Pregnant women’s perceived unhappiness with current pregnancy: an analysis of prevalence, reasons, and associated factors in Quelimane district, Mozambique

Abstract

Background

Pregnant women’s emotional well-being is a major determinant of maternal and pregnancy outcomes. This study aimed to estimate the prevalence of pregnant women’s perceived unhappiness with their current pregnancy, and to examine the reasons and factors associated with such feelings in Quelimane district, Central Mozambique.

Methods

A population-based pregnancy surveillance study was conducted during 2023–2024. Data included pregnant women’s perceived status of unhappiness with their current pregnancy at realization, reasons for such feelings, and perceived reactions of their sexual partners. Variables associated with women’s unhappiness were investigated within a couple-centered conceptual framework that views pregnancy-related unhappiness as the result of interacting socio-demographic, reproductive, and partner-level factors; using descriptive statistics and multivariate logistic regression (three sequentially adjusted models).

Results

Among the 1546 participating women, 24.5% referred being unhappy with their pregnancy, and this was higher among adolescents (42.2%), in daughters/grand-daughters/daughters-sisters-in-law (33.5%), residing in larger households (29.7%), un-married women (36.6%), first trimester of gestation (37.7%), without couple’s communication about childbirth (48.9%), and with unhappy partner (74.5%). The leading reasons for unhappiness were unplanned pregnancy (30.2%), unwanted/undesired pregnancy (26.5%), and fear of consequences (17.7%). In the multivariate logistic regression analysis, the associations of socio-demographic (first model) and reproductive (second model) variables with unhappiness were progressively attenuated as the models were being fitted. In the fully adjusted model (third model), women’s age and gestational age maintained strong association with unhappiness, with adolescents and older women showing higher odds of unhappiness compared to middle-aged adults (OR = 2.18, 95% CI:1.43–3.32) and (OR = 2.21, 95% CI:1.25–3.93), respectively; and women interviewed in the first and third trimester being more likely to be unhappy than those in the second trimester (OR = 2.19, 95% CI:1.10–4.37) and (OR = 1.51, 95% CI:1.08–2.11), respectively. However, partner-related variables emerged as strongly associated with maternal unhappiness, with partner’s unhappiness with pregnancy showing ORs of 8.23 (95% CI:5.04–13.43), lack of couple’s communication (OR = 2.23, 95% CI:1.34–3.74), and partner’s pregnancy intendedness showing ORs of 1.73 (95% CI:1.12–2.65).

Conclusions

Nearly one in four women in Quelimane reported being unhappy with their pregnancy, and unhappiness was particularly frequent among adolescents, non-married, women in first gestational age, those without couple’s communication, and with unhappy partners. While women’s socio-demographic and reproductive factors contributed to unhappiness, partner’s emotional response and couple communication were the most influential factors. These findings highlight the importance of couple-centered approaches in maternal health research and interventions, particularly for adolescents and other vulnerable groups.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1186/s12884-026-09682-1.

Introduction

The overall level of women’s happiness during pregnancy plays an important role in maternal and child health. Women experiencing emotional distress during pregnancy are at increased risk of adverse maternal and perinatal outcomes [1, 2]. The concept of happiness usually refers to an emotional state of joy, satisfaction, contentment, and fulfillment [3]. It refers to how a person reacts psychologically within several ideologies and dimensions of well-being or life satisfaction, leading to positive attitudes both at individual and at broader social welfare [1, 2]. The relationship between happiness and quality of life has been widely documented [1, 4, 5] Unhappiness or low level of happiness, particularly among pregnant women, can lead to adverse maternal and pregnancy outcomes, including neonatal and infant death, and reduced prenatal care service attendance [4, 5]. Studies have reported association between pregnancy unhappiness with antenatal and postpartum depression [6]. However, most existing studies on happiness in pregnant women investigated the general happiness of women during pregnancy, looking at fears, anxiety, distress, and social support [4], but little is known about women’s happiness or unhappiness specifically regarding the fact of being pregnant. Some of these studies used standardized instruments designed to measure general happiness, such as the Oxford Happiness Questionnaire [7], the Satisfaction With Life Scale [8], the Cantril Self-Anchoring Ladder [9], and the Warwick–Edinburgh Mental Well-being Scale [10]. These tools are similar in that they ask respondents to rate a number of statements using scales, and the sum of these response yield numbers where the lowest values represent lowest satisfaction and the highest values, highest satisfaction [8]. However, these scales do not measure happiness or unhappiness specifically in relation to pregnancy, and some recent studies appear to favour measuring unhappiness rather than happiness [11–13].

This knowledge gap is particularly relevant in low- and middle-income countries, where unintended pregnancies, early childbearing, adverse maternal and perinatal outcomes, and limited access to reproductive health services remain prevalent. In sub-Saharan Africa, few studies have explicitly examined women’s perceptions of happiness or unhappiness with pregnancy, and have reported higher levels of unhappiness during pregnancy - for example, 24% in Ethiopia [5]. Other studies focused on adolescent pregnancies and investigated parental perceptions of adolescent pregnancies [14], but to date, scarce data are available on how the pregnant women perceive their pregnancy. In Southern Africa (where Mozambique is located), two studies (one qualitative [15], and one quantitative [16]) on teenagers’ perceptions of their pregnancy in rural South Africa reported higher levels of unhappiness in adolescent girls and their parents [15]. In Mozambique, no studies on women’s happiness related to pregnancy were found. Among the few studies addressing similar topics, a qualitative study on women’s quest to become pregnant in Southern Mozambique highlighted the depression challenges that childless women experience in their marriages and in the society, including disrespect by husbands and in-laws [17]. Another qualitative study on women’s perceptions and decision-making with regard to pregnancy among HIV positive women in Southern rural Mozambique highlighted the role of fertility gender norms and duties, and lack of contraception knowledge on overriding medical recommendations before getting pregnant [18]. Inadequate use of contraceptive methods and gender norms were also reported to lead adolescent girls into unintended pregnancies throughout the country [19].

Building on this evidence, this study adopts a couple-centered conceptual framework that views pregnancy-related unhappiness as the result of interacting individual-level socio-demographic and reproductive factors, and couple-level factors. Distal socio-demographic characteristics (e.g., age, education, and household context) influence reproductive histories and pregnancy intentions, which in turn are shaped by proximal couple dynamics, including partner emotional responses, communication, and acceptance of pregnancy. This framework is informed by the World Health Organization reproductive health framework that emphasizes that fertility and pregnancy experiences are shaped within partner and social contexts [20, 21], and by theories of dyadic coping, which highlight how emotional responses to stressors are shaped within couple relationships [22–24]. Empirical evidence further shows that partner support, communication, and concordance in pregnancy intention are key determinants of maternal emotional well-being during pregnancy [24, 25].

The aim of this study was to assess pregnant women’s perceived unhappiness related to their pregnancy, at pregnancy realization, and to examine the reasons and factors associated with such feelings in rural and urban areas of Quelimane district, Central Mozambique. The results of this assessment contribute to a more comprehensive understanding of pregnancy-related emotional experiences in Quelimane district and similar settings.

Methods

Study design and setting

This study was conducted in Quelimane district, Zambézia province, in the Central region of Mozambique, by the Manhiça Health Research Centre (Centro de Investigação em Saúde de Manhiça - CISM) between June 2023 and May 2024. Since 2018, CISM is running the Child Health and Mortality Prevention Surveillance (CHAMPS) study in Quelimane, a multi-country study that aims to ascertain the causes of death in stillbirths and children under the age of 5 years in seven countries of sub-Saharan Africa and South Asia, using post-mortem minimally invasive tissue sampling technique (MITS) [26]. In 2022, CISM established a Health and Demographic Surveillance System (HDSS) in the district, and in 2023 started the Pregnancy Surveillance study (PregSurv), a population-based surveillance of pregnancies and their outcomes that aims to guide maternal and child health interventions in CHAMPS study areas [26]. Quelimane HDSS covers urban and rural areas of Quelimane district, the capital of Zambézia Province, and is part of CISM’s strategic efforts to expand its geographical scope to include provinces that have some of the country’s poorest health indicators. Zambézia is the second most populated province, with 5 164 732 inhabitants, after Nampula, with 5 758 920 people [27]; with an under-five mortality rate of 109 per 1000 live births, followed by Sofala (89), Nampula (85) [28]. Further, Zambézia has the second highest percentage of unschooled women aged 15–49 years (38.8%), after the northern province of Niassa (40.1%); the lowest prevalence of modern contraceptive use among married/cohabiting women (11.2%), followed by Nampula (13.1%) and Cabo Delgado (13.7%); and the highest prevalence of unmet need in married/cohabiting (33%), followed by Nampula (32%) and Niassa (30%) [29]. The district of Quelimane comprises the city of Quelimane (the largest city in the province) which is the capital of Zambézia province. Thus, although this district includes the surrounding rural areas, it is the most socio-economic privileged compared to the rest of 21 districts. For example, apart from hosting all the top provincial administrative and political authorities with their corresponding district-level authorities, Quelimane district has the largest number of health facilities (24 health facilities), including the Quelimane Central Hospital and Quelimane General hospital. It is also the only district that hosts 5 Universities in the province. Further, Quelimane has experienced 4 cyclones and floodings just before and during the study period (Cyclone Eloise, in January 2021; Ana, in January 2022; Freddy, in March 2023; and Filipo, in March 2024) [30]. These cyclic extreme weather events may have affected participants and their households in different ways.

PregSurv uses standardised questionnaires across the CHAMPS Network to collect quantitative data on maternal health problems and treatments during current pregnancy, access to antenatal, perinatal, and post-partum health care, and pregnancy outcomes, until Day 42 after delivery. The questions include health problems and adverse pregnancy outcomes in the pregnancies preceding the current one. In addition to these questions, CISM added site-specific questions to measure women’s perceptions of unhappiness with these pregnancies, which included women’s socio-demographic background, family planning, pregnancy intentions, couple’s communication about childbirth, unhappiness with current pregnancy, and reasons for such feelings. Similar partner’s data were collected from women, not by asking the men themselves (see data collection Forms in Additional files 1 and 2). Although the data used in this article were collected through this longitudinal surveillance, the study design in this article is of a cross-sectional quantitative survey.

Study population and data collection

PregSurv study uses the HDSS platform to identify pregnant or recently delivered women using a team of female interviewers who approach potential participants at home or at health facilities for enrolment, after signing informed consent. PregSurv was designed to enrol all the women registered by the HDSS as pregnant, even if PregSurv team finds them after delivery. Thus, no sampling was used, and by the time of this analysis (June 2024), 4170 pregnant women had been registered by the HDSS since the beginning of baseline census in January 2022. When PregSurv started in June 2023, the HDSS area had been reduced (after baseline), which excluded, from follow-up, 1374 previously registered women. Of the remaining 2796 women, 2737 were recruited for PregSurv, and 59 were lost to follow-up without registration of motives. Of the 2737 women recruited, 398 were excluded from the analysis because their IDs could not be matched in the HDSS database (and were sent for data cleaning); 742 because they were enrolled after delivery; and 51 because had unknown unhappiness status. Thus, the final sample size was 1546 women enrolled while pregnant, and this was deemed sufficient to estimate a proportion of 50% unhappy women as the main outcome, with a margin of error of 5% and confidence level of 95% [31]. The study population for this analysis were pregnant women aged 12–49 years who were resident members of the HDSS area for at least three months, or had entered the area recently with intention to live for three months or more. The lower-limit age was reduced from the commonly used 15 years to 12 in order to include 12 pregnant girls under the age of 15 years. Participation of pregnant girls under the age of 18 required their informed assent after the consent of their parents or legal guardians. PregSurv was approved by the CISM’s Institutional Scientific Committee, CISM’s Institutional Bioethics Committee, and the National Bioethics Committee for Health. The HDSS was approved by the CISM’s Institutional Scientific Committee and the Institutional Bioethics Committee. The reference numbers for these approvals are presented in the ethics approval and consent to participate section.

The key question to measure perceived unhappiness in PregSurv was “When you noticed this pregnancy how did you feel? Have you perceived this pregnancy to be a happy event or a sad event in your life?”. The responses’ options were “It was a happy event”, “It was not a happy event”, and “Do not know”. This question was repeated to capture the sexual partner’s perception this way: “What about your sexual partner, how did he perceive this pregnancy, to be a happy event or not?”. PregSurv was not designed to enroll sexual partners, thus, all partner’s information were obtained from women’s report. Furthermore, women were asked whether the current pregnancy had been intended or not (same question for the partner); whether she was using any modern contraceptive at pregnancy; whether the partner had intended this pregnancy, and had accepted the pregnancy at pregnancy realization or not; and whether she felt any abandonment by her partner or family members. In addition, they were asked whether they received support from partners and from family members (both economically and emotional/psychologically). Reasons for specific responses were recorded in open-ended questions.

The data were collected using REDCap [32] on Android-based Tablets, and a team of field supervisors, social scientists, demographers and data managers guaranteed the quality of the data. Data analysis was performed using Stata 14.2 software [33].

Methods and measures of data analysis

The women were interviewed at enrolment and all the questions refer to the time she and her partner or family members noticed this current pregnancy. The focus on women’s feelings at pregnancy realization was to minimize recall bias and the effect of the time-changing nature of unhappiness. To minimize the confusion between unhappiness related to other reasons and unhappiness with regards to pregnancy, the questions were framed to insist that the respondent should focus on her being pregnant.

The analyses included descriptive statistics of women’s perceived unhappiness and motives for such feelings in relation to their current pregnancies at the time they noticed the pregnancy. These proportions were compared by key categories such as age groups, marital status, gravidity (number of previous pregnancies that a woman has had), education, place of residence (rural vs. urban), and past adverse pregnancy outcomes. The variable “previous adverse pregnancy outcomes” included any of the following: abortions, stillbirths, pre-term births, and neonatal deaths that the woman experienced in any of the pregnancies preceding the current one. The questionnaire had open-ended questions, mostly those asking why the respondent gave a certain response to a previous question, and spaces to write details to specify the “other” options in closed-ended questions. When asking open-ended questions, the interviewers were instructed to prompt respondents with “anything else?” to capture additional information until exhaustively stated, in order to explore more the information and capture maximum response variability. These text responses were analysed using quantitative content analysis, which permitted assigning lengthy text responses into pre-defined response-options, and to create new ones [34]. A similar approach was used in a quantitative study on perceptions and compliance with anti-COVID-19 measures in Mozambique [35]. All the response categories of the open-ended questions were created at data analysis stage.

The women were divided into four age groups: adolescent (aged 12–19 years), young adults (20–24), older adults (25–34) and the oldest (35–49), to examine unhappiness in adolescents and young mothers as vulnerable population groups with regards to pregnancy and childbirth.

Bivariate and multivariate logistic regression analysis were used to examine the association between explanatory variables and pregnant women’s perceived unhappiness - the measures of association were Odds Ratios (OR) of a woman being unhappy with her pregnancy. The outcome was coded 1 for unhappy and 0 for happy. In 12 households there were 2 women recruited, which violated the assumption of independence between observations required for standard logistic regression models. Thus, at the regression analysis, 1 woman from these cases was randomly selected to remain for the analysis. Given the very small number of such cases relative to the total sample (12 out of 1546), and the random selection applied, it is unlikely that there were systematic differences between the included and excluded women to materially affect the results in the regression model. The grouping of explanatory variables in the regression models was guided by the study’s conceptual framework. Within this framework (visualized in Fig. 1), the variables were grouped into three domains: (i) socio-demographic characteristics, (ii) women’s reproductive and pregnancy-related characteristics, and (iii) partner-related factors. Thus, three multivariate logistic regression models were fitted sequentially (nested) to reflect these domains, allowing the assessment of how the associations between background characteristics and pregnancy-related unhappiness are mediated or attenuated by reproductive and partner-level factors. Following Fig. 1, the first model (socio-demographic characteristics) included age of women, relation to household head, household size, level of education, and whether the woman was studying or not. The second model included the variables in Model I and women’s reproductive and pregnancy-related characteristics, namely, marital status, gestational age at interview, history of past adverse outcome, and knowledge of modern contraceptives (defined as knowledge or awareness of at least one modern method). The third model, included the variables in Models I and II, and added partner-related factors, namely, couple’s communication about childbirth, partner’s pregnancy intendedness, partner’s unhappiness, and partner’s rejection of pregnancy.

Fig. 1.

Conceptual framework of factors associated with women’s unhappiness with current pregnancy

Note on Figure 1: Conceptual framework developed by authors based on the World Health Organization reproductive health framework and by theories of couples dyadic coping [20–24]

Women with no previous pregnancy history were originally coded as missing on the adverse pregnancy history variable because this question was not applicable to them. Thus, to avoid unnecessary exclusion of these respondents in the regression analysis, the questions on previous pregnancy history and on adverse pregnancy outcomes were combined into one variable with three categories. The new variable was “had adverse pregnancy outcomes”, and the options were: (i) had previous pregnancy without adverse outcome; (ii) had no previous pregnancy; (iii) had previous pregnancy with adverse outcome - to account for the survey skip pattern and retain women with no previous pregnancy in the analysis. Both the bivariate and multivariate regressions were conducted using the same complete-case sample of women obtained from multivariate Model III.

Family planning variables and contraceptive methods were defined according to Demographic and Health Surveys (DHS) [36]. The DHS defines intended pregnancy as one that is consciously and clearly desired by the woman and/or her partner at the time of conception. We added the question “Were you using family planning at pregnancy?” to simultaneously measure: (i) use in month(s) prior to pregnancy but stopped, and (ii) irregular or improper usage. This, together with the question “what do you think might have happened, for you to be pregnant while in contraceptives?” helped to better understand the responses on use of contraceptives. The focus was in women’s behavior, rather than contraceptive failures. However, three terms (“unplanned pregnancy”, “unwanted/undesired pregnancy”, and “short birth interval”), together with “wanted no more children”, “fear of consequences”, and “dubious partner” emerged at analysis stage from the open-ended question “why were you unhappy?”. Although one may intuitively link these three terms to their standard meaning in the demographic literature, we did not have ways to check how they reflect or refer to these standard meanings.

The relevance of background variables such as rural-urban place of residence, living arrangements, education, and religion has been extensively demonstrated in the literature on fertility determinants [37, 38] and on marriage dynamics [39–42]. For example, living with parents/in-laws, in rural area, and in extended family is often associated with higher fertility intentions or early childbearing, while, independent living arrangements, urban residency, and co-residence with same-generation relatives tend to correlate with lower fertility rates, with household socio-economic status mediating these effects [43–45]. With regards to marriage, the assortative mating theory (one of the most prominent theories of marriages) states that people seek to marry individuals of similar traits (race, religion, socio-economic status, education) [46, 47] and/or those who can complement them (marrying similar or better off partners) [41, 46]. The economic theory of marriage (another prominent theory) posits that people marry to reduce the costs of living apart and to maximize benefits of living together (reduce housing costs, increase the benefits of love and sexual relations, care, rearing of children, and others) [40]. The socio-economic variables pointed in these theories, together with age, are key for the understanding of most differences in fertility intentions and how people perceive pregnancy events in own or others’ life course. For example, studies have shown that unintended pregnancy is associated with factors such as age, parity, inconsistent and incorrect use of contraception, marital status, partner preferences, poverty, support from family, and access to reproductive health services such as contraception and abortion [24, 48].

All the explanatory variables were used in the logistic regression as binary or categorical variables (nominal). Potential multicollinearity among associated factors was assessed using variance inflation factors (VIF), with all VIF values below 2.2 (mean = 1.5). However, the variable “woman’s pregnancy intention” was excluded from the regression models due to its conceptual overlap with the outcome and evidence of quasi-complete separation, which produced extremely large and unstable odds ratio when included in preliminary models (OR = 96.96, 95% CI: 61.58–152.67).

Results

Socio-demographic characteristics of pregnant women and prevalence of unhappiness

This study included 1546 pregnant women enrolled to PregSurv while still pregnant. Participants’ socio-demographic profile, reproductive variables and knowledge of modern contraceptives are presented in Table 1, including the prevalence of unhappiness and Pearson X² test p-value for comparing the prevalence of unhappiness between categories of women. Overall, 24.5% (n = 378) of women reported being unhappy with their current pregnancy.

Table 1.

Characteristics of pregnant women and prevalence of unhappiness with current pregnancy

Selected variables		Total of women (N, %)		Prevalence of unhappiness (n, %)		X2 P-value*	Selected variables		Total of women (N, %)		Prevalence of unhappiness (n, %)		X2 P-value*
Total (%) PANEL A		1546	100.0	378	24.5		PANEL B
Socio demographic characteristics							Women’s reproductive variables and pregnancy intentions
Age groups (in years)	12–19	334	21.6	141	42.2	0.000	Gestational age at interview	1 st trimester	75	4.9	25	33.3	0.036
	20–24	556	36.0	95	17.1			2nd trimester	490	31.7	104	21.2
	25–34	541	35.0	105	19.4			3rd trimester	889	57.5	228	25.6
	35–49	115	7.4	37	32.2			Missing #	92	6.0	-	- #
Place of residence	Urban	1351	87.4	330	24.4	0.954	Had previous pregnancies	Yes	1246	80.6	264	21.2	0.000
	Rural	195	12.6	48	24.6			No	300	19.4	114	38.0
Relation to head of household	Head	472	30.5	92	19.5	0.000	Had adverse pregnancy outcomes (N = 1246)	Yes	474	38.0	83	17.5	0.013
	Wife/mother	457	29.6	90	19.7			No	772	62.0	181	23.4
	Daught/grand	379	24.5	128	33.8		Know modern Contraceptives	Yes	1405	90.9	329	23.4	0.003
	Daughter-law	162	10.5	53	32.7			No	141	9.1	49	34.8
	Unrelated	76	4.9	15	19.7		Was pregnancy intentional	Intentional	1031	66.7	24	2.3	0.000
Household size	Below median	692	44.8	149	21.6	0.001		Unintentional	515	33.3	354	68.7
	Median size	271	17.5	56	20.7		Timing of unintended pregnancy (n = 515)	Wrong time	325	63.1	204	62.8	0.000
	Above median	583	37.7	173	29.7			Shouldn’t occur	175	34.0	148	84.6
Education of women	Primary/none	179	11.6	46	25.7	0.000		Don’t know	15	2.9	2	13.3
	Secondary	1,016	65.7	277	27.3		Why were you happy (N = 1,168)	Wanted more children	289	24.7	Not applicable. These women are happy. This question applies to happy women only.
	Higher	351	22.7	55	15.7			Blessing/good surprise	333	28.5
Religion of women	Catholic	902	58.3	223	24.7	0.558		Satisfy spouses/family	64	5.5
	Protest/Evan	206	13.3	45	21.8			Desperate for children	46	3.9
	Islamic/Hind	350	22.6	84	24.0			Wanted a child	398	34.1
	None	88	5.7	26	29.5			Don’t know	38	3.3
Woman marital status	Single/never	287	18.6	110	38.3	0.000	Why were you unhappy (N = 378)	Unplanned	114	30.2	The values on the left refer to unhappy women. This question applies to unhappy women only.
	Married/cohb	1,237	80.0	260	21.0			Want no + childr	21	5.6
	Divorced/sep	22	1.4	8	36.4			Short birth interva	31	8.2
								Fear consequence	67	17.7
								Unwantd/undesir	100	26.5
								Don’t know	45	11.9

In the “Relation to head” variable, the option “Daught/grand” stands for “Daughter and granddaughter”. Option “Daughter-law” stands for “Daughter-in-law”

In variable “Religion of women”, option “Protest/Evan” stands for “Protestant and Evangelic”; and option “Islamic/Hind” stands for “Islamic and Hindu”

In variable “Women marital status”, option “Single/never” stands for “Single, never married”; option “Married/cohb” stands for ““Married or cohabiting”; option “Divorced/sep” stands for “Divorced or separated”

In variable “Why were you unhappy”, “Want no + childr” stands for “Want no more children”, option “Short birth interva” stands for “Got pregnant while she had a too young baby”, which we termed “short birth interval”. Option “Unwantd/undesir” stands for “Unwanted or undesired pregnancy”

^#The prevalence of unhappiness was estimated excluding the “missing” category

*Pearson X² test p-value for comparison of prevalence of unhappiness between categories of women. Where covariates have a category with missing or unknown values, the X² was estimated omitting these categories

As shown in Table 1, the majority of women were aged 20–24 years (36.0%) and 25–34 years (35.0%), while 21.6% were adolescents (12–19 years) - age ranging from 13 to 49 years, with mean of 25.2 years. Unhappiness with pregnancy was more prevalent among adolescents (42.2%) and women 35–49 years (32.2%), than in women aged 20–34 years (p < 0.001). Most women resided in urban areas, i.e. in Quelimane City (87.4%), and there was no significant difference in pregnancy-related unhappiness by place of residence (p = 0.954). With regards to household relations, 30.5% of women were heads, 29.6% were wives or mothers of household heads, and this variable was significantly associated with unhappiness, with higher prevalence observed among daughters and grand-daughters (33.8%) and daughters-in-law/sisters-in-law (32.7%), compared to household heads (p < 0.001). In relation to women’s education, more than two thirds (65.7%) had completed secondary education. Women with higher education reported lower prevalence of unhappiness (15.7%) compared to those with secondary (27.3%) and primary or no education (25.7%) (p < 0.001). Household size was also associated with the outcome, with higher prevalence among women living in larger households (29.7%) than in smaller (21.6%) and in median-sized households (20.7%), (p = 0.002). The prevalence of unhappiness did not differ significantly by religion (p = 0.558).

Women’s reproductive variables, contraception and pregnancy intentions

shows that 80.0% of women were married or cohabiting, and marital status was strongly associated with pregnancy-related unhappiness, with single women (38.3%) and divorced or separated (36.4%) reporting higher prevalence compared to married or cohabiting women (21.0%) (p < 0.001). More than a half of women (57.5%) were interviewed in the third trimester of gestation, 31.7% in the second, and 4.9% in the first trimester (mean gestational age of 6.8 months, ranging from 0 to 9 months). Women interviewed during the first trimester reported higher prevalence of unhappiness (33.3%) and those in the third trimester (25.6%) compared to those in the second trimester (21.2%) (p = 0.036). The majority of women (80.6%) had previous pregnancy, of which 38.0% had at least one adverse pregnancy outcome. Women with no previous pregnancies were more likely to report unhappiness (38.0%) compared to those with previous pregnancies (21.2%) (p < 0.001). Knowledge of modern contraceptives was high (90.9%). The most mentioned methods were implant (26.8%), injection (25.8%), condom (21.1%) and oral contraceptive pill (15.6%), but most women (88.4%) were not using contraceptives when they became pregnant. Women without knowledge of modern contraceptives had higher prevalence of unhappiness (34.8%) compared to those with knowledge (23.4%) (p = 0.003). Responses to what they thought were the reasons for getting pregnant while on contraceptives showed that 42.6% had temporarily stopped or abandoned contraceptives, 33.7% did not know why they got pregnant, and 23.7% had failed to correctly use the method. Further, Table 1 shows that 33.3% of pregnancies had been unintended and pregnancy intention was strongly associated with the outcome, with 68.7% of women with unintended pregnancies reporting unhappiness, compared to 2.3% among those with intended pregnancies (p < 0.001). Of the 515 women with unintended pregnancies, 63.1% said that the pregnancy occurred on the wrong time, while 34.0% said that this pregnancy should not have occurred at all. The leading reasons for unhappiness were unplanned pregnancies (30.2%), unwanted/undesired pregnancy (26.5%), followed by fear of consequences (17.7%), short birth interval (8.2%); and the fact that the women did not want any more children (8.2%). Of the 1168 happy women, 34.1% said that were happy because they wanted a child, 28.5% because the pregnancy had been a blessing or a good surprise, 24.7% wanted more children, and 3.9% were already desperate for a child. Additional file 3 presents details of this information by age groups of women).

Partner’s-related variables and prevalence of unhappiness in women

As mentioned in the methods section, partner’s data were collected from women’s responses. Table 2 shows that of the 1237 married women, 74.9% were living with their partners, but 52.1% did not know the age or date of birth of their husbands. About 71.3% of women did not know their partner’s level of education. Table 2 shows that 21.5% of women did not have conversations with their partners about childbearing before the current pregnancy. About a quarter (23.3%) of women said that their partners did not intend to have the current pregnancy; and 16.9% had unhappy partners. Responses to why the partners were unhappy show that 30.2% were unhappy because it was an unwanted/undesired pregnancy, 29.0% were unplanned pregnancies, and 3.4% were pregnancies with short birth interval. Reasons for partner’s happiness included partner wanted a child (43.0%), wanted more children (21.8%), and first time to be a father (12.0%). Partner’s acceptance of pregnancy was high (97.3%). Of the 41 partners who rejected the pregnancies, 39.0% did so because they did not want to take responsibility, and 34.2% said that the pregnancies had not been caused by them, while 26.2% of women did not know why the partner rejected the pregnancy.

Table 2.

Partner- and couple’s-related variables and prevalence of unhappiness (based on women’s report)

Selected variables		Total of women (N, %)		Prevalence unhappiness (n, %)		X2 P-value*	Selected variables		Total of women (N, %)		Prevalence unhappiness (n, %)		X2 P-value*
Partner- and couple’s-related variables							Partner’s and family support to pregnancy (if partner accepted pregnancy N = 1504)
Resid with partner same hhd	Yes	927	74.9	194	20.9	0.657	Partner is helping in this pregnancy	Yes	1,441	95.8	328	22.8	0.606
	No	51	4.1	12	23.5			No	47	3.1	17	36.2
	Missing	259	20.9	54	20.8			Some times	16	1.1	8	50.0
Age of partner, if married	15–24	65	5.3	11	16.9	0.721	Partner is physically present in preg	Yes	1385	92.1	306	22.1	0.693
	25–64	528	42.7	99	18.8			No	75	5.0	31	41.3
	Missing	644	52.1	150	23.3			Some times	44	2.9	16	36.4
Couple have talked about childbirth	Yes	1,213	78.5	215	17.7	0.000	Partner is emotionally present in preg	Yes	1433	95.3	320	22.3	0.627
	No	333	21.5	163	48.9			No	54	3.6	27	50.0
Partner pregnancy intentional	Intentional	1140	73.7	156	13.7	0.000		Some times	17.0	1.1	6	35.3
	Unintentional	360	23.3	195	54.2		Partner is economically present in preg	Yes	1423	94.6	315	22.1	0.806
	Don’t know	46	3.0	27	58.7			No	54	3.6	25	46.3
Partner happy with pregnancy	Yes	1237	80.0	180	14.6	0.000		Some times	27	1.8	13	48.1
	No	262	16.9	183	69.8		His family helping in this pregnancy	Yes	1128	75.0	237	21.0	0.137
	Do not know	47	3.0	15	31.9			No	267	17.8	86	32.2
Why partner happy (N = 1237)	Wanted + kids	270	21.8	28	10.4	0.001		Some times	109	7.3	30	27.5
	Blessing/good surprise	27	2.2	0	0.0		Your family is helping in pregnancy	Yes	1299	86.4	290	22.3	0.240
	Satisfy family	95	7.7	11	11.6			No	130	8.6	37	28.5
	Desper for child	33	2.7	2	6.1			Some times	75	5.0	26	34.7
	Wanted a child	532	43.0	86	16.2		Perception of abandonment, if partner rejected pregnancy (N = 41)
	First time father	148	11.96	32	21.6		Felt abandoned because of this pregnancy	Yes	21	50.0	Values not presented due to small “n”
	Don’t know/mis	132	10.67	21	15.9			No	19	45.2
Why husband unhappy (N = 262)	Unplanned	76	29.0	65	85.5	0.042		Some times	2	4.8
	Want no + kids	3	1.2	1	33.3		Who abandoned women (N = 21)	Partner	14	66.7	Values not presented due to small “n”
	Short birth intrv	9	3.4	7	77.8
	Unwantd/unds	79	30.2	51	64.6			Relatives	7	33.3
	Have kids outsd	3	1.2	0	0.0
	Dubious partnr	3	1.2	2	0.7
	Don’t know/mis	89	34.0	57	64.0
Partner accepted this pregnancy	Yes	1505	97.3	353	23.5	0.101
	No	41	2.7	25	61.0
Why partner rejected (N = 41)	Said was not his	14	34.2	Values not presented due to small “n”
	Did not want	16	39.0
	Don’t know	11	26.8

Variable “Resid with partner same hhd” stands for “Women resid with partner in the same household”

In variable “Why partner was happy”, option “Wanted + kids” stands for “Wanted more children”; option “Blessing/surprise” stands for “The pregnancy represented a blessing, or a good surprise”; option “Satisfy family/ki” stands for “He wanted to satisfy his family or his children”; option “Desperat for kid” stands for “Has been looking for children and was already desperate for having children”; option “Wanted a child” stands for “He wanted a child”; option n “First time father” means “partner was happy because he was going to be a father for the first time in his life”

In variable “Why husband unhappy”, option “Want no + kids” stands for “Wanted no more children”; option “Short birth intrv” stands for “The woman got pregnant while she had a too young baby”, which we termed “short birth interval”; option “Unwantd/unds” stands for “Unwanted or undesired pregnancy”; option “Have kids outsd” stands for “This man is unhappy because has children (or a child) outside this union”; option “Dubious partnr” represents the man who is unhappy because he thinks the pregnancy was not caused by him

^#All the measurement of partner variables are based on female respondent’s report

*Pearson X² test p-value for comparisons of prevalence of unhappiness between categories of women. Where covariates have a category with missing or unknown values, the X² was estimated omitting these categories

Partner-related factors were strongly associated with women’s unhappiness with current pregnancy. Women who reported lack of couple’s communication about childbirth had substantially higher prevalence of unhappiness (48.9%) compared to those who reported communication (17.7%) (p < 0.001). Similarly, women whose partners considered the pregnancy unintended had higher prevalence of unhappiness (54.2%) compared to those whose partners intended the pregnancy (13.7%) (p < 0.001). The highest prevalence was observed among women who reported that their partners were unhappy with the pregnancy (69.8%), compared to 14.6% whose partners were happy (p < 0.001). Partner acceptance of the pregnancy was also strongly associated with the outcome. Among women whose partners rejected the pregnancy, 61.0% reported unhappiness compared to 23.5% whose partners accepted the pregnancy (p < 0.001).

Further, Table 2 shows higher percentages of partners’ and family support to the pregnancy among women whose partners accepted the pregnancies than for women whose partners rejected the pregnancies. It shows that 92.1% of partners that accepted the pregnancies were physically present when needed to provide support to the women, 95.3% were providing emotional and psychologically support, and 94.6% were supporting economically or financially. Half (50.0%) of the 41 women whose partners rejected the pregnancies said that they felt abandoned; of which 66.7% said that were feeling abandoned by their sexual partners, and 33.3% by their parents.

These descriptive findings guided subsequent regression analyses, which were structured according to the conceptual framework to assess the relative contribution of socio-demographic, reproductive, and partner-related factors.

Factors associated with women’s unhappiness with pregnancy

Bivariate logistic regression analyses were conducted as an initial assessment of associations between each explanatory variable and pregnancy unhappiness. The results are presented in Additional file 5. In general, the bivariate findings were consistent with the multivariable analyses, although some associations were attenuated after adjustment for reproductive and partner-related characteristics.

Multivariate regression analysis

Table 3 presents the results from the three sequentially adjusted multivariate logistic regression models as explained in the methods section. In Model I, which included background socio-demographic characteristics, adolescent age (OR = 2.24, 95% CI:1.61–3.13) older age (35–49 years; (OR = 2.65, 95% CI:1.64–4.29) were significantly associated with pregnancy unhappiness. Lower educational attainment also remained significant. Women who are daughters or sisters or grand-daughters of household head had higher odds of being unhappy compared to women who are heads themselves or wives of heads. In Model II, which added women’s reproductive and pregnancy characteristics, the effect of education and age remained at similar levels of association with higher odds of pregnancy unhappiness (as in Model I), while the effects of household size and relationship with head were attenuated. Of the newly added variables, marital status and gestational age showed strong association with unhappiness, with single women (OR = 1.90, 95% CI:1.20–3.02) showing higher odds compared to married or cohabiting women. Women in the first trimester of pregnancy had higher odds of unhappiness compared to those in the second trimester (OR = 2.35, 95% CI:1.30–4.24). In Model III, which further incorporated partner-related variables, age of women, gestational age at interview, and history of past adverse pregnancy outcomes maintained their strong association with unhappiness as in Model II, while the effects of the other socio-demographic and reproductive factors were attenuated. However, partner-related variables emerged as the strongest associated factors with maternal unhappiness. Women who reported lack of couple communication about childbirth had higher odds of unhappiness (OR = 1.82, 95% CI: 1.26–2.63) compared to those who reported couple communication. The strongest association was observed for partner’s unhappiness with the pregnancy, where women had substantially higher odds of reporting unhappiness themselves (OR = 8.23, 95% CI:5.04–13.43) compared to those with happy partners.

Table 3.

Multivariate association of explanatory variables with pregnant women’s unhappiness

Variables and categories		Model I: Background variables		Model II: Model I nested here		Model III: Model I & II nested here
		Background variables		Background and women’s variables		Background, women’s, and partner’s variables
Variables	Categories	Odds Ratio [95% Confidence Interval]	P-value	Odds Ratio [95% Confiden Interval]	P-value	Odds Ratio [95% Confidence Interval]	P-value
Age of woman (in years)	Adults 20–34	Reference
	Adolescents 12–19	2.24 [1.61–3.13]	0.000	2.02 [1.40–2.91]	0.000	2.18 [1.43–3.32]	0.000
	Older adults 35–49	2.65 [1.64–4.29]	0.000	2.63 [1.61–4.31]	0.000	2.21 [1.25–3.93]	0.007
Relation to head	Heads/wives/mothers/aunts	Reference
	Daughters/sisters/cousins	1.53 [1.07–2.21]	0.021	0.97 [0.60–1.57]	0.904	0.99 [0.57–1.71]	0.967
	Daughters&sisters-in-law	1.45 [0.94–2.25]	0.092	1.42 [0.91–2.23]	0.124	1.29 [0.77–2.17]	0.333
	Grand-daughters/nephews	1.98 [1.08–3.66]	0.028	1.47 [0.74–2.90]	0.271	1.68 [0.78–3.63]	0.184
	Unrelated	0.83 [0.44–1.60]	0.585	0.75 [0.38–1.45]	0.386	1.02 [0.50–2.09]	0.962
Household size	Median size (5 members)	Reference
	Below median	1.12 [0.76–1.66]	0.565	1.10 [0.74–1.63]	0.636	1.20 [0.77–1.88]	0.417
	Above median	1.36 [0.92–2.01]	0.122	1.37 [0.92–2.03]	0.121	1.17 [0.74–1.84]	0.501
Education of women	Higher	Reference
	Primary/none	1.74 [1.04–2.90]	0.035	1.78 [1.05–3.02]	0.032	1.41 [0.78–2.56]	0.252
	Secondary/medium	1.78 [1.23–2.55]	0.002	1.73 [1.20–2.50]	0.004	1.42 [0.94–2.14]	0.097
Is woman studying	No	Reference
	Yes	1.20 [0.87–1.64]	0.262	1.10 [0.79–1.43]	0.569	0.97 [0.67–1.42]	0.894
Woman’s reproductive variables and categories
Marital status	Married/cohabiting			Reference
	Single/never married	(See Model I and II)		1.90 [1.20–3.02]	0.006	1.20 [0.70–2.06]	0.509
	Divorced/separated			1.51 [0.47–4.88]	0.487	0.96 [0.26–3.64]	0.958
Gestational age, trimester of interview	2nd trimester			Reference
	1 st trimester	(See Model I and II)		2.35 [1.30–4.24]	0.005	2.19 [1.10–4.37]	0.025
	3rd trimester			1.43 [1.07–1.92]	0.017	1.51 [1.08–2.11]	0.016
Had previous pregn adverse outcome	Had prg, no advers outcome			Reference
	No preg no advers outcome	(See Model I and II)		1.18 [0.80–1.74]	0.397	0.94 [0.60–1.48]	0.798
	Had prg, had adver outcome			0.74 [0.54–1.03]	0.073	0.64 [0.44–0.92]	0.017
Know/aware of modern contraceptives	Yes			Reference
	No	(See Model I and II)		1.05 [0.67–1.64]	0.821	0.74 [0.43–1.25]	0.254
Parnter’s-related variables and categories
Couple had communication about having children together	Yes	Reference
	No		See results in Model III			1.82 [1.26–2.63]	0.001
Partner intended current pregnancy	Intended	Reference
	Unintended		See results in Model III			1.73 [1.12–2.65]	0.013
Partner was happy with current pregnancy	Happy	Reference
	Unhappy		See results in Model III			8.23 [5.04–13.43]	0.000
Partner accepted this pregnancy	Accepted pregnancy	Reference
	Rejected pregnancy		See results in Model III			0.55 [0.02–1.45]	0.226

PANEL B: MODEL DIAGNOSTIC
Measures of model diagnostic	Model I	Model II	Model III
Sample size (N)	1,361	1,361	1,361
Likelihood Ratio χ²	95.6	120.8	372.5
Model degree of freedom	11	18	22
p-value	0.000	0.000	0.000
Log likelihood	−694.47	−681.87	−556.03
Pseudo R²	0.064	0.081	0.251

Likelihood-ratio test between Models			LR χ² difference	Degree of freedom	p-value
Likelihood-ratio test, Model I nested in Model II			25.19	7	0.001
Likelihood-ratio test, Model II nested in Model III			251.68	4	0.000
Likelihood-ratio test, Model I nested in Model III			276.87	11	0.000

Note: The variable named “Had previous pregn adverse outcome” stands for “Woman had previous pregnancy adverse outcome” and the options (i) Had prg, no advers outcome – stands for: had previous pregnancy without adverse outcome; (ii) No preg no advers outcome – stands for: had no previous pregnancy; (iii) Had prg, had adver outcome – stands for: had previous pregnancy and adverse outcome

Overall, the results indicate that while socio-demographic factors play an important role in bivariate associations, their effects are attenuated after adjustment for reproductive and partner-related variables, excepting age of women, which maintained similar levels of effects across the three models. Thus, alongside maternal age, partner emotional responses and couple communication emerge as the most influential determinants of women’s pregnancy-related unhappiness.

Goodness-of-fit test indicated progressive improvement across the three nested models. As shown in Panel B of Table 3, Model was statistically significant (Likelihood Ratio -LR χ² = 95.6, df = 11, p < 0.001). The addition of women’s reproductive and pregnancy characteristics in Model II significantly improved model fit (ΔLR χ² = 25.19, df = 7, p = 0.001). The inclusion of partner and couple relationship characteristics in Model III resulted in a substantial further improvement in fit (ΔLR χ² = 251.68, df = 4, p < 0.001). Overall, the full model fit the data significantly better than Model I alone (ΔLR χ² = 276.87, df = 11, p < 0.001). The magnitude of the LR improvements suggests that partner and relationship characteristics contributed more to explaining pregnancy unhappiness than the reproductive variables added in Model II.

Additional analysis on age-stratified unhappiness

In this study, additional analysis was carried out to assess how pregnancy unhappiness varies by age of woman, comparing three age groups of women: adolescents (12–19 years), adults (20–34), and older adults (35–49). This was in recognition of women’s age as a key variable for understanding differences in fertility intentions and how pregnancy is perceived within individuals’ life course.

Overall, this analysis showed that adolescents were the most disadvantaged group followed by older women in most of the indices examined. For example, Additional file 3 shows that adolescents had the highest percentage of women that had at least one adverse pregnancy outcome (52.6%), followed by 43.7% in adults and 42.3% in older women. Knowledge of modern contraceptives was lowest in adolescents (78.4%); and the percentage of women who had stopped contraceptives was highest among adolescents (54.5%), while failure to comply with contraceptives was highest in older women (50.0%). Additional file 4 shows that the percentage of women lacking communication with their partners about childbearing was highest in adolescents (37.7%) and in older women (21.7%). Unintended pregnancy was more prevalent in adolescents (54.5%) and in older women (38.3%). Adolescents had the highest percentage of unhappy partners (21.3%), followed by older women (15.7%); and, similarly, adolescent had the highest percentage of partners who rejected the pregnancies (5.4%).

These findings suggest that examining pregnancy perceptions among adolescents may be a key research question to pursue in settings such as this, where early pregnancy and marriage remain significant public health concerns and are also linked to human rights issues.

Discussion

This study estimated the prevalence of unhappiness in pregnant women with regards to their current pregnancy and examined the reasons and factors associated with such negative perceptions using a couple-centered analytical framework. The findings show that about one in every four women were unhappy with their pregnancies, and this prevalence was higher in adolescents and older women than in middle-aged women; in single and divorced/separated women compared to married women; during the first trimester of gestation than during the second trimester, in women with unintended pregnancies than those with intended pregnancies, as well as in women who lacked couple’s communication regarding childbirth, and in women whose partners did not intend to have pregnancy. The leading reasons for women’s unhappiness were unplanned pregnancy, unwanted/undesired pregnancy, fear of consequences, and short birth intervals.

Further, the findings demonstrate that although women’s socio-demographic and reproductive characteristics were associated with pregnancy-related unhappiness in bivariate analyses, these associations were substantially attenuated in multivariate models, excepting maternal age and gestational age, which maintained their strong association. In the fully adjusted model, maternal age, gestational age, and partner-related factors particularly partner emotional response and couple communication emerged as the most strongly associated factors with women’s pregnancy-related unhappiness. In the bivariate analysis, younger and older maternal age groups, lower educational attainment, and adverse reproductive histories were significantly associated with higher odds of pregnancy unhappiness. These findings are consistent with previous studies showing that both adolescent and older maternal age can be associated with reduced pregnancy acceptance, often due to socioeconomic constraints, life-course timing, or reproductive expectations. In the sequentially adjusted multivariable models, most socio-demographic associations were attenuated after adjustment for women’s reproductive characteristics and partner-related variables. This suggests that these background factors may operate indirectly through reproductive intentions and couple dynamics rather than exerting a direct effect on maternal emotional responses to pregnancy. Maternal age maintained similar levels of strength of association across models, highlighting the importance of woman’s age as a key factor for understanding fertility intentions and pregnancy perceptions. Gestational age has also remained significantly associated with pregnancy unhappiness in the adjusted models, with women in the first trimester more likely to report unhappiness compared to those in the second trimester. This may reflect adaptation processes during pregnancy, where emotional acceptance increases as pregnancy progresses. Previous studies have reported decrease of unhappiness from pregnancy realization, through the three trimesters of gestation, and the post-partum period [49–51].

The results of this study are similar to those reported in other sub-Saharan African countries. For example, a study in Ethiopia reported a prevalence of unhappiness of 24% among pregnant women [5], and another study in Iran (out of sub-Saharan Africa) reported a prevalence of 31.4% [4] – both studies reported higher levels of unhappiness in adolescents than in older women. Furthermore, the divergent states of unhappiness between partners (and, possibly, family members) about the same pregnancy that our study demonstrated was also reported in a study in Nigeria [52]. Our study shows that more women were unhappy than their partners about the same pregnancy, meaning that for the same pregnancy, partners can be discordant in their happiness about the pregnancy, and even where both partners reported being happy, the reasons for happiness may differ. For example, in this study, the overall 3 top reasons for happiness in women and men coincided only on the first 2 reason as “wanted a child” and “wanted more children”, but were different from the third reason - “blessing/good surprise” among women vs. “first time father” among partners.

A key finding of this study is the dominant role of partner-related variables in shaping women’s pregnancy-related unhappiness. Lack of couple communication about childbirth was independently associated with increased odds of maternal unhappiness, highlighting the importance of shared decision-making and emotional dialogue within couples. Most notably, partner’s emotional response to the pregnancy showed the strongest association with women’s unhappiness. Women whose partners were unhappy with the pregnancy had substantially higher odds of reporting unhappiness themselves. This finding supports the conceptual framework proposed in this study, which emphasizes pregnancy-related emotional experiences as embedded within couple dynamics rather than being solely individual-level phenomena. These results align with theoretical perspectives on dyadic coping and shared reproductive decision-making, which suggest that partner attitudes and emotional reactions significantly shape women’s psychological responses during pregnancy.

The role of partner’s and family context on pregnancy unhappiness was also reported in a study from Nigeria, which recorded very high percentage of parents (93%) who said that extramarital pregnancy in adolescent was an anti-social event; 85.8% said it represented parental failure, and 49.8% would stop the girl’s schooling and send her to the man/boy responsible for the pregnancy [14]. High levels of parental unhappiness were also reported in a similar study in South Africa [15]. In these contexts, it is very likely that women with extra-marital pregnancies may experience higher levels of pregnancy unhappiness. The findings from our study suggest that parental denial of pregnancy in adolescents may also be higher - as about a half of the women with rejected pregnancy said that had been felling abandoned, mostly by their sexual partners, their fathers, and brothers and sisters. This suggests that once the male partner rejects a pregnancy, the burden of the pregnancy and the baby is left to the pregnant woman and, eventually, her relatives. From the programmatic perspective, this finding reinforces the message that preventing unintended pregnancies should include both partners (boys/men and girls/women) [53] [54]. The motives for partners’ rejection of pregnancies reported in this study (not wanting to take the responsibility and saying that the pregnancies were not theirs), highlights the role of unplanned childbirth, and the call for more responsible sexual relations by both women and men. Investigating these within the dyadic approach to understand how each partner influences the other may provide insights for better strategies to prevent not only unwanted or unplanned pregnancies, but most importantly, the consequences of these events.

Building on this study and the existing literature, a variety of future research questions can be pursued on the perceptions and attitudes towards pregnancy, parenting, marriage, and divorce in settings like Mozambique, where extramarital pregnancy is predominant [55]. For example, a study like this, but focusing only on adolescents, is appealing for detecting potentially important differences in pregnancy unhappiness between younger, middle, and older adolescents, particularly because this age group includes a great diversity of life experience, emotional development, and happiness. Further, some studies have argued that marriage is losing its role as a social institution [56–58], yet, rates of marriages and divorces, and of extramarital pregnancies are scarce. Also, it is known that most of extramarital pregnancies trigger marriages, particularly in adolescents [24]. This study suggests that some unplanned pregnancies, particularly the rejected ones, may have contributed to dissolve partnerships – and this needs investigation because although most of the rejected pregnancies (88.9%) were among single and divorced/separated women, these data cannot be used to assess whether pregnancy rejection meant rejection of the pregnant women and dissolution of partnership or not. Thus, examining the role of unplanned pregnancies in anticipating marriages/cohabitation and in dissolving partnership may be a relevant research question to be investigated.

Other research questions can be pursued: (i) how to better measure (un)happiness with pregnancy in settings like this, where socio-cultural norms and public health recommendations appear to promote sexual and reproductive behaviours towards opposite direction?; (ii) What should be measured - happiness or unhappiness, or both? - some recent studies appear to favour measuring unhappiness instead of happiness due to its clear association with negative outcomes [11, 13]. These authors argue that happiness is an elusive and subjective concept than unhappiness because unhappiness has more objective than subjective sources such as unplanned pregnancy, rejection from partner, hunger, illness, school drop-out, death of loved ones, family dissolution and others [13]. Other researchers have measured both happiness and unhappiness, arguing that they are related but not opposites statuses, and no cut-off line can objectively be drawn between them [12]. Such research would include asking the participants the meaning they give to the terms “happiness” and “unhappiness” [12], as well as asking whether the expected baby will be a boy, girl, or twins, as it is known that child’s sex preferences impact fertility intentions.

With regards to happiness scales (positive side) and depression scales (negative side), De Paola et al. [12] mentions that these scales have placed Finland at higher levels of happiness and of depression simultaneously. The tools may be improved by incorporating other tools such as the Evaluation and Nurturing Relationship Issues, Communication, and Happiness (ENRICH) tool, which measures how satisfied people are in their marriage or intimate partnership, examining items such as communication, conflict resolution, emotional closeness, roles and responsibilities, sexual relationship, financial management, and shared values [24] [23]. In addition to adapting the scores of degrees of happiness and of unhappiness for measuring pregnancy unhappiness, the inclusion of motives for unhappiness (like done in this study) appears to add mor insights into the understanding of unhappiness. These insights may be improved by asking respondents to describe the whole couples’ context in which the pregnancy occurred and the context of her/his unhappiness. Improvements may be maximized using longitudinal designs (where possible) to better understand how couple dynamics evolve throughout pregnancy and how these changes influence maternal emotional well-being over time, including some period after delivery (ideally, 1 year, to capture infant health and survival).

Study limitations

This study has several strengths, including the use of a large sample and a structured analytical approach based on nested logistic regression models that allowed for examination of how associations change with sequential adjustment. The integration of partner-level variables represents a major strength and addresses an important gap in reproductive health research.

However, the study has limitations. A potential weakness (source of recall bias) in this study may arise from the time elapsed between pregnancy realization and the interview, particularly when 57.5% and 31.7% of women were interviewed during the third and second trimesters of gestation, respectively. However, the impact of recall bias can hardly be explored in this study. Further, it is unlikely that selection bias may have occurred (i.e. losing women who may be more likely to have abortion or other characteristics) because PregSurv is a population-based surveillance recruiting all pregnant women and those whose pregnancies terminated recently, in hospitals and at home. If they change household of residence before or after enrolment, the fieldworkers will search for them leveraging the HDSS platform, which records cell phone numbers of all members); and, because most people move within the study area, i.e. short distances, as stated in the Ravenstein Laws of Migration [59], they are most likely to be found for recruitment or follow-up.

Another weakness is that this study used male partner’s data obtained from women’s interviews, which may have caused inaccuracies of the information. However, while the data show large percentages of missing or unknowns in age (52.1%) and education (72.0%) of partner among married women, key variables for this analysis appear to have few missing/unknow responses in the overall sample, namely couple’s communication (no missing/unknown), partner’s pregnancy intendedness (3.0%), partner unhappiness (7.6%), pregnancy rejection (no missing/unknown), and reasons for rejection (26.2% - that may be high). Ideally it would have been preferable to collect partner variables directly from them.

Some secondary outcomes had smaller sample sizes to allow robust conclusions even from the perspective of descriptive statistics. For example, there were 41 women whose partners rejected pregnancies – 16 (39.0%) cases because did not want to take responsibility, 14 (34.2%), because they said that it was not their pregnancy, and 11 (26.8%) “do not know”. The sample for this and other secondary outcomes will increase as the surveillance is still ongoing, and they may inform well future analysis latter than now. Also, this study combined all adolescents into a single age group 12–19, which may have limited the ability to detect potentially important differences in pregnancy unhappiness between younger, middle, and older adolescents, particularly because this age group includes a great diversity of life experience, emotional development, and happiness.

The other point worth commenting is related to the 51 women excluded from the analysis because they responded “don’t know” to the question “were you happy?”. They represent a distinct and heterogeneous group. The questionnaire included an open-ended follow-up question (“Can you explain your answer?”), which provided further insight. Among these women, 64.7% were unable to further clarify their response (they said “don’t know”, again), while others expressed concerns such as fear of pregnancy loss (particularly among those with prior adverse outcomes) and fears of loneliness or partner abandonment. While these findings highlight the complexity of emotional responses to pregnancy and warrant further investigation, this group could not be meaningfully incorporated into the binary outcome used in this study (unhappy vs. happy); and given the small proportion (approximately 3% of the sample), their exclusion is unlikely to have materially affected the results.

Conclusions

The prevalence of women unhappy with their pregnancy in Quelimane was high, particularly in adolescents, older women, unmarried women, and during the first trimester of gestation. This study demonstrates that pregnancy-related unhappiness is not solely an individual experience but is strongly embedded within couple dynamics. While socio-demographic and reproductive factors are important in unadjusted analyses, maternal age, gestational age, and partner emotional response and communication patterns are the most influential determinants of maternal pregnancy unhappiness. These findings support the need for couple-centered approaches in maternal health research and interventions. Particular to adolescents (a critical transitional developmental stage in life), for who the negative consequences of pregnancy are most severe, this study suggests that examining unhappiness in adolescents (girls, boys and their parents or legal guardians) can be a strategic research approach in settings like Mozambique, where early marriage and adolescent childbirth are a major problem of public health, that transcend to human rights violation.

Authors’ contributions

AN: Conceived this study and its methods, performed data analysis and interpretation, and wrote this article.FA: Participated in the coordination and supervision of fieldwork, data collection and curation, and contributed to the writing of this article.HA: Participated in the supervision of fieldwork, data collection and curation, and contributed to the writing of this article.TA: Participated in the supervision of fieldwork, data collection and curation, and contributed to the writing of this article DM: Participated in the data management, including data processing and curation.NA: Participated in the data management, including data processing and curation.EX: Contributed to the writing of this article – reviewing and editing.MK: Contributed to the writing of this article – reviewing and editing.RV: Contributed to the writing of this article – reviewing and editing.LG: Contributed to the writing of this article – reviewing and editing.QB: Contributed to the writing of this article – reviewing and editing.IM: Contributed to the writing of this article – reviewing and editing.OA: Participated in the study design, data analysis, interpretation of results, and writing of this article.CA: Participated in the study design, data analysis, interpretation of results, and writing of this article.PA: Participated in the study design, data analysis, interpretation of results, and writing of this article.All authors reviewed the manuscript.

Funding

The HDSS and Pregnancy Surveillance platforms in Quelimane are integral parts of the CHAMPS Network Study, which is funded by the Bill and Melinda Gates Foundation through the Emory University (USA). The core activities of CISM are funded by the Spanish Agency for International Development and Cooperation.

Data availability

The data used in this study can be accessed from two sources: (i) the data from the Standard CHAMPS Pregnancy Surveillance questions are available in the Child Health and Mortality Prevention Surveillance (CHAMPS) data portal, after completing an online Data Request Form at https://champs-redcap.emory.edu/surveys/?s=PCEERX993Y. (ii) the data from Mozambique-specific questions (on unhappiness with pregnancy) are not available in a public web-based repository but can the obtained by a formal request to Dr Charfudin Sacoor (charfudin.sacoor@manhica.net). Dr Sacoor is the Scientific Coordinator at Manhica Health Research Center.

Declarations

Ethics approval and consent to participate

The HDSS was approved by the CISM’s Institutional Scientific Committee (Ref: CCI/007/MAR/2021), and the Institutional Bioethics Committee (Ref: CIBS-CISM/029/2021). The PregSurv study has ethical approval from the Institutional Ethics Review Board for Health at CISM (Ref: CIBS-CISM/080/2022), and from the National Bioethics Committee for Health (Ref:104/CNBS/23). All the participants of the two studies have given and signed an informed consent for each study. Participants aged 12–17 years have signed an assent after the consent of their parents or legal guardians.

Consent for publication

The above-mentioned informed consent included permission to publish the results of the data analysis and sharing anonymous data with researchers who collaborate with CISM.

Competing interests

The authors declare no competing interests.