An institution-based cross-sectional study design was conducted in Boneya Boshe Woreda from 1 October to 30 October 2023. This Woreda is found in the East Wallaga Zone of the Oromia region, which is located 81 km from the capital city of the zone, Nekemte town, and 311 km from the capital city of the country, Addis Ababa. Boneya Boshe Woreda is bordered by Nono Benja Woreda in the south, Gobbu Sayo and Sibu Sire Woreda in the north, Bako Tibe and Ilu Galan in the east, and Wahama Hagalo in the west.

In this Woreda, the total population is 73,227, with 14,925 households in 12 Kebeles. Out of this population, 34,803 are in the reproductive age group, of which 16,206 are women in the reproductive age group. This Woreda has an expected delivery rate of 2,541 per year. The 2022 Woreda Health Bureau report revealed that the total delivery in one year was 2,355. This Woreda has three health centers and 12 health posts, with a total of 68 health professionals to carry out healthcare activities.

All postpartum mothers attending immunization clinics in the first one year after childbirth during data collection period and who presented with their infants whose age was from two weeks after birth to one year were included in the study. However, those mothers who were critically ill during the data collection period and not permanent residents of this Woreda were excluded from the study.

The sample size was determined by using a single population proportion formula based on assumptions of 95% CI, 5% margin of error, and 50% of the proportion of self-harm since there has been no study done on this population in the study area, as follows: n = Z ∝ 2 2 p q d 2 .

Then n = 1.96 2 0.5 x 0.5 0.05 2 = 384.

Finally, by adding 10% of the non-response rate, the final sample size of 423 was arrived at for this study. A systematic random sampling technique was used to select the study participants based on the total number of infants attending immunization clinics. Initially, the total number of infants attending three immunization clinics in this Woreda was obtained from the infant immunization logbooks of each health center. Those health centers are Bilo Health Center, Qidame Health Center, and Jawis Health Center. The number of infants attending immunization clinics were 761 from Bilo Health Center, 692 from Qidame Health Center, and 676 from Jawis Health Center 676. Then it was proportionally allocated, and the allocated mothers were selected by a systematic random sampling method as follows (Figure 1).

The magnitude of self-harm was the dependent variable in this study.

Socio-demographic variables include maternal age, husband’s age, maternal educational level, husband’s educational level, maternal occupation, husband occupation, average monthly income, and family size.

The data collected for the study include the place of delivery, postnatal follow-up, history of adverse pregnancy outcome, maternal history of known diagnosed medical illnesses, history of diagnosed mental health problems during pregnancy, husband’s involvement in maternity and child healthcare, husband’s index of pregnancy during the current baby, husband’s satisfaction with the gender of the current baby, alcohol use by the father and mother of the infant, husband’s khat use, and husband’s smoking habit.

Intimate partner violence during the postnatal period, postpartum depression, maternal social support level, loneliness, and the decision-making power of the mother were explored.

Intentional self-harm is the condition in which an individual tries to hurt themselves purposefully. They could indulge in self-injurious behaviors like cutting, burning, hitting, hanging, overdosing, poisoning, banging head on objects, scratching the body, bloodletting (intentional act of an individual to let blood from them for self-hurt), strangulating the neck (the act of an individual to forcing or damaging one’s own neck or body for self-hurt), and electrocuting by the mother who intended to harm herself. If she has at least attempted one or more items developed provided in the Self-harm Screening Inventory (SHSI) tool to hurt herself during her first year after childbirth, she is considered to have done intentional self-harm, for which code “1” was assigned. For a mother who has not tried any of the items mentioned in the SHSI tool is considered to have not resorted to intentional self-harm, which was coded as “0” (27).

Partner involvement in maternity and child healthcare: This variable was assessed with nine questions. For each question, the responses were scored between 0 and 1. The total score was 9, with a minimum of 0 and a maximum of 9. Hence, husband’s involvement with a score above the median showed that the husband was involved in maternity an child healthcare (28).

Postpartum depression: The women who had an Edinburgh postnatal depression score of greater than or equal to 13 during their first year after childbirth were considered to be depression, and code “1” was given to them; those who scored less than 13 were not considered to have depression symptoms and coded as “0” (29).

Known medical illness: Women diagnosed with chronic medical illnesses like diabetes, cardiac disease, renal disease, hypertension, liver disease, and tuberculosis that are confirmed by doctors at health institutions are considered to have known medical illness (30).

Decision-making power: This was measured by the ability of women to act independently and decide on household activities including their health, their children’s health, freedom of movement, and control over finances without asking permission from another person. Depending on the items designed to assess maternal decision-making power, “2” was given if she decides by herself, “1” will be given if she decides jointly with her husband, and “0” was given if it is decided by others. Based on items designed to assess household decision-making power, women who scored above the median were considered to have good decision-making power and coded “1,” and those who scored below the median were considered to have poor decision-making power and were coded “0” (31).

Maternal social support: Social support was assessed by three items on the OSLO-3 social support scale. Mothers whose score was 3–8 were considered poor on social support and coded “0”. A score of 9–11 signified moderate social support and assigned the code “1”, and 12–14 indicated strong social support and code “2” was given (32).

Postpartum Intimate partner violence (PIPV): This was assessed by 13 items developed from an adapted tool to assess domestic violence against women in low-income country settings. It includes physical violence, sexual violence, and psychological violence. Any mother who is the victim of at least one type of PIPV was coded “1” and who was not the victim was coded “0” (33).

Maternal loneliness: The University of California Los Angeles (UCLA-3) 20-item Loneliness Scale (Version 3) was used to measure maternal loneliness during the first year after childbirth. The UCLA-3 20-item Loneliness Scale was used to gather the total score. Based on this tool, code “0” was given for those mothers who scored <28 (no loneliness) and “1” was given for those who scored ≥28 (had loneliness) (34).

History of adverse pregnancy outcome: Mothers who have a history of abortion, neonates with congenital anomalies that are incompatible with life, stillbirth, and/or neonatal death were considered to have adverse pregnancy outcomes for the purpose of this study.

Average monthly income: Based on this income, the mother was categorized as living above or below the poverty line based on the current classification system by the World Bank group. Earning US$1.90 or 97.85 Ethiopian Birr (ETB) or lower per day meant living below the poverty line. At the current exchange rate of US dollar to Ethiopian Birr, for 30 days, earning 2935.5 ETB or lower indicated that the mother was living below the poverty line (35).

In face-to-face interviewers, structured questionnaires prepared by Kobo Tool Box (an innovative open-source platform for collecting, managing, and visualizing data) was administered to collect the data. The tool has a socio-demographic component, reproductive history, maternal and child healthcare services, and psycho-social factors related to the mother. Postpartum intimate partner violence was assessed by 13 items containing sexual violence, psychological violence, and sexual violence (33). Postnatal depression was assessed by a validated Edinburgh postnatal depression tool that was validated in an Ethiopian context (29).

Maternal social support was assessed by a validated OSLO-3 social support scale that contains three items (32), and loneliness was assessed by the University of California Los Angeles (UCLA-3) 20-item Loneliness Scale (Version 3) (34). The outcome variable was assessed by the SHIS tool, which has 20 items to assess self-harm (27). Husband’s involvement in maternity and child healthcare was assessed by a tool developed from related literature that has nine items (28). Similarly, decision-making autonomy was assessed by nine items developed from the related literature (31).

A pretest was done on 10% of the study participants one month before data collection at Wahama Hagalo Woreda. The internal consistency of the tool was checked and had a Cronbach’s α test of 0.79. Initially, the tool was developed in the English language and translated to Afan Oromo for actual data collection, and the Afan Oromo version was retranslated back to English to cross-check the consistency of the tool.

Data was collected from three diploma holder women midwifery health professionals and supervised by three diploma holder women nurses. Training was given for two days on the objectives of the study for data collectors and the training also included participant safety for both data collectors and supervisors. Supervision was carried out by supervisors daily for the sake of clarity, accuracy, and consistency of the data.

The data collected by Kobo Tool Box was exported to SPSS version 25 software for cleaning, coding, and further analysis. Descriptive statistics were done, and the results were presented using diagrams and tables. Both bivariable and multivariable logistic regression analyses were done to identify factors associated with self-harm during the first year after childbirth. Variables with a p-value <0.25 in binary logistic regression analysis were transferred to multivariable logistic regression.

The crude odds ratio (COR) and the adjusted odds ratio (AOR) with a 95% CI were calculated to show association and strength of association, respectively. In multivariable logistic regression analysis, variables with a p-value of <0.05 were reported as statistically significant. The model goodness-of-fit test was done. Finally, the results were presented both in the narrative, diagram, and table forms.

Among the 423 postpartum mothers selected for this study, 415 mothers participated in the study, at a response rate of 98.10%. The study identified that 31.80% of the participants were in the 18–23 age category and 17.00% were in the 41–45 age category. As revealed by this study, 53.30% of the mothers were housewives and 57.38% of their husbands were farmers. In addition, 13.50% of mothers have no formal education, and 9.60% of their husbands have no formal education. Furthermore, 51.30% of mothers have an average monthly income below the poverty line, and 47.50% have family sizes greater than or equal to six (Table 1).

As identified by this study, 55.18% of the infants were girls, and 52.00% of the husbands were not satisfied with the gender of their infants. Similarly, 48.00% of mothers did not intend to have the baby during the index of pregnancy. Furthermore, 78.31 and 51.80% of the mothers have antenatal care (ANC) and postnatal care (PNC) follow-up, respectively. From the total study participants, 13.50% of mothers gave birth at home and 10.80% of them have a history of adverse pregnancy outcomes. It was revealed that only 45.10% of husbands were involved in maternity and child healthcare.

In addition, 5.10 and 3.62% of mothers have been diagnosed with medical illnesses and mental illnesses, respectively. Similarly, 2.66% of their husbands have been diagnosed with medical illnesses. Furthermore, 3.90% of mothers have a family history of known diagnosed mental illnesses. Out of the total study participants, 36.39% of mothers and 52.80% of their husbands were alcohol users. Finally, 15.90% of the husband participants consume the khat (Table 2).

Out of the total study participants, 13.50 and 32.30% of mothers suffered from postnatal depression and postpartum intimate partner violence. In addition, 41.90 and 17.10% of the mothers enjoyed poor and moderate social support, respectively, during their postnatal period. Similarly, 29.40% of the mothers were lonely and only 31.81% of the mothers have decision-making power over household activities (Table 3).

This study revealed that the magnitude of self-harm among postnatal mothers in this study setting was 12.53% (95% CI: 9.33, 15.73) (Figure 2).

From the total variables fitted for binary logistic regression at a p-value less than 0.25, six variables were significantly associated in the multivariable logistic regression model at a p-value less than 0.05 at a 95% CI. Variables significantly associated with self-harm in the final model were husband’s involvement in maternity and child healthcare, depression, loneliness, postnatal follow-up, postpartum intimate partner violence, and average monthly incomes.

This study identified that the odds of self-harm among mothers whose husbands do not participate in maternity and child healthcare were 1.90 times higher than those whose husbands participate in the care (AOR = 1.90; 95% CI: 1.12, 2.10). In addition, the odds of self-harm among mothers who have depression were 2.79 times higher than their counterparts (AOR = 2.79; 95% CI: 2.14, 6.94), and the odds of self-harm among mothers who do not have postnatal follow-up were 2.72 times higher than those who follow their postnatal care (AOR = 2.72; 95% CI: 1.28, 5.80).

Similarly, the odds of self-harm among mothers who have loneliness were 2.49 times higher than their counterparts (AOR = 2.49; 95% CI: 1.15, 5.40), and the odds of self-harm among mothers who suffered from postpartum intimate partner violence were 2.15 times higher than those who did not face the problem (AOR = 2.15; 95% CI: 1.01, 4.54). Finally, the study identified that the odds of self-harm among mothers with average monthly income below the poverty line were 3.70 times higher than those who were above the poverty line (AOR = 3.70; 95% CI: 2.17, 10.50) (Table 4).

Since it was a cross-sectional study, it does not identify cause-and-effect relationships. In addition, since it was an institution-based study, it might lead to an underestimation of the magnitude of self-harm thoughts. The study focused on the neglected areas that hold great significance for reducing preventable causes of maternal deaths and assessed important factors that have a great impact on maternal health.