Midwifery in New Zealand: Government Policies, Provider Choice, and Health Outcomes

Presenter: Andrea Menclova, University of Canterbury

Abstract

In New Zealand, about 80% of deliveries are fully taken care of by a midwife. This is at least partly due to government policies of the 1990’s financially favouring midwives. Following a series of reforms, a lead maternity carer (LMC) system was introduced in 1996. Under this system, each pregnant woman receives a fixed-dollar voucher from the government and chooses an LMC for her pregnancy and delivery: a midwife, a general practitioner (GP), or a specialist. An LMC is supposed to be selected at the beginning of pregnancy (but can later be changed) and plays an important role in the prenatal period as well as at delivery and in the postpartum period of up to four weeks. While a specialist can charge extra for his/her services, midwives and GPs have to accept a uniform fee.

This paper investigates the health care labour market changes following the introduction of LMCs (using data from 1998-2004) and evaluates the impact on birth outcomes (in 2003-2006). Importantly, the data identifies the LMC at first registration, i.e., the carer selected at the beginning of pregnancy. This is analogous to the intent-to-treat approach and removes much of the non-random selection in provider choice. Any remaining endogeneity is addressed in instrumental variable analyses using the gender composition of the medical workforce (and the fact that pregnant women tend to prefer a female provider, ceteris paribus).

The findings indicate that the reforms of the 1990s lead to an increase in the number of direct-entry midwives and a reduction in the number of GPs providing maternity care. Different types of women seek care from midwives and GPs. Controlling for observable individual and regional characteristics, GPs have a significantly greater percentage of very low birth weight babies but a lower neonatal mortality rate. When selection of LMCs along unobservable characteristics is controlled for using instrumental variable methodology, the GPs’ detrimental effects disappear but the beneficial effects in terms of reduced neonatal deaths persist. In particular, the preferred model indicates that having a general practitioner as an LMC reduces the neonatal death rate by 10%. Given New Zealand’s social, economic, and demographic characteristics, lessons learned from the local natural experiment are applicable far beyond the region.