He Named My Superspreader Network Model After Himself in the WHO Report — Then the Outbreak Required My Contact-Tracing Algorithm

The phylogenetic tree printout was unfolded on her desk — the A3 sheet flattened with the heel of her hand against the left crease, the right crease still showing where the paper wanted to return to its folded position. Dr. Valentina Russo had underlined the critical branch in red pen eight months ago, the day she had printed it for Carmichael’s briefing. The red ink ran along the branch that separated the hospital cases from the community cases, crossing the fold crease at the midpoint. When the printout was folded, the red line was still visible on both sides of the fold if you held it to the light.

Emilio was at the adjacent desk. He had run the sequence alignment on the 47 samples — 72 hours of compute time, each sample aligned against the reference genome, the phylogenetic tree constructed from the pairwise distance matrix. She had supervised the analysis. She had identified the branching topology that showed a genomic distance of zero between eight of the hospital cases.

Zero.

Not near zero. Not statistically negligible. Zero — the same pathogen, same mutation profile, same source event. Eight cases sharing the same genomic distance as identical samples taken from the same individual. That topology meant one thing: they had all acquired the infection from a single contact in a single location. Not community spread. Not independent acquisitions from a common environmental reservoir. One source.

She had looked at the hospital porter’s movement log — eight wards in four days, transporting patients between oncology and orthopedics, checking supply rooms on the respiratory unit, carrying laundry through the cardiac care corridor. The network model had shown it immediately: his movement pattern was the hub. Remove his node from the transmission graph and the 47 cases became 39 community cases with no epidemiological link, plus eight hospital cases floating in isolation. Add his node back: the transmission chain resolves from one end to the other.

She unfolded the printout fully. She pointed at the branch.

“The genomic distance at this node,” she said to Emilio. “What does zero distance between eight cases tell you?”

He said: “Single source event.”

“Single source event in a single location,” she said. “Now look at the next branching point — the community cluster. The genomic distance between the hospital group and the community group. Read me that distance.”

He looked at the tree. “Four mutations.”

“Four SNPs of separation. These cases are related — same outbreak, same region, same pathogen lineage — but they acquired it independently, at different times, from different exposures. The hospital cluster and the community cluster are cousins, not identical twins. The porter is the link between them. He moved between the community cases before the hospital cases. The direction of transmission is clear from the branch topology.”

She folded the printout along the familiar creases — first to A4, then to A4 again. She placed it in the corner of her desk, the red-underlined side up.

“The hospital source,” Emilio said. “The porter’s ward-to-ward movement.”

“The porter’s ward-to-ward movement,” she said. She opened the network model on her screen. She had 18 new cases to add from the past week’s surveillance data. She began entering the coordinates.

—

The week the cluster had been confirmed — eight weeks before, when Carmichael had reviewed her network model and phylogenetic analysis — he had called her from the regional authority’s incident command center.

He had said: “The network model was right. The porter’s movement pattern explains the cluster.”

She had said: “The phylogenetic analysis confirms it — the genomic distance between the hospital cases is zero. They couldn’t have come from community exposure.”

He had said: “This changes the response. We’re isolating the ward, stopping the porter’s cross-ward duties, and contact-tracing every patient he moved.”

She had said: “The 47 cases all trace to the same source event. You can see it in the branching topology — the hospital cluster branches from the community cluster at the point where the porter’s exposure would have been.”

He had said: “Remarkable work. This is a significant finding.”

She had said: “Thank you.”

He had gone to brief the response team.

She had folded the printout and put it on the corner of her desk. She had opened the network model. She had added the new case data.

“Significant finding.” He had said it on the phone before hanging up. She had noted it: specific words from a specific call at a specific time. Not “your finding.” Not “you found it.” “This is a significant finding” — the discovery presented as a phenomenon that existed independently of the work that had identified it.

She had noted it and continued entering the coordinates.

—

The WHO case report had been submitted to the WHO regional office through Carmichael’s office — he was the regional authority’s WHO liaison. She had known the report was being prepared. She had submitted her analysis documentation to his office six weeks ago: the network model files, the phylogenetic analysis, the genomic distance matrix, the transmission chain reconstruction.

She had read the published report in her office, on her computer screen. The phylogenetic printout was in the corner of her desk, folded.

She read: “Carmichael Cluster Analysis — epidemiological investigation led by Regional Medical Director C. Carmichael.”

She read: “network analysis support: Dr. Valentina Russo, WHO-EPID-VR-5517.”

She opened the network model file registry. WHO-EPID-VR-5517 — her WHO epidemiologist identification number, issued when she had joined the WHO-affiliated institute seven years ago, embedded in every model file she had ever produced. The registry showed 41 model files under her WHO ID. The hospital cluster model was file 41.

She looked at the phylogenetic printout on the corner of her desk.

She picked it up. She unfolded it to the red-underlined branch — first crease, second crease, A3 fully flat. She looked at the red ink. Four SNPs of separation between the hospital cluster and the community cluster. Zero distance within the eight hospital cases. One source.

She folded it again. She put it back on the corner of her desk.

She opened the network model. She had 18 new cases to enter.

She began entering them.

The International Epidemiology Symposium was held in a conference center in Geneva — three days, 600 delegates, eight parallel tracks. Carmichael presented in the plenary session on the second day: “Hospital Cluster Identification in Outbreak Response: A Network Methodology Case Study.”

His presentation had 14 slides. Slide 7 was Valentina’s network diagram — the transmission graph she had built from the case contact data, the porter’s node highlighted, the 47 cases as nodes colored by exposure type. Hospital cases in red. Community cases in blue. The porter at the center, red edges radiating outward. She had built the graph in her network modeling software at 2 AM on the night the case data had been complete enough to run the model. It had taken four hours to prepare for the briefing.

He said: “Our network methodology identified the hospital transmission source through contact network reconstruction. The 47-case cluster, initially attributed to community spread, was resolved to a single point of origin through graph analysis of the case data.”

He said “our network methodology.”

He said “graph analysis of the case data.”

He did not show the phylogenetic tree. He did not describe the genomic distance of zero between the eight hospital cases. He did not name WHO-EPID-VR-5517. He did not say: the transmission chain was confirmed by phylogenetic analysis conducted by Dr. Valentina Russo, whose WHO epidemiologist ID is on every model file.

He showed the slide. He described the cluster.

Valentina was not at the conference. She was in her office, entering the monthly surveillance data into the network model. She read the conference abstract the week before. She had noted “network analysis support” in the authors list and had continued working.

—

The WHO outbreak review board’s request arrived in her institutional email from Dr. Hans Bjerkness, WHO Scientific Officer, on a Tuesday morning.

“Dr. Russo — The WHO Outbreak Review Board has been convened to assess the response methodology used in the hospital cluster investigation and to develop protocol recommendations for similar outbreak scenarios. The board requires, for verification of methodology reproducibility: (1) the original network model files, (2) the phylogenetic analysis files, and (3) the WHO epidemiologist ID of the originating analyst. The documentation is required within 21 days. Please confirm your availability to present the methodology to the board.”

She read “WHO epidemiologist ID of the originating analyst.”

She read “present the methodology.”

She opened the network model file registry. WHO-EPID-VR-5517. File 41: hospital cluster, 47 cases, porter identification, phylogenetic confirmation. Every file in the registry under the same ID.

She picked up the phylogenetic printout from the corner of her desk. She unfolded it — first crease, second crease. She looked at the red-underlined branch. The four SNPs. The zero distance. The porter.

She folded it.

She did not call Carmichael.

She opened a new email to Dr. Bjerkness. She confirmed her availability. She began preparing the documentation package: the network model files, the phylogenetic analysis, the genomic distance matrix, the WHO ID certificate, and a three-page methodology summary explaining the network reconstruction process and the phylogenetic confirmation protocol.

She sent the confirmation that evening.

She opened the surveillance model. The monthly data entry was not complete. She finished entering it.

—

Carmichael’s administrative assistant had forwarded him the WHO review board notice the same morning. He had read it in his office between meetings. He was pleased: a WHO review board producing protocol recommendations from his cluster investigation — the findings would improve outbreak response globally. He had forwarded the notice to his public health team with a note: “Support the board’s work. Provide all documentation they require.”

His public health team had gone to the case archive.

—

She had not told Carmichael she was responding to the review board directly. Dr. Bjerkness had addressed the request to her — he had identified WHO-EPID-VR-5517 in the model documentation and had contacted the registered epidemiologist. The request was for the originating analyst’s documentation. She was the originating analyst.

The idea that she should have looped in Carmichael before responding — that the appropriate protocol was to notify the regional authority’s WHO liaison before the WHO’s scientific officer before responding to a direct WHO request — had not presented itself as a question while she was composing the email. The request had come through official WHO channels to the registered epidemiologist. She had answered it.

She had included the methodology summary because the board would need to understand the network reconstruction process and the phylogenetic confirmation protocol before the review sessions. Three pages: section one, the network model architecture; section two, the phylogenetic analysis pipeline; section three, the epidemiological interpretation framework. She had written all three sections from memory, from the analysis she had conducted eight months ago, from the methodology she had built and refined over seven years of WHO-affiliated work.

The documentation had taken her six hours to prepare.

She had sent it and gone back to the surveillance data.

His public health team had looked for the network model in the case archive. The archive contained the final cluster report, the response timeline, Valentina’s analysis PDF summaries — three documents, each summarizing a phase of the analysis — and the WHO case report submission correspondence. It did not contain the network model files. It did not contain the phylogenetic analysis data.

The team lead had gone to Carmichael.

“The review board needs the original model files and the WHO ID of the analyst who produced them. The archive has the PDF reports. The model files are on Dr. Russo’s system. The WHO ID is hers — WHO-EPID-VR-5517.”

Carmichael had said: “What do you mean the WHO ID is hers?”

The team lead had said: “The WHO outbreak review board needs the WHO epidemiologist ID of the originating analyst. The regional medical director credential doesn’t include a WHO epidemiologist ID. Dr. Russo’s ID is embedded in every model file she produced. WHO-EPID-VR-5517 is her registration.”

He had looked at the WHO case report on his screen. “Carmichael Cluster Analysis.”

He had sat for a long time.

He had been the regional authority’s WHO liaison for nine years. He had submitted case reports, outbreak summaries, and response evaluations. He had always been the named authority on submissions because he was the institutional interface — the person who had operational authority over the response, who had made the decisions to deploy resources, to isolate wards, to initiate contact-tracing. Those decisions were his. They were real decisions that had required him to take the analysis to the Minister and say: we need to act on this today.

But the review board was not asking about the response decisions. The review board was asking about the methodology that had made the response possible. The network model. The phylogenetic analysis. The genomic distance calculation that had shown zero — that specific number, in that specific transmission chain, identified by the analyst who had built the model that could see it.

He had said “significant finding” on the phone. He had said “remarkable work.” He had understood those words as recognition of the work — of Valentina’s contribution to the finding. He had not understood them as all that he had done to acknowledge what she had produced.

He had taken the analysis to the Minister. He had ordered the ward isolation and the porter’s reassignment. Those decisions had been his. But the finding that made the decisions necessary — the finding that the 47 cases were not community spread, that they were a hospital cluster with a single human source whose movement pattern could be identified in the contact network — that had been hers.

He had described it at Geneva as “our network methodology.” As if the methodology were a shared capability of the regional authority. As if “our” could encompass the seven years of network modeling expertise, the WHO registration, the phylogenetic analysis that Emilio had run for 72 hours under Valentina’s supervision.

He had never thought about what “our” could encompass. He had thought about the response decisions, which had been his.

He reached for the phone to call her. He set it down. She had already responded to the review board. She had sent the documentation the previous evening. He had been eating dinner.

He opened the WHO case report. He looked at “network analysis support: Dr. Valentina Russo.”

He opened the draft for the amendment.

—

The phylogenetic printout was on the corner of her desk, folded to A4. She had checked the red-underlined branch when she had assembled the review board documentation package — had unfolded it, verified the branch was clearly readable in the photocopy she was including, refolded it. The crease ran through the underline. Both were still clean. She had put it back on the corner of the desk and returned to the documentation.

She had not thought about why she had checked the red underline. She had checked it because it was the most important element of the printout — the branch that showed the zero distance, the evidence that separated the hospital cluster from the community exposure hypothesis. When including a document in a methodology package, you confirm that the key element is legible. She had confirmed it. She had continued working.

The WHO outbreak review board convened its first session in Geneva on a Wednesday morning. Valentina had the phylogenetic printout in her folder — folded to A4, the red-underlined branch visible at the top edge of the fold crease.

Dr. Bjerkness opened the session. Seven board members from six countries: two infectious disease physicians, a biostatistician, a hospital infection control specialist, a public health policy researcher, and two epidemiologists. The board had reviewed the documentation package she had submitted: the network model files, the phylogenetic analysis, the genomic distance matrix, the WHO ID certificate. They had questions.

She unfolded the printout and placed it on the table. She had done this before — in Carmichael’s briefing room, eight months ago, unfolding it on the conference table and pointing at the branch. The crease ran through the red underline. She pointed at the branch.

“The branching point here represents the separation between the hospital transmission cluster and the community exposure cluster,” she said. “The genomic distance at this node — between the eight hospital cases — is zero. What that means epidemiologically is that these eight patients acquired their infection from a single contact event in a single location. This is not consistent with community spread, which would show a distance of at least two to three mutations between cases acquired from an environmental reservoir.”

The biostatistician said: “And the phylogenetic analysis was conducted by your team.”

“By Emilio Ferretti under my supervision,” she said. “He ran the sequence alignment. I identified the branching topology and interpreted the epidemiological significance. The analysis files are registered under WHO-EPID-VR-5517.”

The hospital infection control specialist said: “The porter’s movement log — how was that obtained and correlated with the network model?”

She explained: the movement log from the hospital’s patient transport records, cross-referenced with the case exposure timeline, entered into the network model as a potential transmission source node. The network model had identified the porter’s node as the hub before the phylogenetic analysis was complete — the epidemiological signal preceded the genomic confirmation. The phylogenetic analysis had confirmed what the network model had already shown.

The board worked through the methodology for two sessions — four hours across two days. She answered every question. She pointed at the branch seven times. She explained the genomic distance to four different board members who needed different entry points into the same conclusion.

On the second afternoon, Dr. Bjerkness said: “Dr. Russo — the board has completed the methodology review. The network model and phylogenetic analysis meet the WHO’s reproducibility standards. We will be citing this methodology in the global protocol for hospital cluster identification. The protocol will credit the originating methodology: Dr. Valentina Russo, Spatial Epidemiologist, WHO-EPID-VR-5517.”

She said: “Thank you.”

He said: “We’ll be in contact about the protocol publication.”

She folded the printout along the familiar creases. She put it in her folder.

—

The WHO review board record was published in the restricted WHO documentation system — accessible to member state health ministries: “Epidemiologist of record: Dr. Valentina Russo, WHO-EPID-VR-5517. Network model and phylogenetic analysis: Dr. Valentina Russo, WHO-EPID-VR-5517. Global protocol for hospital cluster identification: Methodology — Russo Network Model, confirmed phylogenetically by Emilio Ferretti under Dr. Russo’s supervision.”

Emilio read the board decision on the institute’s internal server. He came to her doorway.

He said: “WHO-EPID-VR-5517 in the global protocol.”

She said: “Yes.”

He said: “The phylogenetic tree.”

She said: “The branch.”

She picked up the printout from her desk. She unfolded it to the red-underlined branch. She looked at it. She folded it again.

—

Carmichael called the morning the review board record was published.

He said: “The global protocol. Your model is going to improve hospital cluster identification worldwide.” She said: “The network model is well documented.” He said: “Yes. I’ve amended the WHO case report to correctly credit your analysis — Dr. Valentina Russo, WHO-EPID-VR-5517, network model and phylogenetic analysis — and I’m updating the regional attribution policy for all WHO submissions.” She said: “Yes.” He said: “Good work, Valentina.” She said: “Thank you.” She folded the printout. She put it on the desk. She opened the network model.

—

Dr. Bjerkness had emailed her after the second board session concluded.

“Dr. Russo — Your network model and phylogenetic confirmation are the clearest documentation of a hospital superspreader cluster I have reviewed at the board level. The methodology has direct implications for how health ministries approach hospital outbreak investigations globally. The new protocol will go to all 194 WHO member states. It will be cited as the Russo Network Methodology. Your WHO ID is the reference number in the protocol documentation.”

She read “Russo Network Methodology.”

She read “194 WHO member states.”

She printed the email. She filed it in the cluster project folder — in the same folder as the original WHO case report with “network analysis support.” Both documents. She had not annotated either of them. She had not written anything in the margins.

She had closed the folder and returned to the new case data.

She took the phylogenetic tree printout from her folder and unfolded it — the A3 sheet folded twice to A4, the familiar crease running through the branching point. She had underlined the branch in red pen eight months ago, the first time she had printed it for Carmichael’s briefing. The crease ran through the underline. The red ink was still legible on both sides of the fold. The WHO review board’s second session was beginning — they were working through the methodology’s reproducibility for the new global protocol. Emilio had the supplementary network diagrams ready at the adjacent table. The WHO protocol document was in review: “Methodology: Dr. Valentina Russo, Spatial Epidemiologist, WHO-EPID-VR-5517.” Dr. Bjerkness indicated she should begin. She unfolded the printout and set it on the table. She pointed at the red-underlined branch. She explained the genomic distance.

—

The amended WHO case report arrived by email from Carmichael’s office the following week: “Network model and phylogenetic analysis: Dr. Valentina Russo, WHO-EPID-VR-5517.” She read it. She filed it in the cluster project folder — alongside the original WHO case report with “network analysis support” in the credits, alongside the review board documentation package she had submitted, alongside the global protocol citation. All of them in the same folder. She had not separated them.

The amended regional attribution policy arrived the same day: all WHO submissions from the regional authority would now credit named epidemiologists for analytical contributions, not operational roles. She read it. She filed it.

—

The WHO guidance circular to 14 national health ministries — the one issued before the review board convened — had cited “Carmichael’s analysis” as the model approach for hospital cluster identification. Nine of the 14 ministries had already implemented their hospital surveillance protocols based on the circular. She had the circular reference number: WHO/OPR/2025-047-B. The review board’s protocol update would be issued to the same distribution list. She had confirmed this with Dr. Bjerkness.

The nine ministries that had implemented protocols would receive the updated citation — “Russo Network Model, WHO-EPID-VR-5517” — alongside their existing implementation documents. Whether they would update their internal training materials, their case study files, their departmental records, she did not know. She had the circular reference. She had the confirmation of redistribution.

That was the extent of what was correctable from her end. The original circular was in the archive. It was there.

—

She was at her desk with the surveillance model open — a new region, a new cluster scenario, different pathogen but the same network analysis framework. She had added the phylogenetic printout to the project documentation file for the new case. It was not on the corner of her desk. It was in the new case folder, its appropriate place.

She was building the transmission network graph from the initial contact data — 22 cases, early analysis, no clear hub yet. She entered the coordinates for cases 1 through 8. She ran the first model pass.

Emilio came to her doorway. He said: “The protocol is published.”

She said: “Yes. Dr. Bjerkness sent the link.”

He said: “Russo Network Model. The citation is in the protocol document.”

She said: “Yes.”

He said: “The phylogenetic tree.”

She said: “The branch.” She did not look up from the model. She was entering coordinates for case 9.

She had 13 more cases to enter.

The new case was a respiratory cluster in a care facility — 22 cases, three wings of the building, no confirmed source yet. The network model’s first pass showed clustering in the east wing but no clear hub. She had three candidate source nodes: a staff member who worked across wings, a shared ventilation system, and a visiting family member with a travel exposure. She would need the movement logs for the staff member and more contact interview data from the east wing cases before she could distinguish between them.

She entered case 9. She ran the updated model. The east wing clustering tightened. The staff member’s node gained two new connections.

She entered case 10.

She had 12 more cases to enter. The model would run after all 22 were in. She entered case 11.