To better understand these patterns, infection prevention teams are turning to hotspot mapping.
However, despite rapid advances, evidence for hotspot mapping remains largely confined to well-resourced health care networks.
Mechanisms Underlying Hospital AMR HotspotsUnderstanding why AMR hotspots emerge within hospitals is the first step toward mapping them across health care networks.
Mapping Resistance Across Health Care NetworksMapping AMR across an entire health care network requires integrating regularly collected data, such as antibiotic prescribing records, laboratory antibiograms, and patient movement data.
As more health care networks integrate these data systems, AMR hotspot mapping is likely to become a standard part of infection prevention.
Antimicrobial resistance (AMR) is a major global public health challenge that could cause 10 million deaths annually by 2050, according to the World Health Organization (WHO).1 Within a hospital, resistant organisms aggregate in specific wards, sinks, and patient populations. Across health care networks, they often follow the same pathways as patients, moving between emergency departments, intensive care units, and outpatient clinics. To better understand these patterns, infection prevention teams are turning to hotspot mapping. Hotspot mapping is an approach that integrates clinical, laboratory, and patient movement data to pinpoint where resistant pathogens emerge and how they spread through a health care system. However, despite rapid advances, evidence for hotspot mapping remains largely confined to well-resourced health care networks.
Mechanisms Underlying Hospital AMR Hotspots
Understanding why AMR hotspots emerge within hospitals is the first step toward mapping them across health care networks. Health care settings are ideal environments for the emergence and persistence of AMR due to extensive antibiotic use and large populations of vulnerable patients. The WHO estimates that around 42.7 million of the 421 million people hospitalized worldwide each year develop a health care-associated infection.2
Surface contamination surveys support these findings. In one Italian hospital, 33.6% of bacterial isolates collected from surfaces carried resistance to at least one antimicrobial, with high-touch surfaces such as bed rails and door handles the most heavily contaminated.2 Indoor air and wastewater further contribute to the spread of AMR. In one Chinese hospital, inpatient wards contained more than 100 antibiotic resistance gene copies per cubic meter of air, and outdoor air near ventilation outlets carried approximately twice the resistance-gene load of the surrounding urban environment.2 Together, these findings suggest that hospitals are not only the places where resistant infections are treated but also ecosystems that can promote the emergence and spread of AMR.
Mapping Resistance Across Health Care Networks
Mapping AMR across an entire health care network requires integrating regularly collected data, such as antibiotic prescribing records, laboratory antibiograms, and patient movement data. A recent study analyzed electronic health records from 60,551 patients with confirmed bacterial infections across a network of 4 public hospitals, 13 primary care centers, and 5 diagnostic laboratories between 2017 and 2022.3 Linking antibiotic prescribing data with laboratory antibiograms showed how prescribing patterns influenced resistance over time. In many cases, increases in the use of certain antibiotics were followed by corresponding increases in resistance within just 1 week.
The network-wide analysis also highlighted the clinical consequences of AMR. Across the dataset, patients with resistant infections had approximately 2.18 times higher risk of death than those with susceptible infections.3 The study also revealed a lasting impact of the COVID-19 pandemic. Changes in antimicrobial prescribing persisted for up to 2 years after the pandemic began, and several resistance trends that accelerated during that period had yet to return to prepandemic levels by the end of the study.3
Emerging Trends in AMR Research
Beyond hospital mapping, studies also show how AMR research is evolving. A bibliometric analysis of nearly 2,500 publications on artificial intelligence (AI) and AMR published between 2014 and 2024 found that annual output grew from just 4 papers in 2014 to 549 in 2023.1 The research centered on 6 major themes: sepsis, artificial neural networks, antimicrobial resistance, antimicrobial peptides, drug repurposing, and molecular docking. The field’s most cited studies include AlphaFold's protein structure prediction model and a deep learning system that screened more than 107 million compounds to identify a novel antibiotic active against multidrug-resistant pathogens.1
A separate bibliometric review of hospital antimicrobial stewardship research revealed a similar shift. Studies are increasingly centered on machine learning, clinical decision support systems, and electronic health records, with annual publication output reaching a record 49 papers in 2024.4
Current Limitations to Hotspot Mapping
Despite advances in data science, translating AMR hotspot mapping into system-wide practice remains challenging. A review of 57 AMR intervention studies across the 10 member states of the Association of Southeast Asian Nations (ASEAN) found that 86% focused exclusively on the human health sector, with 82.5% conducted in hospital settings. In contrast, only 8.8% adopted a true One Health approach that integrated human, animal, and environmental sources of antimicrobial resistance.5 Moreover, none of the included studies examined the environmental sector alone, despite growing evidence that wastewater and agricultural runoff contribute to the spread of resistance. Also, the review was concentrated in 3 countries, Thailand, Singapore, and Vietnam, which together accounted for more than 70% of the studies reviewed.5
A comparison of AMR surveillance systems across several European countries revealed a similar challenge. Most national surveillance programs monitor antimicrobial resistance and antimicrobial use in either the human health or livestock sector, but rarely integrate both.6 As a result, for infection prevention teams, the data required to build a complete hotspot map, spanning clinical, laboratory, environmental, and veterinary sources, are often present in separate systems that are not designed to work together.
Future Directions for AMR Hotspot Mapping
Together, these findings suggest that AMR mapping is most effective when it integrates multiple data sources, including antibiotic prescribing records, laboratory antibiograms, patient movement, and environmental sampling. Health care network systems that achieve such a level of integration have been able to detect relations between prescribing, resistance, and mortality that single-hospital surveillance would often overlook.
Linking pharmacy, laboratory, patient movement, and bed management systems can enable the identification of resistance hotspots in near real time. At the same time, the evidence base must extend beyond the relatively small number of high-income health systems that currently dominate the literature and embrace a One Health perspective that incorporates environmental and animal reservoirs of resistance. As more health care networks integrate these data systems, AMR hotspot mapping is likely to become a standard part of infection prevention.
References
