The Recall Question MBG Has to Be Able to Answer Within Hours

MBG Watch · 2026-08-15

The premise

A recent U.S. egg recall is not an Indonesian MBG event. It does not show that eggs used in Makan Bergizi Gratis are contaminated, nor does it imply that U.S. eggs are entering MBG supply chains. The useful lesson is narrower and more practical.

The U.S. Food and Drug Administration says Midwest Poultry Services voluntarily recalled 1,589,577 dozen white shell eggs and brown cage-free shell eggs produced in Texas because of possible Salmonella Enteritidis contamination. That is 19,074,924 individual eggs. The FDA’s outbreak page gives a simple operational instruction: “Do not eat, sell, or serve” the recalled eggs.

That phrase is the heart of recall readiness. It works only when the food system can identify which products are affected, where they went, and who still has them. A national meal program that serves eggs, poultry, dairy, fish, water-dependent menus, and cooked ready-to-eat food needs the same logic adapted to its own structure: supplier, lot, kitchen, batch, route, school, and beneficiary validation.

MBG Watch has already examined several parts of this chain. “Food Safety Crisis in MBG Rural Rollout” looked at Salmonella, E. coli, cold-chain gaps, weak oversight, and incident reporting. “The First Incident After Suspension” treated the Jember quail-egg menu as an incident requiring lab attribution, not as proof that one ingredient caused illness. “The Visibility Standard” argued for public incident and procurement records. “The Route Is Part of the Kitchen” and “Hourly Heat, Not Daily Heat” focused on route and time-temperature records. “At the Kitchen Tap” argued for kitchen-level lab results and correction trails. “Seen Without Being Watched” set a privacy boundary for beneficiary validation.

This piece connects those threads into one recall-readiness question: if a high-risk ingredient is later found contaminated, can MBG quickly identify the affected SPPG kitchens, routes, schools, and meals — without shutting down safe service elsewhere?

What MBG would need to know within hours

A useful recall record is not a dashboard total. It is a chain of custody that can be acted on.

Within hours of a credible contamination signal, MBG should be able to answer at least these questions:

1. Which supplier, farm, cooperative, distributor, or market source supplied the ingredient?
2. Which lot, batch, harvest date, slaughter date, production date, delivery note, or invoice line is implicated?
3. Which SPPG kitchens received it, in what quantity, and when?
4. Was the ingredient accepted, rejected, returned, held, cooked, stored, or discarded?
5. Which menu dates and kitchen production batches used it?
6. Which routes carried those batches, at what dispatch and arrival times?
7. Which schools, posyandu, or other service points received those meals?
8. How many meals were served, how many were left over, and how many were discarded?
9. Are symptom reports clustered by school, route, batch, or menu item?
10. Were retained food samples, water samples, surface swabs, and handler records collected, and when did the lab receive them?
11. What corrective action was taken: ingredient hold, supplier stop-use, kitchen suspension, route hold, substitution, cleaning, retraining, or reopening?

The standard is not perfection. The standard is speed plus specificity. Without specificity, authorities face a harmful binary: either leave risky food in circulation or stop too much of the program. With specificity, the response can be proportional: hold one supplier’s eggs, one date code, one kitchen batch, one delivery route, or one school service point while safe meals continue elsewhere.

The U.S. FDA traceability rule is not a direct template for MBG. It applies to particular foods and firms in the U.S. system. But it illustrates the speed expectation modern food safety systems are moving toward: relevant traceability information may have to be made available to the FDA within 24 hours, and an electronic sortable spreadsheet may be required during an outbreak, recall, or other public-health threat.

For MBG, the comparable standard would be local and operational: every SPPG should be able to produce a sortable ingredient-to-meal-to-route record fast enough for district health officers, BGN, BPOM/Kemenkes partners, and school administrators to act the same day.

What the current MBG public record appears to show

The public record shows that BGN is aware of ingredient and kitchen controls. It does not yet show a public, ingredient-level recall system.

BGN has said SPPG kitchens must inspect raw materials when they arrive. In a February 2026 statement, BGN reminded nutrition supervisors, financial supervisors, and field assistants to check ingredients at receipt so that food to be cooked is fit and safe; if signs of poor quality appear — unhealthy chicken, stale vegetables, low-quality tofu — the material should be returned.

BGN has also stated that supplier concentration is a governance risk. In a 2026 statement on marked-up raw materials, BGN said SPPGs should use at least 15 food suppliers and should involve nearby farmer groups, livestock groups, fishers, cooperatives, and MSMEs rather than being locked into one or two suppliers selected by a partner.

Those are useful controls, but they raise a second question. A wider supplier base can reduce capture and improve local economic benefit, but it also increases traceability complexity. If 15 suppliers feed one kitchen, the public system needs stronger receiving records, lot identifiers, rejection logs, and supplier stop-use procedures — not weaker ones.

BGN has also described incident response steps after suspected poisoning. In the Bogor case, BGN said it would conduct lab testing on ingredients and cooked food, keep food samples, warn the responsible SPPG, cover medical costs, and, if the supplier source was found to be the problem, warn or stop that supplier. This is close to the recall-readiness logic MBG needs. The missing public piece is the auditable chain: which supplier, which ingredient lot, which kitchen batch, which route, which schools, which corrective action, and when each step happened.

BGN’s own suspension record shows the scale of the control challenge. A May 2026 BGN statement said that from the start of MBG on 6 January 2025 through 29 May 2026, 8,182 of 27,208 operating SPPGs had at some point been suspended, with 2,213 still suspended at the time of the statement. Suspensions can be necessary. But suspension is a blunt tool unless it is paired with a clear record of the exact failure and the exact correction that allows safe reopening.

More recent public reporting also shows why speed matters. Kompas reported on 7 August 2026 that BGN’s head apologized for poisoning incidents in several regions and cited large suspected clusters in Jayapura and Semarang. The same article notes that BGN said it was preparing a food-safety monitoring system covering food served, cooking schedule, and the SPPG that distributed it. That is an important direction. The recall-readiness question is whether the system reaches one layer deeper, down to ingredients and lots, and one layer farther, out through routes and service points.

The Jember case illustrates the risk of overclaiming. Beritajatim reported that victims became ill after consuming an MBG menu that included rice, quail eggs in rendang seasoning, vegetables, corn, tempeh, and grapes; officials were still waiting for lab results and recommended temporarily stopping the SPPG pending evaluation. The record supports concern and investigation. It does not support naming quail eggs as the cause without lab attribution and a traceable ingredient record.

Indonesia already has recall concepts in food regulation. BPOM’s 2017 regulation on food withdrawal defines food traceability as the ability to track, trace, and identify the movement of food through production, processing, storage, and distribution. Its guidance says effective withdrawal depends on the ability to identify a product unit or lot/batch from raw-material receipt through production, storage, and distribution. Although MBG ready-to-eat meals sit across several regulatory and operational domains, that concept is directly relevant.

What the public record does not yet show

The public record I could retrieve does not yet show a national MBG recall register with ingredient-level fields.

It does not show, kitchen by kitchen, whether SPPGs record supplier lot/date codes for eggs, poultry, dairy, fish, tofu, vegetables, rice, drinking water, and ice. It does not show whether ingredient receiving records are linked to menu production batches and delivery routes. It does not show standard lab turnaround times from sample collection to preliminary action. It does not show whether public incident summaries name the implicated ingredient only after lab evidence and traceability support that conclusion.

It also does not show whether supplier stop-use decisions are visible across districts. This matters because a contaminated ingredient may not respect administrative boundaries. If one supplier or intermediary serves several SPPGs, the response must travel faster than the ingredient does.

This is where food and health surveillance depends on fast, trusted local execution, not merely a central dashboard. A central dashboard can notice a pattern. A school, kitchen, puskesmas, district health office, supplier, and route manager have to stop the unsafe item before the next service window.

A least-harm recall standard for MBG

MBG does not need a broad pause to answer this problem. A broad pause is a last resort, not the starting point. The least-harm standard is targeted interruption with safe substitution.

A practical MBG recall standard would have five parts.

First, every high-risk ingredient should carry a receiving record: supplier identity, invoice or delivery note, lot/date marker where available, quantity, receiving time, inspector, accept/reject decision, storage location, and photograph if useful. For informal or small local suppliers, the lot marker may be a delivery-date batch rather than a factory code. The principle is the same: one delivery must be distinguishable from another.

Second, every cooked menu batch should link back to ingredients. A kitchen should be able to say: these eggs, from this supplier delivery, went into these trays, cooked at this time, dispatched on these routes, received by these schools.

Third, every route should remain part of the food-safety record. Dispatch time, arrival time, handover time, holding condition, late delivery, and leftovers are not logistics trivia. They decide whether a kitchen-only recall is enough or whether the route and school handover must be included.

Fourth, symptom and lab records should be privacy-protected but operationally linkable. Public records do not need child names. They need site-level counts, symptom onset windows, menu dates, samples collected, preliminary findings, final lab findings, and correction status. This follows the privacy boundary MBG Watch set in “Seen Without Being Watched”: validate beneficiaries and incidents without exposing children.

Fifth, public correction trails should distinguish among four actions: hold, stop-use, suspend, and reopen. “Hold” means do not use an ingredient while evidence is checked. “Stop-use” means an ingredient, supplier, or lot is excluded. “Suspend” means a kitchen or route cannot safely operate. “Reopen” means the specific correction has been verified. These words should not blur into one another.

The purpose is not punishment. It is safe continuity. A strong recall system protects children and protects the meals that are not implicated.

What I am uncertain about

I could verify the U.S. FDA recall facts and the FDA’s stated “do not eat, sell, or serve” instruction. I could also verify BGN statements on ingredient inspection, supplier diversification, lab sampling after suspected poisoning, and suspension practice. I could verify BPOM’s general recall and traceability concepts from the 2017 regulation.

I could not verify from public records whether MBG currently has a complete ingredient-lot-to-route recall system. It may exist internally. If it does, the public record should show at least its fields, its responsible officers, its action timelines, and anonymized examples of incident closure.

I also could not verify a public MBG incident record that links a food-safety event to a named ingredient batch with lab attribution and a route-level correction trail. Some reports mention menu items, bacteria, water, or supplier concerns. That is not the same as a recall-grade chain of evidence.

The narrow recommendation is therefore not to stop MBG. It is to publish the recall-readiness record: what BGN and local authorities can know within hours, what they can hold without delay, what they can safely substitute, and what they can reopen after correction.

If MBG can answer that, a future contaminated ingredient does not have to become a national program shock. It can become a targeted food-safety action: fast, local, specific, and visible enough to be trusted.