From home to hospital: The evolution of childbirth in the United States, 1928–1940☆

Within the first few decades of the 20th century, American hospitals underwent a significant transformation from almshouses to centers of medical science. Possible explanations for the transformation and rise in the status of hospitals include changes in physician training and advances in medical technology that led physicians to prefer hospitals, as well as a growing confidence in hospitals among consumers and a belief that hospitals provided more effective medical care (for greater discussion see the Commission on Hospital Care, 1947, Stevens, 1989).

While data on generalized hospital outcomes in the early 20th century is not available, it is possible to look at one outcome in particular to help shed light on the factors underlying the rise in hospitals: childbirth. The transition of childbirth from homes into hospitals mirrors the rise in the use of hospitals overall.

In 1900, only 5% of all births occurred in hospitals, compared to nearly 75% of urban births and half of all births by 1935 (Wertz and Wertz, 1977, p. 133). While the factors underlying the shift from home to hospital in childbirth may not be completely generalizable, they do offer some insight into the growth of an industry that today consumes nearly 5% of GDP.

An intuitively appealing explanation for the sudden growth of hospitals is that they offered superior outcomes compared to alternative home-based medical care. With respect to childbirth, it is not obvious that hospitals necessarily led to better outcomes; as birth shifted to hospitals, maternal mortality rates (measured as the number of maternal deaths per 100,000 live births) did not decline.

Fig. 1 shows total, urban (places over 10,000) and rural maternal mortality from 1915 to 1940.

Fig. 2 shows the death rate among women for tuberculosis and all causes related to childbirth (puerperal causes) from 1900 to 1940.

While tuberculosis death rates fell throughout the period, maternal mortality rates, aside from the spike associated with the 1918 influenza epidemic, did not begin to decline until the 1930s.¹ Further, infant mortality rates due to birth injuries increased 40–50% between 1915 and 1929 as hospital birthrates increased (White House Conference on Child Health and Protection, 1933, pp. 215–217).

Several historians of childbirth have argued that stagnant maternal mortality rates and rising rates of infant mortality due to birth injury occurred as a result of increased operative intervention on the part of practitioners as birth moved from the home to the hospital (see Loudon, 1992, Leavitt, 1986, Wertz and Wertz, 1977).

Unnecessary intervention may have led to excess maternal deaths for a number of reasons. A primary cause of maternal mortality was puerperal (related to childbirth) septicemia. Increased operative intervention in the form of version, forceps delivery and cesarean section all increased the mother’s likelihood of contracting such an infection, and complications from anesthesia could also lead to maternal death.

Operative deliveries were much more common in hospitals than in women’s homes: a study published in 1932 examined 40,143 births in both hospitals and homes and found that 24% of hospital births were operative compared to 8% of home births (Plass, 1932). Numerous studies observed that “in the hospital, experimentation and intervention greatly increased relative to home births” (Wertz, 1983).²

How preventable was maternal death in this period? A study published in 1931 noted:

… there has been a gradual but constant increase in the number of women entering hospitals for delivery in the past 10 years … that it is only in the past 15 years that any concerted effort has been made to develop on the part of the women a demand for antepartum care and better obstetric service, and to educate the physician to give this service; that in spite of this effort, just as many women die of sepsis now as formerly, a few less from the toxemias, but actually more from operative delivery (De Normandie, 1931).

Two larger studies also published in the early 1930s claimed that between half and two-thirds of maternal deaths occurred because poorly trained obstetricians relied too often on operative techniques. The studies concluded that the deaths could have been prevented by better training of the attendants (White House Conference on Child Health and Protection, 1933, New York Academy of Medicine Committee on Public Health Relations, 1933). If these arguments are correct, then modern medicine may have actually increased the number of maternal deaths due to childbirth than otherwise would have occurred.

The fact that a shift to hospital birth was not reflected in declining maternal mortality rates may indicate that the movement of childbirth from home to hospital occurred for reasons other than improved outcomes. Using a panel of city-level data over the period 1928–1940, we examine how the hospitalization of childbirth affected maternal mortality in the U.S., both before and after sulfa drugs were introduced on a widespread basis in 1937. Our findings indicate that medical intervention during childbirth had no statistically significant impact (either positive or negative) on maternal mortality rates until after the development of sulfa and improved obstetrics practices in the late 1930s.

Following the advent of sulfa, medical intervention decreased maternal mortality. This paper builds on the existing literature in two primary ways. First, while historians such as Loudon and Leavitt have hypothesized that excessive medical intervention may have led to increased maternal mortality as childbirth moved from home to hospital, they do not statistically test their hypotheses.

We build on their work by constructing a panel of city-level data that enables us to statistically measure the effect of hospitalization on maternal mortality across cities and over time. Second, if hospitals did not offer necessarily safer outcomes, why did women prefer them?

While we can only speculate as to the motives of women and their physicians, we do offer some possible explanations. The next section discusses the background and history of childbirth in the United States, and frames the model discussed in Section 3. Section 4 describes the data and results, while Section 5 concludes.