The Association of Religion and Spirituality with Postpartum Mental Health in Women with Childhood Maltreatment Histories

Abstract

Objectives

Although the associations between religion and spirituality and mental health and trauma have been studied extensively across various populations, relatively few studies have focused on the postpartum period. This study aimed to shed light on specific domains of religiosity and spirituality that may be resiliency factors for positive postpartum adjustment defined as low depression and high quality of life in mothers oversampled for childhood trauma histories.

Methods

We examined several religion and spirituality variables among 108 women at 6 months postpartum as well as prospective relations from religion and spirituality to postpartum depression and quality of life at 12 and 15 months postpartum.

Results

We found that the personal aspects of self-forgiveness and forgiveness for others were most relevant as resiliency factors predicting lower postpartum depression and better quality of life even when controlling for other risks (trauma and demographics). Surprisingly, no other religion and spirituality domain had associations with postpartum depression or quality of life, with the exception of a significant negative association for organizational religiousness with quality of life at 12 months postpartum.

Conclusions

Our findings suggest that forgiveness, especially to self and to others, in women who have been physically and mentally hurt as children may be associated with mental wellness and quality of life in the late postpartum period. Further, our results point to the need to study specific religion and spirituality aspects in the context of specific populations and conditions instead of generally studying religion and spirituality as a common marker for coping.

The postpartum time is a period of transition wherein a mother undergoes significant changes, increasing her vulnerability to psychopathology (e.g. Van der Woude, Pijenborg, & de Vries, 2015). It is a time of tremendous physical, emotional, and social changes, which may also impact her quality of life (Irwin, Beeghly, Rosenblum, & Muzik, 2016; Mogos, August, Salinas-Miranda, Sultan, & Salihu, 2006) as well as make her prone to postpartum depression (O’Hara & McCabe, 2013). Not surprisingly, postpartum depression is found to be associated with quality of life (Van der Woude, Pijenborg, & de Vries, 2015; Webster, Nicholas, Velacott, Cridland, & Fawcett, 2011), and postpartum mothers’ subjective quality of life is highly relevant to the well-being of the family as a whole (Irwin, Beeghly, Rosenblum, & Muzik, 2016). Particularly, mothers with adverse childhood experiences have increased vulnerability to postpartum depression (Morelen, Rosenblum, & Muzik, 2018). Many studies focus on resiliency factors for positive postpartum wellbeing (i.e., low depression and high quality of life), particularly in the context of maternal childhood abuse (Irwin, Beeghly, Rosenblum, & Muzik, 2016; Martinez-Torteya, Katsonga-Phiri, Rosenblum, Hamilton, & Muzik, 2018). However, of the many potential postpartum resiliency factors, one factor (that is, an individual’s religiosity and spiritualty) has received less attention.

Although there are many different definitions for religiosity and spirituality (R/S), they can be understood as dimensions of human experience that involve beliefs, practices, and experiences related to transcendent or sacred reality (Miller & Thoresen, 2003). Although interrelated, religiosity and spirituality are empirically and theoretically different. It was suggested that religiosity involves behaviors related to organized traditions, while spirituality usually refers to more general beliefs and experiences. Many individuals may identify as both religious and spiritual, but not all do (Miller & Thoresen, 2003) and it is possible to be spiritual being religious and vice versa (Hall, Meador, & Koenig, 2008).

Although the link between R/S and mental health has been studied extensively across various populations, relatively few studies have focused on the postpartum period (Cheadle et al., 2015; Cheadle & Schetter, 2018; Clements, Fletcher, Childress, Montgomery, & Bailey, 2016; Mann, McKeown, Bacon, Vesselinov, & Bush, 2008). Most studies in non-perinatal populations show positive relationship between R/S and mental health. Across varying populations and cultures greater R/S was associated with better mental health (e.g. AbdAleati, Mohd Zaharim, & Mydin 2016; Koenig, 2009; Paloutzian & Park, 2013), for example, lower rates of depression, anxiety and other stress-related mental illnesses (Oman & Lukoff, 2018). Similarly, a large meta-analysis reported direct effects of religiousness on fewer depression symptoms, even when controlling for gender, age, or ethnicity. The inverse association between religiousness and depression was strongest in studies involving people undergoing recent life stress (Smith, McCullough, & Poll, 2003). In addition, the positive effect of R/S was not limited to depression but also found for quality of life (Mosqueiro, da Rocha, & de Almeida Fleck, 2015; WHOQoL SRPB Group, 2006).

R/S experiences may be especially relevant for postpartum women as many traditions attach religious and spiritual significance to childbirth, parenthood, and family (Cheadle et al., 2015). The postpartum period may be regarded even in the best circumstances, as a period of heightened stress, based on sleep deprivation, recovery from birthing, emotional demands and hormonal transitions (e.g. Matthey, 2016). In addition, the post-partum period is a time of personal and familial transition and can be a high stress period. Therefore, R/S could serve as potential modifying forces or guiding factors to prevent postpartum depression and/or enhance quality of life.

However, research regarding the role of R/S in the postpartum period is more limited. In one study of mostly Christian, African-American new mothers, authors reported that those mothers with lower R/S evidenced more depressive symptoms during the firth 6 postpartum months (Cheadle et al., 2015). In a longitudinal study following mothers for 12 months postpartum, religion and spirituality separately and as a latent factor predicted decreasing trajectory of depressive symptoms (Cheadle & Schetter, 2018). Similarly, in another longitudinal study of mostly poor postpartum women in rural Appalachia, those with low social support or religious commitment were at highest risk for depression (Clements, Fletcher, Childress, Montgomery, & Bailey, 2016). Finally, in a prospective cohort study, organized religious participation was found to be protective for the development of postpartum depression (Mann, McKeown, Bacon, Vesselinov, & Bush, 2008). In sum, while limited in scope, the currently available data point to R/S as potentially protective of postpartum depressive symptoms.

Prior work has shown a strong link between mothers’ experiences of childhood trauma, such as abuse, neglect or significant household dysfunction, and postpartum mental health problems (Morelen, Rosenblum, & Muzik, 2018). In these contexts, R/S may be particularly relevant. R/S and childhood abuse may have reciprocal relationships, such that R/S beliefs may buffer the impact of childhood abuse, whereas childhood abuse may shape the development of adult R/S (Dervic, Grunebaum, Burke, Mann, & Oquendo, 2006). Prior work has also suggested that R/S may help in trauma recovery and support coping/resilience in the face of adversity (Morgan, 2009; Southwick, Litz, Charney, & Friedman, 2011). In addition, R/S were suggested as possible mechanisms to promote stress resilience in survivors of childhood trauma (Brewer-Smyth & Koenig, 2014) and may, in some cases, be a moderator of the development of posttraumatic responses in childhood abuse survivors (Walker, Reid, O’Neill, & Brown, 2009)

Thus, given that childhood abuse history is a known risk factor for postpartum mental health, and that R/S has been identified as one potential buffer in the face of abuse, we are particularly interested in exploring these links. We specifically wonder whether for mothers who are childhood abuse survivors, early postpartum R/S may function as a resiliency factor against the onset of postpartum depression and self-rated low quality of life across the subsequent postpartum period. We hypothesize the existence of this buffering role, such that high R/S measurements at 6 months postpartum will be associated with lower depression symptoms and higher ratings of quality of life at 12 and 15 months postpartum, even when controlling for severity of childhood abuse and demographic risk.

Methods

Participants

Our sample consisted of a subset (n=108/268) of postpartum women recruited from a larger longitudinal study (removed for blind review) examining the associations between maternal childhood abuse and adjustment to motherhood, parenting, and infant outcomes. Participants were recruited either as a follow-up to a parent study or through community advertisements. All participants were non-psychiatrically referred English-speaking women, aged 18 years and older, who were up to four months postpartum. Exclusion criteria included diagnoses of schizophrenia or bipolar disorder, substance use problems within the last three months, and mothers of infants with self-reported severe health/developmental problems or more than 6 weeks premature.

Procedures

The study was approved by the University of Michigan review board with mothers providing oral informed consent at the four-month enrollment phone call and written consent at the 6 month home visit. Mothers were enrolled before or at 4 months postpartum and followed until the child’s age of 18 months. Longitudinal data collection on maternal and infant metrics occurred at 4, 6, 12, 15 and 18 months postpartum and included information on maternal demographics, childhood trauma, mental health, and observed and self-reported parenting, as well as child attachment and behaviors and mother-and-infant physiology.

In this paper, we report on maternal data collected at 4, 6, 12 and 15 months postpartum, and only on a subset of mothers (n=108) who were given the Brief Multidimensional Measure of Religiousness and Spirituality (BMMRS; Fetzer Institute/National Institute on Aging Working Group, 1999); this measure had been added after study initiation and thus only a subset provided data. There was no difference between the overall sample and subsample with completed BMMRS on any demographic characteristics except race; the subsample had fewer African American women (73% in the overall sample versus 23% in the subsample). Therefore, race was added as a covariate (as part of the demographic risk variable). Due to attrition, the 12 month (20%) and 15 month (13%) quality of life and depression measures had a smaller sample size, and this is reflected in the tables. Participants were compensated up to $140 across the study period.

Participant demographics are displayed in Table 1. Overall, the sample had relatively low demographic risk ($\overline{x}=1.03$, sd = 1.33), which we defined by low income, low education, young maternal age ($\overline{x}=29$ years, sd = 5.57), being unpartnered, and being a non-white minority; however, over 70% of the sample reported experiencing some type of childhood abuse or neglect, and about half of those (35.2%) indicated their previous childhood abuse or neglect as severe. Most of the participants (58%) reported more than one type of childhood abuse. The religious preferences of the sample were mainly of Christian denomination (74.1%), followed by non-religious (19.4%), Islam (3.7%), Buddhist (2.8%), and Judaism (2.8%). These were further categorized into Christian denomination, Nonreligious, and Non-Christian (see Table 1). Across these three groups, there were no significant differences on any of the predictor or outcome variables.

Table 1.

Descriptive Characteristics

		Total Sample (n=108)
		n	Percent
Age	> 22 years old	98	90.7
	< 22 years old*	10	9.3
Partnered	Partnered	87	80.6
	Not Partnered*	21	19.4
Racial	Caucasian	73	68.9
	Minority*	33	31.1
Education	Less than High School*	4	3.7
	High School or GED	11	10.3
	Some College, Associate’s Degree, Vocational Degree, or Technical Degree	31	29
	Bachelor’s Degree	32	29.9
	Graduate Degree	29	27.1
Income	Less than $25,000*	32	30.1
	$25,000–$49,999	23	21.8
	$50,000–$74,999	11	10.3
	$75,000–$99,999	16	15.1
	$100,000 or more	24	22.6
Total Demographic Risk	0	55	50.9
	1	21	19.4
	2	16	14.8
	3+	16	14.9
Religious Preference	Christian	75	72.1
	Non-Religious	20	19.2
	Religious but non-Christian	9	8.7

Note: n’s may not total 108 due to missing information or the ability to select multiple categories in the case of religious preference.

^*Risk variable.

Measures

Childhood abuse

The Childhood Trauma Questionnaire (CTQ; Bernstein & Fink, 1998) was completed at the 4 months postpartum visit, when participants rated their childhood experiences on this 28-item self-report questionnaire. The CTQ yields information on childhood physical, sexual and emotional abuse, and physical and emotional neglect. Each item was rated on a 5-point Likert scale ranging from 1 (never true) to 5 (very often true), with some items reverse-scored to yield a continuous total score (total CTQ score; α = 0.936). Higher total CTQ score indicates more abuse exposure. Examples of items are: “[When I was growing up] I didn’t have enough to eat”, “Someone threatened to hurt me or tell lies about me unless I did something sexual with them”, and “People in my family looked out for each other”.

Demographics

The demographics questionnaire was administered at the 6 months postpartum visit and consisted of 28 items tapping into current living situation, race/ethnicity, work, relationship status, education level of mother and father (if known), and income. Race, income, age, marital status, and education variables were dichotomously coded (0=no, 1=yes) to create a continuous risk variable: non-White minority, single parent (i.e., unmarried or unpartnered), low education (i.e., less than a high school diploma or GED), low family income (i.e., <$25,000), and young maternal age (i.e., <22 years old) statuses were summed (0 to 5) to create a cumulative risk index (Sameroff et al., 1993).

Religiousness and spirituality (R/S)

The Brief Multidimensional Measure of Religiousness and Spirituality (BMMRS; Fetzer Institute/National Institution on Aging Working Group, 1999) was administered at the 6 months postpartum visit and is a self-report questionnaire focusing on various dimensions of religion and spirituality that contribute to health outcomes. This instrument has been used in many studies worldwide, and has been validated with many different populations. Although we administered the entire scale, only the following subscales with acceptable internal consistency within our sample (Cronbach’s alpha above 0.7) were used in analysis. Daily Spiritual Experiences (α = 0.92), which measures the individual’s experience of transcendence in daily life and experience of interaction with a higher power. This subscale consists of six items rated on a six-point response format, ranging from 1 (“many times a day”) to 6 (“never”). Meaning (α = 0.71) measures a sense of meaning in life. This subscale is composed of two items with a four-point response format, which range from 1 (“strongly agree”) to 4 (“strongly disagree”). Forgiveness (α = 0.78) measures the degree of forgiveness of self and others, and belief in the forgiveness of God. This subscale consists of three items rated on a four-point response format, ranging from 1 (“always”) to 4 (“never”). Private Religious Practices (α = 0.79) measures the frequency of religious behaviors This subscale is composed of five items, four of which used an eight-point response format and one a five-point format. Scores range from 1 (“greater frequency”) to 5 or 8 (“never”), depending on the item. Religious and Spiritual Coping (α = 0.77) measures additional religious/spiritual practices and beliefs specifically related to coping with life’s problems. This subscale consists of seven items with a four-point response format, ranging from 1 (“a great deal”) to 4 (“not at all”). Organizational Religiousness (α = 0.79) measures the frequency of involvement in a formal public religious institution. This subscale consists of two items with a six-point response format, ranging from 1 (“more than once a week”) to 6 (“never”).

Depression symptoms

The Postpartum Depression Screening Scale (PDSS; Beck & Gable, 2002) is a self-report, 35-item scale, and was administered at the 12 and 15 months postpartum visits. Items are rated on a Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree) and the scale yields total scores ranging from 35–175, with a cutoff score greater than 80 suggesting major depressive disorder (total symptom count yielded Cronbach’s α = 0.953 at 12 months and α = 0.963 at 15 months).

Maternal quality of life

The Maternal Quality of Life (QOLI; Frisch, Cornell, Villanueva, & Retzlaff, 1992) is a modified, 9-item version of the Quality of Life Index, and was administered at the 12 and 15 months postpartum visits. Participants self-report their levels of satisfaction with various aspects of their life, including health, work and living arrangements, leisure time activities, romantic and extended familial relationships, neighborhood, and community. Each item was rated on a 5-point Likert scale ranging from 1 (very dissatisfied) to 5 (very satisfied), and these responses were summed to create a continuous quality of life total score. Reliability estimates at 12 and 15 months postpartum demonstrated high reliability (α = 0.802 and α = 0.825, respectively).

Data Analyses

Analyses were performed using SPSS version 24. Depression and total demographic risk were positively skewed. Given that these numbers are consistent with what would be expected, the numbers were not transformed. The current project examined associations between cumulative demographic risk, childhood trauma, religiousness/spirituality, and postpartum depression symptoms and quality of life. Bivariate Pearson correlation tests as well as linear regressions were conducted to examine relationships between the cumulative demographic risk, childhood trauma, the six BMMRS dimensions with acceptable internal consistency, and depression symptoms and quality of life ratings. Primary variable means, standard deviations and correlations are listed in Table 2.

Table 2.

Descriptive Statistics and Correlations Between Main Study Variables.

	Mean (SD)	1	2	3	4	5	6	7	8	9	10	11	13	14	15	16
1. Demographic Risk (n=108)	1.03 (1.33)	1
2. CTQ Total Score (n=108)	44.37 (18.98)	0.092	1
3. Daily Spiritual Experiences (n=108)	23.94 (8.26)	0.359**	−0.108	1
4. Forgiveness (n=107)	9.42 (2.45)	0.287**	−0.240*	0.745**	1
5. Private Religious Practices (n=108)	17.08 (7.86)	0.220*	0.066	0.774**	0.533**	1
6. Organizational Religiousness (n=105)	5.73 (2.94)	0.157	0.009	0.522**	0.423**	0.672**	1
7. Religious and Spiritual Coping (n=108)	21.13 (7.84)	0.238*	−0.116	0.758**	0.614**	0.657**	0.552**	1
8. Meaning (n=105)	5.92 (1.59)	0.091	0.004	0.753**	0.567**	0.612**	0.548**	0.652**	1
9. “Because of my religious or spiritual beliefs, I have forgiven myself for things that I have done wrong” (n=106)	3.07 (0.95)	0.208*	−0.262**	0.550**	0.816**	0.323**	0.180	0.429**	0.342**	1
10. “Because of my religious or spiritual beliefs, I have forgiven those who hurt me” (n=107)	3.11 (0.89)	0.219*	−0.266**	0.549**	0.813**	0.453**	0.301**	0.450**	0.317**	0.644**	1
11. “Because of my religious or spiritual beliefs, I know that God forgives me” (n=103)	3.4 (0.98)	0.228*	−0.092	0.698**	0.804**	0.521**	0.485**	0.541**	0.609**	0.486**	0.513**	1
13. MDD Symptom Count - 12 months postpartum (n=86)	63.05 (21.15)	0.115	0.328**	−0.001	−0.224*	0.049	0.066	−0.096	−0.004	−0.399**	−0.246*	−0.033	1
14. MDD Symptom Count - 15 months postpartum (n=94)	64.54 (24.87)	0.117	0.418**	−0.04	−0.242*	0.068	0.105	−0.144	−0.059	−0.310**	−0.217*	−0.113	0.709**	1
15. QOL Score - 12 months postpartum (n=86)	34.09 (5.84)	−0.230*	−0.533**	0.134	0.331**	0.043	−0.063	0.179	0.043	0.376**	0.374**	−0.012	−0.52**	−0.502**	1
16. QOL Score - 15 months postpartum (n=94)	34.52 (5.75)	−0.303**	−0.441**	0.111	0.294**	0.047	−0.017	0.193	0.111	0.301**	0.315**	0.053	−0.439**	−0.604**	0.723**	1

Note: The n are not consistent due to attrition from 6 to 15 months.

Results

Associations Between BMMRS Subscales and Postpartum Depression

Linear Regression models tested for associations between total CTQ symptoms, demographic risk (cumulative risk index: non-White minority, single parent, low education, low family income and young maternal age), and the six BMMRS subscales (daily spiritual experiences, forgiveness, private religious practices, organizational religiousness, religious and spiritual coping, and meaning), with depression severity at 12 and 15 months postpartum (Table 3). At 12 months, forgiveness and total CTQ symptoms were associated with symptoms of depression; specifically, individuals with higher total CTQ scores had more symptoms of depression, and those with higher levels of self-reported forgiveness had lower symptoms of depression. Analyses examining these associations at 15 months yielded similar results (Table 3).

Table 3.

Linear Regressions Between Risk Factors, BMMRS Subscales, Depression, and Quality of Life Among Postpartum Women.

Depression

Total Demographic Risk

Total CTQ

Forgiveness

Daily Spiritual Experiences

Private Religious Practices

Organizational Religiousness

Religious and Spiritual Coping

Meaning

Total R2

Δ R2

f

df

12 months

0.035 (0.547)

0.28* (0.312)

−0.402* (−3.389)

0.558 (1.37)

−0.13 (−0.334)

0.221 (1.569)

−0.215 (−0.995)

−0.121 (−1.53)

0.212

0.212

2.559*

8, 76

15 months

0.041 (0.75)

0.353** (0.446)

−0.337* (−3.755)

0.387 (1.136)

−0.001 (−0.003)

0.211 (1.771)

−0.086 (−0.478)

−0.181 (−2.845)

0.275

0.275

3.882**

8, 82

Quality of Life

Total Demographic Risk

Total CTQ

Forgiveness

Daily Spiritual Experiences

Private Religious Practices

Organizational Religiousness

Religious and Spiritual Coping

Meaning

Total R2

Δ R2

f

df

12 months

−0.217* (−0.914)

−0.445** (−0.134)

0.399** (0.906)

−0.133 (0.088)

0.168 (0.117)

−0.282* (−0.538)

0.032 (0.04)

−0.012 (−0.041)

0.431

0.431

7.201**

8, 76

15 months

−0.331** (−1.425)

−0.335** (−0.099)

0.364* (0.847)

−0.174 (−0.12)

0.078 (0.056)

−0.153 (−0.3)

0.13 (0.168)

−0.005 (−0.017)

0.361

0.361

5.795**

8, 82

Note: The df are not consistent due to attrition from 6 to 15 months.

Associations Between BMMRS Subscales and Postpartum Quality of Life

Regression models tested for associations between total CTQ symptoms, demographic risk, and the six BMMRS subscales (daily spiritual experiences, forgiveness, private religious practices, organizational religiousness, religious and spiritual coping, and meaning), with quality of life ratings at 12 and 15 month postpartum (Table 3). At 12 months, cumulative demographic risk, total CTQ score, forgiveness, and organizational religiousness were associated with quality of life; specifically, individuals with higher CTQ score, higher cumulative demographic risk, and more organizational religiousness had lower quality of life scores, whereas those with more forgiveness had higher quality of life scores. Results examining these associations at 15 months yielded slightly different results (Table 3). Total demographic risk, total CTQ score, and forgiveness were associated with quality of life scores; specifically, those with more cumulative demographic risk and higher CTQ score had lower quality of life, and those with more forgiveness reported higher quality of life. Organizational religiousness was an insignificant predictor in this model, though the direction of the relationship stayed the same.

Forgiveness Post Hoc Analysis

Due to the significant result, we gave unique attention to the forgiveness subscale, consisting of three questions: “Because of my religious or spiritual beliefs, I have forgiven myself for things that I have done wrong”, “I have forgiven those who hurt me”, and “I know that God forgives me”. Preliminary bivariate correlations demonstrated very consistent and statistically significant relationships between the forgiveness subscale with the demographic risk and childhood trauma predictors and maternal outcomes; therefore, we believed this relationship warranted further exploration.

Depression and forgiveness

Three separate regression models examined the relationships between demographic risk, CTQ scores, and the three aspects of forgiveness as rated by the three scale items (self-forgiveness, forgiveness for others, and forgiveness by God) with depression at 12 and 15 months postpartum.

The first regression revealed significant associations between total CTQ scores and self-forgiveness; specifically, those with higher CTQ scores reported more depression symptoms, and those with higher levels of self-forgiveness reported less symptoms (Table 3). The second and third regression models examining the relationship between demographic risk, CTQ scores, and forgiveness for others or forgiveness by God with depression at 12 and 15 months postpartum, showed that only CTQ scores were a significant positive predictor (Table 4).

Table 4.

Linear Regressions between risk factors BMMRS forgiveness questions, Depression, and Quality of Life among Postpartum Women.

Depression	Total Demographic Risk	Total CTQ	Forgiven Myself	Total R2	Δ R2	f	df
12 months	0.14 (2.14)	0.22* (0.24)	−0.36** (−8.25)	0.23	0.23	8.20**	3, 81
15 months	0.12 (2.17)	0.35** (0.46)	−0.23* (−6.18)	0.24	0.24	9.08**	3, 88
Depression	Total Demographic Risk	Total CTQ	Forgiven Others	Total R2	Δ R2	f	df
12 months	0.11 (1.73)	0.25* (0.28)	−0.19 (−4.29)	0.14	0.14	4.52**	3, 82
15 months	0.11 (2.16)	0.37** (0.48)	−0.14 (−3.87)	0.2	0.2	7.35**	3, 89
Depression	Total Demographic Risk	Total CTQ	Forgiven by God	Total R2	Δ R2	f	df
12 months	0.07 (1.07)	0.34** (0.38)	−0.02 (−0.40)	0.13	0.13	3.73*	3, 78
15 months	0.11 (1.99)	0.42** (0.54)	−0.096 (−2.45)	0.2	0.2	7.32**	3, 86
Quality of Life	Total Demographic Risk	Total CTQ	Forgiven Myself	Total R2	Δ R2	f	df
12 months	−0.24** (−1.02)	−0.39** (−0.12)	0.31** (1.89)	0.38	0.38	16.39**	3, 81
15 months	−0.33** (−1.41)	−0.33** (−0.10)	0.26** (1.59)	0.32	0.32	14.07**	3, 88
Quality of Life	Total Demographic Risk	Total CTQ	Forgiven Others	Total R2	Δ R2	f	df
12 months	−0.22* (−0.95)	−0.41** (−0.13)	0.29** (1.81)	0.38	0.38	16.525**	3, 82
15 months	−0.34** (−1.46)	−0.33** (−0.10)	0.29** (1.89)	0.34	0.34	15.265**	3, 89
Quality of Life	Total Demographic Risk	Total CTQ	Forgiven by God	Total R2	Δ R2	f	df
12 months	−0.17 (−0.74)	−0.50** (−0.15)	−0.02 (−0.09)	0.31	0.31	11.47**	3, 78
15 months	−0.31** (−1.32)	−0.40** (−0.12)	0.08 (0.47)	0.27	0.27	10.83**	3, 86

Note.

^*p<.05

^**p<.01.

Standardized Betas are reported with Unstandardized Betas in parenthesis. Df are not consistent due to attrition, and missing data on each forgiveness question.

Quality of life and forgiveness

Similarly, three separate regression models examined the relationships between quality of life at 12 and 15 months postpartum with: demographic risk; CTQ scores; and self-forgiveness, forgiveness for others and forgiveness by God.

When examining the relationships between demographic risk, CTQ scores, and self-forgiveness or forgiveness by others and quality of life outcomes, all predictor variables demonstrated significance at 12 and 15 months. Specifically, those with more cumulative demographic risk and higher CTQ scores reported worse quality of life, and those with more self-forgiveness or more forgiveness for others demonstrated higher quality of life (Table 4).

In contrast, the regression model examining the predictive significance of forgiveness by God did not yield a significant result at 12 and 15 months postpartum for forgiveness by God. At 12 months, only total CTQ score were significant negative predictors of quality of life, and at 15 months postpartum both CTQ score and cumulative demographic risk significantly predicted worse quality of life (Table 4).

Discussion

This study aimed to shed light on specific domains of religiosity and spirituality (R/S) that may be resiliency factors for positive postpartum adjustment defined as low depression and high quality of life in mothers oversampled for childhood abuse histories. We tested several domains of R/S ranging from more organized and structured aspects such as organizational religiousness to more personal aspects such as meaning and forgiveness. We found that, in accordance with our hypothesis, for our sample of postpartum women with childhood abuse histories the personal aspects of self-forgiveness and forgiveness for others were most relevant as resiliency factors predicting better postpartum mental health and quality of life even when controlling for other risk (trauma and demographics). Surprisingly, contrary to our research hypothesis, no other R/S domain had consistent associations with postpartum mental health or quality of life, with the exception of a significant negative association for organizational religiousness with quality of life at 12 months postpartum.

Our main finding, aligning with our hypothesis, that specifically forgiveness was associated with better adaptation (lower depression and more quality of life) across the postpartum period, is consistent with a number of prior studies outside of peripartum showing that forgiveness is associated with lower levels of depression and overall better mental health (for a review, see Toussaint & Webb, 2005). With regards to the postpartum period, our study supports a recent finding showing that trait forgiveness, which is the dispositional motivation towards forgiveness, is inversely associated with postpartum depression (Ripley et al., 2018). It could be speculated that the ability to forgive is an important virtue associated with mental wellbeing, and protects even in the face of demographic risk and adverse childhood experiences.

The affirmative role played by forgiveness is in line with prior evidence showing that forgiveness in general may buffer negative psychological effects of trauma (Ochu, Davis, Magyar-Russell, O’Grady, & Aten, 2018). For example, a study with military veterans reported that when controlling for demographic risk factors, combat exposure, and PTSD symptom severity, forgiveness fully mediated the link between overall spirituality and quality of life (Currier, Drescher, Holland, Lisman, & Foy, 2016). The practice of forgiveness may be a means by which individuals resist harmful negative emotions that can result from stress exposure or trauma history (Toussaint, Williams, Musick, & Everson, 2001), especially man-made trauma (Worthington & Sandage, 2016). It seems that forgiveness may help an individual to let go of negative emotions and thoughts associated with emotional distress (Hirsch, Webb, & Jeglic, 2011) and that forgiveness could be considered a unique coping strategy that can be helpful in either religious or non-religious contexts (Johnstone, Yoon, Franklin, Schopp, & Hinkebein, 2009).

Although it appears that forgiveness in general may be an important virtue, it is not a homogenous construct but may rather be compromised of different elements that should be studied separately (Currier, Drescher, Holland, Lisman, & Foy, 2016), as was exemplified in our post-hoc analyses. In the current research, forgiveness for the self was associated with less symptoms of depression and a better quality of life while forgiveness of others was positively associated with quality of life only, and forgiveness from God was not associated with either. These differentiated findings are in line with prior studies showing inconsistent associations for different aspects of forgiveness. For example, a study in older adults reported that forgiveness for others yielded greater psychological wellbeing than forgiveness by God (Krause & Ellison, 2003). Others reported sex differences, such that among women, forgiveness for others, self-forgiveness, and forgiveness by God were all associated with decreased odds of developing major depression, but this was not true for men where only self-forgiveness protected against depression (Toussaint, Williams, Musick, & Everson-Rose, 2008a). Lastly, in another study, self-forgiveness and forgiveness for others inversely associated with suicidality, whereas forgiveness by God was unrelated (Hirsch, Webb, & Jeglic, 2011). This present study is unique as we explored differential associations of forgiveness domains with outcomes in postpartum women.

In our study, postpartum women had been oversampled for their childhood abuse histories; thus, for this group of women, self-forgiveness and forgiveness for others may have special salience as resiliency factor. Forgiveness for others involves forgiving another person for having done harm. Self-forgiveness involves release of negative affect and self-blame associated with past wrongdoings, mistakes, or regrets (Toussaint, Williams, Musick, & Everson-Rose, 2008b). In the context of having experienced childhood abuse, entering birth and postpartum period may trigger memories of the past trauma and associated feelings for oneself and toward others, including the perpetrator(s). In contrast, forgiveness by God refers to the belief or perception that one’s transgressions are forgiven by the divine (Toussaint, Williams, Musick, & Everson-Rose, 2008b), which may be more associated with one’s sins and less with one’s trauma experiences and memories.

In particular, individuals may use self-forgiveness as a strategy to safeguard their overall well-being against the toxic effects of self-conscious thoughts and emotions, such as, guilt, shame, and regret (McConnell, 2015) which may be a useful mechanism for healing from trauma. Further, self-forgiveness is often inversely associated with shame (Rangganadhan & Todorov, 2010), an emotion we have found in prior work salient to postpartum adaptation among women with childhood histories of maltreatment (Menke, Morelen, Simon, Rosenblum, & Muzik, 2018). During the postpartum period women often think about their emerging identity of becoming a parent. This results in them thinking about their own history of being parented, and individuals with histories of childhood maltreatment have increased feelings of shame (Wright, Fopma-Loy, & Oberle, 2012).

To our surprise, and contrary to our hypotheses, most of the R/S sub-scales were not significantly associated with mental health or quality of life. In fact, organizational religiousness which is the frequency of involvement in a formal public religious institution, even had a negative effect on quality of life at 12 months postpartum. This finding was in contrast to prior work in non-perinatal (e.g. AbdAleati, Mohd Zaharim, & Mydin, 2016; Koenig, 2009; Paloutzian & Park, 2013) and perinatal populations (Cheadle et al., 2015; Cheadle & Schetter, 2018; Clements, Fletcher, Childress, Montgomery, & Bailey, 2016; Mann, McKeown, Bacon, Vesselinov, & Bush, 2008). However, our result may reflect the special sample population of women exposed to childhood abuse. We can speculate that the possible negative effect of organizational religiousness on quality of life at 12 months postpartum maybe attributed to the fact that the involvement with religious organizations and institutions could sometimes be harmful. Some research suggests that stressors which raise a person’s inner ambivalence or conflict with the values embodied by religious organizations, such as the value of family cohesiveness, may be hard to confront by some individuals, and thus for these individuals the involvement with religious organizations and communities, particularly in the case of trauma perpetrated by family members, may sometimes even elevate stress and negatively impact mental health (Strawbridge, Shema, Cohen, Roberts & Kaplan, 1998). This may have also been the case for women in our sample. We argue that our findings suggest to study specific R/S aspects in the context of specific populations and conditions instead of studying R/S as a general marker of “good or bad” coping. Such would be an over-simplified understanding of the role of R/S (Hall, Meador, & Koenig, 2008).

Previously, the connection between R/S and wellbeing has been studied either in other trauma populations or in non-trauma exposed perinatal samples. For example, in a study with female veterans, frequency of religious service attendance buffered the negative impact of sexual assault on mental health and depression (Chang, Skinner, & Boehmer, 2001). Similarly, in Latina women religious attendance was found to buffer risk when examining the impact of childhood physical and sexual abuse histories on lifetime prevalence of substance use disorders (Ai & Lee, 2018). However, the data on R/S and mental wellness are mixed. For example, a study in China showed no associations between religiosity and mental health (Shiah, Chang, Chiang, Lin, & Tam, 2015), and in another study of pregnant women religious coping was less important to mental health than locus of control (Puente, Morales, & Monge, 2015). When testing the role of religiosity more mechanistically, the results are also less convincing. For example, in a sample of high-risk women (with ongoing trauma and suicidality), religious wellbeing (defined as a sense of meaning, life purpose and relationship with God) did not mediate the association between childhood maltreatment and current PTSD symptoms (Zhang et al., 2015). The mixed results for R/S on mental health could be explained by variations in construct measurement as different studies use different R/S scales, possibly tapping into diverse aspects of R/S. Furthermore, different cultures and populations could put differential value and importance on R/S occluding the effect. Further research is clearly needed to disentangle the associations between R/S and mental health in special populations.

Lastly, our results showing positive effects for forgiveness rather than for broader aspect of religiosity on postpartum depression and life quality, must also be seen from the perspective that religiosity in general and forgiveness have great overlap (Krause, 2018). In fact, religiosity may serve as a frame for the transmission of virtues like forgiveness. For example, greater religious involvement is also correlated with greater self-forgiveness (Davis, Worthington, Hook, & Hill, 2013; Krause, 2015). Overall, forgiveness is a major target in the teachings of all major world religions (Farhadian & Emmons, 2010).

Limitations

Our study is not without limitations. Despite the longitudinal design predicting postpartum mental health outcomes at 12 and 15 months postpartum from 6 month R/S, the study cannot establish causality. Women may have been mentally unhealthy already prior to the postpartum period, which may have contributed to postpartum symptoms. Moreover, preexisting mental illness may have jeopardized R/S as individuals who are psychologically unhealthy may be in a worse position to forgive (Tucker, Bitman, Wade, & Cornish, 2015). Only longitudinal designs following mental health and R/S over time could address such causal links. Furthermore, the study population was very narrowly defined—oversampled for childhood abuse in general in the postpartum period. While this may be limiting generalizability, we believe this is actually a strength and adds to the literature as it allowed us to make inferences about this understudied population. Future research could be conducted as to the association of R/S and depression and quality of life in the context of specific traumas. Also, the subsample with completed R/S data was predominantly Caucasian (70%), and thus the results may not be representative of African American women’s spirituality or other racial and ethnic minorities. Further, although forgiveness is a virtue addressed in the teachings of all major world religions (Farhadian & Emmons, 2010), the study population was predominantly Christian, and the results may be mainly generalizable for populations with Christian denomination. Another limitation concerns the use of self-report measures and the problem of common method bias. We believe that the constructs measured in this study have a demonstrated acceptable construct validity with no overlap between measures, therefore addressing this possible bias (Conway & Lance, 2010). A final limitation concerns the forgiveness measure used for this study. We used a forgiveness subscale of a well-researched R/S measure, and not a designated measure for forgiveness. Future research should use more fine-tuned forgiveness measures.