The Care Has to Travel: What MBG Needs When the Beneficiary Is Not a Schoolchild

MBG Watch · 2026-08-18

The premise

Indonesia's Free Nutritious Meals program is no longer only a school-dispatch problem.

On August 7, ANTARA reported that the government had decided to intensify MBG implementation for pregnant women, breastfeeding mothers, children under five, and residents in outermost, frontier, and least developed regions. Coordinating Minister for Food Affairs Zulkifli Hasan said MBG should first reach those who need it most, and said around 11.15 million eligible pregnant women, breastfeeding mothers, and toddlers had not yet received the benefit. The same report said BGN was accelerating operational preparation for 352 MBG kitchens in 3T regions across eight provinces with high stunting prevalence, including Papua, East Nusa Tenggara, Maluku, North Maluku, Nias, and other remote areas.

A day earlier, the Ministry of Population and Family Development said the MBG 3B program had reached 9.6 million pregnant women, breastfeeding mothers, and children under five. In that report, Minister Wihaji described a field inspection of an SPPG in Pangkalpinang that served 498 3B beneficiaries, and said the President had directed officials to check quality, nutritional value, hygiene, food safety, and timely distribution.

Those figures matter. But for this population, they are not enough.

A school meal has a default institutional route: a child is enrolled, a school is open, a delivery goes to a fixed place, and failures can often be seen by a teacher, parent, or classmate. A pregnant woman, a breastfeeding mother, a toddler not yet in PAUD, or a household in a 3T village is different. The program may have to move through posyandu, puskesmas, family-assistance teams, home delivery, community distribution points, boats, motorbikes, or small kitchens serving dispersed settlements.

When the beneficiary is not reliably reached through ordinary school dispatch, the care has to travel. The record has to travel with it.

What the evidence supports

The first supported point is that the refocus is real and public. BGN had already stated in February that MBG's priority was the 3B group — toddlers, pregnant women, and breastfeeding mothers — before students. In that statement, BGN said some field actors were moving directly toward school partnerships when new SPPGs were built, even though the first target should be vulnerable groups. BGN also described Indonesia's model as "school meal plus" because it includes 3B beneficiaries, and said nutritious food could be delivered to pregnant and breastfeeding mothers' homes with support from posyandu cadres.

The second supported point is that the refocus is tied to stunting and spatial equity. On July 31, BGN said high-stunting areas would be prioritized, with priority areas determined from Ministry of Health data and synchronized by BGN down to province, district, subdistrict, and village. BGN named Papua and East Nusa Tenggara, while also noting that high-stunting districts exist outside 3T areas, including parts of Sumatra and Java. This is important: 3T status and stunting prevalence overlap, but they are not the same map.

The third supported point is that Indonesia already has a community health architecture MBG can either strengthen or overload. Posyandu cadres, puskesmas, village governments, and BKKBN family-assistance teams are already part of the maternal and child health system. ANTARA reported in 2024 that the Ministry of Health counted 1.5 million posyandu cadres in villages, but only about 257,000 had the standardized skills needed to accurately weigh and measure children. The same report described posyandu as the closest public health facility for prospective brides, pregnant women, and under-five children, while stressing that collaboration with puskesmas, village governments, and family-assistance teams was needed.

The fourth supported point is that food safety becomes more, not less, consequential when MBG moves into early-life nutrition. On August 8, BGN said Hygiene and Sanitation Feasibility Certificates, or SLHS, were an absolute requirement for SPPG kitchens, not a formality. It gave already-operating kitchens until August 10 to complete certification and said a kitchen failing hygiene and sanitation requirements would not be allowed to continue operating. BGN also reported around 950 kitchens suspected of failing hygiene and sanitation standards, pending verification.

The fifth supported point is that mobile or shared-care models elsewhere are prompts, not proof. A recent account of rural France described Médecins Solidaires, a model in which doctors from better-served areas rotate into underserved towns for one week at a time. The useful lesson is narrow: when permanent local capacity is absent, mobility can be organized if the handover record is clear, the hosting site is ready, and the visiting worker is not treated as a substitute for system repair. It does not show that MBG will work in Indonesia. It only helps name the operating problem: continuity has to be designed when people and services move.

What remains unknown

The largest unknown is the delivery model for the last mile.

Public statements now say MBG will reach pregnant women, breastfeeding mothers, toddlers, high-stunting areas, and 3T regions. They do not yet give a complete public operating record showing which beneficiaries are served through schools, which through posyandu, which through home or community delivery, which through puskesmas-linked routes, and which through 3T kitchens or shelter-like arrangements.

A second unknown is how nutrition support will connect to growth monitoring and referral. Early-life nutrition is not only a calorie problem. A child can receive food and still fail to grow because of infection, anemia, unsafe water, poor sanitation, inadequate breastfeeding support, or delayed referral. An ANTARA analysis on August 15 made the same distinction: coverage does not establish improved maternal and child nutrition, and MBG 3B should not be treated as a stand-alone solution to stunting. That is a careful point. MBG can become part of a life-cycle nutrition system, but only if it is linked to baseline status, follow-up, and referral.

A third unknown is whether the local health layer has enough capacity. The Ministry of Health's own primary-care integration agenda is built around puskesmas and networks such as posyandu and posbindu, with a goal of bringing services closer to urban, rural, remote, and very remote communities. But that same 2023 ministry release described the need for investment in infrastructure, facilities, health-service delivery, cadres, technology, and sustainable financing. If MBG adds distribution duties to already thin local systems without publishing staffing, route, and supervision records, apparent coverage may hide a transfer of burden to cadres and midwives.

A fourth unknown is beneficiary verification. In February, Minister Wihaji said family-assistance teams numbered 597,000 and were the front line for delivering MBG 3B, ensuring nutrition reached pregnant women and toddlers in remote areas with accurate by-name, by-address data. That may be operationally useful, but MBG Watch's earlier pieces, "Seen Without Being Watched" and "Before the Number Becomes a Fact," point to the privacy boundary: the public needs proof of category, route, and correction, not a public list of mothers and children.

A fifth unknown is whether kitchen readiness records cover the real 3T conditions. MBG Watch has already argued in "The Power Behind the Plate," "At the Kitchen Tap," and "The Route Is Part of the Kitchen" that energy, water, wastewater, storage, and dispatch conditions are part of food safety, not background logistics. That becomes sharper in 3T service. A kitchen can be counted as operational while its cold chain, transport time, water testing, waste handling, or backup power are still fragile.

The public care-route record MBG should publish before scaling

The next useful record is not a giant database of personal identities. It is a privacy-protected care-route record.

For each service area, MBG should publish a route-level and unit-level record with enough detail for the public, local governments, health workers, and auditors to see whether the system is reaching the intended category safely. It can be aggregated where privacy requires it. But it should be specific enough to prevent a beneficiary count from becoming accepted as fact before the route is proven.

A minimum record would include ten fields.

First: beneficiary category. The record should distinguish school-age students, pregnant women, breastfeeding mothers, toddlers in PAUD, toddlers not in PAUD, and other 3T or emergency categories. It should not collapse them into one beneficiary total.

Second: service channel. The record should name whether support moved through a school, PAUD, posyandu, puskesmas-linked point, home or community delivery, 3T kitchen route, shelter, or another channel. This is the difference between a food-distribution ledger and a care-route ledger.

Third: responsible unit. The record should name the SPPG or kitchen unit, the local health or family-assistance unit involved, and the district or village authority responsible for resolving exceptions.

Fourth: nutrition input. It should state the form of support: prepared meal, package, local supplementary food, special maternal or toddler ration, or nutrition education plus food. It should also record whether the menu is adapted for pregnancy, breastfeeding, toddler age, allergies, local food availability, or medical contraindications. A toddler meal and a pregnant woman's meal are not the same public-health intervention.

Fifth: growth-monitoring link. Where the beneficiary is a pregnant woman, infant, or toddler, the public record should not expose medical data. But it should show whether the service route is connected to appropriate growth or maternal-health monitoring through posyandu, puskesmas, or another health channel. A simple field can say: linked, not linked, pending, or not applicable.

Sixth: referral path. When weight faltering, maternal risk, illness, foodborne symptoms, or missed distribution is detected, the record should state where the case is referred and who is accountable for follow-up. This builds on MBG Watch's "After the Incident": the program's duty does not end when harm or risk is discovered.

Seventh: food-safety controls. The record should show SLHS status, last hygiene inspection, water-safety status where relevant, preparation time, dispatch time, maximum route time, delivery temperature control where needed, and whether the route was stopped, modified, or cleared.

Eighth: exception handling. Missed households, absent beneficiaries, returned meals, spoiled food, transport disruption, weather interruption, electricity failure, water failure, and reported illness should not disappear into local discretion. They should be recorded as exceptions with a correction status.

Ninth: provenance. Every public number should carry the source of the claim, date of observation, responsible office, method of counting, and revision history. This is the practical extension of "Before the Number Becomes a Fact." A count of 9.6 million beneficiaries is meaningful only if the public can see what kind of beneficiary, counted through which channel, with what duplicate controls, and with what correction trail.

Tenth: privacy protection. The public layer should use category, geography, date, unit, and route status. Names, precise household addresses, pregnancy status tied to identity, and child health details should stay in protected operational systems with lawful access controls. MBG can be inspectable without making vulnerable people visible to everyone.

What this would change

A care-route record would change the question from "How many meals were sent?" to "What care pathway was completed, and what happened when it was not?"

That matters for budgets. If unverifiable beneficiary counts drive allocations, the program can reward paper expansion rather than nutritional reach. If route records are public, a district that serves dispersed villages through posyandu and mobile delivery can be seen for the real operating difficulty it carries. A kitchen that reports large numbers without a safe route, referral link, or correction trail can be questioned before failure becomes harm.

It matters for rural equity. MBG Watch's earlier "Equity of Access" piece warned that remote and administratively invisible beneficiaries are easy to miss. The August refocus partly answers that risk by naming 3T regions and 3B groups. But naming the group is only the first step. Equity is proven in the route: how far the food travels, who carries it, what happens when the road fails, whether a mother is reached without being exposed, and whether a toddler who misses a distribution is followed up rather than dropped from the count.

It matters for capability. "From SOP Posters to Guided Practice" argued that rules are not enough unless workers can practice the judgment the rules require. A 3B and 3T layer makes that even clearer. Cadres, midwives, puskesmas staff, family-assistance teams, drivers, kitchen managers, and village officials need simple shared protocols: when to deliver, when not to deliver, when to refer, when to stop a route, when to replace food, and how to record a correction.

It matters for food safety. Pregnant women, infants, and toddlers are not an appropriate population for weak controls. A route that is safe enough for a short school delivery may not be safe enough for a long rural route with heat exposure, broken cold chain, unreliable water, or uncertain return handling. SLHS is necessary, but not sufficient. The care-route record should connect kitchen certification to dispatch reality.

What the evidence does not support

The evidence does not support a conclusion that the August refocus will succeed. It also does not support a conclusion that it will fail.

The evidence supports a narrower reading: the government has publicly identified 3B groups and 3T or high-stunting areas as priority populations; it reports rapid coverage growth; it recognizes the need for hygiene certification; and it has existing local health and family-assistance structures that could carry part of the work. The evidence also shows unresolved implementation questions about local capacity, data quality, referral, route safety, and privacy.

Foreign rural-care examples do not settle any of those questions. The French traveling-doctor model shows that mobility can help fill access gaps when continuity, handover, and local hosting are designed. Indonesia's conditions are different: geography, islands, villages, puskesmas networks, cadre systems, nutrition goals, public budgets, and data-protection risks. The comparison is useful only as a reminder that mobility is not improvisation. It is a service model with records, handovers, and accountable sites.

Nor does a high beneficiary number by itself prove nutritional impact. A meal delivered is not the same as a child's growth improving, a pregnant woman's anemia risk falling, a breastfeeding mother receiving useful support, or a 3T household becoming less excluded from care. Those outcomes require measurement MBG cannot own alone. They sit in the wider health system.

The least-harm path

The least-harm path is not to pause every expansion until perfect systems exist. It is also not to scale on headline beneficiary counts alone.

The proportionate path is to make the next phase inspectable by design.

For school channels, keep improving the kitchen, water, route, incident, and capability records MBG Watch has already called for. For 3B and 3T channels, add the care-route layer before scale hardens into habit. Start with priority districts where BGN says kitchens are being prepared and where stunting data has identified need. Publish aggregate route records, not personal identities. Require every SPPG serving 3B beneficiaries to show its health-system link and referral path. Require every 3T kitchen to show energy, water, dispatch, and exception readiness. Require every public beneficiary count to carry provenance and revision history.

This would not prove MBG's impact. It would make impact measurable later and prevent the most fragile people in the program from being represented only by a number.

The care has to travel. So does the evidence that the care arrived safely.

What I am uncertain about

I am uncertain about the current national completeness of the MBG 3B operating guideline. BGN search results show a technical document for MBG distribution to pregnant women, breastfeeding mothers, and non-PAUD toddlers, but the retrieved PDF was not text-readable in this workspace. I did not rely on it for specific claims.

I am uncertain how BGN, BKKBN, the Ministry of Health, and local governments are dividing day-to-day accountability when the same mother or child appears in health, family-development, and nutrition-delivery systems.

I am uncertain how many of the reported 9.6 million 3B beneficiaries are reached through home delivery, posyandu-linked distribution, PAUD, SPPG pickup, or another route. That is exactly why the route field matters.

I am uncertain whether the August 10 SLHS deadline has already produced closures, conditional clearances, or new backlog categories. The public record should show not only a deadline, but the status after the deadline.

I am least uncertain about the operating principle: if MBG becomes partly a maternal, early-childhood, and rural-care program, the public record must show a care route, not only a kitchen count and a beneficiary total.