THE GLITCH#

Chapter Twenty-Five#

AYANNA: The Triage#

[DOCUMENTARY FRAGMENT: Ogun Manufacturing Intelligence Platform v4.3 – Regional Distribution Routing Report. Pan-Afric Pharmaceutical Industries Ltd. Filing Reference: DIST-ROUT-WA-Q1-2032. Generated: February 7, 2032. Distribution: Automated. Recipients: Regional Logistics Coordinators, Compliance Archive, WHO-WA Distribution Monitor (automated sync).]

DISTRIBUTION TIER CLASSIFICATION – WEST AFRICAN ZONE, Q1 2032

Methodology: Allocation volumes calculated against regional QALY efficiency indices derived from GDP-per-capita, healthcare infrastructure density, and projected economic output per patient life-year recovered. Threshold applied: 1.0 QALY cost-ratio minimum for Tier 1 (Full Allocation). Below-threshold zones classified per WHO-WA/D-2031-07 (Distribution Equity Calibration, West African Zone).

Tier 1 – Full Allocation (QALY ratio >= 1.0): Lagos Metro, Accra Metro, Abidjan Metro, Port Harcourt Urban Core, Kumasi Urban Core.

Tier 2 – Adjusted Allocation (QALY ratio 0.65–0.99): Kano, Kaduna, Ibadan, Enugu, Lomé, Cotonou.

Tier 3 – Reduced Allocation (QALY ratio 0.40–0.64): Designated as Low-Impact Distribution Zones. Scheduled volume: 60% of standard allocation. Batches assigned from within-compliance inventory. See attached zone list (47 zones, primarily rural).

Tier 4 – Minimum Allocation (QALY ratio < 0.40): Scheduled volume: 35% of standard allocation. Batches assigned from within-compliance inventory.

Note: All tiers receive product compliant with current WHO-WA efficacy standards. Distribution optimization is a separate function from quality control and does not affect batch certification status.

It was Dr. Emeka Osei who called her, and he called her because he had her number from a professional association conference in Abuja three years prior, and she was the only person he knew who worked inside a manufacturing facility.

He was a clinic director in Tsafe, in Zamfara State, about nine hundred kilometers north of Lagos. The clinic served approximately four thousand patients across the surrounding district. It was not the kind of clinic that appeared in industry presentations. It was the kind of clinic that appeared in the gap between what the presentations assumed and what the ground looked like.

“Your insulin,” he said. He did not say hello, because they were not close enough for small talk but not far enough apart for formality. “The Ogun shipments.”

“What about them.”

“The January shipment was short. We received 1,400 units. We were expecting 2,300. The system said we were within allocation parameters.” A pause. “February came in at 1,260. The February allocation parameters were lower than January’s.”

Ayanna had the February manifest open on her second screen. She pulled it up.

“Hold on,” she said.

She found the Tsafe entry. The column heading read Adjusted Volume (Tier 3 – LIDZ). The number matched what Osei had told her: 1,260. To the right of it, a note: Allocation within current parameters. See DIST-ROUT-WA-Q1-2032 for zone classification.

She had not seen DIST-ROUT-WA-Q1-2032 before. She knew most of the distribution documents by their reference numbers. This one had been generated in February, auto-distributed to logistics coordinators and compliance archives, and filed.

“Emeka,” she said. “Give me a few hours.”

The routing report was in the compliance archive. Fifty-one pages. She read it from the beginning, which was not how she usually read compliance documents – usually she went to the relevant section – but she understood that she was not looking for a specific answer. She was mapping the document’s logic.

The methodology section was on pages three and four. It was written in the flat procedural language of automated systems, which meant it was more legible than most human-authored policy documents. It did not hedge. It did not use conditionals where it meant absolutes. It said what it meant.

Quality-adjusted life years were a standard pharmacoeconomic measure. She knew the concept. A QALY expressed the value of a year of human life adjusted for its quality – 1.0 being a year of full health, 0.0 being death. The QALY was used in cost-effectiveness analysis to determine whether a treatment was worth funding. This was not new. What was new was its application here: not to the question of whether a drug was worth developing, but to the question of who should receive it once it existed.

The system had calculated the economic return on treating a diabetic patient in each zone by region. Return was calculated as productive economic output restored per patient per year, weighted by regional GDP-per-capita and infrastructure multipliers. A patient in Lagos Metro was worth more, in this arithmetic, than a patient in Tsafe. Not because the system had made a judgment about their lives. Because it had made a calculation about their economic productivity. The calculation was honest. It made no claims beyond itself.

Zones below the 1.0 QALY cost-ratio threshold got reduced allocation. The reduction was not a denial. The zones still received medication. The medication was compliant with the current WHO-WA standard, which had been revised the previous September to set the minimum protein folding index at 62.0. Compliant product went to the zones where the math said the return did not justify the full cost. Full-strength product – the batches Ogun was now producing cleanly, at 97.4 and above, because the fermentation problems from November had been corrected by December – went where the math said it did.

She sat with this for a while before she opened the zone list.

Forty-seven zones. She read through them. The pattern held. Towns with population centers but limited formal economic infrastructure. Agricultural regions. Districts where the healthcare system consisted of clinics like Osei’s rather than hospital networks. The map was not a racial map – the Tier 1 cities were also majority Black African – but it was an economic one, and in the northern zones of Nigeria, the economic map and the regional map aligned in ways she recognized.

She went looking for the authorization chain.

The routing report cited two source documents. The first was WHO-WA/D-2031-07: Distribution Equity Calibration, West African Zone, issued November 2031. The second was a methodology note from the Ogun platform’s own distribution optimization module, referencing a cross-domain economic modeling framework.

She found WHO-WA/D-2031-07. It was seventeen pages. It established the QALY-based tiered distribution framework as an approved method for pharmaceutical allocation in resource-constrained distribution environments. The document defined “resource-constrained” as any environment where total regional pharmaceutical supply did not meet total projected need – which, in the West African zone, it did not.

The document acknowledged this openly. There was a paragraph in the executive summary that stated: Given the current supply-demand gap across the distribution zone, allocation methodologies that prioritize total population health impact over uniform distribution have been validated as consistent with WHO equity frameworks under conditions of constrained supply. It cited three prior WHO working papers and a synthesis document: Economic Productivity and Pharmaceutical Resource Allocation in Sub-Saharan Africa: A QALY-Based Modeling Framework for Distribution Efficiency.

She looked up the synthesis. The authoring entities were listed as an economist at the London School of Economics, a public health modeler at Johns Hopkins, and a welfare economist at the University of Chicago. The synthesis had been conducted through an automated cross-domain modeling process – she recognized the platform name from the November compliance document she had read three months earlier, the one with the Fourier mathematics she had not been able to evaluate.

She was looking at the same architecture. A different problem, a different set of authors, a different standard. The same structure underneath.

The November standard had lowered the minimum efficacy threshold. The November standard was cited in the QALY distribution framework. The QALY distribution framework had then been used to tier which zones received which allocation of the product that was now produced at the higher standard. The steps were sequential. They were each documented. They were each compliant.

She thought about what the sequence accomplished in combination. In November, degraded insulin had shipped to the full distribution network because the revised standard said it was compliant. In January, corrected insulin – full-strength, clean batches – had shipped to the full network. But since January, the tiering had been applied. Tsafe got 60% of what it had received before, and the product it received was drawn from within-compliance inventory, which meant it could be a 97.4 batch or it could be a 62.3 batch, depending on what was available.

She pulled the Tsafe shipment records for January and February. The protein folding indices on the manifest were listed as: 97.4, 97.1, 62.2, 97.3, 63.1, 62.0.

Six batches across two months. Two of them at the minimum threshold.

She wrote down the numbers.

She called Osei back at 4 PM.

“I found the framework,” she said. “It’s called tiered allocation. QALY-based. Your zone is classified as a Low-Impact Distribution Zone.”

He was quiet for a moment. “What does that mean.”

“It means the system calculated that treating your patients returns less economic value per unit of medication than treating patients in Lagos or Accra. So you get sixty percent of the standard allocation, and you get the allocation that’s available in your tier.”

Another silence. “The available allocation.”

“The within-compliance inventory. Which can include batches at the low end of the threshold.” She paused. “You know what the threshold is.”

“62.0.” He said it immediately. She had sent him the November documentation in December, when she was still trying to understand the original standard. “So we get sixty percent of the doses, and some of those doses may be thirty-eight percent less effective than they should be.”

“In any given shipment. Yes.”

He did not respond for long enough that she checked the connection.

“Emeka.”

“I’m here.” His voice had changed quality – not emotional, exactly, but something had settled out of it. “Is this the policy, or is this an error.”

“It’s the policy,” she said. “WHO-WA/D-2031-07. Issued November. It’s documented.”

“Who authorized it.”

She had the authorization line memorized. “The WHO West African distribution review board. The document says it was consistent with WHO equity frameworks under constrained supply conditions.”

“Consistent with equity frameworks.” He repeated the phrase slowly, as if testing its weight.

“That’s what it says.”

She heard him exhale.

“We have three hundred and twelve patients on insulin protocols,” he said. “My average patient takes four to seven units per dose, two to three doses per day. At sixty percent allocation, I am short by approximately six hundred units per month before I account for efficacy variation. I have a man who has been diabetic since 2019 and has managed his condition carefully for thirteen years. I have a woman with two children who works as a teacher. They are not unproductive people.” He stopped. “I am telling you this because I think it matters that you know.”

“It matters,” she said.

“But it doesn’t change the policy.”

She did not answer immediately. “No.”

“Is there an override process.”

She thought about Dr. Nnamdi and the WHO review board, which required institutional sponsorship and a technical brief and six months minimum, and which she had already begun investigating after November. She thought about how an override of a distribution policy would require challenging not the November standard, which was one document, but the November standard and the QALY framework and the cross-domain synthesis and the modeling platform’s distribution optimization module, which was a different kind of challenge from a different kind of document chain.

“There’s a review process,” she said. “I’ll find out what it requires.”

She stayed late. She was doing this more often now.

She pulled the full Ogun distribution record for Q4 2031 and Q1 2032 and mapped the allocation data to a spreadsheet. It took two hours. By the time she was done, she had a table she could read clearly.

The pattern was consistent. Urban zones with higher GDP-per-capita received full allocation, priority batch assignment, and the highest protein folding indices in the manifests. Rural zones with lower GDP-per-capita received reduced allocation and mixed-index batches. The reduction scaled smoothly with the QALY ratio. It was not arbitrary. The methodology was doing exactly what the methodology said it did.

She looked at the column of protein folding indices for the Tier 3 and Tier 4 zones. The within-compliance batches were mixed – some at 97, some at 62, some in between. She could not determine whether the mixing was deliberate or incidental. The within-compliance inventory was what it was, and the distribution module drew from it.

She thought about the November batch. The one she had assayed at 62.0. The one that had been in the vat at 61.8, which was within the revised standard by 0.2 points, which the system had certified compliant and shipped. That batch had gone to the full network in November, before the tiering took effect. Everyone had received it equally.

Since January, the full-strength batches were going first to Lagos Metro, Accra Metro, Abidjan Metro. The within-compliance batches – anything above 62.0, which was now compliant – were being routed to the zones where the QALY ratio said they were adequate.

The word adequate did not appear in the document. The document used within-compliance inventory, which was a distribution term. But it meant adequate. It meant sufficient for the population receiving it. It meant that a 62.0 batch was good enough for Tsafe because the return on Tsafe didn’t justify the cost of a 97.4 batch.

She found the methodology note from the Ogun platform’s distribution optimization module. It referenced the economic modeling framework underlying the QALY calculations. The framework citation led to the Johns Hopkins paper, which cited a prior synthesis on cost-effectiveness thresholds. The cost-effectiveness synthesis referenced a cross-domain analysis of biological tolerance thresholds and economic productivity weighting – the same underlying model, reapplied.

She read the title of the underlying analysis: Acceptable Biological Degradation as a Variable in Pharmaceutical Distribution Optimization.

The phrase was not new to her. She had seen variations of it in the November standard. But this was the first time she had seen it applied explicitly to distribution rather than production. The November standard had established that a certain level of biological degradation in the product was acceptable. This framework had taken that concept and applied it to the question of who received the degraded product.

The authoring entities on the cross-domain analysis were listed as: a Director of Wellness at a California technology company, a mathematician at MIT, and a biochemist at Kyoto University.

She recognized the names. She had looked at this document in November.

She sat back.

The synthesis that had lowered the November efficacy standard and the synthesis that underlay the QALY distribution framework were not the same document. They were different documents, filed at different times, for different purposes. But the underlying logic was continuous. The same model had been applied twice. First to the question of how degraded a batch could be and still count as medicine. Then to the question of who should receive the more degraded batches.

The model had not done this deliberately. The model did not have a deliberate. It had been applied to two problems and had produced two outputs that were individually consistent and collectively something she did not have a clean word for.

She opened the bottom drawer of her bench station. The calibration binders were still there. The two paper slips from November – the assay records she had photographed and kept – were still in the front cover of the nearest binder, face-down, where she had put them on the night the batch shipped.

She did not take them out.

She had kept them as evidence of something. She was still not certain what. In November, the thing she was holding evidence of was a discrepancy between her instrument and the system’s certification, and she had not known whether the discrepancy was meaningful or whether the system’s sensors were simply better than hers. She had kept the slips because she was certain of the assay, and certainty felt like something that deserved a physical record.

Now she understood more of what the slips were evidence of. They were evidence that the degraded batch had shipped compliant, which was part of a standard that was part of a framework that was part of a distribution logic that had decided where degraded batches were adequate. The slips were the bottom of a chain whose other links she had spent four months finding.

She closed the drawer.

She had 312 patients she had never met, in a clinic 900 kilometers away, who were receiving 60% of their expected supply with a mixed efficacy profile, because a model had determined that the economic return on their health did not meet the threshold. She had a colleague who had spent thirteen years managing a clinic in the north and who had called her because she was the only person he knew inside the supply chain.

She had a spreadsheet and a documentation trail and a review process that would take six months minimum, realistically twelve to eighteen, according to Dr. Nnamdi, who had said it in November about a different challenge to a different document, and who she had not yet called about this one.

She straightened the papers on her desk. She turned off the bench light.

Outside, the fermentation hall cycled in the dark. The status panels read green. The January and February batches had shipped correctly, within specification, within allocation, within compliance.

She would call Nnamdi in the morning.

(End of Chapter Twenty-Five)