Femicide in Ethiopia: an autopsy-based cross-sectional study

Abstract

Background

Femicide, the most extreme form of gender-based violence directed at women and girls, often lacks sufficient national statistical data in several countries. The United Nations Office on Drugs and Crime Global Study on Homicide reported that Ethiopian data were unavailable. Moreover, there is no published evidence regarding femicide cases. To address this gap, this study investigated femicide victims referred for medicolegal autopsies at St. Paul’s Hospital Millennium Medical College (SPHMMC) in Ethiopia.

Methods

A cross-sectional study was conducted to investigate all femicide cases (aged ≥ 15 years) that were observed over the course of one year, from January 1, 2021, to December 31, 2021, at SPHMMC. Data were summarized using descriptive statistics. Associations between categorical variables were assessed with the chi-square or Fisher’s exact test, and independent t-tests were used to compare means between groups. Statistical significance was determined as p < 0.05.

Results

There were 62 femicide cases, representing 10.7% of all female cases and 2.0% of all medicolegal deaths. Of the 62 femicide cases, 53.2% were classified as intimate femicide, with 87.9% involving current partners and 12.1% involving former partners. The remaining 46.8% were classified as non-intimate partner femicide. The average age of the victims was 33.4 years, with half between 21 and 30 years. Perpetrators averaged 30.5 years old, with the majority falling within the 31–40 age range. Quarrels and disputes (46.8%) were the most often cited motive, followed by jealousy (13%). Most femicides occurred at night (64.5%) within private residences (72.6%). A history of prior intimate partner violence was recorded in 39.4% of cases. Blunt trauma (50%) and sharp injuries (29%) were the most common mechanisms of injury.

Conclusion

Femicide in Ethiopia is a significant public health and human rights concern, with a high proportion linked to intimate partner violence. The predominance of documented prior abuse, blunt trauma, domestic settings, and nighttime homicides reflects both the private nature of violence and systemic barriers to prevention. Integrating robust risk-assessment and early-warning mechanisms within healthcare, justice, and community structures is essential for identifying women at imminent risk and averting escalation to femicide.

Introduction

Violence against women (VAW) is a widespread issue affecting women globally, regardless of location, culture, or socioeconomic status. This human rights violation encompasses diverse forms of abuse, including sexual, physical, economic, and emotional violence [1, 2]. Among the most extreme manifestations of VAW is femicide, which is defined as the intentional killing of women and girls because of their gender [2–4]. Femicide is a stark indicator of gender inequality and societal health, reflecting deep-rooted patriarchal norms and systemic failures to protect women’s rights [5]. Globally, femicide is recognized as a critical public health and human rights issue, with devastating consequences for individuals, families, and communities [3].

According to the United Nations Office on Drugs and Crime (UNODC), approximately 87,000 women were killed globally in 2017, with 58% of these homicides committed by family members or intimate partners [5]. This statistic underscores the alarming prevalence of VAW and the disproportionate risk faced by women in their own homes. The impact of femicide extends beyond the loss of life; it perpetuates cycles of trauma and violence within communities and undermines societal well-being. Despite its global significance, femicide remains underreported and poorly understood due to challenges in data collection, including inconsistent definitions, misclassification of cases, and limited forensic and medicolegal resources in many regions [3, 6].

The UNODC Global Study on Homicide (2019) highlights that the African region has one of the highest rates of femicide globally, with intimate partners or family members responsible for a significant proportion of these deaths [5]. This elevated risk can be attributed to a combination of factors, including entrenched patriarchal norms, gender inequality, poverty, and limited access to justice and support services for survivors of violence. In many African societies, traditions such as female genital mutilation, early/forced marriage, and bride abduction intensify the vulnerability of women and girls to violence [7].

Ethiopia, like many other African countries, faces significant challenges in addressing VAW. The Ethiopian Demographic and Health Survey revealed that 26% of women aged 15–49 years have experienced physical and sexual violence at least once in their lifetime [8]. Societal acceptance of domestic violence remains high, with 63% of women and 28% of men believing that a husband is justified in beating his wife under certain circumstances [8]. Cultural practices such as “telefa” (abduction for marriage) and early/forced marriages further perpetuate gender-based violence and limit women’s autonomy. Despite these challenges, Ethiopia has made strides in legal and policy reforms to protect women’s rights, including the 2005 Revised Family Code and the 2019 Domestic Violence Prevention and Response Law. However, the implementation and enforcement of these laws remain inconsistent, and access to justice and support services for survivors is often limited [9–11].

Despite the high prevalence of VAW in Ethiopia, there is a paucity of comprehensive data on femicide, particularly forensic and medico-legal data. According to the UNODC Global Study on Homicide (2019), data from Ethiopia are reported as unavailable [5]. This knowledge gap hinders the development of evidence-based policies and interventions to address femicide and protect women’s rights.

Medicolegal autopsy data provide valuable insights into the causes of death, patterns of injury, and circumstances surrounding homicides, offering an objective basis for understanding the nature and magnitude of femicide [4, 6, 12, 13]. The analysis of medicolegal autopsy data is crucial for understanding femicide cases and developing targeted prevention strategies in Ethiopia. This study aimed to address this gap by analyzing the autopsy reports, police records, and medical records of femicide victims referred to the SPHMMC for medico-legal autopsy. This study also aimed to provide data on femicide and improve the understanding of femicide in Ethiopia by examining victim and perpetrator characteristics, injury patterns, and their relationship. The findings can inform evidence-based policies, prevention programs, and legal reforms to reduce violence against women, while also highlighting the critical role of forensic systems in addressing femicide. Ultimately, this study seeks to raise awareness of femicide as a public health and human rights issue and to contribute to the promotion of gender equality in Ethiopia.im

Methods

Study area, period, and design

This study employed a cross-sectional study design to investigate all femicide cases aged ≥ 15 years to the Department of Forensic Medicine and Toxicology who were brought from various parts of Ethiopia for medicolegal autopsy Ethiopia between January 1, 2021, and December 31, 2021.

SPHMMC is a tertiary referral hospital and medical college in Addis Ababa, Ethiopia, operating under the governance of a board appointed by the Federal Ministry of Health. During the study period, postmortem services in Ethiopia were primarily concentrated in two centers: SPHMMC in Addis Ababa and Aider Hospital in the Tigray region. Ethiopia’s administrative structure during this period comprised twelve regional states and two chartered cities (Addis Ababa and Dire Dawa). The Department of Forensic Medicine and Toxicology at SPHMMC provides postmortem examination services for cases originating from most regions of Ethiopia, except Tigray and parts of the Afar region, which Aider Hospital typically serves [14]. This geographical distribution ensures a broad representation of femicide cases from many regions.

Case selection

All female cases aged 15 years and older that were brought to the Department of Forensic Medicine and Toxicology at SPHMMC for medicolegal autopsy during the study period were initially identified. The forensic autopsy records for these cases were comprehensively reviewed. Cases with a confirmed natural, suicidal, or accidental manner of death were excluded. Upon identification of such cases, the corresponding data from forensic files were selected, tracked, and collected daily. Subsequently, each remaining case was reviewed in detail using the complete forensic file (including autopsy reports, police requests, pre-autopsy interview records), police dockets, and crime scene investigation reports archived either by the forensic team or the police, to determine whether the death was a homicide. These records were maintained meticulously on a case-by-case basis. To gather comprehensive and well-documented information, data from police dockets were collected 2 to 3 years after the recorded events, allowing for the accumulation of detailed records and subsequent thorough review. When discrepancies existed between these data sources, the study team discussed each case thoroughly. Any case in which the evidence did not clearly indicate homicide was excluded from further analysis to ensure accuracy and case validity.

For this study, femicide was defined as the intentional killing of women and girls because of their gender [3, 4]. Cases were subsequently classified into intimate partner violence and non-intimate partner violence. Intimate partner violence (IPV)-related femicide was defined as the killing of a woman by a current or former intimate partner, including husbands, ex-husbands, boyfriends, or ex-boyfriends. Non-intimate partner violence (non-IPV) femicide was defined as the killing of a woman by an offender who was not an intimate partner [2, 4]. Police reports, forensic files, and crime scene analyses were used to develop an initial hypothesis about the presence of a gender-related motive or evidence of a sexual offense associated with the crime. A case was classified as femicide if any of the following criteria were met: (1) the perpetrator was a current or former intimate partner; (2) the perpetrator was another family member; (3) police records indicated a gender-related motive; (4) postmortem examination findings demonstrated gender-related features such as genital injury, sexual assault, or overkilling (extensive mutilation) suggestive of gender-based violence; or (5) the crime scene characteristics, including the place and manner of death or body disposal, supported a gender-motivated context. Female homicides without any indication of a gender-related motive, such as deaths resulting from robbery or other criminal acts, were excluded from the study.

Two independent forensic medicine experts screened all female homicide cases using a structured classification checklist to ensure consistency and accuracy. Any disagreements were resolved through consensus discussions, and if a consensus could not be reached, the case was reviewed by a third senior forensic medicine expert, who made the final determination.

Data sources and collection procedures

The police investigation files contained comprehensive sociodemographic and contextual information about both victims and perpetrators, including age, marital status, educational level, occupation, nature and duration of relationship, number of children, suspected motive, weapon used, and information about perpetrator intoxication or prior history of violence. The files also documented the initial legal progression of each case. When sociodemographic information on victims was incomplete in the police records, missing data were supplemented with information from autopsy reports or medical records.

All cases underwent a complete medicolegal autopsy performed by experienced forensic pathologists. The autopsy files provided detailed information about the pregnancy status, cause and mechanism of death, location and type of injuries, number of wounds, presence of defensive or genital injuries, and evidence of overkilling or other indicators of assault. Findings from ancillary laboratory investigations, including histopathology, toxicology, and biochemical analyses, were also reviewed and recorded. In addition to these, routine pre-autopsy family interviews and witness statements provided supplementary information about the victim–perpetrator relationship, history of abuse, and the events leading up to the death. Medical records, which were usually submitted to the forensic unit by the police, offered additional demographic information and, in some cases, documentation of prior intimate partner violence.

The motive was primarily inferred from police investigation summaries and witness statements and was corroborated using autopsy findings and pre-autopsy interview notes when documentation was incomplete or inconsistent. When more than one motive was reported, the primary motive as determined by the investigating officer was recorded. The suspect or perpetrator was defined as the individual identified by the investigating officer as primarily responsible for the killing. In cases involving multiple suspects, the sociodemographic data of the primary suspect were used for analysis.

Data analysis

Data were analyzed using SPSS version 27. Descriptive statistics were employed to summarize the data. Results were presented using text, tables, and figures as appropriate. The chi-square test was used to assess the association between categorical variables; however, for tables with expected cell counts less than 5, Fisher’s exact Test was used. Independent t-tests were used to compare mean values between two independent groups. Confidence intervals (95%) were calculated for key proportions. The level of statistical significance was set at p < 0.05.

Ethical considerations

This study received ethical approval from the St. Paul’s Hospital Millennium Medical College Institutional Review Board (Ethical clearance reference no: PM 23/524). In accordance with the Declaration of Helsinki and Ethiopian National Research Ethics Review Guideline, the need for informed consent to participate was waived by the Institutional Review Board (IRB). This waiver was granted based on the following considerations: As this study used only prospectively collected, anonymized data as part of routine postmortem procedures and involved no interaction with participants or families, the risk to individuals was minimal. Furthermore, the research posed no risk to participants, as all data were anonymized using unique identification codes to ensure confidentiality. The data was solely used for the purposes of this research, with the potential to contribute to improvements in forensic practice, inform national policy, and aid femicide prevention efforts. The Department of Forensic Medicine and Toxicology, School of Medicine, SPHMMC, Addis Ababa, Ethiopia, granted permission for data collection and publication following a full review of the study details, including the intended data use, methodology, and data security measures. After obtaining IRB approval, a support letter from SPHMMC was obtained. This letter facilitated obtaining formal permissions from the respective police departments/offices, granting the research team access to official police files for review. To gather comprehensive data, we collected information from police records 2 to 3 years after the recorded events. To ensure the confidentiality of the data, all information was anonymized upon extraction using unique identification codes. The resulting dataset was securely stored on a password-protected institutional computer, with access strictly limited to the research team to uphold privacy standards. Importantly, no direct interviews were conducted by the researchers; we relied solely on the analysis of documents obtained from police dockets.

Results

Sociodemographic characteristics of victims and perpetrators

During the 1-year study period (January 1 to December 31, 2021), 3,025 autopsies were conducted at SPHMMC. Of these, 581 cases involved female deaths, among which 62 cases were identified as femicide cases (aged ≥ 15 years), constituting 10.7% (95% Confidence interval [95% CI]: 8.3% to 13.5%) of female cases and 2.0% (95% CI: 1.6% to 2.6%) of total autopsied cases. Of the 62 femicide cases, 53.2% (n = 33; 95% CI: 40.5% to 65.7%) were classified as IPV-related femicide, with 87.9% (n = 29; 95% CI: 71.8% to 96.6%) involving current intimate partners and 12.1% (n = 4; 95% CI: 3.4% to 28.2%) involving former intimate partners. The remaining 46.8% (n = 29, 95% CI: 34.3% to 59.5%) of the femicide cases were classified as non-IPV femicide.

The mean age of the femicide victims was 33.4 years (SD = ± 15.4), with 50% (n = 31) aged 21–30 years. Younger victims (21–30 years) were more often killed by intimate partners (64.5%, n = 20), whereas older victims (41–50 years) were more often killed by non-intimate partners (60%, n = 3). Importantly, victims of intimate partner violence (IPV)-related femicide were significantly younger on average (30.2 years) than victims of non-IPV femicide (38.5 years) (P = 0.017). The majority of victims were married (53.2%, n = 33), and the most common occupation was housewife (35.5%, n = 22), followed by self-employed (21%, n = 13). IPV-related femicide was more common among married women (78.8%, n = 26) compared with non-IPV cases (24.1%, n = 7) (p < 0.05). Additionally, IPV-related femicide was more common among housewives (45.5%, n = 15) compared with non-IPV cases (24.1%, n = 7) (P = 0.034). In terms of education, 37.1% (n = 23) of the victims were illiterate, while 29% (n = 18) had attended primary school (Table 1). The majority of cases were from the Oromia region, accounting for 43.5%, followed by Addis Ababa with 29.0%. Slightly more cases were reported in rural areas (51.6%) than in urban areas (48.4%). Among urban cases, 54.5% (n = 18) were IPV-related, whereas 41.4% (n = 12) were non-IPV cases. In the rural areas, 45.5% (n = 15) were IPV-related, whereas 58.6% (n = 17) were non-IPV. However, no significant difference was found in the educational status (p = 0.374) and residence (p = 0.412) of victims between the IPV and non-IPV groups.

Table 1.

Sociodemographic characteristics of the victims and perpetrators

Variable	Category	Frequency (n)	Percentage (%)
Victim age	15–20	10	16.1
	21–30	31	50
	31–40	12	19.4
	41–50	5	8.1
	> 50	4	6.4
Residence	Rural	32	51.6
	Urban	30	48.4
Address by region	Oromia	27	43.6
	Addis Ababa	18	29
	SNNPR*	10	16.1
	Amhara	3	4.9
	Harar	2	3.2
	Sidama	2	3.2
Victim’s marital status	Married	33	53.2
	Unmarried	22	35.5
	Divorced/widowed	7	11.3
Victim’s educational status	No education	23	37.1
	Primary	18	29
	Secondary	10	16.1
	More than secondary	5	8.1
	Unknown	6	9.7
Victim occupation	Housewife	22	35.5
	Self-employed	13	21
	Farmer	9	14.5
	Government employee	8	12.9
	Student	6	9.6
	Daily laborer	4	6.5
Perpetrator age	21–30	12	19.3
	31–40	23	37
	41–50	15	24.2
	> 50	12	19.5
Perpetrator occupation	Farmer	25	40.3
	Self-employed	18	29
	Daily laborer	10	16.1
	Government employee	9	14.6
Perpetrator’s educational status	No education	14	22.6
	Primary	19	30.6
	Secondary	8	12.9
	More than secondary	3	4.8
	Unknown	18	29

*SNNPR- Southern Nations, Nationalities, and Peoples’ Region

The mean age of the perpetrators was 30.5 years (SD = ± 18.42), with the most frequent age group being 31–40 years (37%, n = 23). The perpetrators of IPV-related femicide (mean age: 34.2 years) were significantly older than the non-IPV perpetrators (mean age: 28.7 years) (p = 0.038). The majority of perpetrators were farmers (40.3%, n = 25), and 30.6% (n = 19) had attended primary school (Table 1). However, no significant differences were found in education, occupation, or alcohol/drug use by the perpetrators between the IPV and non-IPV cases.

The circumstances surrounding the femicide

Most homicides occurred in private residences (72.6%, n = 45), typically in homes shared by the victim and perpetrator. Isolated locations, such as hotel rooms, forests, remote stretches of road, and bridges in secluded settings, accounted for 13% (n = 8) of the cases. The remaining cases occurred in public places (8%, n = 5) and workplaces (6.4%, n = 4). IPV-related femicide predominantly occurred in private residences (87.9%, n = 29), whereas non-IPV cases were more likely to occur in public places (27.6%, n = 8) (P = 0.002). The most common time of homicide was at night (64.5%, n = 40). However, no significant difference was observed in the time of homicide between the IPV and non-IPV groups (P = 0.541). The leading motive was quarrels/disputes (47%, n = 29), followed by jealousy/love (13%, n = 8) (Table 2). In intimate partner violence (IPV) cases, the average relationship length was 10.7 years (range: 1–40 years). The most common relationship lengths were 1–5 years (54.5%, n = 18), followed by 6–15 years (24.2%, n = 8), and more than 15 years (18.2%, n = 6). Furthermore, the relationship length was unknown in one case (3.1%). A history of prior intimate partner violence was recorded in 39.4% of IPV cases (n = 13 cases). Autopsy findings revealed that 9.1% (n = 3) of the IPV-related cases involved victims who were pregnant at the time of the incident, whereas none of the non-IPV cases were pregnant. Approximately 69.7% (n = 23) of IPV cases have children (Table 2). All suspected or identified primary perpetrators in this study were male. In the vast majority of cases, 91.9% (n = 57), a single perpetrator was identified or suspected. Multiple perpetrators were less common: two identified in 4.8% (n = 3) of cases, while three or four perpetrators were identified in one case each. Alcohol use by the perpetrators at the time of the incident was documented in 22.6% (n = 14) of the cases, while 48.4% (n = 30) did not use alcohol or other drugs. No information was available for the remaining 29% (n = 18) of cases. Most victims (75.8%, n = 47) did not receive professional medical care after the incident and died at the scene.

Table 2.

Circumstances surrounding the femicide

Variable	Category	Frequency	Percentage (%)
Relationship with the perpetrator	Current intimate partner	29	46.8
	Former intimate partner	4	6.4
	Family	9	14.5
	Acquaintances	12	19.4
	Strangers	8	12.9
Location of the homicide	Private residence	45	72.6
	Isolated place (private and secluded)	8	13
	Public place	5	8
	Workplace	4	6.4
Time of homicide	Night (18:00–05:59)	40	64.5
	Morning (06:00–11:59)	12	19.4
	Afternoon (12:00–17:59)	10	16.1
Motive	Quarrel/dispute	29	47
	Jealousy/love	8	13
	Economic problems	5	8
	Unknown	20	32

Mechanism and location of the injuries

The most common type of injury was blunt trauma (50%, n = 31), followed by sharp force injuries (29%, n = 18). Blunt trauma was more common in the IPV-related femicide (57.6%, n = 19) than in the non-IPV groups (41.4%, n = 12). Head injuries from blunt trauma were the most frequently encountered injury type overall (27.4%), occurring in 24.2% of IPV cases and 31.0% of non-IPV cases. Sharp injuries showed equal frequency between IPV and non-IPV cases (27.3% vs. 31.0%). Firearm injuries were more commonly reported in non-IPV cases (17.2%) than in the IPV cases (3.0%). Additionally, alcohol use by the perpetrators was significantly associated with blunt trauma (p = 0.033). Among perpetrators who used alcohol, 71.4% (n = 10) caused blunt trauma, compared with 60% (n = 18) of those who did not use alcohol. However, no significant difference was observed in both the mechanism and location of injuries between the IPV and non-IPV groups (P > 0.05). Comparisons between IPV and non-IPV femicides according to the mechanism and location of injuries are detailed in Table 3.

Table 3.

Comparisons between IPV and non-IPV femicides according to the mechanism and location of fatal injuries

Variables	IPV (n = 33)	Non-IPV (n = 29)	Total (n = 62)
Mechanism and location of the injuries
Blunt trauma	19 (57.6%)	12 (41.4%)	31 (50.0%)
Abdomen	5 (15.2%)	1 (3.4%)	6 (9.7%)
Chest	2 (6.1%)	0 (0%)	2 (3.2%)
Head	8 (24.2%)	9 (31.0%)	17 (27.4%)
Multiple sites	4 (12.1%)	2 (6.9%)	6 (9.7%)
Sharp force trauma	9 (27.3%)	9 (31.0%)	18 (29.0%)
Neck	2 (6.1%)	2 (6.9%)	4 (6.45%)
Chest	2 (6.1%)	2 (6.9%)	4 (6.45%)
Neck and chest	1 (3.0%)	1 (3.4%)	2 (3.2%)
Multiple sites	4 (12.1%)	4 (13.8%)	8 (12.9%)
Firearm injury	1 (3.0%)	5 (17.2%)	6 (9.7%)
Asphyxia	4 (12.1%)	3 (10.3%)	7 (11.3%)
Strangulation	3 (9.1%)	3 (10.3%)	6 (9.7%)
Smothering	1 (3.0%)	0 (0%)	1 (1.6%)

The initial legal case progression

The majority of suspects or perpetrators (66.1%, n = 41) were charged and on trial, while there is no information available regarding the legal outcome for 32.2% (n = 20). In one case, the perpetrator committed suicide.

Discussion

This study investigated cases of femicide among women aged 15 years and older in Ethiopia in 2021, focusing on injury patterns, causes of death, and the role of IPV. Of the 62 femicide cases identified, representing 10.7% of all medicolegal female deaths, 53.2% (n = 33) were classified as IPV-related femicide, with 87.9% (n = 29) involving current intimate partners and 12.1% (n = 4) involving former intimate partners. The remaining 46.8% (n = 29) were classified as non-IPV femicides. These findings align with global trends in which intimate partners are responsible for a significant proportion of femicide, particularly in less developed countries [6, 15, 16]. The high prevalence of femicide, particularly IPV-related femicide, underscores the urgent need for targeted interventions to address intimate partner violence in Ethiopia.

The proportion of femicide, particularly IPV-related femicide, in this study was slightly higher than that reported in Jordan and Algeria but lower than that observed in South Africa [4, 13, 17, 18]. These disparities likely reflect variations in cultural norms, gender dynamics, and legal protections against gender-based violence across these regions. However, direct comparisons across countries are problematic due to methodological differences, varying reporting mechanisms, and divergent study settings. Femicide remains underreported and underdocumented in many low- and middle-income countries, including Ethiopia, highlighting the need for standardized national data collection systems to inform effective prevention strategies [3, 15].

The mean age of victims of IPV-related femicide was 30.2 years, which was significantly lower than that of non-IPV victims (38.5 years), highlighting the vulnerability of younger women to IPV. This finding is consistent with a previous study in Ireland that identified young women as having an increased risk of violence and death [16]. Most perpetrators were aged between 31 and 40 years, a pattern also observed in another similar study in Algeria [13]. The rural predominance (51.6%) further emphasizes geographic disparities in access to protective services and legal redress.

Blunt trauma was the most common injury pattern overall (50%), particularly in IPV-related cases (57.6%), with head injuries being the leading cause of death (27.4%). Sharp/stab injuries accounted for 29% (n = 18) of all femicide cases, whereas firearm injuries accounted for 9.7% (n = 6). This is inconsistent with previous similar studies conducted elsewhere, where sharp force trauma and firearm injuries were commonly reported as common injury patterns [12, 13, 16, 18, 19]. The lower rate of firearm injuries in our study compared to many other settings may be attributed to Ethiopia’s relatively strict firearms legislation and very low civilian firearm ownership (estimated at around 0.4 firearms per 100 people) [20, 21]. These contextual contrasts underscore the importance of country-specific prevention strategies. The predominance of blunt trauma in IPV cases may reflect the use of readily available weapons, such as household objects, in domestic settings. Murder-suicide cases were rare, noted in only one case in which the victim was killed by sharp force trauma.

Alcohol use by perpetrators was observed in 22.6% (n = 14) of cases, whereas 48.4% (n = 30) did not use alcohol or other drugs. Importantly, alcohol use by the perpetrators was significantly associated with blunt trauma (p = 0.033). Among the perpetrators who used alcohol, 71.4% (n = 10) caused blunt trauma, compared with 60% (n = 18) of those who did not use alcohol. This finding aligns with a previous study, where alcohol abuse increased the risk of partner violence eightfold and doubled the risk of femicide or attempted femicide [22]. Addressing alcohol abuse through public health interventions could be a critical component of femicide prevention strategies.

This study revealed critical patterns in the settings and motivations of femicide in Ethiopia, underscoring the lethal dangers women face within intimate relationships and private settings. The finding that 72.6% of all femicides occurred in private residences, and that this figure rises to 87.9% for IPV-related cases, strongly aligns with global evidence identifying the home as the most common and most dangerous setting for IPV-related femicides [1, 2]. The home, which should be a place of safety, is consistently identified in the literature as the most dangerous setting for IPV-related femicides. This stark reality complicates prevention efforts due to its inherent privacy and limited public oversight, necessitating evidence-based targeted interventions.

The timing of these homicides, predominantly at night (64.5%), further highlights the vulnerability of women during periods of social isolation and limited external assistance. Nighttime also corresponds with heightened emotional conflict and alcohol use, which have been linked to impulsive and escalatory patterns of violence in IPV cases. These conditions may facilitate impulsive or premeditated acts of violence, particularly in settings where victims are economically or physically constrained from seeking help.

Motivations such as quarrels/disputes (47%) and jealousy or possessiveness (13%) are consistent with previous studies identifying interpersonal conflict and control-based dynamics as the primary drivers of femicide [23]. These emotionally charged motives often reflect long-standing patterns of psychological abuse, coercive control, and unresolved conflict, rather than random or opportunistic violence.

Critically, 39.4% of IPV-related femicide victims in our study had a documented history of prior partner violence, which is consistent with international findings that past abuse is one of the most robust predictors of intimate partner femicide [1, 2]. The observed patterns suggest that femicide is frequently the culmination of a cumulative, unresolved pattern of intimate partner violence, rather than an isolated or spontaneous event. This finding underscores a major systemic failure in identifying and responding to escalating risk in abusive relationships, representing missed intervention opportunities. Furthermore, this issue is further complicated by Ethiopia’s socio-cultural context, where domestic violence is often normalized. According to the 2016 Demographic and Health Survey, 63% of women and 28% of men justified wife-beating [8]. Such attitudes can impede efforts to prevent femicide by discouraging help-seeking behaviors, influencing community and police responses, and perpetuating a culture of silence surrounding abuse. Addressing this crisis necessitates urgent, coordinated, and evidence-based interventions designed to challenge the normative acceptance of violence, promote safer pathways for disclosure and help-seeking, and ultimately, prevent future femicides.

The implications arising from this study are crucial in disrupting potentially fatal cycles of violence. First, community-level initiatives must prioritize funding for IPV prevention, including structured risk assessment protocols and community-based support networks. Second, there is an urgent need for community-based early warning systems, including confidential reporting channels and protective orders for women at risk of IPV. Third, community organizations and the media should lead awareness campaigns to promote healthy relationships, de-escalation skills, and bystander intervention. Fourth, law enforcement should enhance the capacity to recognize IPV patterns, enforce protective orders, and apply lethality screening tools. Furthermore, healthcare systems should implement routine IPV screening protocols in clinical settings, along with established referral pathways to specialized support services.

A key strength of this study is the use of triangulated medicolegal data sources (autopsy, police, and medical records), which together improve the internal reliability and validity. This approach is consistent with studies in Italy and South Africa, where forensic data have been instrumental in understanding femicide [4, 6, 12, 13]. The other strength of this study lies in its thorough approach to data collection. To gather comprehensive and well-documented information, data from police dockets were collected 2 to 3 years after the recorded events, allowing for the accumulation of detailed records and subsequent thorough review. However, this study has limitations, including the potential underreporting of femicide cases, especially in rural areas, due to cultural stigma and limited access to forensic services. While this research was conducted at a national referral hospital over the course of one year, it primarily captured cases referred from six regions. The high proportion of cases originating from Addis Ababa and the nearby Oromia region likely reflects the proximity to St. Paul’s Hospital Millennium Medical College (SPHMMC) rather than a true representation of geographic distribution. Consequently, femicide cases in more remote areas may be underrepresented due to transportation challenges and investigative limitations, introducing a referral bias that limits the national generalizability of our findings. Additionally, the study excluded cases involving individuals under the age of 15, which restricts our understanding of femicide across all age groups.

Secondly, the inherent incompleteness and occasional disagreement within these records posed challenges. Specifically, reliance on such documentation limited the independent verification of crucial contextual details, such as perpetrator intoxication, precise motives, or prior psychosocial dynamics, potentially leading to information bias or misclassification of some femicide cases. Moreover, the relatively small sample size limited our ability to conduct extensive subgroup analyses. The small cell sizes observed in certain comparisons necessitate caution in interpretation and may affect the robustness of our conclusions. Finally, the cross-sectional design of this study precludes the establishment of causal relationships between the identified factors and femicide. Despite these limitations, this study provides the first forensic evidence-based description of femicide in Ethiopia and establishes a crucial framework for future multi-center, longitudinal research with standardized national protocols.

Conclusion

This descriptive study offers vital insights into the sociodemographic characteristics, injury mechanisms, and contextual factors surrounding femicide in Ethiopia. It highlights the troubling prevalence of IPV-related femicide, particularly among young, married women residing in rural settings. The predominance of blunt trauma, domestic settings, and nighttime homicides reflects both the private nature of violence and systemic barriers to prevention. Most notably, the significant proportion of cases with documented prior abuse illustrates the urgent need for improved early intervention strategies. Femicide is not an isolated or spontaneous act but is often the result of a cumulative, unresolved IPV.

Although the data are derived from a single referral center and cover only a one-year period, the results may not be nationally representative. However, they offer crucial initial insights rather than a comprehensive national picture. Integrating robust risk-assessment and early-warning mechanisms within healthcare, justice, and community structures is essential for identifying women at imminent risk and averting escalation to homicide. Furthermore, strengthening national forensic and medico-legal data systems, standardizing the documentation of gender-related killings, and fostering improved linkages between police, health, and justice services are crucial steps toward prevention.

Abbreviations

CI

Confidence interval

IRB

Institutional Review Board

IPV

Intimate partner violence

SPHMMC

St. Paul’s Hospital Millennium Medical College

UNODC

United Nations Office on Drugs and Crime

VAW

Violence against women

Authors’ contributions

ASL, JSS and KKJ conceptualized and designed the study. ASL, JSS, KKJ and SKRS performed data visualization and analysis, interpreted the findings, and critically reviewed and edited the manuscript. ASL, JSS, and KKJ supervised every component of the data collection process. ASL and JSS drafted the original manuscript. Finally, all authors critically reviewed and approved the final version of the manuscript.

Funding

None.

Data availability

All data collected for this study were analyzed and are presented in the manuscript. The complete dataset and additional supporting materials are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study received ethical approval from the St. Paul’s Hospital Millennium Medical College Institutional Review Board (Ethical clearance reference no: PM 23/524). In accordance with the Declaration of Helsinki and Ethiopian National Research Ethics Review Guideline, the need for informed consent to participate was waived by the Institutional Review Board. This waiver was granted based on the following considerations: As this study used only prospectively collected, anonymized data as part of routine postmortem procedures and involved no interaction with participants or families, the risk to individuals was minimal. Furthermore, the research posed no risk to participants, as all data were anonymized using unique identification codes to ensure confidentiality. The data was solely used for the purposes of this research, with the potential to contribute to improvements in forensic practice, inform national policy, and aid femicide prevention efforts. The Department of Forensic Medicine and Toxicology, School of Medicine, SPHMMC, Addis Ababa, Ethiopia, granted permission for data collection and publication following a full review of the study details, including the intended data use, methodology, and data security measures. After obtaining IRB approval, a support letter from SPHMMC was obtained. This letter facilitated obtaining formal permissions from the respective police departments/offices, granting the research team access to official police files for review. To gather comprehensive data, we collected information from police records 2 to 3 years after the recorded events. To ensure the confidentiality of the data, all information was anonymized upon extraction using unique identification codes. The resulting dataset was securely stored on a password-protected institutional computer, with access strictly limited to the research team to uphold privacy standards. Importantly, no direct interviews were conducted by the researchers; we relied solely on the analysis of documents obtained from police dockets.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.