Economic and personal factors affecting women's use of nurse-midwives in Michigan

The proportion of American women who use a nurse-midwife rather than an obstetrician to deliver their baby is increasing. Relatively little is known, however, about the determinants of a midwife-assisted birth.

Logistic regression analyses using birth-certificate data on 149,437 Michigan births in 1990 examined the characteristics associated with midwife-attended births.

Women who paid for childbirth with Medicaid were 3.5 times more likely than those paying with private insurance to use a certified nurse-midwife, net of controls for maternal characteristics; this effect varied significantly by race, with Medicaid payment increasing the odds of midwife use threefold among whites and nearly fivefold among nonwhites. The effect of education on midwife use also varied by race: A college education significantly increased the likelihood of midwife use among white women (odds ratio of 2.1), but higher education decreased that probability among nonwhite women (odds ratio of 0.74). Father's education and age, were also significantly associated with the likelihood of a midwife-attended birth; the babies of college-educated fathers had higher odds of being delivered by a nurse-midwife, as did the babies of men in their 30s.

The relationship between socioeconomic status and the use of midwives may not be as straightforward as previously thought. The patients of nurse-midwives are a diverse group whose socioeconomic characteristics and probable reasons for choosing a midwife over a physician vary widely.