in teenage pregnancy have typically achieved a lower level of education, have higher rates of unemployment, are more financially dependent, lower in socioeconomic status, have more behavioural problems such as smoking, drinking and illicit drug use, have more simultaneous sexual partners and sexually transmitted infections, engage in more aggressive behaviour and have more adverse early life experiences [3–12]. Therefore, teenage fathers are potentially an at-risk subgroup of the population in terms of experiencing psychological symptomatology.
In the present study we planned to explore the levels of psychological symptomatology in fathers in the setting of teenage pregnancy, as compared to fathers in the setting of pregnancy where the pregnant woman was over 20 years of age. The secondary aimwas to explore for factors that had an independent association with high levels of psychological symptomatology in fathers.
Method
A cross-sectional cohort study was undertaken. Institutional ethics committee approval and individual informed consent from study participants were obtained.
Access to fathers occurred after obtaining consent from the pregnant women. Pregnant women aged less than 20 years were defined as teenage mothers and the father of their baby was defined as a father in the setting of teenage pregnancy (Teenage). Of note, the fathers were not all necessarily teenagers themselves, but were partnered to a teenage woman. Those women over 20 years were defined as nonteenage mothers and the father of their baby was defined as a father in the setting of a control pregnancy (Control). For the father to be eligible for participation, the mother of the baby had to: (i) have no plans to abort or relinquish the baby; (ii) know the name and contact details of the father of the baby; (iii) be nulliparous; (iv) have no known fetal anomalies in the pregnancy; and (v) provide consent for the father to be contacted. All subjects were recruited through a large metropolitan tertiary referral hospital.
Men were interviewed by a research staff member during the antenatal period at a mean gestation of 22 weeks (SD=3.5) and completed several questionnaires. The first questionnaire asked about common demographic variables and also included specific questions on contact with psychiatric or counselling services, exposure to death of a parent and predicted age of death.
The Hospital Anxiety and Depression Scale (HADS; [14]) and General Health Questionnaire-28 (GHQ-28; [15]) were completed. The HADS has been used extensively in over 747 clinical studies since it was developed in 1983 by Zigmond and Snaith and is designed for use among patients in non-psychiatric hospital clinics [14], [16]. It is divided into an anxiety subscale (HADS-A) and a depression subscale (HADSD), both consisting of seven items. The HADS is a valid and reliable screening instrument for assessing separate dimensions of anxiety and depression and detecting cases of both disorders in patients from nonpsychiatric hospitals. It gives clinically meaningful results in clinical group comparisons and in correlational studies with various aspects of disease and quality of life [16]. Validation studies have found satisfactory internal consistencies of mean 0.83 in HADS-A and mean 0.82 in HADS-D [16]. Retest reliability shows a high correlation, r<0.80, for up to 2 weeks. The GHQ-28 is a widely used 28-item self-administered screening test designed to detect psychiatric disorders among respondents in primary care or general medical outpatients [15]. It is sensitive to very transient disorders and declines in normal function.
Primary outcome and power calculation
The primary hypothesis was that men, in the setting of teenage pregnancy (Teenage), would be significantly more likely to report high levels of symptomatology on the HADS-A and HADS-D (i.e. a cut-off score greater than 8 on each HADS subscale) compared to those not in the setting of teenage pregnancy (Controls). It was a priori estimated that 45 men in each arm of the study would result in greater than 80% power to detect as significant (at a two-sided error of 0.05), a difference in the incidence of meeting this case criteria for HADS-A or HADS-D in the teenage and control subgroups of 50% and 20%, respectively. Interpreting the HADS as a continuous, rather than dichotomous, measure required an even smaller sample size. Therefore, the study required 90 men but to allow for possible non-compliance 100 were recruited.
Secondary outcomes
There were several secondary hypotheses. They were that men, in the setting of teenage pregnancy, compared to controls, would have higher total scores on the HADS and have higher GHQ-28 total and subscale scores. In addition they would have greater utilization of psychiatric or counselling services in the past, and as a result of unrecognized depressive symptomatology, would have amore pessimistic viewon life and, therefore, predict that they would die at any earlier age than control fathers. Multivariate analysis, using factors significant on univariate analysis, was planned to look at independent associations with a high HADS score in fathers.
Statistical analysis
Data were entered onto a database and analysed using Minitab (Minitab Academic, version 14 ceanet). Within-group comparisons of discrete data, the chi-squared or Fischer's exact tests were applied according to cell size. The Student's t-test or Mann–Whitney U-test were applied to generate p-values for continuous data with normal and skew distributions, respectively. Multivariate analysis was undertaken with the model including all variables significant at a p<0.01 level on univariate analysis. A p-value of 0.05 was considered significant.
Results
Of 56 consecutive eligible pregnant teenage women approached to participate in the study (teenage group), informed consent was obtained from 50 (89% response). All partners subsequently provided consent for interview. The principle reason for the mother declining participation in the study was a fear of a breach of confidentiality. Of 60 consecutive eligible pregnant women over 20 years of age approached to participate in the study (control group), informed consent was obtained from 50 (83% response). Again, all approached men subsequently provided consent for interview. The principle reasons for women declining participation were disinterest in the study and concern over confidentiality. Men were heavily influenced by their partners' decision to support their involvement in the study and provided this as their main reason for accepting study involvement.
The mean age of the teenage fathers group was 20.7 years compared to 29.6 years in older controls (p<0.0001). The mean age difference between teenage mothers and their partners was greater than that seen in older controls (age difference: teenage=3.2 years, control=2.5 years, p<0.0001). There were no significant racial differences between the two groups (p<0.05). After adjusting for age, the educational background of the teenage group was significantly lower than controls (p<0.0001). Family income was also significantly lower in the teenage group compared to controls (p<0.0001). Nearly onethird of the fathers in the setting of teenage pregnancy described themselves as being homeless or as living in very unstable households, compared to no fathers in the control subgroup (homeless/very unstable household: teenage=36%, control=0%, p<0.0001). Teenage fathers were significantly less likely to have plans for a job in the future compared to controls (plan to participate in the workforce in the future be it in an unskilled, semiskilled or professional capacity: teenage=56%, control=88%, p<0.0001). Teenage fathers were also significantly less likely to state that they had a religious belief than controls (teenage=38%, control=82%, p<0.0001). Teenage fathers were more likely to be exposed to family violence as a young child <5 years (teenage=22%, control=2%, p<0.0001).
Table 1 summarizes HADS and GHQ-28 data for the two subgroups of fathers. The HADS-total scores, as well as HADS-A and HADS-D subscale scores were significantly higher in fathers in the setting of teenage pregnancy compared to controls. This was reflected in both an absolute increase in scores, as well as in an increase in the proportion of fathers in the setting of teenage pregnancy meeting the case criteria in all subscales. The magnitude of effect was greater with symptoms relating to anxiety compared to depression (OR anxiety=7.33, OR depression=4.93). Likewise, GHQ-28 total and all subscale scores were significantly higher in fathers in the setting of teenage pregnancy, as compared to control fathers. The effect was again most marked in the anxiety subscale.
Hospital Anxiety and Depression Scale (HADS) and General Health Questionnaire-28 (GHQ-28) scores in the two subgroups of fathers
Despite the higher levels of symptomatology in teenage fathers, there were no differences in the prevalence of utilization of previous psychiatric or counselling services, which was low at 8% and 2% for teenage and control fathers, respectively (p=0.05).
Fathers in the setting of teenage pregnancy predicted that they would have significantly shorter life expectancies compared to controls, with a median difference of 15 years (teenage=70 years (IQR=65–85), control=85 years [IQR=77.5–95], p=0.0003).
When family histories were taken of their exposure to parental death, a striking difference was observed in relation to loss of a male parent. Although overall loss rates were similar, seven men in the setting of teenage pregnancy (14%) had lost a father as a child, compared to only a single control father (2%; Fisher's exact test: one-tailed p=0.028, two-tailed p=0.065). Other control fathers had experienced loss of a father, but the loss was experienced as an adult. These findings are summarized in Table 2.
Exposure to parental loss as a child in the two subgroups of fathers
Multivariate analysis was conducted using total HADS scores as the dependent factor. The independent factors in the model were: age at onset of fatherhood; smoking; drinking alcohol; using illicit drugs during the partner's pregnancy; level of schooling adjusted for age; level of personal and family income; perceived impact of the pregnancy on their life and lifestyle; career ambitions; religious belief; presence of their own mother or father being alive; and exposure to domestic violence or parental divorce under 5 years of age. The results showed that only age of onset of fatherhood (p=0.026, OR=1.10, 95% CI=1.01–1.19), exposure to domestic violence as a child less than 5 years (p=0.050, OR=2.64, 95% CI=1.00–6.97) and their own father still being alive (p=0.030, OR=0.38, 95% CI=0.16–0.91), retained significant associations with total HADS score in the multivariate analysis. Thus, being a father at a younger age and being exposed to domestic violence as a child aged less than 5 years increased the likelihood of a high HADS score, and the presence of a living father was protective and was associated with a lower HADS score.
Discussion
This is one of the first studies to specifically explore the levels of psychological symptomatology in fathers in the setting of teenage pregnancy. The younger age of onset of fatherhood was independently associated with higher HADS and GHQ-28 scores in a multivariate analysis. Symptoms in anxiety domains had a stronger association compared to those in depressive domains in both the HADS and GHQ-28. Similar to teenage mothers [2], fathers in the setting of teenage pregnancy may have high levels of unrecognized psychological symptomatology.
The cross-sectional design in the present study means that it is not possible to determine whether the teenage father's symptomatology predated their partner's pregnancy or was a consequence of it. Interviews were held at a mean gestation of 22 weeks in the antenatal period. The findings do suggest that fathers in the setting of teenage pregnancy are more socially disadvantaged than older control fathers, as has been reported elsewhere 3– [12]. It is therefore possible that such disadvantage may, in turn, increase the likelihood of anxiety or depressive symptomatology and of becoming a teenage father. This sequence would favour the former hypothesis (i.e. symptomatology predating pregnancy).
However, we have previously surveyed 1546 Australian adolescents' levels of idealization about the consequences of pregnancy and parenthood. We reported that young men had significantly higher levels of idealization than young women, tending to overestimate the positive consequences and underestimate (or ignore) the negative ones [17]. It is therefore equally possible that when confronted with the reality of an actual pregnancy, disillusionment may emerge, whichmay contribute to depression and anxiety in the antenatal period. This would favour the consequence hypothesis.
Of concern, in the present study, is the finding that fathers in the setting of teenage pregnancy held a view that their ultimate lifespan would be 15 years shorter on average, compared to fathers in the setting of older pregnancies. This could be secondary to the high levels of anxiety and depressive symptomatology experienced by the teenage subgroup of fathers making them more pessimistic about lifespan. Of more concern may be the fact that these fathers are predicting a truth and that in fact they will have a significantly shorter lifespan compared to other fathers. The reports of personal exposure to the death of their own father as a child may impact upon these beliefs. Of note, there is now a substantial literature in both animal and human research linking emotional stress and early life adversity with trends toward early production of offspring and shorter lifespan 18– [21]. These data culminate in the disposable soma theory that suggests that longevity is determined through the setting of longevity assurance mechanisms so as to provide an optimal compromise between investments in somatic maintenance (including stress resistance) and in reproduction [21]. The study raises several hypotheses that warrant further exploration within the context of this theory. These include the possibility that the early loss of a father may predispose to seeking fatherhood or to early sexual relationships.
Few of the symptomatic fathers had contact with psychiatric or counselling services. The consequences of unrecognized and untreated anxiety and depressive symptomatology in young men are potentially serious, including increased risk-taking behaviour and suicide [22], [23].
We also found an independent association between psychiatric symptomatology and childhood exposure to domestic violence. Evidence of a linkage between childhood witnessing of violence and adult mental illness in women has been established. Our findings support a similar consequence in men. The findings support concerns about the intergenerational impact of domestic violence [24], [25].
In this study, fathers were accessed through their pregnant partners and there were a number of exclusion criteria. Therefore, the findings may not be generalizable to fathers in the setting of teenage pregnancy where the mother of the baby had plans to abort or relinquish the baby, did not know the name and contact details of the father of the baby or did not provide consent for the father to be contacted. We had uniform participation from fathers once maternal approval of the study had been obtained. Most of the fathers commented that they were participating because their partner felt the research was important. It was a strict requirement of our institutional ethics committee that fathers could be approached only if the mother first provided consent. These recruiting restrictions may result in our sample containing a greater proportion of fathers who were in tune to the thoughts or wishes of their partners than might otherwise be expected.
This study has a number of other limitations. Participants were recruited from a single institution, which may limit generalizability. Although events enquired into are highly likely to be accurately recalled, the study includes retrospective data and the potential for recall bias is present, and this could influence findings. The crosssectional design does not enable cause/effect relationships to be determined. However, a longitudinal study that assesses differences between fathers in the setting of teenage and non-teenage pregnancy would enable clarification of whether the difference between the two groups found in this study increase or decrease as the fatherhood role is assumed.
We conclude that greater attention needs to be paid to fathers in the setting of teenage pregnancy, in addition to mothers, as these vulnerable young men may have high levels of anxiety and depressive symptomatology that may benefit from assessment and intervention.