Protective and risk factors for women’s mental health after a spontaneous abortion*

Objective: to examine personal and contextual protective and risk factors associated with women’s mental health after a spontaneous abortion. Method: a cross-sectional study was carried out where 231 women who had experienced spontaneous abortions in the past 4 years answered a self-reporting online questionnaire to assess their mental health (symptoms of depression, anxiety, perinatal grief) and to collect personal as well as contextual characteristics. Results: women who had experienced spontaneous abortions within the past 6 months had higher scores for depressive symptoms than those who had experienced spontaneous abortions between 7 and 12 months ago, while anxiety level and perinatal grief did not vary according to the time since the loss. Moreover, low socioeconomic status, immigrant status, and childlessness were associated with worse mental health after a spontaneous abortion. In contrast, the quality of the conjugal relationship and the level of satisfaction with health care were positively associated with women’s mental health. Conclusion: women in vulnerable situations, such as immigrants, women with a low socioeconomic status, or childless women are particularly vulnerable to mental health problems after a spontaneous abortion. However, beyond those personal and contextual factors, the quality of the conjugal relationship and the level of satisfaction with health care could be important protective factors.


Introduction
In Western societies, it has been estimated that approximately 20% of the pregnancies end in spontaneous abortions (also known as miscarriages) within the first 22 weeks (1)(2) . The exact numbers are not known, as most countries only collect statistics for later perinatal deaths, that is, after the 24 th gestational week or when the foetus weighs more than one pound (3) . Despite this high prevalence, bereavement associated with spontaneous abortions has received much less attention from the scientific and professional communities than that associated with any other type of death (4)(5) . The repercussions of bereavement related to early perinatal loss are nevertheless a significant issue for society, particularly in terms of public health, since a number of studies have indicated its important deleterious effects on women's mental health (6)(7)(8) . Those findings raise an important question about protective and risk factors that may exacerbate or diminish the negative effects of spontaneous abortions on women's mental health (9) . Answers to this question are crucial in order to better target women who are more at risk of developing mental health problems after a spontaneous abortion and to offer appropriate support.
There are personal and contextual variables influencing women's mental health after a miscarriage.
A few authors have examined different personal and contextual variables that may influence women's mental health after a spontaneous abortion and which act as protective or risk factors. Regarding the personal variables, some findings indicated that sociodemographic factors (e.g., age, socioeconomic status) were not associated with women's mental health after a spontaneous abortion (3,6) . However, the majority of the studies on the experience of spontaneous abortion were conducted among white middle-class women living in a marital relationship (3) . Women living in conditions of vulnerability (low income, less schooling, or immigrant status) (3) have received less research attention; as such, there is a need to assess the influence of these variables among more heterogeneous samples. In this vein, a study carried out with a large sample of Australian women found that the schooling level was positively associated with mental health after a spontaneous abortion (10) . The contextual variables of childlessness, advanced gestational age, and spontaneous abortion history were identified as factors that intensify perinatal bereavement (6,11) . However, except for childlessness, the results concerning these variables are inconsistent.
Indeed, some researchers found that advanced gestational age and prior loss were associated with worse mental health (11)(12)(13) , while others found no significant association (11) .
The support offered to the parents by family members and friends or by health care professionals is another variable of interest. The result of an earlier study indicates that support from their spouse was especially important in helping women cope with spontaneous abortions (14) . Thus, it may be that the quality of the conjugal relationship has more influence on women's mental health after a spontaneous abortion than the simple fact of being in a couple. The level of satisfaction with health care may also be important to mitigate the deleterious effects of spontaneous abortions on women's mental health. Indeed, health care professionals are key actors in the health pathway of couples living through the experience of a spontaneous abortion (4,15) . However, even though the experience and repercussions of bereavement have been studied, women's experiences of professional interventions-or the absence thereofhave received minimal research attention.
Findings from the few qualitative studies on the women's experience of spontaneous abortions have abortion was classified as within the past 6 months, between 7 and 12 months, between 1 and 2 years, and between 2 and 4 years, while the number of spontaneous abortions was classified as 1, 2, or 3 or more. Finally, the participants indicated whether they had any living children of whom they were biological parents (yes/no). The participants could also indicate "not applicable" if they had not received the service. The scores for each service received were averaged to obtain a global satisfaction score (α = .95).
The analyses were performed using SPSS v.22 (IBM). There were no missing variables for the measures of mental health and quality of conjugal relationship.
For the level of satisfaction with health care, personal and contextual variables, there were few missing data per variable (from 1 to 7 per variable -less than 4%).
Since these data were also randomly distributed, they were not imputed, and listwise deletion was used for main analysis (28) . Descriptive statistics (mean, standard deviation, proportion) was conducted to characterize the sample. One-way analyses of variance (ANOVAs) for the categorical variables, and correlations for the continuous variables, were then performed to examine associations between personal and contextual factors Rev. Latino-Am. Enfermagem 2020;28:e3350. and women's mental health. Tukey post hoc tests were performed when the ANOVAs indicated a significant difference and the independent variable comprised more than two groups (e.g., income, schooling).
Hierarchical regression analyses were performed to examine the association between the quality of the conjugal relationship and the level of satisfaction with health care and mental health, while controlling for personal and contextual factors. In the first step, the contextual and personal variables that were significantly associated with mental health in the ANOVAs and the correlations were entered into the analysis. The quality of the conjugal relationship was entered in a second step, followed by the level of satisfaction with health care services in a third step. The statistical power analysis performed with G*power (29) indicate that, to achieve a statistical power of 95%, with an average effect size, and 8 independent variables (quality of the conjugal relationship, level of satisfaction with health care, and up to six control variables), a sample of 160 participants is needed. Thus, the sample of the current study (n=231) provides a satisfactory statistical power. Figure 1 illustrates the study variables included in the analysis.

Contextual Characteristics
Childlessness, number of weeks of foetal gestation, number of miscarriages, and time elapsed since miscarriage

Personal Characteristics
Age, socioeconomic status (schooling, income), immigrant status

Protective factors
Quality of the conjugal relationship Satisfaction with health care women were not asked if they were pregnant at the time of data collection, which could be a limitation of this study. Regarding age, income and schooling, the sample is representative of the women in the region (18) .

Discussion
The main objective of the present study was to identify personal and contextual variables that can represent risk factors for women's mental health after a spontaneous abortion. Specific attention was also given to two potential protective factors: quality of the conjugal relationship and satisfaction with health care.
The results showed a high rate of depression among the women in the study. Indeed, half of them could be classified as possibly depressed. This proportion is much higher than the rate of postpartum depression among Canadian women overall (8.69%) (31) .
The results also indicated that the women who had Along these lines, a previous study (12) had found that the association between prenatal loss and depression and anxiety did not differ significantly over time and could persist for up to 3 years. A recent study (6) indicated that depressive and grief symptoms persist  (6,11) , and the use of We also wish to thank the participating mothers for their generosity with their time in a demanding period of their life.