WGS-Enabled Surveillance Improves Detection Of Transmission Events Within A Large Tertiary Care Hospital Trust In London

This study demonstrates that integrating whole-genome sequencing with patient movement data significantly improves the detection of carbapenem-producing Enterobacterales transmission events in a London hospital compared to traditional surveillance, offering earlier outbreak identification and substantial economic savings.

The Gist

Imagine a large hospital as a bustling, high-tech city. In this city, invisible "criminals" (superbugs called CPEs) are trying to spread from person to person. The hospital's security team (Infection Prevention and Control, or IPC) has a job: catch these criminals before they start a riot (an outbreak).

For years, the security team has used a very old, slightly blurry map to do their job. Their rule is simple: "If two people get sick with the same bug, in the same room, within 7 days of each other, we assume they caught it from each other."

This paper is like a report card showing that while this old map works sometimes, it misses a lot of the action and sometimes points the security team in the wrong direction. The researchers decided to upgrade the system by adding Whole-Genome Sequencing (WGS).

Think of WGS as giving every single bacterium a unique, high-definition fingerprint. Instead of just guessing who met whom based on time and place, the security team can now look at the fingerprints and say, "Yes, these two bacteria are identical twins," or "No, these two are strangers who just happen to look similar."

Here is the breakdown of what they found, using some everyday analogies:

1. The Old Map vs. The High-Tech Scanner

The researchers tested their new "Fingerprint Scanner" against the old "Time-and-Place Map" using data from two different outbreaks in a London hospital.

• The Missed Criminals (False Negatives): The old map missed 80% of the actual transmission events!

  • The Time Blindspot: Sometimes, Patient A gets sick on Monday, and Patient B gets sick 25 days later in the same room. The old map says, "Too much time has passed, they aren't connected." But the fingerprints showed they were connected. The bacteria were just hiding in the shadows.
  • The Place Blindspot: Sometimes, Patient A is in Ward 10 and Patient B is in Ward 20. The old map says, "Different rooms, no connection." But the fingerprints showed they were linked. The bacteria traveled via a doctor's hands or a shared piece of equipment between the wards.
  • The Shape-Shifter Blindspot: This is the coolest part. Sometimes the bacteria change their "species" (like a criminal changing their disguise). A Klebsiella might swap a weapon (a plasmid) with an E. coli. The old map only looks for the same species, so it ignores this. The fingerprint scanner sees the shared weapon and catches the link.

• The False Alarms (False Positives): The old map also cried wolf too often. It flagged 71% of its "suspects" as connected when the fingerprints proved they were actually unrelated. This wasted the security team's time and resources chasing ghosts.

2. The "Superpower" of Early Detection

The most exciting finding is about time.

• The old map usually only sounded the alarm after the outbreak had already grown.
• The new WGS system acted like a smoke detector instead of a fire extinguisher. It spotted the transmission 25 to 47 days earlier than the old methods.
• Analogy: Imagine a fire starting in a kitchen. The old method waits until the smoke fills the hallway (7 days later) to call the fire department. The new method smells the smoke the moment the first spark flies (Day 1), allowing them to put it out before it burns the whole building down.

3. The Money Talk (Economics)

You might think, "But isn't sequencing bacteria expensive?"
The researchers did the math, and the answer is a resounding yes, it saves money.

• The Cost of Ignorance: When the hospital misses an outbreak, they have to isolate entire wards, cancel surgeries, and treat sick patients for weeks. This costs millions.
• The Cost of the Scanner: Sequencing the bacteria costs money, but it's a tiny fraction of the cost of an outbreak.
• The Result: By catching the real criminals early and ignoring the false alarms, the hospital could save between £20,000 and £3.6 million per year.
• Analogy: It's like buying a high-quality security system for your house. It costs a few hundred pounds a year, but it saves you from losing your entire house to a burglary. The return on investment is huge.

The Bottom Line

This paper argues that hospitals need to stop relying on "guessing games" based on time and location. By using genomic fingerprints combined with patient movement data, hospitals can:

1. See the invisible: Catch bacteria that move between different wards or change species.
2. Act faster: Stop outbreaks before they become disasters.
3. Save money: Stop wasting resources on false alarms and prevent expensive outbreaks.

The authors conclude that this technology is ready to become the "standard of care," turning infection control from a reactive game of whack-a-mole into a proactive, precise science.

Technical Summary

1. Problem Statement

Current Infection Prevention and Control (IPC) surveillance methods in hospitals rely heavily on spatial and temporal proximity (e.g., patients in the same ward within 7 days) to detect bacterial transmission events. This approach suffers from significant limitations:

• Low Resolution: It frequently misattributes or misses transmission events, leading to delayed outbreak containment.
• Blind Spots: It fails to detect cross-species transmission mediated by mobile genetic elements (plasmids) and is insensitive to transmission events occurring outside strict time/ward windows.
• Economic Burden: Misattributed events trigger unnecessary clinical interventions, while missed events allow outbreaks to spread, costing the NHS billions annually.
• AMR Threat: Carbapenem-producing Enterobacterales (CPEs) are resistant to key antibiotics, making rapid and accurate detection critical.

While Whole-Genome Sequencing (WGS) offers higher resolution, its routine integration is hindered by data silos, long turnaround times, and analytical complexity. This study evaluates the value of integrating WGS data with patient movement data to overcome these barriers.

2. Methodology

The study employed a retrospective evaluation using two distinct datasets from a large tertiary care hospital trust in London:

• Dataset A (CPE Collection): 103 genomes from a diverse carbapenemase outbreak (Jan 2021–Mar 2021), including K. pneumoniae, E. coli, and others.
• Dataset B (IMP Collection): 82 genomes of Imipenem-Hydrolysing $\beta$-lactamase-positive CPEs (Jun 2016–Oct 2019).

Technical Pipeline:

1. Genomic Processing: Raw reads were quality-controlled (FastQC), trimmed (Trimmomatic), and assembled (SPAdes). Species identification was performed using Kraken2/Bracken.
2. Similarity Analysis:
   • Within-species: Core genome alignments were generated (Panaroo), and pairwise Single Nucleotide Polymorphism (SNP) distances were calculated (snp-dists).
   • Cross-species/Plasmid: Plasmid backbones were reconstructed (Mobsuite). Similarity was quantified using PLING (a network-based approach) to detect identical plasmid sharing across different species.
3. Data Integration: Genomic data was mapped to pseudanonymized patient movement data (ward-day records) to create contact–genome pairs.
4. Evaluation Logic:
   • Clinical Criteria (Baseline): Transmission defined as same species, same ward, $\le$ 7 days apart, with post-admission acquisition.
   • Genomic Thresholds (Gold Standard): SNP distance $\le$ 10 (within species) and PLING distance = 0 (plasmid sharing).
   • Classification: Pairs were classified as True Positives (TP), False Positives (FP), True Negatives (TN), and False Negatives (FN) to calculate sensitivity, specificity, and predictive values.
5. Economic Modeling: A cost-benefit analysis compared the cost of WGS sequencing (£100/genome) against savings from preventing unnecessary interventions ("Rule-Out") and preventing secondary infections from missed cases ("Rule-In").

3. Key Contributions

• Integrated Framework: Developed a methodology combining high-throughput WGS analysis with granular patient movement data to create a unified view of transmission.
• Cross-Species Detection: Demonstrated the ability to detect transmission events mediated by plasmids between different bacterial species (e.g., E. coli to K. pneumoniae), which standard SNP-based methods miss.
• Quantification of "Blindspots": Systematically categorized missed transmission events into Temporal (>7 days), Spatial (different wards/buildings), and Mechanistic (cross-species plasmid) blindspots.
• Economic Validation: Provided a robust economic evaluation showing that WGS integration yields a positive Return on Investment (ROI) in most scenarios, even under conservative cost assumptions.

4. Key Results

Detection Performance:

• Current IPC Methods: Across 3,423 contact–genome pairs, standard clinical criteria detected only 20.5% of genomically confirmed transmission events (Sensitivity) while maintaining high specificity (98.5%).
• Missed Events (False Negatives):
  • CPE Collection: 45 missed events. 24 were temporal (>7 days), 18 were spatial (different wards/buildings), and 3 were mechanistic (cross-species plasmids).
  • IMP Collection: 48 missed events. 9 were temporal, 15 were spatial, and a significant 23 were mechanistic (cross-species plasmid sharing), representing nearly 48% of missed events in this dataset.
• False Positives: Current methods triggered unnecessary investigations for 50–71% of patient pairs identified as potential contacts.

Timeliness:

• WGS-enabled surveillance provided an earlier detection window of 25–47 days compared to current methods, allowing intervention before outbreaks escalate.

Economic Impact:

• Rule-Out Savings: Preventing unnecessary clinical actions for genomically unrelated cases saved between £34k and £715k (CPE dataset).
• Rule-In Savings: Preventing secondary infections from missed cases added further savings.
• ROI:
  • CPE Dataset: Annualized savings ranged from £126k to £3.6 million, with an ROI between 2.69 and 76.51.
  • IMP Dataset: Annualized savings ranged from £366 to £66k, with an ROI between 0.13 and 22.63.
  • In 7 out of 8 cost scenarios, the ROI exceeded 2-fold.

5. Significance and Conclusion

This study provides compelling evidence that integrating WGS data with patient movement analytics significantly outperforms traditional IPC surveillance.

• Clinical Impact: It transforms outbreak detection from a reactive, low-resolution process into a proactive, high-resolution tool capable of identifying transmission across time, space, and species barriers.
• Economic Viability: The analysis demonstrates that the cost of sequencing is outweighed by the savings from avoided outbreaks and unnecessary interventions, making a strong business case for adoption.
• Policy Implication: The authors argue for the adoption of WGS-enabled IPC surveillance as a standard-of-care tool to mitigate AMR-driven outbreaks.

Limitations & Future Work:
The study is retrospective and relies on short-read sequencing, which may underestimate complex plasmid diversity. Future work must focus on real-time implementation, addressing sequencing turnaround times, and potentially utilizing long-read sequencing to better resolve plasmid structures. Despite these limitations, the findings strongly support the operationalization of WGS in hospital infection control.