*2.4. Training and Procedures for Data Collection*

Training was provided by the local team members/co-authors (Brazil C.R.L.A., Malawi L.M.S., Pakistan M.A.R.), with a consistency check visit by one of the other co-authors (Brazil P.K., Malawi M.G., Pakistan V.C.). Training took place over a period of 2–3 days at each site and was conducted in person at each site. The cost of the training depended on the cost of the trainer visit, hiring a room and transport costs to the training location for children and families attending. The training procedures were monitored so a formal "training the trainer" package could be developed. Training materials are available through the IYCD website (https://ezcollab.who.int/iycd, accessed on 2 June 2021).

Once trained, the country teams conducted assessments with caregivers. Generally, the IYCD assessment was conducted first, followed by the physical measurement of the child. In seven cases in Brazil, three in Malawi and none in Pakistan, the IYCD assessment had to be ceased and reprised on another occasion. The majority of the interviews were recorded using tablets run on Open Data Kit software [20].

#### 2.4.1. Reliability Testing

At the initial assessment, a random subsample (*n =* 96; 32 from each of the three sites) was invited to come back for a second assessment to test consistency of measurement. Inter-rater reliability (two different assessors within the same day at a different time and test-retest reliability (same assessor 3–7 days apart) were conducted in the same setting where the child was seen originally (in Brazil and Malawi, this was done in the health centers, and in Pakistan this was done in homes).
