n.V1-7.15 | PROFESSIONAL HOME VISITING
01| The mirror image of the previous two chapters. The Nurse-Family Partnership sends registered nurses — credentialled professionals, expensive ones — into the homes of first-time low-income mothers, from pregnancy through the child's second year. It has been evaluated in multiple randomised trials across different populations, with effects on parental care, child injuries, subsequent maternal pregnancies, workforce participation and reliance on public assistance, and in the first trial on adolescent outcomes fifteen years after the visits ended.
02| So highly professional delivery, sustained relational work, durable multi-domain effects. And lay grandmothers on a bench in Harare also work.
03| Both cannot be evidence for the proposition the sector argues about most — that local, non-professional deliverers reach people professionals cannot, or conversely that only trained professionals produce real outcomes. If credentialled and lay delivery both produce durable effects, then the credential is not the variable that determines success.
04| What both models share is what the failed scale-ups lack: a defined protocol, deliberate selection of who delivers it, real training, ongoing supervision, and a bounded caseload. Those five things are present in every programme in this family that has produced a result, and absent or degraded in every one that has disappointed. The structure is the variable. The credential is an input to the structure, not a substitute for it.
05| That reading has an uncomfortable corollary for anyone drawn to lay delivery on cost grounds. The reason to use non-professionals is that there are more of them and they cost less. But if the structure is what produces the effect, the saving is smaller than it looks, because the structure — selection, training, supervision, bounded caseload — is most of the cost and does not go away. A model that hires lay workers and skips the structure has not built a cheaper version of the programme. It has built something else.
06| The experiment that would settle this has not been run: the same protocol, same population, delivered by professionals in one arm and local lay workers in another, measured on the same outcomes. Until someone runs it, the relative contribution of credential and structure is inferred from a comparison across studies with different populations, different countries and different outcomes — which is exactly the kind of cross-study comparison this book part elsewhere warns against.
07| Home visiting also carries a specific limitation worth naming. The effects are concentrated on first-time mothers and their children, over a defined window, in a domain — infant health and parenting — where the right actions are known and the constraint is genuinely knowledge, confidence and attention. That is unusually favourable ground. It does not follow that a home visitor is a general-purpose instrument, and the programme has never claimed to be one.
08| Open, and cheap to answer relative to its value: what caseload a visitor can carry before effects degrade, and whether degradation is linear or a cliff. Neither this literature nor the community health worker literature has produced a caseload–effectiveness curve, which means every costing of every programme in this family rests on an assumption nobody has tested.
