7 Overview of the results
7.2 Early parental mental health trajectories
Our first aim was to identify distinct early maternal and paternal mental health subgroups (i.e. latent developmental trajectory groups) according to timing and course of depression and psychological distress symptoms across the transition to parenthood.
7.2.1 Maternal trajectories (Article I)
Different information criteria and likelihood ratio tests of maternal mental health were giving somewhat different results. Based on theoretical meaningfulness and BIC as a highly reliable criteria (Nylund et al., 2007; Tolvanen, 2007), we selected the eight-group solution. The average predicted posterior probabilities for group membership in that solution were acceptable, ranging from .81 to 1.0. However, only four trajectory groups had distinct timing and course of mental health problems.
These groups were also theoretically and conceptually meaningful, and represented 96% of the data. The remaining four groups consisted each of only few subjects and were highly heterogeneous in the course of the maternal mental health: mothers in these groups reported high levels of problems in more than one assessment points, and their symptom courses did not fit any of the four larger groups. Although perhaps representing meaningful subpopulations of mothers, our sample was not large enough to make statistical inferences. However, we did not want to leave them out of the subsequent analyzes, which is why we combined them into one new trajectory group, named as Heterogeneous high levels of mental health problems.
Therefore, five maternal mental health trajectory groups were identified that differed systematically in the timing and course of symptoms across the pre- and postnatal period, revealing distinct peaks in pregnancy, early postpartum or late postpartum. The mean levels of maternal GHQ-psychological distress and BDI-depressive symptoms are illustrated in Figure 3 according to the identified trajectories at T1 pregnancy, T2 two months postpartum, and T3 twelve months postpartum.
Figure 3. Means of reported GHQ-psychological distress and BDI-depressive symptoms according to mothers’ early mental health trajectory group. GHQ = General Health Questionnaire (36-item); BDI = Beck’s Depression Inventory (13-item). N = 763.
The contents of the five trajectories were the following: I Stable low levels of mental health symptoms was the largest trajectory group, comprising 75% of the data.
Characteristic to mothers in this group were low symptom levels of psychological distress and depression throughout the pre- and postnatal period. The trajectory (average) did not exceed the clinically significant cut-off points for GHQ-psychological distress and BDI-depressive symptoms at any time point. II Prenatal
mental health problems –trajectory group involved 6% of the mothers in our sample.
They reported high level of mental health problems in pregnancy that then decreased to low or moderate levels in the early postpartum and remained fairly stable until the late postpartum. In pregnancy the trajectory crossed the clinically significant cut-off points for both GHQ-psychological distress and BDI-depressive symptoms (mild depression). III Early postpartum mental health problems –group involved 9% of the mothers in our sample. Their symptom levels were high in early postpartum, when the child was two months old, but relatively low in pregnancy and late postpartum.
In the early postpartum, the trajectory exceeded the clinically significant cut-off point for GHQ-psychological distress, but not for BDI-depressive symptoms. IV Late postpartum mental health problems –trajectory group comprised 6% of mothers in our sample. They showed a high level of symptoms in the late postpartum, when the child was twelve months old, but low or moderate level of symptoms in pregnancy and the early postpartum. At twelve months the trajectory exceeded the clinically significant cut-off points for both GHQ-psychological distress and BDI-depressive symptoms (mild depression). V Heterogeneous high levels of mental health problems -post-hoc trajectory group combined four small trajectories, comprising in total 4% of the mothers in our data. They showed either chronically high levels of mental health problems or had a highly variable profile. The trajectory exceeded the clinically significant cut-off points for GHQ-psychological distress and BDI-depressive symptoms at every time point (T1-T3).
7.2.2 Paternal trajectories (Article II)
The information criteria and statistical tests of the paternal analysis gave somewhat conflicting results. Based on the highest reliability of BIC and BLRT as statistical criteria (Nylund et al., 2007), the 9-group solution appeared to be the best.
Importantly, however, log-likelihoods of the solutions with nine and ten groups could be only rarely replicated even with as many as 1000 starting values, indicating instability for these solutions. BIC and BLRT suggested the eight-group solution to be the best of stable solutions. When visually comparing the mean courses and group sizes in the 8- vs. 9-group solutions, we found only minor differences: one small trajectory group was split into two, and rest of the groups remained fairly stable.
Therefore, we selected the 8-group solution. The average posterior probabilities for group membership were acceptable, ranging from .81 to 1.00. The solution involved four trajectory groups that were representative of the sample, covering 96% of the
data, and theoretically meaningful. The remaining four groups comprised each only a few fathers, and were highly heterogeneous in the timing and course of paternal mental health. Consequently, they could neither be combined into one group nor included in any of the larger groups. Similarly to maternal trajectory groups we combined them into one new trajectory group, named as Heterogeneous high levels of mental health problems.
Figure 4 displays the courses of fathers’ GHQ-psychological distress and BDI-depressive symptoms in each mental health trajectory group from pregnancy (T1) to child being two months (T2) and twelve months (T3) old.
Figure 4. Means of reported GHQ-psychological distress and BDI-depressive symptoms according to fathers’ early mental health trajectory group. GHQ = General Health Questionnaire (36-item;
Goldberg & Hiller, 1979); BDI = Beck’s Depression Inventory (13-item; Beck et al., 1961). N = 763.
The identified five trajectory groups were as follows: I Stable low levels of mental health symptoms -group was the largest, involving 79% of fathers in our sample. They showed low levels of psychological distress and depressive symptoms throughout the pre- and postnatal period. The trajectory (average) did not exceed the clinically significant cut-off point for GHQ-psychological distress or BDI-depressive symptoms at any time point. II Moderate increasing levels of mental health symptoms –group involved 9% of fathers in our sample. Their symptom levels started out low during pregnancy, but gradually increased towards early and late postpartum. The trajectory did not exceed the clinically significant cut-off point for GHQ-psychological distress or BDI-depressive symptoms at any time point. III Prenatal mental health problems – group involved 5% of fathers in our sample. They reported a relatively high level of mental health problems during pregnancy, but the symptoms then decreased to low or moderate levels towards the postpartum period. The trajectory crossed the clinically significant cut-off points for GHQ-psychological distress and BDI-depressive symptoms in pregnancy, but not during the postpartum period. IV Mental health problems in early fatherhood –group comprised 3% of fathers in our sample. They showed a peak in mental health problems when the child was two months old, whereas in pregnancy and at twelve months the symptom levels were relatively low.
The trajectory exceeded the clinically significant cut-off point for GHQ-psychological distress, but not for BDI-depressive symptoms, at the child’s age of two months. V Heterogeneous high levels of mental health problems –post-hoc group involved 4% of fathers in our sample. Some of them showed high and increasing course of problems; others had extremely high levels of problems during the pregnancy and early postpartum, but not anymore in the late postpartum; yet others suffered chronically high levels of problems. The trajectory exceeded the clinically significant cut-off points for both GHQ-psychological distress and BDI-depressive symptoms in pregnancy and at twelve months, and for GHQ-psychological distress only at two months.
7.2.3 Impact of former infertility
Results showed similar distribution of early mental health problems between ART and NC-groups among both mothers and fathers. This indicated that infertility history and fertility treatments had no effect on the timing and course of parental pre- and postnatal mental health symptoms.