Attacks on healthcare workers and healthcare facilities during the 2024 war on Lebanon: a geospatial information system (GIS) mapping study

Abstract

Introduction

Healthcare systems, including healthcare workers (HCWs), healthcare facilities and medical transport, are increasingly affected by violence in armed conflicts worldwide, despite being protected under International Humanitarian Law (IHL). Reported attacks on HCWs and facilities have been widely documented in the ongoing Israel–Lebanon war. This study examines the characteristics and geographic/temporal distribution of reported attacks affecting the Lebanese healthcare system during the 2024 Israeli war on Lebanon.

Method

This retrospective observational study analysed data on reported attacks affecting Lebanon’s healthcare system between 1 January and 5 December 2024. Incident records compiled by Insecurity Insight were retrieved from multiple open sources, cleaned, classified, geocoded and cross-referenced with national data from the Lebanese Ministry of Public Health. A descriptive geospatial information system mapping and monthly trend analysis were conducted to examine the geographic distribution, concentration and temporal pattern of documented incidents.

Results

A total of 496 unique incidents of reported attacks affecting Lebanon’s healthcare system were included in the final analytic dataset. These incidents were associated with 729 deaths and 590 injuries among HCWs. Descriptive spatial analysis showed the largest concentration of documented incidents and HCW casualties in southern Lebanon, including South and Nabatiyeh governorates, with additional incidents documented in the Beqaa, Mount Lebanon and Beirut. A temporal distribution of monthly incident counts showed a marked increase in documented incidents during September and October 2024.

Conclusion

The findings document a substantial number of reports of HCWs killed and healthcare facilities damaged or destroyed by Israeli attacks in the 2024 war. These findings urgently call for stronger measures to safeguard HCWs and systems in conflict zones, alongside systematic documentation efforts to inform future accountability.

WHAT IS ALREADY KNOWN ON THIS TOPIC.

WHAT THIS STUDY ADDS

• This study documents 496 unique reported Israeli attacks affecting Lebanon’s healthcare system between 1 January and 5 December 2024, resulting in 729 deaths and 590 injuries among healthcare workers.

• Descriptive geospatial mapping showed the largest concentration of documented incidents and healthcare worker casualties in southern Lebanon, including South and Nabatiyeh governorates.

• A temporal distribution of monthly incident counts showed a marked increase in documented incidents during September and October 2024.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• The findings highlight the need for stronger healthcare protection measures, real-time monitoring of attacks on healthcare and systematic documentation to support preparedness, recovery and future accountability efforts.

Introduction

Healthcare systems represent the cornerstone of civilian infrastructure and are explicitly protected under International Humanitarian Law (IHL), including the Geneva Conventions and their Additional Protocols.¹ IHL establishes fundamental protections for healthcare infrastructure during armed conflicts through three core principles: distinction between civilian and military objects, proportionality in assessing anticipated military advantage against civilian harm and precaution in minimising damage to protected sites.^{2 3} The Geneva Conventions and their Additional Protocols provide specific protections for medical personnel, medical units and medical transport during armed conflicts.^{1 2} Central to these protections is the principle of medical neutrality, which prohibits interference with the provision of healthcare in times of conflict.^{4 5} The WHO defines attacks on healthcare as ‘any act of verbal or physical violence or obstruction or threat of violence that interferes with the availability, access, and delivery of curative and/or preventive health services during emergencies’.⁶ Such acts violate both legal protections and fundamental humanitarian principles. Yet, despite the prohibitions established under IHL, attacks on healthcare in contemporary conflicts remain widespread and are increasing globally.⁷

In recent decades, attacks on healthcare, including violence against healthcare facilities, patients and healthcare workers (HCWs), have become common in armed conflicts.^8–13 In 2023 alone,

the Safeguarding Health in Conflict Coalition documented 2562 reported incidents of violence against HCWs or obstruction of healthcare across 30 countries or territories, representing a 25% increase from 2022; these reports documented 487 health workers killed and 625 incidents in which health facilities were damaged or destroyed.¹⁴

Repeated attacks targeting hospitals, clinics and HCWs across multiple settings resulted in the collapse of essential services, restricted access to care and weakening of emergency and referral systems.^{10 15–17} Such attacks not only lead to immediate destruction, but also cause displacement of HCWs, shortages in medical supplies and disruption of transportation systems.^{11 12} Prolonged conflict further obstructs the continuity of chronic disease care, disrupts preventive services and impedes recovery well beyond periods of active hostilities.¹⁸

During the sustained assault on Gaza, which numerous experts and human rights organisations have termed as a genocide,^19–24 hospitals have been destroyed as part of a systematic dismantling of the health system and other means necessary to sustain life for Palestinians in Gaza. A geospatial analysis, assessing damage to medical and non-medical buildings, revealed no significant difference in degree of damage and raised concerns about lack of distinction, proportionality and precaution in dealing with these protected institutions in potential violation of IHL.¹⁵

Lebanon has, for decades, endured the compounded effects of political instability, armed conflicts and economic crises.²⁵ During the 2006 Israeli war on Lebanon, healthcare delivery was severely disrupted by strikes affecting ambulances, hospitals and rescue teams.²⁶ During the current Israel war on Lebanon (beginning November 2023), reported attacks on healthcare have been geographically concentrated in specific regions including South Lebanon, which suffered the highest number of attacks, followed by the southern suburbs of Beirut and the Beqaa Valley. Attacks resulting in damage to healthcare facilities and injury or death to healthcare personnel have strained an already fragile healthcare system and depleted its limited resources.²⁷ Official reports document attacks affecting 68 hospitals, 63 primary healthcare centres (PHCs) and 237 emergency medical team units, leading to the closure of 8 hospitals and 58 PHCs and causing at least 217 deaths among HCWs and patients.²⁸ According to the United Nations (UN), attacks affecting healthcare in Lebanon were proportionally more lethal than those reported in other contemporary conflicts, including Ukraine and Gaza, with 47% of attacks resulting in the death of at least one HCW or patient.²⁹

The present study aims to describe the characteristics of reported attacks affecting healthcare facilities, HCWs and medical transport in Lebanon. It further seeks to map the geographic and temporal distribution and concentration of these documented attacks. Systematic documentation of attacks affecting HCWs and facilities is critical for strengthening emergency preparedness, guiding recovery efforts, enhancing the protection of healthcare and informing future accountability mechanisms.

Methods

Study design

This retrospective, observational study examined reported attacks affecting healthcare facilities, medical transport and HCWs in Lebanon during the 2024 war. The study period spanned from 1 January to 5 December 2024, which corresponded to the escalation of the war in Lebanon, during which Israeli military operations and airstrikes were concentrated in multiple regions, particularly South Lebanon, the southern suburbs of Beirut and the Beqaa Valley. The end date of 5 December was selected as it occurs 1 week after the announced cessation of hostilities on 27 November 2024. This date represented the latest date of reported incidents available in the Insecurity Insight dataset at the time of data extraction.

Within the context of this study, the WHO definition of attacks on healthcare⁶ was applied to reported attacks affecting healthcare facilities, healthcare personnel and medical transport, as recorded in the Insecurity Insight dataset. Healthcare facilities included hospitals, clinics and civil defence buildings, while medical transport included ambulances, fire trucks and medical supply vehicles. Other types of attacks on healthcare included in the broader WHO definition, such as denial of access to services, psychological violence, obstruction of care delivery, sexual assault and armed searches of healthcare assets, were not analysed because they were not reported in the Insecurity Insight dataset.

Data source

Incident data were obtained from Insecurity Insight’s publicly available datasets, Attacks on Health Care Data dataset,³⁰ which aggregate reports from multiple verified sources including non-governmental organisations (NGOs), UN agencies, the Lebanese Ministry of Public Health (MoPH), media outlets, satellite imagery and verified eyewitness documentation.

Attribution of incidents to Israeli military forces was based on the dataset’s ‘Reported Perpetrator Name’ variable and was treated as a characteristic of the source data rather than as a pre-specified inclusion criterion. The analysis focused on documenting reported incidents and their health-system impacts, without making legal determinations regarding intent, proportionality or responsibility.

Insecurity Insight records were cross-referenced with official MoPH records to enhance accuracy and eliminate duplication. As part of the analytical process, damage classification was applied specifically to transport assets, including ambulances, fire trucks and civil defence vehicles, where ‘damaged’ referred to vehicles that sustained partial but repairable damage, while ‘destroyed’ indicated total loss of functionality. Vehicle categories were based on the designations reported in the source dataset and cross-referenced records. ‘Ambulances’ referred to vehicles described as dedicated emergency medical response vehicles, ‘rescue vehicles’ referred to vehicles described as civil defence or technical rescue vehicles and ‘medical transport’ referred to vehicles described as transporting patients or medical resources. Information on official markings, emblems, lights, sirens or aerial visibility was not consistently available and therefore could not be independently verified.

The Insecurity Insight dataset was used as the primary analytical dataset. Official MoPH records documenting attacks on healthcare were reviewed and cross-referenced with Insecurity Insight records using incident date, location, affected health resource, reported outcomes and geolocation data. All 368 incidents obtained from MoPH were confirmed to have already been recorded in the Insecurity Insight dataset and were therefore used for validation instead of being incorporated as new incidents. The final analytic dataset included 496 unique documented incidents.

Data analysis

Insecurity Insight data were extracted into tables using Microsoft Excel version 2108 (Microsoft Corporation, Redmond, WA, USA) for cleaning and categorisation by location, date, weapon type, affected health resource, damaged/destroyed facility and vehicle, and recorded outcomes (injuries and deaths). Attacks were classified according to affected health resources, including: (1) facilities (hospitals, civil defence centres, medical facilities and related medical infrastructure) and (2) vehicles (ambulances, civil defence vehicles, motorcycles and other rescue units). The primary outcomes were unique reports of Israeli attacks affecting healthcare, including HCW deaths and injuries, healthcare facility damage or destruction and medical transport damage or destruction. Patient and civilian casualties were not included in the HCW casualty totals reported in this analysis. Each documented incident was counted once in the total number of unique incidents. However, when a single incident affected more than one healthcare asset or personnel category, it could contribute to more than one descriptive category, such as facility damage, vehicle damage or HCW casualties. Therefore, facility, transport and casualty categories should be interpreted as event characteristics rather than mutually exclusive incident counts. Reported casualties included only HCWs, such as physicians, nurses, paramedics, hospital staff, civil defence personnel and volunteers. Where available, casualties were further classified by HCW role or affiliation. Casualties without a reported role or affiliation were categorised as ‘HCW role not available’. Information on weapon type was also recorded when available. Missing or unspecified information was retained as ‘weapon type unknown’ and reported descriptively. Incident dates were aggregated by calendar month to describe the temporal distribution of documented incidents during the study period. December 2024 was interpreted as a partial month because the dataset included incidents through 5 December 2024 only.

All spatial analyses and coordinate-based processing were conducted using WGS 1984 UTM Zone 36N (EPSG:32636), which is appropriate for Lebanon and provides metric units for spatial measurement and analysis. The geocoded healthcare incident locations were stored and analysed in this projected coordinate reference system.

For map production, the analytical layers were displayed over the ArcGIS Online basemap, which uses WGS 1984 Web Mercator Auxiliary Sphere (EPSG:3857). ArcGIS Pro projected the analytical layers on the fly to align with the basemap for visualisation. Therefore, EPSG:32 636 was the working CRS used for spatial analysis, while EPSG:3857 was used only for final cartographic display. Location coordinates and spatial attributes were obtained from the Insecurity Insight dataset and were projected automatically during visualisation to align with the basemap coordinate system. The Lebanon administrative boundary was obtained from the default basemap and administrative layers available within ArcGIS Pro/ArcGIS Online. Mapped point locations therefore represent the geographic coordinates reported in the Insecurity Insight dataset. The dataset included a geo-precision variable with three reported levels: 25 km precision, district/municipality/commune-level precision and country-level precision. No records had missing latitude or longitude values; however, coordinate precision varied across records. Both facility-related incidents and vehicle/medical transport-related incidents, including ambulances, civil defence vehicles, motorcycles, rescue vehicles and other medical transport assets were included in the descriptive geospatial information system mapping. Each incident was mapped as a point location using the latitude and longitude coordinates provided in the Insecurity Insight dataset. For transport-related incidents, the mapped point represented the reported incident location rather than the full route or movement path of the vehicle. Therefore, coordinates were interpreted cautiously and were not treated as exact incident locations. The maps were used to describe the geographic distribution and concentration of documented incidents. Given the variation in geographic precision across source data, formal hotspot analyses, such as kernel density estimation or Getis-Ord Gi*, were not conducted because we were cautious that the spatial uncertainty may produce unreliable or misleading results.

Patient and public involvement

Patients and members of the public were not involved in the design, conduct, analysis or reporting of this study. The study relied on secondary incident-level data from publicly available datasets and official records.

Reporting guidelines

This study was reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines where applicable.

Results

A total of 496 unique reported incidents of Israeli attacks affecting the Lebanese healthcare system were included in the final analytic dataset between 1 January and 5 December 2024 (figure 1). These incidents resulted in 729 deaths and 590 injuries among HCWs (figure 1). Role classifications for HCW deaths and injuries are presented in table 1.

Figure 1. Geographic distribution of documented incidents affecting healthcare facilities and vehicles in Lebanon, 1 January–5 December 2024. Boxed areas indicate the regional maps presented in figure 3 and online supplemental figures 1 and 2.

Table 1. Impacts on HCWs, healthcare facilities and medical transport during Israel’s war on Lebanon (1 January–5 December 2024).

	Count (N)	Percentage
Type of HCW killed in attacks on healthcare facilities and medical transport
International HCWs outside the aid system	2	0.3
National staff of local NGOs	4	0.5
HCWs within the national health system	140	19.2
HCWs within the local health system	97	13.3
HCW role not available	486	66.7
Total	729	100
Type of HCW injured in attacks on healthcare facilities and medical transport
HCWs within the national health system	133	22.5
National staff of international aid agencies	36	6.1
HCW role not available	421	71.4
Total	590	100
Healthcare facility damage/destruction events by outcome
Damaged	195	93.8
Destroyed	13	6.2
Total	208	100
Medical transport damage/destruction events by outcome
Damaged	228	97.9
Destroyed	5	2.1
Total	233	100

All reported deaths and injuries refer to HCW casualties. ‘HCW role not available’ indicates HCW casualties for whom the specific role or affiliation was not reported in the source data.

HCW, healthcare worker; NGO, non-governmental organisation.

A temporal distribution of monthly incident counts showed relatively low numbers of documented incidents from January to August 2024, followed by a marked increase in September (N=170) and October (N=171). Incidents decreased in November (N=37) and early December (N=7), although December should be interpreted cautiously because data were available only through 5 December 2024 (figure 2).

Figure 2. Temporal distribution of monthly documented incidents of reported Israeli attacks affecting healthcare in Lebanon, 1 January–5 December 2024. December includes incidents documented through 5 December 2024 only.

Of the 496 unique documented incidents, 308 incidents (62%) were facility-related or involved healthcare institutions, while 188 incidents (38%) were vehicle/medical transport-related. Together, these two incident-level categories accounted for the full analytic sample. Within the 308 facility-related incidents, 208 incidents resulted in documented damage to or destruction of healthcare facilities, including partial damage (N=195, 93.8%) and complete destruction (N=13, 6.2%). Within the 188 vehicle/medical transport-related incidents, 233 medical transport damage/destruction events or affected assets were recorded, reflecting that more than one vehicle or transport asset could be affected in a single incident. These included partial damage to 228 vehicles/assets (97.9%) and complete destruction of 5 vehicles/assets (2.1%) (table 1). Among the 188 vehicle/medical transport-related incidents, the reported affiliation/service category was emergency medical services (EMS) in 171 incidents (91.0%), health and medical institutions in 6 (3.2%), rescue and civil protection in 5 (2.7%), humanitarian and relief operations in 2 (1.1%), fire and firefighting services in 1 (0.5%) and not available in 3 (1.6%) (table 2).

Table 2. Classification of affected medical transport by affiliation/service category during Israel’s war on Lebanon (1 January–5 December 2024).

Affiliation/service category	Count (N)	Percentage
Emergency medical services (EMS)	171	91.0
Health and medical institutions	6	3.2
Rescue and civil protection	5	2.7
Humanitarian and relief operations	2	1.1
Fire and firefighting services	1	0.5
Affiliation/service category not available	3	1.6
Total	188	100

Descriptive spatial analysis revealed a geographic concentration of documented attacks and HCW casualties in southern Lebanon, the Beqaa region and Beirut including its suburbs (figure 1). The South region, including South and Nabatieh Governorates, accounted for 559 deaths (76.7%) and 497 injuries (84.2%; figure 3). The Beqaa region, including Beqaa and Baalbek-Hermel Governorates, accounted for 133 deaths (18.2%) and 91 injuries (15.4%; online supplemental figure 2), while the Beirut region and its suburbs recorded 21 deaths (2.9%) and 2 injuries (0.3%; online supplemental figure 1). The remaining 16 deaths were documented in different regional locations across Lebanon.

Figure 3. Geographic distribution of documented incidents affecting healthcare facilities and vehicles in the South region, including South and Nabatiyeh Governorates.

The majority of incidents involved aerial bombings (eg, drones and aircraft; N=262, 52.8%). Artillery and shelling (N=12, 2.4%), firearm assaults (N=7, 1.4%) and explosives, including remote-controlled improvised explosive devices (N=6, 1.2%), were also reported. Weapon type was unknown for 209 incidents (42.1%).

Discussion

This study examined and characterised reported Israeli attacks affecting HCWs (HCWs), healthcare facilities and medical transport during the 2024 war on Lebanon by mapping their spatial patterns. Although IHL provides special protection for healthcare personnel, infrastructure, medical units and transport, findings from this study document repeated incidents affecting these protected health resources across various settings. Frontline HCWs and emergency responders are often exposed to heightened risks because of their presence in affected facilities and their involvement in emergency response activities. These patterns highlight persistent challenges in protecting healthcare in armed conflict and in ensuring the effective implementation and enforcement of international legal protections.

Consistent with our findings, similar patterns have been reported across multiple conflict settings, including Syria, Ukraine, Yemen, Afghanistan and Palestine, where volunteer HCWs and those registered in national health systems face a substantial burden of attacks.^31–34 In Lebanon, the WHO reported that 47% of attacks affecting healthcare resulted in the death of at least one HCW or patient.²⁹ The fixed locations and physical size of hospitals, together with the visibility of medical emblems and vehicles, may increase their exposure during military operations.^{8 11 35} In Lebanon, a large proportion of documented healthcare attacks affected medical transport, particularly ambulances. This pattern mirrors trends observed in Yemen, where strikes frequently resulted in mass casualties and disrupted emergency response, suggesting the use of highly destructive weaponry.^{11 14 33 34} Similarly, research from Lebanon and Palestine has documented ‘double-tap’ tactics, defined as sequential strikes during rescue operations.^{14 36} These repeated incidents involving emergency responders in conflict zones substantially underscore the persistent risks they face while delivering life-saving care, which weaken health system capacity by increasing harm to frontline healthcare personnel and disrupting rescue operations.^{37 38}

Similar to patterns observed in Ukraine, Palestine and Syria,^39–41 this study found that many attacks affecting healthcare infrastructure in Lebanon caused partial yet operationally significant damage rather than complete destruction. These incidents included damage to windows, solar panels, oxygen systems, sensitive medical equipment and other critical infrastructure, which directly compromised essential patient care services such as intensive care and chemotherapy. This aligns with documented conflict trends where attacks on health services can disrupt medication cold chains and create barriers to critical, acute care, thereby weakening the overall health system and interrupting service delivery for local populations during prolonged crises.¹¹

The temporal pattern observed in this study was characterised by a marked increase in documented incidents during September and October 2024. It is consistent with evidence showing that attacks affecting healthcare often rise during broader escalations in hostilities, particularly when military activity intensifies.⁴² Reports from Lebanon describe a similar accumulation of attacks affecting healthcare during the 2024 escalation, including incidents involving emergency responders and medical transport.⁴³ These temporal patterns are particularly relevant as attacks during escalation periods can place additional pressure on already strained health systems by disrupting emergency response, referral pathways and continuity of care.

The geographic clustering of attacks in southern Lebanon reflects its position as a border frontline and principal zone of active conflict, where destruction of essential infrastructure during Israeli military operations obstructed access to healthcare. However, due to the lack of denominator data on the regional distribution of healthcare facilities, EMS vehicles, population exposure and conflict intensity, these findings are restricted to the interpretation of the geographic concentration of documented incidents rather than as estimates of regional risk. In Syria, concentrated strikes in opposition-held territories not only caused mass casualties but also accelerated health system collapse by forcing skilled personnel to flee.⁴⁴ Likewise, Yemen frontline governorates like Taiz faced repeated strikes affecting medical infrastructure, resulting in widespread disruption that extended beyond the immediate conflict.⁴⁵ Also in Palestine, particularly in Gaza, medical centres near active conflict areas experienced substantial damage during Israeli hostilities, leading to reduced healthcare capacity locally and increased strain on facilities in less affected areas.^{13 46} In Lebanon, attacks affecting healthcare also extended across multiple governorates, further weakening national healthcare infrastructure.

Collectively, these patterns underscore the gap between the legal protections provided to healthcare in conflict and their practical enforcement. Evidence from multiple conflict zones highlights that despite robust documentation by humanitarian agencies, accountability for such attacks remains rare, with limited legal prosecution.⁴⁷

This study has some limitations. Although the Insecurity Insight dataset, which includes incidents from secondary sources such as media reports, NGO updates and UN briefings, was compiled and validated, the data are prone to under-reporting (missing incidents), especially in high-risk areas with restricted access or disrupted communications. Some incident characteristics, including HCW role/affiliation and weapon type, were missing or not specified in the source data. The large proportion of HCW casualties with unavailable role information may reflect the diversity of personnel operating in healthcare roles during conflict, including private sector HCWs, informal volunteers, ambulance drivers, community members assisting with care or other individuals functioning in healthcare support roles outside formal institutional reporting systems. Therefore, the role/affiliation breakdown should be interpreted as characterising the documented subset only, and the true distribution across HCW categories cannot be determined from the available data. Weapon type missingness may also not be random, as incidents in areas with restricted media access or involving less distinctive weapon signatures may be disproportionately missing. As such, the distribution of reported weapon types likely represents a conservative estimate for some categories. These limitations may have led to underestimation or misclassification of specific patterns.

Furthermore, although latitude and longitude were available for all records, the geographic precision of coordinates varied and depended on the quality of original reporting. In addition, information on vehicle markings, emblems, lights, sirens or aerial visibility was not consistently available, limiting our ability to assess whether affected vehicles were officially marked or recognisable as medical or rescue vehicles from above. Triangulation with MoPH data improved completeness and strengthened validity, yet MoPH’s dependence on municipal reporting likely led to under-representation of incidents occurring in rural or hard-to-access areas, creating a reporting bias towards facilities.

Conclusion

This study provides systematic documentation and descriptive geospatial and temporal mapping of reported Israeli attacks affecting Lebanon’s healthcare system during the 2024 war. The findings document repeated incidents affecting HCWs, healthcare facilities and medical transport, with the largest concentration of documented incidents and HCW casualties observed in southern Lebanon. These findings underscore the need for stronger measures to protect healthcare in conflict settings, improve real-time monitoring of attacks on healthcare and support documentation efforts that may inform future accountability. Systematic documentation and geospatial analysis of reported Israeli attacks affecting HCWs and healthcare facilities are critical for strengthening preparedness for future emergencies, improving healthcare protection and informing accountability efforts. Further research should integrate qualitative accounts from community HCWs and local emergency teams to capture unreported attacks and contextualise their operational impact. Ultimately, protecting HCWs and healthcare facilities in conflict zones is both a legal obligation and a fundamental public health concern, requiring comprehensive research and coordinated action.

Data availability statement

Data are available in a public, open access repository.