Table 2 Would more testing have made a difference?

"Yes, absolutely. Legal constraints prevented proper implementation of the supply-side intervention. This constraint should have been detected (and resolved) before the [start of the trial]"

"...the number of young children eligible for the program and tracked by the information system degraded over time. Proper piloting of both the information system and the public communication campaign should have reduced this attrition."

"Piloting of recruitment and assessment and compliance rates would have been helpful but no funding for time to do this."

"I think there are ways that we could have improved the assistance package, but it was ready to be tested at the time we did the intervention trial."

"A full pilot phase would have alerted us to some of problems encountered such as inadequate team working."

"Almost certainly. The more the better really – in particular in depth interviews with providers and recipients of the pilot intervention or piloting in more than one setting could have resulted ultimately in an even more robust intervention."

"With hindsight what might have been useful would have been more data on exactly what happened in the black boxes to see which components were responsible for any benefits gained."

"Answer has to be yes but only if one had enough funds and time to do multiple tests."