Two patients on the same ward test positive for the same organism in the same week. Are those infections traceable to a common source within the hospital, or merely a coincidence independently acquired from outside? Most hospitals cannot definitively answer that question—even though the technology capable of doing so has been available for more than a decade.
Hospitals and the clinicians who staff them operate at capacity, and existing clinical laboratory workflows are designed around what a bedside physician needs in the moment: identifying the organism and selecting the right treatment. These systems were built to address an individual patient’s acute needs, not to trace the origin of an infection. Consequently, an outbreak goes undetected until it rises above the background noise—an unusual organism spreading through a ward or a conspicuous pattern of illness that has already caused significant harm.
Compounding the problem, antimicrobial resistance continues to accelerate, rendering many outbreaks increasingly difficult to treat. Between 2019 and 2023, carbapenemase-producing Enterobacterales—a class of bacteria resistant to certain last-resort antibiotics—increased by 69 percent across 29 states. Candida auris, a fungal pathogen with inherently higher levels of resistance, surged from 766 documented US cases in 2020 to over 6,304 in 2025. Yet as resistance rates climb, the cost of DNA sequencing continues to decline, offering a powerful countermeasure.
Identity Is Not Linkage
The resolution of conventional diagnostics is limited to the species level. A standard identity and antibiotic susceptibility profile can determine what an organism is and which agents should be effective against it, but it cannot reveal where it came from or how it relates to other infections. Without that information on relatedness, investigators are left to count similar cases above an expected baseline—risking either the omission of genuine transmission chains or the overprediction of false ones.
Relatedness is fundamentally a question of the genome. When two pathogen genomes differ at only a few sites, very little time has elapsed for those mutations to accumulate. The most recent common ancestor of those genomes must therefore be quite recent, strongly suggesting that both infections were seeded by the same source. Conversely, thousands of divergent sites point to a distant shared ancestry—and independent colonizations that merely happen to look alike.
Without genomic data, case counting alone can fail in either direction. Two patients may be colonized in the emergency department and subsequently transferred to separate units, where their infections manifest weeks apart. Ward counts rise independently, and the transmission link is entirely missed.
The reverse scenario is equally problematic. Several cases of the same species with identical susceptibility profiles cluster on a single unit, prompting an investigation and the implementation of contact precautions. But genomic analysis reveals the organisms are unrelated, the sources are distinct, and the assumed transmission was never real.
Genomic linkage corrects both failures. It surfaces connections that would otherwise be invisible, and it clears coincidences that might otherwise be mistaken for clear evidence of transmission.
The Forensic Era
In 2011, whole-genome sequencing by Snitkin et al. reconstructed the transmission map of a carbapenem-resistant Klebsiella pneumoniae outbreak at the NIH Clinical Center, an event that affected 18 patients and claimed the lives of 11. The causative organism was a carbapenemase-producing Enterobacterales of the type described above.
That retrospective analysis established both the power of the method and the manner in which it would initially be deployed. Sequencing became epidemiology’s forensic laboratory—summoned only after an outbreak had already become apparent. It remained in that role because sequencing was expensive, and costly tests get rationed to the cases that most clearly demand them.
The Economics of Continuous Surveillance
When a measurement becomes sufficiently inexpensive, it transitions from being a test one orders to a signal one monitors. A continuous glucose monitor is no more accurate than a periodic blood draw. It simply alerts you to a problem sooner, before symptoms fully emerge.
Since the initial investigation by Snitkin et al., sequencing costs have fallen dramatically. For routine isolate sequencing, however, the binding constraint has been per-sample economics rather than the headline cost per base. Those economics are, at their core, a function of volume—the more isolates processed simultaneously, the faster per-unit costs decline.
Sequencing also represents only part of the total expense. Linkage analysis requires someone to interpret what a given number of genetic differences means in the context of a particular ward, timeline, or organism. That expertise constitutes a fixed cost that a thin stream of isolates does not justify on its own.
This work has, however, been successfully operationalized. Foodborne pathogen surveillance has already made precisely this transition. FDA and CDC genome-based tracking of foodborne pathogens generated an estimated return of approximately $500 million annually by 2019, against an investment of roughly $22 million per year. A significant portion of that investment supported the sequencing of more than 100,000 pathogen genomes. The obstacle preventing a similar shift in hospitals is the siloed nature of individual institutions and their insufficient volume of isolates.
What Continuous Surveillance Looks Like in Practice
A small number of centers meet that volume threshold independently, and they provide the earliest evidence for the approach. Researchers at UPMC Presbyterian sequenced clinical isolates as a routine practice rather than as a targeted investigation. Their efforts identified 172 outbreaks involving 476 distinct pathogens, a substantial proportion of which would have been entirely missed by conventional surveillance methods. They detected clusters between two patients that might otherwise have gone unnoticed for a year or more, along with numerous additional infections. Over a two-year period, they estimated that 62 infections and approximately five deaths were avoided, yielding a 3.2-fold return on investment. UPMC’s isolate volume is atypical, but nothing about the outcome depends on it being theirs. Their capacity for continuous surveillance was purchased through volume—a threshold that aggregation can meet for every institution.
It is important to emphasize that whole-genome sequencing does not replace traditional epidemiology or infection prevention. Rather, it strengthens both with the highest possible quality of evidence and the greatest confidence in the links it identifies. A genome can establish whether two infections share a common source; it cannot walk the ward, locate the contaminated sink, or change the practice that placed it there. That essential work remains where it has always been.
From Counting to Knowing
Antibiotics added decades to modern life expectancy, and infection control helps preserve those gains. Hospitals are where the most drug-resistant organisms face the highest selective pressure and where patients are least able to withstand them. Containing these threats within hospital walls remains one of the most powerful levers available in public health—and we now possess the tools to do so far more effectively. The source of the mystery infection affecting two patients on one ward in a single week is now an answerable question.