Continental scale geographic analysis of hospital exposure to natural hazards in the Americas

Abstract

Natural hazards—including earthquakes, hurricanes, floods, landslides, and volcanic activity—pose substantial threats to healthcare systems in the Americas. This study aims to evaluate the exposure of primary, secondary, and tertiary hospitals across the Americas to natural hazards and to identify the most affected areas and hospital facilities. This study assembled a harmonized inventory of hospitals with emergency services (2017–2021) and quantified geographic exposure via GIS overlays. Hospital point locations were intersected with the ZC-NASA-CU multi-hazard zoning and hazard-specific layers (earthquakes, hurricanes, floods, landslides, volcanoes). We summarized exposure by deciles, proximity to hazard features, and multi-hazard overlap, and classified facilities by their highest exposure decile. No site-level engineering, vulnerability, or damage modeling was performed. Across 51 countries and territories, 20,396 hospitals were identified; 88.1% are exposed to ≥ 1 hazard. By hazard, floods potentially affect 85.1%, hurricanes 28.3%, and earthquakes 15.1% of hospitals. High-exposure zones encompass 13.6% of the continental area; 42.5% of hospitals fall within the highest exposure deciles (8–10), and 38.1% intersect ≥ 2 hazards. Notable clusters occur in Central America and the Caribbean, along the U.S. East Coast (hurricane–flood), and along the Pacific margin/Andes (earthquake–volcano–landslide). This study shows that hospitals across the Americas face substantial exposure to natural hazards, underscoring the urgent need to strengthen disaster risk reduction. A harmonized, continental-scale exposure map—built on reliable GIS data—enables evidence-informed screening, prioritization, and strategic investment, while detailed risk and vulnerability assessments are best conducted through site-specific studies.

Introduction

Disasters have shaped human history since its origin. The natural forces of earthquakes, volcanic eruptions, hurricanes, floods, wildfires, and other cyclic phenomena such as tornadoes, heat waves, or extreme weather fluctuations come with planetary changes and significant costs to human settlements, cities, countries, and populations¹. Since the beginning of civilization, disasters caused by natural phenomena have killed millions of people and displaced many more. Despite advances in climatology and geology, important knowledge gaps persist among institutional stakeholders, resulting in delays in resilient infrastructure planning². Although precise quantification is challenging, hydrometeorological and geophysical events account for considerable mortality and morbidity each year, with a large, often unmeasured burden of disability and displacement^{3, 4, 5–6}.

The Americas are among the most exposed regions globally^7,8, not only because of hazard frequency and intensity but also due to social and political vulnerability—marked high inequalities, political instability, and incomplete adherence to international disaster risk reduction (DRR) frameworks and recommended building safety standards^{9, 10–11}. Between 2000 and 2021, 2,791 disasters were recorded in the Americas, 73.7% linked to natural hazards¹². Hydrometeorological events were the most frequently reported, including 772 floods, 723 storms, 77 landslides, 94 earthquakes, and 42 volcanic events. These disasters resulted in approximately 276,011 deaths, 2,739,113 injuries, and affected ~ 284,348,599 people¹².

Hospitals are critical infrastructure that must sustain emergency care while maintaining routine services. Therefore, they must be accessible and safety standards substantially higher than conventional buildings^{13, 14–15}. Proper and coordinated hospital operations during and after a disaster can reduce loss of life and minimize the impact within the community¹⁶. Identifying hazards in hospital surroundings is essential to ensure continuous functionality during emergencies^17,18.

Understanding disaster risk—considering vulnerability, exposure of people and goods, environmental knowledge, and hazard characteristics—is a key priority of the Sendai Framework for DRR (2015–2030)¹⁹. Exposure assessment is a foundational first step that enables more granular vulnerability analyses. Collecting reliable spatial data such as geographic coordinates, hazard intensity maps, and health facility characteristics to be analyzed using Geographic Information Systems (GIS) is essential for estimating hospital risk and implementing evidence-based mitigation interventions²⁰.

Hospitals across North, Central, and South America, as well as the Caribbean, are exposed to a wide range of hazards, with significant differences in regional risk contexts. In North America, over 57% of hospitals are exposed to earthquakes, hurricanes, and floods²¹. In Mexico City, approximately half of public and two-fifths of private hospitals are exposed to seismic hazards²². The Caribbean region, with its high exposure to hurricanes and volcanic activity, exemplifies the challenges posed by overlapping hazards. However, the limited integration of geospatial data across countries has hindered regional analyses of hospital exposure. While some national assessments exist, a standardized continental-scale approach remains scarce.

Healthcare facilities can be severely affected by hazards—such as earthquakes, volcanic eruptions, excessive rainfall, river flooding, or tsunamis—even when risk has been assessed and hospitals are operating at full capacity, if adequate mitigation measures have not been implemented. The loss of electricity, breakdown of communication systems, or physical inaccessibility due to road destruction can compromise hospital functionality and worsen the consequences of disasters^13,23,24. Because the exact timing of earthquakes, volcanic eruptions, flash floods, and hurricanes cannot be predicted with precision²⁵, risk reduction therefore emphasizes identifying hazard zones and configuring a safe, connected hospital network. Despite available environmental data and local studies, systematic use in planning remains uneven, often constrained by financial and governance barriers^16,26.

Although site-level mitigation requires local engineering data, a continental-scale exposure map adds decision-relevant scientific and policy value: it (i) enables macro-prioritization by consistently flagging hospitals in top exposure deciles; (ii) reveals transboundary hotspot corridors where multiple hazards co-occur; (iii) supports network-level resilience (hub-and-spoke redundancy, mutual aid) rather than siloed facility upgrades; (iv) establishes a baseline for longitudinal monitoring under climate and demographic change; and (v) aligns with DRR frameworks that call for risk understanding at multiple scales to guide investments^{18,27, 28–29}. Accordingly, we present a continental, multi-hazard geospatial analysis of hospital exposure across the Americas as a screening tool to guide subsequent site-specific assessments. Primary question: To what extent are hospitals across the Americas exposed to natural hazards, as determined by their spatial proximity to zones of high hazard intensity? Sub-questions: (a) What is the spatial distribution of hospitals exposed to natural hazards across subregions? (b) To what degree do hospitals intersect multi-hazard zones (earthquakes, hurricanes, floods, landslides, volcanoes)? (c) Are exposure patterns associated with hospital categories (small/medium/large)? (d) How robust are exposure classifications when validated against hazard-specific historical datasets?

Materials and methods

Aim

The objective was to analyze and describe the geographic exposure of hospitals across North, Central, and South America, and the Caribbean, to natural hazards, in order to inform DRR planning.

Study design

The hospital dataset spans a consistent five-year period (2017–2021), during which facility-level characteristics were collected, and hospitals were georeferenced and mapped in a geographic information system (GIS). Hazard exposure was assessed by overlaying hospital locations onto standardized hazard zoning layers. Spatial analyses included distance calculations from hazard sources (e.g., earthquake fault lines, hurricane tracks, floodplains), identification of high-exposure clusters, and classification of hospitals based on their proximity to high-risk zones. Hazard datasets were covered varying historical periods depending on the phenomenon: earthquakes (1900–2020), hurricanes (1851–2020), floods (1985–2021), landslides (1996–2020), and volcanoes (Holocene activity).

Identification and characterization of hospitals

The standardized Pan American Health Organization (PAHO) database was used to identify hospitals and their emergency services, categorized hierarchically based on service level and management structure.

Inclusion criteria

Hospitals with emergency, surgery, and/or intensive care unit (ICU) services located in North, Central, and South America, as well as the Caribbean. These facilities must operate 24/7 and be listed in the standardized PAHO database (2017–2021).

Exclusion criteria

Records lacking precise geographic location data, missing classification information in the PAHO database, or corresponding to facilities that do not provide emergency or critical care services were excluded.

Sample

The sample includes 20,396 hospitals: 11,741 in North America, 7,644 in South America, 594 in the Caribbean, and 417 in Central America. All hospitals were identified using a database developed by the Health Emergencies Department of the PAHO (Fig. 1).

Fig. 1 [Images not available. See PDF.]

Spatial distribution of hospitals with emergency, surgery, and ICU services in the Americas. Figure created by the authors in ArcGIS Pro 2.9.0 (Esri; https://pro.arcgis.com). Hospital locations from the PAHO standardized database (2017–2021). Equidistant projection appropriate for continental scale.

Hazard datasets

Critical zones of natural phenomena (ZC-NASA-CU)

This zoning system, developed collaboratively by NASA and Columbia University, is based on the integration of historical and modeled data for various natural hazards^30,31. It uses geospatial analysis to assess exposure, incorporating parameters such as hazard frequency, intensity, and spatial distribution. These parameters are combined into a composite index that reflects the overall risk level for each area. The zones are classified into deciles, where deciles 1–7 indicate low to moderate exposure, and deciles 8–10 identify critical zones with the highest risk levels. This methodology ensures a standardized approach to identifying areas of elevated vulnerability across multiple hazard types.

Digital cartography, climate, and geological data

This dataset was used to validate and complement ZC-NASA-CU exposure estimates by providing updated hazard-specific information for earthquakes, hurricanes, floods, landslides, and volcanoes (Table 1). Sources include specialized environmental and geophysical archives covering earthquakes (1900–2020)³², hurricanes (1851–2020)³³, floods (1985–2021)³⁴, landslides (1996–2020)^35,36, and Holocene volcanic activity³⁷.

Table 1. Summary of secondary data sources.

Dataset name	Source	Hazard type	Period covered	Purpose in study
ZC-NASA-CU Critical Zones	NASA/Columbia University	Multi-hazard	1985–2003	Baseline dataset for multi-hazard exposure analysis, providing composite indices for various hazard types
Dartmouth Observatory Flood Data	Dartmouth Observatory	Floods	1985–2021	Used to validate ZC-NASA-CU flood exposure data and incorporate historical flood characteristics
NOAA NHC Cyclone Data	NOAA National Hurricane Center	Cyclones	1851–2020	Incorporated to complement ZC-NASA-CU cyclone exposure, focusing on historical cyclone intensity and tracks
NASA Landslide Susceptibility	NASA	Landslides	1996–2020	Used alongside ZC-NASA-CU data to refine landslide susceptibility and enhance accuracy
USGS Earthquake Catalog	United States Geological Survey	Earthquakes	1900–2020	Provided earthquake magnitude (≥ 5.0) and proximity data, complementing ZC-NASA-CU seismic exposure indices
Holocene Volcano Activity	Global Volcanism Program (GVP)	Volcanic Activity	Holocene	Used to identify hospitals within proximity to active volcanoes and analyze volcanic exposure

To identify hospitals with emergency services and characterize their level of exposure to natural hazards, the standardized PAHO database (compiled from global and national public sources between 2017 and 2021) was used^30,31. Additionally, historical records and updated data for each type of phenomenon were added, including earthquakes³², hurricanes^33,38, floods³⁴, landslides^35,36, and volcanic activity³⁹. These hazards were selected due to their prevalence in the Americas and their significant potential to disrupt healthcare infrastructure, compromise hospital functionality and threaten public health during emergencies.

Scenario and selection directory of hospital units

The selected hospitals were georeferenced⁴⁰, and their health service offerings were classified hierarchically according to the type of managing institution (e.g., public sector, private sector, social security, or non-governmental organization). The categorization was based on a collaborative care model between several countries and PAHO, generating three categories⁴¹: small hospitals, which offer general surgery and operate continuously; medium-sized hospitals, which include at least two basic specialties—such as maternity services or comprehensive maternal and child care centers; and large hospitals, which provide inpatient care in highly specialized medical disciplines and subspecialties. The final directory consisted of 20,396 hospital records and 77 variables, with specific exposure data.

Data analysis

Geographic analysis

Hospital exposure to natural hazards was examined with GIS-based overlays and proximity analyses⁴². To ensure consistent spatial measurement at continental scale, we applied an equal-distance map projection appropriate for long-range comparisons. This allowed reliable comparison of hospitals’ proximity to hazard zones across diverse latitudes and longitudes.

To address spatial uncertainty, raster-vector geoprocessing techniques were applied. Hospital locations (as points) were overlaid onto hazard zoning maps, which are vector or raster-based. A semi-quantitative scaling approach was used to harmonize data from the ZC-NASA-CU model with hazard-specific datasets, ensuring consistency in exposure classification. Overlay and proximity analyses incorporated complementary datasets such as authoritative earthquake catalogs, historical hurricane track archives, landslide susceptibility datasets, and flood event databases⁴³.

To ensure consistency between the multi-hazard ZC-NASA-CU exposure data and hazard-specific datasets, we implemented a harmonization step prior to overlay and proximity analyses. For example, historical flood records (1985–2021) were used to validate and refine flood exposure classifications. Similarly, cyclonic exposure was refined using both ZC-NASA-CU zoning and NOAA/NHC historical hurricane track and intensity data^33,38.

Landslide exposure was evaluated by overlaying ZC-NASA-CU landslide zones with hospital locations and integrating susceptibility data (type, severity, date)^35,36. For seismic hazard, exposure values from ZC-NASA-CU were complemented by historical earthquake data (M ≥ 5.0) from the USGS³², which allowed for distance-based risk estimation relative to epicenters. Volcanic exposure was assessed by identifying active volcanoes and delineating concentric impact bands (≤ 100 km) informed by ashfall and pyroclastic flow criteria^39,44.

Hospitals were then classified into three standard exposure categories—low, medium, and high based on ZC-NASA-CU zoning thresholds. This terminology aligns with established risk assessment frameworks and provides a consistent basis for spatial comparison. Exposure frequency and severity were analyzed for each hazard type, and results were visualized through detailed maps and tabular summaries to support DRR strategies. While mitigation remained the core focus of this study, preparedness considerations such as geographic isolation or proximity to multiple hazard types were also acknowledged as key elements influencing hospital vulnerability.

To enhance robustness and minimize selection bias, the dataset was independently downloaded and initially analyzed by a team from PAHO and subsequently reanalyzed by the UDLA team to confirm accuracy and reproducibility.

Finally, limitations of individual datasets including variability in resolution, historical coverage, and geolocation precision are explicitly addressed in the Discussion section. The complementary use of multiple hazard datasets was emphasized to enhance exposure accuracy and to address spatial and temporal variability across the continent.

Results

General analysis

Following the inclusion criteria, a total of 20,396 hospitals were identified across the Americas, with the majority located in North America (11,741), followed by South America (7,644), the Caribbean (594), and Central America (417). Among these, 5,070 hospitals offer highly specialized services, 11,598 provide up to two specialties, and 3,728 are small facilities with general surgery. Sector-wise, 10,587 hospitals belong to the private sector, 7,088 to the public, 890 to social security, and 1,831 to NGOs or other organizations.

Floods were the most widespread hazard (36.0%), followed by hurricanes (9.1%) earthquakes (5.4%), while landslides (2.3%) and volcanic eruptions impacted 4.7% of the case study area. High-exposure zones span 5.8 million km² (13.6% of the continent), with significant overlap in hydrometeorological threats in the northern hemisphere and along the Pacific coast. Identified hazard areas include 14,287 earthquake epicenters (M ≥ 5), 1,163 hurricane tracks, 1,275 flood zones, 5,558 landslide zones, and 270 active Holocene volcanoes.

Exposure to natural hazards in the Americas

Among the 20,396 hospitals in the Americas, 88.1% (n = 17,979) are exposed to any kind of natural phenomena (ZC-NASA-CU: 1-10). Regional analysis shows that 89.6% of hospitals in North America, 85.0% in South America, 93.4% in the Caribbean, and 99.8% in Central America are exposed to any hazard. Additionally, 38.1% of the identified hospitals are exposed to two or more natural hazards. In areas categorized as high exposure (ZC-NASA-CU: 8-10), 42.5% of hospitals are included.

Floods

Floods cover extensive continental areas, exposing 85.1% (17,354) of facilities to this hazard. Central America stands out, with 70% of its hospitals at a high level of exposure (Table 2). A total of 1,907 hospitals fall within critical flood zones (deciles 8–10) (Fig. 2, Map 2C).

Table 2. Hospitals exposed to floods by subregion and country, based on ZC-NASA-CU risk exposure zoning.

Country	Not exposure		Medium to low		High exposure		Total
	n	%	n	%	n	%	n
Central America	2	0.5%	123	29.5%	292	70.0%	417
El Salvador	0	0.0%	0	0.0%	66	100.0%	66
Guatemala	0	0.0%	3	3.3%	89	96.7%	92
Honduras	0	0.0%	3	4.5%	63	95.5%	66
Nicaragua	1	1.8%	17	30.9%	37	67.3%	55
Costa Rica	0	0.0%	30	54.5%	25	45.5%	55
Panama	0	0.0%	58	82.9%	12	17.1%	70
Belize	1	7.7%	12	92.3%	0	0.0%	13
South America	1,190	15.6%	3,247	42.5%	3,207	42.0%	7,644
Colombia	4	0.9%	67	15.0%	377	84.2%	448
Ecuador	6	1.5%	76	19.1%	315	79.3%	397
Paraguay	0	0.0%	33	32.4%	69	67.6%	102
Bolivia (Plurinational State of)	9	4.1%	115	52.3%	96	43.6%	220
Chile	41	14.7%	116	41.7%	121	43.5%	278
Venezuela (Bolivarian Republic of)	12	2.2%	306	55.9%	229	41.9%	547
Peru	2	1.0%	116	57.7%	83	41.3%	201
Argentina	122	15.6%	356	45.6%	302	38.7%	780
Brazil	994	22.1%	1,940	43.1%	1 562	34.7%	4 496
Uruguay	0	0.0%	122	69.7%	53	30.3%	175
Caribbean	99	16.7%	253	42.6%	242	40.7%	594
Haiti	0	0.0%	5	6.3%	74	93.7%	79
Jamaica	0	0.0%	2	7.1%	26	92.9%	28
Dominican Republic	0	0.0%	35	20.1%	139	79.9%	174
Cuba	11	20.4%	40	74.1%	3	5.6%	54
Puerto Rico	0	0.0%	115	100.0%	0	0.0%	115
Antigua and Barbuda	0	0.0%	5	100.0%	0	0.0%	5
Aruba	0	0.0%	5	100.0%	0	0.0%	5
Barbados	0	0.0%	5	100.0%	0	0.0%	5
Dominica	0	0.0%	4	100.0%	0	0.0%	4
Guadeloupe	0	0.0%	3	100.0%	0	0.0%	3
Trinidad and Tobago	4	13.3%	26	86.7%	0	0.0%	30
Saint Lucia	2	66.7%	1	33.3%	0	0.0%	3
Guyana	24	77.4%	7	22.6%	0	0.0%	31
British Virgin Islands	1	100.0%	0	0.0%	0	0.0%	1
Sint Maarten	1	100.0%	0	0.0%	0	0.0%	1
Anguilla	1	100.0%	0	0.0%	0	0.0%	1
Curacao	1	100.0%	0	0.0%	0	0.0%	1
Montserrat	1	100.0%	0	0.0%	0	0.0%	1
Saint Martin	1	100.0%	0	0.0%	0	0.0%	1
Bonaire, Sint Eustatius and Saba	3	100.0%	0	0.0%	0	0.0%	3
Cayman Islands	3	100.0%	0	0.0%	0	0.0%	3
St. Kitts and Nevis	4	100.0%	0	0.0%	0	0.0%	4
Grenada	4	100.0%	0	0.0%	0	0.0%	4
French Guiana	4	100.0%	0	0.0%	0	0.0%	4
Martinique	5	100.0%	0	0.0%	0	0.0%	5
Bahamas	6	100.0%	0	0.0%	0	0.0%	6
U.S. Virgin Islands	6	100.0%	0	0.0%	0	0.0%	6
Saint Vincent and the Grenadines	6	100.0%	0	0.0%	0	0.0%	6
Suriname	11	100.0%	0	0.0%	0	0.0%	11
North America	1,751	14.9%	6,324	53.9%	3,666	31.2%	11,741
Mexico	719	14.9%	2,231	46.3%	1,866	38.7%	4,816
United States of America	639	10.1%	3,919	61.6%	1,800	28.3%	6,358
Bermuda	0	0.0%	4	100.0%	0	0.0%	4
Canada	393	69.8%	170	30.2%	0	0.0%	563
Total	3,042	14.9%	9,947	48.8%	7,407	36.3%	20,396

Fig. 2 [Images not available. See PDF.]

Flood hazard and hospital exposure in the Americas. Comparative map integrating: (A) Flood exposure zones according to ZC-NASA-CU data (1985–2003) and flood frequency and magnitude based on Dartmouth Observatory data (1985–2021); (B) Number of hospitals exposed to flood-prone zones based on PAHO hospital data (2021). Figure created by the authors in ArcGIS Pro 2.9.0 (Esri; https://pro.arcgis.com). Data sources: ZC-NASA-CU Global Multihazard Frequency and Distribution (1985–2003); Dartmouth Flood Observatory—Global Active Archive of Large Flood Events (1985–2021); PAHO hospital locations (2017–2021). Equidistant projection.

Hurricanes

In total, 28.3% (5,762) of hospitals are exposed to hurricanes (Fig. 3A, B). The highest exposure is in the Caribbean, with 88% (4,610) of the units, and 29.6% of these located in high exposed areas. In contrast, exposure in South America is nearly zero (Table 3, Fig. 3C). Additionally, 98 hospitals that have experienced hurricanes and floods are located less than 1 km from the nearest coast, with an altitude of less than 10 m and a slope of less than 2 degrees. Of these, 28% (27) are on the east coast of North America. The maps in Fig. 2 show a similarity between the ZC-NASA-CU zonation (Map 3A) and the 120-year hurricane tracks (Fig. 3A, B). In Fig. 2C presents the hospitals exposed to cyclonic activity by exposure level, with six hospitals at the maximum level.

Fig. 3 [Images not available. See PDF.]

Hurricane hazard and hospital exposure in the Americas. Comparative maps: (A) Hurricane exposure zones based on ZC-NASA-CU data (1985–2003); (B) Tracks and intensity of Category 1–5 hurricanes near the Americas from NOAA NHC (1852–2020); (C) Hospitals exposed to cyclone hazard zones and proximity to hurricane paths, PAHO data (2021). Figure created by the authors in ArcGIS Pro 2.9.0 (Esri; https://pro.arcgis.com). Data sources: ZC-NASA-CU cyclone exposure (1985–2003); NOAA/NHC IBTrACS (1851–2020); PAHO hospital locations (2017–2021). Equidistant projection.

Table 3. Hospitals Exposed to hurricanes by subregion and country, based on ZC-NASA-CU risk exposure zoning.

Country	Not exposure		Medium to low exposure		High exposure		Total
	n	%	n	%	n	%	n
Caribbean	73	12.3%	345	58.1%	176	29.6%	594
Puerto Rico	0	0.0%	0	0.0%	115	100.0%	115
U.S. Virgin Islands	0	0.0%	0	0.0%	6	100.0%	6
Antigua and Barbuda	0	0.0%	0	0.0%	5	100.0%	5
Dominica	0	0.0%	0	0.0%	4	100.0%	4
St. Kitts and Nevis	0	0.0%	0	0.0%	4	100.0%	4
Guadeloupe	0	0.0%	0	0.0%	3	100.0%	3
Anguilla	0	0.0%	0	0.0%	1	100.0%	1
British Virgin Islands	0	0.0%	0	0.0%	1	100.0%	1
Montserrat	0	0.0%	0	0.0%	1	100.0%	1
Saint Martin	0	0.0%	0	0.0%	1	100.0%	1
Sint Maarten	0	0.0%	0	0.0%	1	100.0%	1
Cuba	0	0.0%	36	66.7%	18	33.3%	54
Bahamas	0	0.0%	4	66.7%	2	33.3%	6
Bonaire, Sint Eustatius and Saba	0	0.0%	2	66.7%	1	33.3%	3
Cayman Islands	0	0.0%	2	66.7%	1	33.3%	3
Trinidad and Tobago	23	76.7%	4	13.3%	3	10.0%	30
Haiti	0	0.0%	73	92.4%	6	7.6%	79
Dominican Republic	0	0.0%	171	98.3%	3	1.7%	174
Jamaica	0	0.0%	28	100.0%	0	0.0%	28
Saint Vincent and the Grenadines	0	0.0%	6	100.0%	0	0.0%	6
Barbados	0	0.0%	5	100.0%	0	0.0%	5
Martinique	0	0.0%	5	100.0%	0	0.0%	5
Grenada	0	0.0%	4	100.0%	0	0.0%	4
Saint Lucia	0	0.0%	3	100.0%	0	0.0%	3
Curacao	0	0.0%	1	100.0%	0	0.0%	1
Aruba	4	80.0%	1	20.0%	0	0.0%	5
French Guiana	4	100.0%	0	0.0%	0	0.0%	4
Suriname	11	100.0%	0	0.0%	0	0.0%	11
Guyana	31	100.0%	0	0.0%	0	0.0%	31
North America	6,759	57.6%	4,264	36.3%	718	6.1%	11,741
Bermuda	0	0.0%	0	0.0%	4	100.0%	4
United States of America	3,242	51.0%	2,586	40.7%	530	8.3%	6,358
Mexico	3,140	65.2%	1,498	31.1%	178	3.7%	4,816
Canada	377	67.0%	180	32.0%	6	1.1%	563
Central America	166	39.8%	237	56.8%	14	3.4%	417
Belize	0	0.0%	2	15.4%	11	84.6%	13
Honduras	6	9.1%	57	86.4%	3	4.5%	66
Nicaragua	3	5.5%	52	94.5%	0	0.0%	55
Costa Rica	5	9.1%	50	90.9%	0	0.0%	55
Guatemala	20	21.7%	72	78.3%	0	0.0%	92
El Salvador	64	97.0%	2	3.0%	0	0.0%	66
Panama	68	97.1%	2	2.9%	0	0.0%	70
South America	7,636	99.9%	8	0.1%	0	0.0%	7 644
Venezuela (Bolivarian Republic of)	542	99.1%	5	0.9%	0	0.0%	547
Colombia	445	99.3%	3	0.7%	0	0.0%	448
Paraguay	102	100.0%	0	0.0%	0	0.0%	102
Uruguay	175	100.0%	0	0.0%	0	0.0%	175
Bolivia (Plurinational State of)	220	100.0%	0	0.0%	0	0.0%	220
Chile	278	100.0%	0	0.0%	0	0.0%	278
Ecuador	397	100.0%	0	0.0%	0	0.0%	397
Argentina	780	100.0%	0	0.0%	0	0.0%	780
Brazil	4,496	100.0%	0	0.0%	0	0.0%	4,496
Peru	201	100.0%	0	0.0%	0	0.0%	201
Total	14,634	71.7%	4854	23.8%	908	4.5%	20,396

Landslides

In the Americas, 4.2% (852) of hospitals are in areas threatened by landslides, according to criteria from the ZC-NASA-CU study. Central America tops the list, with 4% of its units in areas of greatest exposure. In other regions, this exposure is less than 1% (Table 4). The maps in Fig. 4 show a partial overlap between ZC-NASA-CU slide zones (Fig. 4A) and the most recent NASA susceptibility models (Map 3B). They are similar on the western coast of the continent but do not coincide on the eastern coast, except on the Caribbean islands (Fig. 4C).

Table 4. Hospitals Exposed to Landslides according to subregion and country, based on ZC-NASA-CU risk exposure zoning.

Country	Not exposure		medium to ow exposure		High exposure		Total
	n	%	n	%	n	%	n
Central America	260	62.4%	139	33.3%	18	4.3%	417
Guatemala	49	53.3%	33	35.9%	10	10.9%	92
Costa Rica	17	30.9%	35	63.6%	3	5.5%	55
El Salvador	12	18.2%	51	77.3%	3	4.5%	66
Nicaragua	49	89.1%	5	9.1%	1	1.8%	55
Panama	63	90.0%	6	8.6%	1	1.4%	70
Honduras	57	86.4%	9	13.6%	0	0.0%	66
Belize	13	100.0%	0	0	0	0.0%	13
South America	7,311	95.6%	253	3.3%	80	1.0%	7,644
Ecuador	233	58.7%	97	24.4%	67	16.9%	397
Peru	186	92.5%	11	5.5%	4	2.0%	201
Colombia	339	75.7%	101	22.5%	8	1.8%	448
Argentina	778	99.7%	1	0.1%	1	0.1%	780
Bolivia (Plurinational State of)	207	94.1%	13	5.9%	0	0.0%	220
Chile	263	94.6%	15	5.4%	0	0.0%	278
Venezuela (Bolivarian Republic of)	532	97.3%	15	2.7%	0	0.0%	547
Paraguay	102	100.0%	0	0.0%	0	0.0%	102
Uruguay	175	100.0%	0	0.0%	0	0.0%	175
Brazil	4,496	100.0%	0	0.0%	0	0.0%	4 496
Caribbean	557	93.8%	35	5.9%	2	0.3%	594
Dominica	3	75.0%	0	0	1	25.0%	4
Dominican Republic	164	94.3%	9	5.2%	1	0.6%	174
Guadeloupe	2	66.7%	1	33.3%	0	0.0%	3
Haiti	60	75.9%	19	24.1%	0	0.0%	79
Trinidad and Tobago	28	93.3%	2	6.7%	0	0.0%	30
Jamaica	27	96.4%	1	3.6%	0	0.0%	28
Puerto Rico	112	97.4%	3	2.6%	0	0.0%	115
Anguilla	1	100.0%	0	0.0%	0	0.0%	1
Curacao	1	100.0%	0	0.0%	0	0.0%	1
British Virgin Islands	1	100.0%	0	0.0%	0	0.0%	1
Montserrat	1	100.0%	0	0.0%	0	0.0%	1
Saint Martin	1	100.0%	0	0.0%	0	0.0%	1
Sint Maarten	1	100.0%	0	0.0%	0	0.0%	1
Bonaire, Sint Eustatius and Saba	3	100.0%	0	0.0%	0	0.0%	3
Saint Lucia	3	100.0%	0	0.0%	0	0.0%	3
Cayman Islands	3	100.0%	0	0.0%	0	0.0%	3
Grenada	4	100.0%	0	0.0%	0	0.0%	4
French Guiana	4	100.0%	0	0.0%	0	0.0%	4
St. Kitts and Nevis	4	100.0%	0	0.0%	0	0.0%	4
Saint Vincent and the Grenadines	6	100.0%	0	0.0%	0	0.0%	6
Antigua and Barbuda	5	100.0%	0	0.0%	0	0.0%	5
Aruba	5	100.0%	0	0.0%	0	0.0%	5
Barbados	5	100.0%	0	0.0%	0	0.0%	5
Martinique	5	100.0%	0	0.0%	0	0.0%	5
Bahamas	6	100.0%	0	0.0%	0	0.0%	6
U.S. Virgin Islands	6	100.0%	0	0.0%	0	0.0%	6
Suriname	11	100.0%	0	0.0%	0	0.0%	11
Guyana	31	100.0%	0	0.0%	0	0.0%	31
Cuba	54	100.0%	0	0.0%	0	0.0%	54
North America	11,416	97.2%	317	2.7%	8	0.1%	11,741
Mexico	4,530	94.1%	281	5.8%	5	0.1%	4,816
United States of America	6,323	99.4%	32	0.5%	3	0.0%	6,358
Canada	559	99.3%	4	0.7%	0	0.0%	563
Bermuda	4	100.0%	0	0.0%	0	0.0%	4
Total	19,544	95.8%	744	3.6%	108	0.5%	20 396

Fig. 4 [Images not available. See PDF.]

Landslide hazard and hospital exposure in the Americas. Comparative maps: (A) Landslide exposure zones from ZC-NASA-CU (2000); (B) Landslide susceptibility, type, and severity using NASA landslide catalog data (1996–2020); (C) Hospitals exposed to severe landslide areas, PAHO hospital dataset (2021). Figure created by the authors in ArcGIS Pro 2.9.0 (Esri; https://pro.arcgis.com). Data sources: ZC-NASA-CU landslide zones (1985–2003); NASA Global Landslide Susceptibility (1996–2020); PAHO hospital locations (2017–2021). Equidistant projection.

Earthquakes

The proportion of hospitals exposed to seismic activity in the Americas is 15.1% (3,082). In areas with greater ZC-NASA-CU exposure, there are 3.8% (779) hospitals. According to the subregions, 47.5% of Central American hospitals are in critical areas (Table 5). The maps in Fig. 5 show a similar distribution of seismic activity between ZC-NASA-CU (Fig. 5A) and M ≥ 5 earthquakes (Fig. 5B), aligning along the tectonic plate boundaries on the western coast of the continent and in the Caribbean Island arc.

Table 5. Hospitals Exposed to earthquakes by subregion and country, based on ZC-NASA-CU risk exposure zoning.

Country	Not exposure		Medium to low exposure		High exposure		Total
	n	%	n	%	n	%	n
Central America	128	30.7%	91	21.8%	198	47.5%	417
El Salvador	5	7.6%	4	6.1%	57	86.4%	66
Costa Rica	1	1.8%	9	16.4%	45	81.8%	55
Guatemala	14	15.2%	14	15.2%	64	69.6%	92
Nicaragua	24	43.6%	17	30.9%	14	25.5%	55
Honduras	34	51.5%	18	27.3%	14	21.2%	66
Panama	38	54.3%	28	40.0%	4	5.7%	70
Belize	12	92.3%	1	7.7%			13
South America	6,292	82.3%	1,066	13.9%	286	3.7%	7,644
Chile	44	15.8%	136	48.9%	98	35.3%	278
Peru	35	17.4%	112	55.7%	54	26.9%	201
Ecuador	67	16.9%	239	60.2%	91	22.9%	397
Colombia	196	43.8%	243	54.2%	9	2.0%	448
Venezuela (Bolivarian Republic of)	325	59.4%	201	36.7%	21	3.8%	547
Argentina	715	91.7%	53	6.8%	12	1.5%	780
Bolivia (Plurinational State of)	152	69.1%	67	30.5%	1	0.5%	220
Brazil	4,481	99.7%	15	0.3%	0	0.0%	4,496
Paraguay	102	100.0%	0	0.0%	0	0.0%	102
Uruguay	175	100.0%	0	0.0%	0	0.0%	175
Caribbean	301	50.7%	273	46.0%	20	3.4%	594
Antigua and Barbuda					5	100.0%	5
Dominican Republic	48	27.6%	112	64.4%	14	8.0%	174
Trinidad and Tobago	4	13.3%	25	83.3%	1	3.3%	30
Saint Vincent and the Grenadines	0	0.0%	6	100.0%	0	0.0%	6
Aruba	0	0.0%	5	100.0%	0	0.0%	5
St. Kitts and Nevis	0	0.0%	4	100.0%	0	0.0%	4
Dominica	0	0.0%	4	100.0%	0	0.0%	4
Anguilla	0	0.0%	1	100.0%	0	0.0%	1
Montserrat	0	0.0%	1	100.0%	0	0.0%	1
Sint Maarten	0	0.0%	1	100.0%	0	0.0%	1
Saint Martin	0	0.0%	1	100.0%	0	0.0%	1
British Virgin Islands	0	0.0%	1	100.0%	0	0.0%	1
Saint Lucia	1	33.3%	2	66.7%	0	0.0%	3
Puerto Rico	48	41.7%	67	58.3%	0	0.0%	115
U.S. Virgin Islands	3	50.0%	3	50.0%	0	0.0%	6
Jamaica	15	53.6%	13	46.4%	0	0.0%	28
Haiti	55	69.6%	24	30.4%	0	0.0%	79
Bonaire, Sint Eustatius and Saba	2	66.7%	1	33.3%	0	0.0%	3
Guadeloupe	2	66.7%	1	33.3%	0	0.0%	3
Cuba	53	98.1%	1	1.9%	0	0.0%	54
Curacao	1	100.0%	0	0.0%	0	0.0%	1
Cayman Islands	3	100.0%	0	0.0%	0	0.0%	3
Grenada	4	100.0%	0	0.0%	0	0.0%	4
French Guiana	4	100.0%	0	0.0%	0	0.0%	4
Martinique	5	100.0%	0	0.0%	0	0.0%	5
Barbados	5	100.0%	0	0.0%	0	0.0%	5
Bahamas	6	100.0%	0	0.0%	0	0.0%	6
Suriname	11	100.0%	0	0.0%	0	0.0%	11
Guyana	31	100.0%	0	0.0%	0	0.0%	31
North America	10,593	90.2%	873	7.4%	275	2.3%	11,741
Mexico	4,132	85.8%	535	11.1%	149	3.1%	4,816
United States of America	5,911	93.0%	321	5.0%	126	2.0%	6 358
Canada	546	97.0%	17	3.0%	0	0.0%	563
Bermuda	4	100.0%	0	0.0%	0	0.0%	4
Total	17,314	84.9%	2,303	11.3%	779	3.8%	20,396

Fig. 5 [Images not available. See PDF.]

Seismic hazard and hospital exposure in the Americas. Comparative maps: (A) Seismic hazard zones from ZC-NASA-CU (1976–2002); (B) Historical earthquake epicenters from USGS catalog (1900–2020) overlaid with tectonic plate boundaries; (C) Hospitals exposed to seismic risk areas based on proximity to fault lines and seismic intensity, PAHO (2021). Figure created by the authors in ArcGIS Pro 2.9.0 (Esri; https://pro.arcgis.com). Data sources: ZC-NASA-CU seismic exposure (1985–2003); USGS Earthquake Catalog (M ≥ 5, 1900–2020); PAHO hospital locations (2017–2021). Equidistant projection.

Volcanoes

Within 100 km concentric rings from the 270 active volcanoes studied, 16.0% (3,265) hospitals were identified. Central America has the highest percentage of hospitals in areas of high ZC-NASA-CU volcanic exposure at 33.6%, followed by South America at 2.5% and North America at only 1.2%. In the Caribbean, 5.4% of hospitals are close to active volcanoes, but at a lower level of exposure (Table 6).

Table 6. Nearby hospitals Active volcanoes by subregion and country, based on ZC-NASA-CU risk exposure zoning.

Country	Not exposure		Low or moderate exposure		Critical exposure		Total
	n	%	n	%	n	%	n
Central America	152	36.5%	125	30.0%	140	33.6%	417
Costa Rica	5	9.1%	7	12.7%	43	78.2%	55
Guatemala	13	14.1%	28	30.4%	51	55.4%	92
El Salvador	0	0.0%	39	59.1%	27	40.9%	66
Nicaragua	9	16.4%	28	50.9%	18	32.7%	55
Honduras	42	63.6%	23	34.8%	1	1.5%	66
Belize	13	100.0%	0	0.0%	0	0.0%	13
Panama	70	100.0%	0	0.0%	0	0.0%	70
South America	7 100	92.9%	354	4.6%	190	2.5%	7 644
Ecuador	191	48.1%	41	10.3%	165	41.6%	397
Colombia	293	65.4%	139	31.0%	16	3.6%	448
Chile	125	45.0%	144	51.8%	9	3.2%	278
Peru	189	94.0%	12	6.0%	0	0.0%	201
Argentina	762	97.7%	18	2.3%	0	0.0%	780
Paraguay	102	100.0%	0	0.0%	0	0.0%	102
Uruguay	175	100.0%	0	0.0%	0	0.0%	175
Bolivia (Plurinational State of)	220	100.0%	0	0.0%	0	0.0%	220
Venezuela (Bolivarian Republic of)	547	100.0%	0	0.0%	0	0.0%	547
Brazil	4,496	100.0%	0	0.0%	0	0.0%	4,496
North America	9,317	79.4%	2,287	19.5%	137	1.2%	11,741
Mexico	2,541	52.8%	2,207	45.8%	68	1.4%	4,816
United States of America	6,217	97.8%	72	1.1%	69	1.1%	6,358
Canada	555	98.6%	8	1.4%	0	0.0%	563
Bermuda	4	100.0%	0	0.0%	0	0.0%	4
Caribbean	562	94.6%	32	5.4%	0	0.0%	594
Saint Vincent and the Grenadines	0	0.0%	6	100.0%	0	0.0%	6
Antigua and Barbuda	0	0.0%	5	100.0%	0	0.0%	5
Martinique	0	0.0%	5	100.0%	0	0.0%	5
Dominica	0	0.0%	4	100.0%	0	0.0%	4
St. Kitts and Nevis	0	0.0%	4	100.0%	0	0.0%	4
Guadeloupe	0	0.0%	3	100.0%	0	0.0%	3
Saint Lucia	0	0.0%	3	100.0%	0	0.0%	3
Montserrat	0	0.0%	1	100.0%	0	0.0%	1
Grenada	3	75.0%	1	25.0%	0	0.0%	4
Anguilla	1	100.0%	0	0.0%	0	0.0%	1
Curacao	1	100.0%	0	0.0%	0	0.0%	1
British Virgin Islands	1	100.0%	0	0.0%	0	0.0%	1
Saint Martin	1	100.0%	0	0.0%	0	0.0%	1
Sint Maarten	1	100.0%	0	0.0%	0	0.0%	1
Bonaire, Sint Eustatius and Saba	3	100.0%	0	0.0%	0	0.0%	3
Cayman Islands	3	100.0%	0	0.0%	0	0.0%	3
French Guiana	4	100.0%	0	0.0%	0	0.0%	4
Aruba	5	100.0%	0	0.0%	0	0.0%	5
Barbados	5	100.0%	0	0.0%	0	0.0%	5
Bahamas	6	100.0%	0	0.0%	0	0.0%	6
U.S. Virgin Islands	6	100.0%	0	0.0%	0	0.0%	6
Suriname	11	100.0%	0	0.0%	0	0.0%	11
Jamaica	28	100.0%	0	0.0%	0	0.0%	28
Trinidad and Tobago	30	100.0%	0	0.0%	0	0.0%	30
Guyana	31	100.0%	0	0.0%	0	0.0%	31
Cuba	54	100.0%	0	0.0%	0	0.0%	54
Haiti	79	100.0%	0	0.0%	0	0.0%	79
Puerto Rico	115	100.0%	0	0.0%	0	0.0%	115
Dominican Republic	174	100.0%	0	0.0%	0	0.0%	174
Total	17,131	84.0%	2 798	13.7%	467	2.3%	20,396

The PAHO exposure estimation incorporates geospatial data on hospital locations and their proximity to volcanic activity zones, considering the potential impact of eruptions, such as ash fallout and pyroclastic flows. This estimation complements the ZC-NASA-CU analysis by providing an institutional perspective tailored to health emergency preparedness and response (Fig. 6). The figure shows an alignment of active volcanoes along the western coast of the continent and in the Caribbean Island arc, parallel to the tectonic plate boundaries, similar to the distribution of earthquakes.

Fig. 6 [Images not available. See PDF.]

Volcanic hazard and hospital exposure in the Americas. Comparative maps: (A) zones of exposure to volcanic activity ZC-NASA-CU 1979–2000; (B) Active Holocene volcanoes GVP -10,000 and location with respect to tectonic plate boundaries; (C) estimate of hospitals exposed to volcanic activity, PAHO, 2021. Figure created by the authors in ArcGIS Pro 2.9.0 (Esri; https://pro.arcgis.com). Data sources: Smithsonian GVP—Holocene Volcanoes; ZC-NASA-CU volcanic exposure; PAHO hospital locations (2017–2021). Hospitals shown within ≤ 100 km of active Holocene volcanoes; equidistant projection.

Exposure of hospitals to multiple hazards in critical areas

Among the 8,667 hospitals located in critical hazard areas, 98 are situated in regions simultaneously exposed to both hurricanes and floods. These facilities are in high-risk zones due to their geographic location within 1 km of the coastline, at elevations below 10 m, and on terrain with slopes under 2 degrees. These exposure-related characteristics increase the potential for compound impacts such as storm surges and persistent flooding. Notably, 28% (27) of these hospitals are located along the east coast of North America. On the Pacific coast, hospitals are predominantly exposed to tectonic and volcanic activity. Central America accounts for 100 hospitals in these high-risk zones, with significant clusters in Costa Rica (38), Guatemala (36), and El Salvador (21). Additionally, 87 hospitals in South America face compounded risks from landslides and flooding, with the steep terrain and intense precipitation amplifying their vulnerability.

In the Caribbean, the Dominican Republic stands out as the only country with hospitals exposed to both landslides and severe cyclonic activity. Finally, in areas with overlapping volcanic and landslide risks, Ecuador emerges as the most affected country, with 70 hospitals situated in these dual-hazard zones. These findings highlight the interplay between geographic features and hazard exposure, emphasizing the need for targeted mitigation strategies that consider local topography and hazard types.

Discussion

The results of our investigation indicate that the Americas are highly exposed to numerous hydro meteorological and geological hazards. This exposure arises from the continent’s geodynamics and climatology—multiple interacting tectonic plates (North American, South American, Caribbean, Cocos, Nazca, and portions of the Pacific) and recurrent weather- and climate-related extremes. Major fault systems (e.g., the San Andreas Fault) and subduction along the Pacific Ring of Fire underpin the regional seismic and volcanic context; hospitals are located within this multi-hazard environment and are consequently exposed, rather than being a cause of exposure.

The reported impact of natural phenomena has increased globally, with three times as many disasters documented at the beginning of the twenty-first century compared to the 1980s⁴. However, this apparent rise may partly reflect improved reporting systems, greater global connectivity, and enhanced disaster monitoring, rather than an absolute increase in event frequency.

This increase primarily affects developing countries, where 97% of human losses are recorded, highlighting the vulnerability of Latin American nations to future events⁴⁵.

While the ZC-NASA-CU dataset was used for its comprehensiveness and accessibility, we acknowledge the limitations of its temporality and resolution compared to more advanced hazard models^43,46,47. Incorporating newer datasets, such as the European Commission Joint Research Centre’s flood hazard model or the Global Earthquake Model, could refine the results by providing higher-resolution spatial data and hazard estimates based on current conditions⁴³. These models represent significant advancements in hazard mapping and could enhance the precision of local and regional assessments. Future analyses could benefit from the integration of more recent hazard zoning datasets such as the Global Seismic Hazard Map or the JRC Global Flood Model, which offer higher spatial resolution and updated risk modeling.

Up to date this study concluded, it demonstrated that half of the continental area of the Americas is exposed to natural phenomena, and more than 40% of all hospitals are located in high exposure zones. Among these, Central America stands out for its high hospital exposure to floods, landslides, earthquakes, and volcanic activity. Between 1990 and 2011, Latin America and the Caribbean recorded a total of 83,000 disasters, with an average of 3,772 events per year⁴⁸.

While short-term timing is inherently uncertain for many hazards—especially earthquakes—forecast skill varies by phenomenon. Weather-related events (e.g., hurricanes) can often be forecast in time and space with useful skill, although impact intensity and local effects remain uncertain; by contrast, earthquakes lack reliable short-term prediction, and risk management relies on long-term probability models and preparedness²⁵. Accordingly, our results emphasize screening exposure and prioritizing preparedness and mitigation. They also often follow certain geographic trends and show interrelated patterns²⁵. Hurricanes, for example, predominate north of the equator, influenced by atmospheric and oceanic circulation, seasonality, and the distribution of continental masses in the Americas. Earthquakes and volcanoes align with the edges of tectonic plates along the Pacific Ocean coast which constitute the Ring of Fire and in the Caribbean island arc⁷.

We also recognize that hazard-specific impacts on buildings vary substantially. For example, damage patterns from seismic events differ markedly from those of hurricanes or floods. While our study maps geographic exposure, it does not estimate hazard-specific vulnerability or potential structural damage. Thus, we caution against interpreting our results as a full disaster-risk assessment.

Our findings have important policy implications. At the screening stage, intersecting high-exposure deciles (8–10) with facility role enables macro-prioritization of hospitals whose disruption would cause disproportionate system-wide losses. In our sample, 42.5% of hospitals fall within high-exposure deciles and 38.1% face multiple hazards. Such overlaps are especially salient in Central America (clusters in Costa Rica, Guatemala, and El Salvador); along the U.S. East Coast (98 hospitals with combined hurricane–flood exposure within 1 km of the shoreline and < 10 m above mean sea level); across the Caribbean arc (e.g., the Dominican Republic with cyclones–landslides); and along the Pacific margin and the Andes (combinations of earthquake/volcano/landslide, with Ecuador notably hosting 70 hospitals under dual exposure). These patterns point to the need for integrated intervention packages (e.g., wind–flood; earthquake–landslide) and justify cross-border coordination (mutual aid, shared caches, interoperable protocols) where geography and health-system networks demand it^28,29,49.

More broadly, we propose a simple, reproducible screening model for the region. First, prioritize facilities located in high-exposure deciles (8–10) that also play a critical role in the referral network. Second, map each hazard to standard mitigation measures (e.g., tiered seismic retrofits; access elevation, improved egress, drainage, and flood barriers; protection against wind-borne debris) using publicly available unit-cost ranges (per m², per bed, per linear meter). Third, scale these unit costs with readily available proxies (gross floor area, bed count, footprint of critical systems) to derive order-of-magnitude cost envelopes per hospital. Fourth, where exposure is extreme, compare retrofit envelopes with indicative relocation/new-build envelopes, explicitly considering access and equity. Finally, aggregate envelopes across the prioritized set to size multi-year investment programs. Although the figures produced by this approach are intentionally approximate, they enable macro-prioritization and sequencing of interventions for high-consequence hospitals under substantial risk, while reserving site-specific design and engineering estimates for subsequent local studies^{50, 51–52}.

Our continental screening, by revealing critical clusters, provides the foundation for subsequently identifying alternative referral hubs located outside the same high-exposure tiles and along redundant transport corridors (primary highways, airfields/airports and, where relevant, ports). On this basis, contingency routing for large-scale patient transfers can be designed should compound events compromise specific areas. National authorities can operationalize and stress-test these routes using transport databases, seasonal considerations, and logistical assumptions, incorporating aeromedical and naval options where geography warrants^{53, 54–55}.

Looking ahead, warming oceans and shifts in precipitation regimes may alter the frequency, intensity, and spatial configuration of hydrometeorological extremes and compound events, potentially reshaping hospitals’ exposure patterns across the region^{56, 57–58}. A continental baseline will enable trend detection, scenario screening, and adaptive reprioritization as hazard layers (e.g., updated floodplains, cyclone projections) and health-system information are refreshed, thereby strengthening preparedness and the strategic targeting of investments over time^58,59.

The characterization presented in this research was possible thanks to the use of large-scale datasets and georeferenced systems that translate environmental information into numerical databases. This underscores the value of multidisciplinary work in public health. Information on the frequency and location of natural hazards is available and has been organized by experts from national and international agencies. Such publicly available data can be analyzed with appropriate tools, such as GIS, to enhance preparedness and decision-making.

In this context, this work is relevant because it helps identify areas where risk is both evident and potentially avoidable. By highlighting zones of greater hazard exposure, mitigation and preparedness efforts can be directed more effectively. These insights can benefit communities by providing up-to-date evidence to improve disaster planning and response. Furthermore, this approach can serve as a model for future research and public policies aimed at reducing risks and protecting public health in the region.

This study provides a continental perspective on the extent of natural hazards that pose disaster risk in the Americas. It should be understood as a preliminary screening that estimates the number of hospitals located in adverse environmental conditions. The ZC-NASA-CU study cells, which cover up to 250 km², may lack the precision required for local analysis; therefore, detailed national and subnational studies remain essential.

Reliance on the ZC-NASA-CU datasets, developed nearly two decades ago, introduces inherent limitations due to their coarse resolution and the use of historical frequency data. While these datasets offer a broad and consistent framework for continental-scale analysis, they are less suitable for fine-grained assessments of specific assets, such as individual hospitals.

Unfortunately, due to the lack of publicly available and standardized data across countries, our analysis could not incorporate hospital construction types, building materials, structural resilience, or patient capacity. These are critical variables for assessing physical vulnerability, and we explicitly identify this as a limitation. We encourage further research to integrate such data at national and subnational levels.

Additionally, the classification of hospitals varies by country. To address this, hospitals were grouped into broad categories and included in a generalized model for the entire region. The hospital database will undergo continuous revision and updates to enhance its accuracy and relevance and include for further research health facilities at the primary care level.

Lastly, we acknowledge that validation of results against historical damage data would significantly strengthen this study. However, our current analysis focuses on exposure, aligning with the institutional standards of the PAHO emergency response team. As such, historical validation was not performed for this study. We propose this as an area for future research to assess the predictive value of exposure analyses and enhance the applicability of results in DRR efforts.

Overall, this analysis offers a replicable spatial framework for identifying hospital locations in hazard-prone areas. It should be interpreted as a foundation for more detailed local studies that incorporate building data, real-time hazard models, and functional capacity assessments.

Conclusions

This study shows that hospitals across the Americas face substantial exposure to natural hazards, underscoring the urgent need to strengthen disaster risk reduction (DRR). Of the 20,396 hospitals analyzed, 88.1% are exposed to at least one hazard and 38.1% to multiple hazards. By hazard, floods potentially affect 85.1% of hospitals (notably in Central America), hurricanes 28.3% (predominantly in the Caribbean), and earthquakes 15.1% (with a marked concentration in Central America). Although landslides and volcanic activity affect a smaller share overall, they pose significant localized risks.

A harmonized, continental-scale exposure map—built on reliable GIS data and multi-hazard zoning—enables evidence-informed screening, prioritization, and strategic investment, while detailed risk and vulnerability assessments remain the remit of subsequent site-specific studies. The ZC-NASA-CU framework provides a useful continental baseline, and the classification of hospitals should be regularly updated to improve accuracy and to expand, where feasible, to primary care facilities.

Given projected changes in extreme events and persistent data gaps, robust health disaster risk management is imperative to protect facility integrity and functionality. The information presented here identifies where exposure is most consequential, offers actionable inputs for vulnerability/risk analyses, and can support public policy and investment planning aimed at reducing risk and safeguarding public health in the region.

Acknowledgements

The authors would like to thank Universidad de las Americas for providing the funding related to the publication fee.

Author contributions

Conceptualization: LH, PNH, AC, DA, CU, EP; Methodology: LH, PNH, PY, AC, JIC, EP, Software: AC, JIC, EOP, DA; software: PNH; Validation: DA, JCS; Investigation: PNH, AC, ABG, JCS, MCA; Resources: LH, AC; Data curation: PY, PNH, AC; Writing—original draft preparation: PNH, LH, PY, AC, DA; Writing—review and editing: PE, JIC, EOP, DA, JCS; Visualization: PNH, JIC, EOP; Supervision: AC, LH, EOP; Project administration: AC; Funding acquisition: EOP. All authors have read and agreed to the published version of the manuscript.

Funding

None to declare.

Data availability

The data and information, including the list of hospitals, sources, and databases used to develop this study, are publicly accessible via the following DOI: 10.6084/m9.figshare.26835751. For any further inquiries or requests regarding the data, please contact Dr. Esteban Ortiz-Prado at [email protected].

Declarations

Competing interests

The authors declare no competing interests.

Ethical Consideration

Data included in this study does not involve any human participant or vulnerable information; therefore, and according to the UDLA CEISH and the PAHO guidelines no additional ethical approval was needed for this study.

Consent for publication

Not applicable.

Abbreviations

GIS

Geographic information systems

GVP

Global volcanism program

NOAA

National oceanic and atmospheric administration

PAHO

Pan-American Health Organization

USD

United States dollar

USGS

United States geological survey

ZC-NASA-CU

Critical zones of natural disasters-NASA and Columbia University

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

1. Chang, S. E. & Chang, S. E. Socioeconomic Impacts of Infrastructure Disruptions. in (Oxford University Press, 2016). https://doi.org/10.1093/acrefore/9780199389407.013.66

2. Shivanna, KR. Climate change and its impact on biodiversity and human welfare. Proc. Indian Natl. Sci. Acad.; 2022; 88, pp. 160-171. [DOI: https://dx.doi.org/10.1007/s43538-022-00073-6]

3. Hidalgo, J; Baez, AA. Natural disasters. Crit. Care Clin.; 2019; 35, pp. 591-607. [DOI: https://dx.doi.org/10.1016/j.ccc.2019.05.001] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31445607]

4. Leaning, J; Guha-Sapir, D. Natural disasters, armed conflict, and public health. N. Engl. J. Med.; 2013; 369, pp. 1836-1842. [DOI: https://dx.doi.org/10.1056/NEJMra1109877] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/24195550]

5. Ortiz-Prado, E. et al. El Niño-Southern Oscillation: a call to action for public health emergency preparedness and response. Lancet Region. Health Am.27, (2023).

6. Ortiz-Prado, E; Izquierdo-Condoy, JS; Salazar-Santoliva, C; Vasconez-Gonzalez, J. Leptospirosis deaths in children in the Amazon: Syndemic inequities. The Lancet; 2025; 405, 2275. [DOI: https://dx.doi.org/10.1016/S0140-6736(25)01204-8]

7. Strahler, A. N. 1918-. Geografia fisica. 1, (Omega, 1984).

8. Office of Coordination of Humanitaria Affairs (OCHA-UN). Desastres Naturales en América Latina y el Caribe, 2000 - 2019. (2020). at <https://www.unocha.org/publications/report/world/desastres-naturales-en-am-rica-latina-y-el-caribe-2000-2019>

9. Burton, I. The Environment as Hazard. 6, (Guilford Press, 1993).

10. Ismail-Zadeh, A. Natural hazards and climate change are not drivers of disasters. Nat. Hazards; 2022; 111, pp. 2147-2154. [DOI: https://dx.doi.org/10.1007/s11069-021-05100-1]

11. Izquierdo-Condoy, JS et al. Knowledge, perceptions, and practices on risks and disasters among medical students. A multicenter cross-sectional study in 9 Latin American and Caribbean countries. Adv. Med. Educ. Pract.; 2023; 14, pp. 225-235. [DOI: https://dx.doi.org/10.2147/AMEP.S394331] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/36941904][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10024465]

12. Centre for Research on the Epidemiology of Disasters - CRED. The Emergency Events Database (EM-DAT) - The international disaster database. CRED (1988). at <https://www.emdat.be/>

13. Contreras, D; Bhamidipati, S; Wilkinson, S. Social vulnerability and spatial inequality in access to healthcare facilities: The case of the Santiago Metropolitan Region (RMS) Chile. Socio-Econ. Plann. Sci.; 2023; 90, [DOI: https://dx.doi.org/10.1016/j.seps.2023.101735] 101735.

14. Weiss, DJ et al. Global maps of travel time to healthcare facilities. Nat. Med.; 2020; 26, pp. 1835-1838. [DOI: https://dx.doi.org/10.1038/s41591-020-1059-1] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32989313]

15. Pu, Q; Yoo, E-H; Rothstein, DH; Cairo, S; Malemo, L. Improving the spatial accessibility of healthcare in North Kivu, Democratic Republic of Congo. Appl. Geogr.; 2020; 121, [DOI: https://dx.doi.org/10.1016/j.apgeog.2020.102262] 102262.

16. World Bank, United Nations. Natural Hazards, Unnatural Disasters: The Economics of Effective Prevention. World Bank (2010). 10/11/15/natural-hazards-unnatural-disasters-the-economics-of-effective-prevention.

17. Organización de las Naciones Unidas. Informe del grupo de trabajo intergubernamental de expertos de composición abierta sobre los indicadores y la terminología relacionados con la reducción del riesgo de desastres. (2016). at <https://www.undrr.org/media/76541/download?startDownload=20240603>

18. Pan American Health Organization. Hospitales Seguros. Pan American Health Organization / World Health Organization (2021). at <https://www.paho.org/es/emergencias-salud/hospitales-seguros>

19. Organización de las Naciones Unidas. Marco de Sendai para la Reducción del Riesgo de Desastres 2015–2030. (2015). at <https://www.unisdr.org/files/43291_spanishsendaiframeworkfordisasterri.pdf>

20. Pan American Health Organization & World Health Organization - International Agency for Research on Cancer. Plan de acción para la reducción del riesgo de desastres 2016–2021. Pan American Health Organization / World Health Organization (2016). at <https://www.paho.org/es/node/55341>

21. Iglesias, V et al. Risky development: Increasing exposure to natural hazards in the United States. Earths Future; 2021; 9, e2020EF001795.2021EaFut..901795I [DOI: https://dx.doi.org/10.1029/2020EF001795] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/34435071][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8365714]

22. Morán-Rodríguez, S; Novelo-Casanova, DA. Spatial analysis of the level of exposure to seismic hazards of health facilities in Mexico City, Mexico. Geomat. Nat. Haz. Risk; 2012; 3, pp. 279-292. [DOI: https://dx.doi.org/10.1080/19475705.2011.633105]

23. Pan American Health Organization Institutional Repository for Information Sharing. Mitigation of Disasters in Health Facilities. Evaluation and Reduction of Physical and Functional Vulnerability. General Issues. (1993). at <https://www.unisdr.org/preventionweb/files/2016_VL206713.pdf>

24. Gobierno de México. CENAPRED. Atlas de Riesgos del Estado de Hidalgo. CENAPRED (2009). at <http://www.atlasnacionalderiesgos.gob.mx/AtlasEstatales/?&NOM_ENT=Hidalgo&CVE_ENT=13>

25. Hyndman, D. W. & Hyndman, D. Natural Hazards & Disasters. 1, (CA : Brooks/Cole CENGAGE Learning, 2014).

26. Pan American Health Organization & Naturales, D. I. para la R. de los D. Hacia un mundo más seguro frente a los desastres naturales: la trayectoria de América Latina y el Caribe. 1, (Pan American Health Organization - World Health Organization, 1994).

27. Raymond, C et al. Understanding and managing connected extreme events. Nat. Clim. Chang.; 2020; 10, pp. 611-621.2020NatCC.10.611R [DOI: https://dx.doi.org/10.1038/s41558-020-0790-4]

28. United Nations Office for Disaster Risk Reduction. Sendai Framework for Disaster Risk Reduction 2015–2030 | UNDRR. UNDRR (2015). at <https://www.undrr.org/publication/sendai-framework-disaster-risk-reduction-2015-2030>

29. Hallegatte, S; Rentschler, J; Rozenberg, J. Lifelines: The resilient infrastructure opportunity. Grupo Banco Mundial; 2019; 1, pp. 1-36.

30. Dilley, M. et al. Natural disaster hotspots: a global risk analysis. (World Bank, 2005).

31. Center For Hazards And Risk Research-CHRR-Columbia University, Center For International Earth Science Information Network-CIESIN-Columbia University, & International Bank For Reconstruction And Development-The World Bank. Global Multihazard Frequency and Distribution. (2005). https://doi.org/10.7927/H45718Z5

32. United States Geological Survey. Earthquake Hazard Program—Search Earthquake Catalog. United States Geological Survey (2020). at <https://earthquake.usgs.gov/earthquakes/search/>

33. National Oceanic and Atmospheric Administration & National Hurricane Center. International Best Track Archive for Climate Stewardship (IBTrACS) Project, Version 4. NOAA Historical Hurricane Tracks. NOAA (2020). at <https://services2.arcgis.com/FiaPA4ga0iQKduv3/ArcGIS/rest/services/IBTrACS_ALL_list_v04r00_lines_1/FeatureServer/0>

34. Brakenridge, G. Global Active Archive of Large Flood Events. Dartmouth Flood Observatory (2021). at <https://floodobservatory.colorado.edu/Archives/index.html>

35. NASA Earth Observatory. A Global View of Landslide Susceptibility. Landslides @ NASA / Landslide Susceptibility Map. NASA Earth Observatory (2020). at <https://gpm.nasa.gov/landslides/projects.html#susceptibility>

36. NASA Earth Observatory. Global Landslide Catalog (GLC). Landslides @ NASA—Landslide Team Projects—Global Landslide Catalog (GLC). NASA Earth Observatory (2015). at <https://gpm.nasa.gov/landslides/coolrdata.html%20.>

37. Smithsonian Institution. Global Volcanism Program. Volcanoes of the World, v.5.1.0—Database—Holocene Volcano List. (2023). https://doi.org/10.5479/si.GVP.VOTW5-2023.5.1

38. Levinson, D. H., Knapp, K. R., Kruk, M. C., J., H. & Kossin, J. P. in Indian Ocean Tropical Cyclones and Climate Change (ed. Charabi, Y.) 215–221 (Springer Netherlands, 2010). https://doi.org/10.1007/978-90-481-3109-9_26

39. Siebert, L., Simkin, T. & Kimberly, P. Volcanoes of the World. (Smithsonian Institution, 2011). at <https://www.ucpress.edu/ebook/9780520947931/volcanoes-of-the-world>

40. International Organization for Standardization. Standards. ISO 3166-2 Online Browsing Platform - Codes for Subdivision. ISO (2017). at <https://www.iso.org/obp/ui/#search>

41. Ministerio de Salud Pública y Asistencia Social (MSPAS) de Guatemala. Modelos de Atención Integral a la Salud (MAIS). MSPAS Guatemala (2010). at <https://www.paho.org/gut/dmdocuments/Modelo_de_atencion_integral_en_salud_MSPAS.pdf#gsc.tab=0>

42. Environmental Systems Research Institute. ArcGIS Pro 2.9.0. ArcGIS Pro. (2021). at <https://pro.arcgis.com/es/pro-app/latest/get-started/download-arcgis-pro.htm>

43. Figueiredo, R et al. A semi-quantitative multi-hazard risk assessment framework for European coastal urban areas. Geomat. Nat. Haz. Risk; 2024; 15, 2378994. [DOI: https://dx.doi.org/10.1080/19475705.2024.2378994]

44. Brown, SK; Jenkins, SF; Sparks, RSJ; Odbert, H; Auker, MR. Volcanic fatalities database: Analysis of volcanic threat with distance and victim classification. J. Appl. Volcanol.; 2017; 6, 15. [DOI: https://dx.doi.org/10.1186/s13617-017-0067-4]

45. Datosmacro. Indice de Desarrollo Humano - IDH 2021 |. Datosmacro.com (2022). at <https://datosmacro.expansion.com/idh>

46. Earth Science Data Systems, N. Global Flood Hazard Frequency and Distribution | NASA Earthdata. (2024). at <https://www.earthdata.nasa.gov/data/catalog/sedac-ciesin-chrr-ndh-flood-hfd-1.00>

47. Johnson, K. et al. Global Seismic Hazard Map. (2023). https://doi.org/10.5281/zenodo.8409647

48. Mena, R., Salazar, R. & Jaime, H. Impacto de los desastres en América Latina y el Caribe, 1990–2011. (2013). at <https://eird.org/americas/noticias/Impacto_de_los_desastres_en_las_Americas.pdf>

49. Pan American Health Organization. Safe Hospitals - PAHO/WHO. Pan American Health Organization (2024). at <https://www.paho.org/en/health-emergencies/safe-hospitals>

50. United Nations Office for Disaster Risk Reduction. Global Assessment Report (GAR) on Disaster Risk Reduction 2025. UNDRR (2025). at <https://www.undrr.org/gar/gar2025>

51. Fung, JF; Sattar, S; Butry, DT; McCabe, SL. The total costs of seismic retrofits: State of the art. Earthq. Spectra; 2021; 37, pp. 2991-3014. [DOI: https://dx.doi.org/10.1177/87552930211009522]

52. World Health Organization & Pan American Health Organization. Hospital Safety Index. Guide for Evaluators. (Organizacion Panamericana de la Salud, 2019).

53. Treber, M. Federal Patient Movement Overview Fact Sheet. (2017).

54. Pan American Health Organization. EVAC-H: Evacuation of hospitals and their critical areas. Pan (2025). at <https://www.paho.org/en/documents/evac-h-evacuation-hospitals-and-their-critical-areas>

55. Homeland Security. Planning Considerations: Evacuation and Shelter-in-Place - Guidance for State, Local, Tribal and Territorial Partners (2019).

56. Zscheischler, J et al. A typology of compound weather and climate events. Nat. Rev. Earth Environ.; 2020; 1, pp. 333-347.2020NRvEE..1.333Z [DOI: https://dx.doi.org/10.1038/s43017-020-0060-z]

57. World Meteorological Organization. State of the Global Climate 2023. World Meteorological Organization (2024). at <https://wmo.int/publication-series/state-of-global-climate-2023>

58. Climate Change 2022: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change.1, (IPCC Sixth Assessment Report, 2022).

59. WHO Team Climate Change and Health (CCH). WHO guidance for climate resilient and environmentally sustainable health care facilities. World Health Organization (2020). at <https://www.who.int/publications/i/item/9789240012226>