Patient selection and outcomes for out-of-hospital births in one family practice

Journal of Family Practice, August, 1990 by Louise S. Acheson, Stanley E. Harris, Stephen J. Zyzanski

From the Department of Family Medicine, Case Western Reserve University, Cleveland, Ohio. Presented in part at the Annual Meeting of the North American Primary Care Research Group, Baltimore, Md, April 14, 1986. Requests for reprints should be addressed to Louise S. Acheson, MD, Department of Family Medicine, Case Western Reserve University, 2078 Abington Rd, Cleveland, OH 44106,

Long a tradition or necessity in some rural areas, the opportunity for childbirth outside the hospital resurged in urban America as a response to the "natural childbirth" movement and more recently has been curtailed by concerns about litigation.[1] Yet out-of-hospital childbirth is a lower-cost alternative in a time of rising medical care costs and decreased insurance coverage. Furthermore, the hypothesis that low-risk births have good outcomes in low-technology systems of care has found support from numerous case series and a few retrospectively controlled studies, and has attracted the attention of policymakers.[1-15] Thus, a continued examination of the choice to deliver outside the hospital is needed.

This study describes the outcomes of pregnancies cared for in a family practice offering a choice of settings for childbirth, with a philosophy of avoiding unnecessary medical interventions and of fostering maximum involvement of patients in decisions about their care. The study differs from most other published case series by recording more prenatal variables, thereby allowing analysis of outcomes to control for obstetric risk status. The following questions were addressed: How do those patients who planned out-of-hospital birth (OHB) differ from others in the practice? Judging by outcomes, was selection of the childbirth setting appropriate? In the subgroup of women who planned OHB, could the eventual decision for transfer to the hospital have been predicted?

METHODS

The subjects for this study were all of the pregnant patients followed to delivery during a 4-year period from 1980 to 1984 in one family medicine group practice (two or three family physicians and a nurse practitioner) located in a suburb of a large Pacific Northwest city. The office suite (clinic) adjacent to a community hospital contained birth rooms where selected patients could give birth with a nurse and physician in attendance. Patients apparently at low risk, living close to the clinic, and able to pay a higher fee could choose to give birth at home. Only women prepared for natural childbirth were eligible for clinic or home birth. Intravenous fluids, oxygen, local (but not systemic or regional) analgesia, and drugs and equipment for resuscitation were available for all clinic and home births. Patients giving birth at the clinic could be accompanied by friends and family members of their choice and were usually discharged to home about 2 hours after the birth. Hospital deliveries occurred if the patient so requested or because of a need for a more intensive level of care. Most of these occurred in the local hospital with private practice obstetricians and anesthesiologists as consultants, although a small number of patients were referred to a tertiary care facility.

Each pregnancy record summarized the patient's and family's initial requests about the birth and their stated reasons for these requests. After a 1- to 2-hour semistructured initial interview, physical examination, and discussion, the planned location of the birth was recorded in each chart. The physicians attempted to follow the patient's requests and to provide adequate information to patients and families for an informed choice. The physicians made the final decisions about location of birth and transfer during labor, taking into account past history, medical factors, patient's and family's motivation and ability to prepare for childbirth, financial factors, and social support. No formal risk-scoring system was used. In this study these decisions about birth location were analyzed in relation to outcomes of pregnancy.

Patients were included in this study if they had at least one initial obstetric visit. Out of a total of 1004 such patients during the study period, 119 either left the practice or aborted; for an additional 66, charts were not available for analysis, leaving 819 pregnancies followed to delivery that could be examined in detail. Twenty-four twin pregnancies were omitted from analyses of birth location, resulting in a study sample of 795 singleton pregnancies followed to delivery. Data on planned or actual birth location were missing for five of these.

The mother's charts were abstracted by one author; demographic, prenatal, delivery, and postpartum variables were independently coded for computer analysis. Each of the variables was routinely charted during pregnancy care; the frequency of missing data on single variables ranged from 0% to 13%. Smoking was divided into three categories according to amount smoked at the time of the initial visit: none, less than one pack per day, and one or more packs per day. The obstetric risk-scoring systems of Halliday, Hobel, and Goodwin were retrospectively applied to the data, with minor modifications, as previously described. [16] Since the Goodwin et al score[17] was the best at predicting complications, it has been used in this study.