A MUAC referral alone did not change children's outcomes near the cut-off

Design

Prospective regression discontinuity analysis nested in a cluster-randomised trial

Population

65,554 children aged 6–59 months screened by MUAC in Burkina Faso; 226 below 11.5 cm

Primary outcome

MUAC and death by the next census

Effect

MUAC difference 0.21 cm (−0.22 to 0.65); mortality OR 0.95 (0.03 to 31.37)

This analysis was nested in a cluster-randomised trial in Burkina Faso, where 65,554 children aged 6 to 59 months had mid-upper arm circumference measured at twice-yearly censuses. A MUAC below 11.5 cm triggered a recommendation to attend a nutrition programme. The researchers compared children just below the threshold with those just above it, a design called regression discontinuity.

At the cut-off there was no detectable difference in MUAC at the next census (0.21 cm, 95% CI −0.22 to 0.65) or in death (OR 0.95, with an interval from 0.03 to 31). Only 226 children ever met the threshold, so the mortality estimate is almost uninformative.

The point is not that treatment of severe acute malnutrition does not work; that was not measured. It is that telling a family to go somewhere is not the same as the child being treated. In Indian community screening through anganwadi workers, the step that matters is the one after the tape measure: whether the child reaches and completes care.

• Confirm that a child flagged on MUAC actually reached a nutrition programme; a referral slip is not treatment.
• Follow up a missed referral at the next contact rather than waiting for the next screening round.
• Check for oedema as well as MUAC; bilateral pitting oedema is severe acute malnutrition whatever the arm measures.
• Link with the anganwadi worker or community health worker who did the screening.

Why it matters

It suggests that screening programmes lose children between the referral and the treatment.

Don't overread it

This tested eligibility for referral, not treatment; it says nothing against treating severe acute malnutrition.

The statistics, in plain English

A mortality odds ratio of 0.95 with a 95% interval from 0.03 to 31 means the data fit anything from a large benefit to a large harm. Too few children were near the cut-off to answer the mortality question.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.