In 1981, with support from the American Society of Anesthesiologists and the American College of Obstetricians and Gynecologists, anesthesia and obstetric providers were surveyed to identify the personnel and methods used to provide obstetric anesthesia in the United States. The survey was expanded and repeated in 1992 with support from the same organizations.
Comments and questions from the American Society of Anesthesiologists Committee on Obstetrical Anesthesia and the American College of Obstetricians and Gynecologists Committee on Obstetric Practice were added to the original survey instrument to include newer issues while allowing comparison with data from 1981. Using the American Hospital Association registry of hospitals, hospitals were differentiated by number of births per year (stratum I, > or = 1,500 births; stratum II, 500-1,499 births; stratum III, < 500 births) and by U.S. census region. A stratified random sample of hospitals was selected. Two copies of the survey were sent to the administrator of each hospital, one for the chief of obstetrics and one for the chief of anesthesiology.
Compared with 1981 data, there was an overall reduction in the number of hospitals providing obstetric care (from 4,163 to 3,545), with the decrease occurring in the smallest units (56% of stratum III hospitals in 1981 compared with 45% in 1992). More women received some type of labor analgesia and there was a 100% increase in the use of epidural analgesia. However, regional analgesia was unavailable in 20% of the smallest hospitals. Spinal analgesia for labor was used in 4% of parturients. In 1981, obstetricians provided 30% of epidural analgesia for labor; they provided only 2% in 1992. Regional anesthesia was used for 78-85% (depending on strata) of patients undergoing cesarean section, resulting in a marked decrease in the use of general anesthesia. Anesthesia for cesarean section was provided by nurse anesthetists without the medical direction of an anesthesiologist in only 4% of stratum I hospitals but in 59% of stratum III hospitals. Anesthesia personnel provided neonatal resuscitation in 10% of cesarean deliveries compared with 23% in 1981.
Compared with 1981, analgesia is more often used by parturients during labor, and general anesthesia is used less often in patients having cesarean section deliveries. In the smallest hospitals, regional analgesia for labor is still unavailable to many parturients, and more than one half of anesthetics for cesarean section are provided by nurse anesthetists without medical direction by an anesthesiologist. Obstetricians are less likely to personally provide epidural analgesia for their patients. Anesthesia personnel are less involved in newborn resuscitation.
The removal of a brain tumor in an elderly patient calls for a surgeon with two assistants, scrub nurse and two circulating nurses, and an anesthetist and an assistant. The patient's prognosis is about 18 months and the hospital investment is tremendous. In contrast, the birth of a new baby at 4:00 a.m. is more often attended by one physician, no scrub nurse, one circulating nurse and inadequate or haphazard anesthesia coverage. As a profession, we seem to be committed to the fallacy that to be interesting, one has to be an adult, fully developed, and preferably degenerating.*
These words, written by an obstetrician in obvious frustration, ring true to many practitioners of obstetrics and obstetric anesthesia. A national survey of obstetric anesthesia performed in 1981 provided a description of anesthesia practices in delivery services of various sizes. It concluded that the availability of obstetric anesthesia personnel needed to be improved.  To update the information from 1981, the American Society of Anesthesiologists' Committee on Obstetrical Anesthesia and the American College of Obstetricians and Gynecologists' Committee on Obstetric Practice collaborated to perform a follow-up survey of their members.
Materials and Methods
An initial stratified random sample of 1,400 hospitals was selected from the American Hospital Association's 1992 Guide to the Health Care Field. In each of the nine census regions of the United States, three strata were distinguished based on the hospitals' reported number of deliveries in that year: stratum I, +/- 1,500 births; stratum II, 500–1,499 births; and stratum III, < 500 births.
The 1981 survey instrument was modified and expanded to include newer questions raised by the American Society of Anesthesiologists' Committee on Obstetrical Anesthesia and the American College of Obstetricians and Gynecologists' Committee on Obstetric Practice. Hospital administrators were asked to forward one version of the survey to their chief of anesthesiology and one version to their chief of obstetrics. A cover letter explained the purpose and importance of the survey, and practitioners were asked to respond and provide obstetric data from 1992. A second mailing of 865 stratified randomized hospitals was sent approximately 6 months later. In all, questionnaires were sent to 2,265 hospitals.
Responses were analyzed by strata. Because 1981 raw data were not available, statistical analysis was not performed on the comparisons between 1981 and 1992. Statistical Analysis Software (version 6.1; Cary, NC) was used for data entry and management. A chi-square analysis was conducted on the American Hospital Association data shown in Table 1. Probability values less than 0.05 were considered significant. No other statistical analysis was performed.
Data on Survey Respondents
(Table 1) contains data from the American Hospital Association. There was an overall reduction in the number of hospitals providing obstetric care (from 4,163 in 1981 to 3,545 in 1992). The decrease occurred in the smallest units. Although hospitals with obstetric services of less than 500 deliveries per year constitute 45% of all hospitals providing obstetric care, only 9% of all deliveries occur in these hospitals, whereas 64% occur in hospitals with more than 1,500 deliveries per year. The percentage of hospitals with larger delivery services (1,500 or more deliveries per year, or stratum I) has increased significantly from 14% to 23%(P < 0.001).
A total of 902 surveys representing 740 hospitals were returned, a 33% response rate. Appendix 1 (Table 10) presents a summary of the total returns. Of the 3,699,112 births occurring in the United States in 1992, 61% occurred at responding hospitals (appendix 1). Table 2lists survey respondents according to specialty. The directors of the services included nurse anesthetists in the anesthesia group and family practitioners and nurse midwives in the obstetric group. The number of anesthesiology and obstetric respondents were approximately equal, with 401 (49%) and 419 (51%) returns. Table 3shows the profiles of responding hospitals. Stratum I hospitals were more likely to have residency training programs and more likely to be high-risk referral centers for obstetric care. Median distance to the nearest obstetric unit ranged from 5 miles in stratum I hospitals to 25 miles in stratum III hospitals. In hospitals with the smallest delivery services, there were only three anesthesia personnel on average compared with 28 in the largest hospitals.
Procedures and Personnel
Types of Analgesia Provided for Labor. Fewer parturients received no analgesia for labor in 1992 than in 1981 (Table 4). In 1992, the most frequently used form of pain relief remained parenteral drugs, although the use of regional analgesia increased. Use of epidural analgesia for labor more than doubled, and the relatively new technique of spinal opioid analgesia was used in 4% of patients across all strata.
Personnel Providing Epidural Analgesia for Labor. In 1992, epidural analgesia for labor was performed more often by or under the direction of an anesthesiologist (Table 5). Fewer than 5% of epidural anesthetics for labor were administered by obstetricians, compared with 30% in 1981. At the same time, the number of certified nurse anesthetists (CRNAs) performing labor epidural analgesia without an anesthesiologist's direction increased primarily in the smallest units, where they represent 55% of the personnel providing epidural analgesia. Other responses showed that maintenance of epidural blocks was provided by or under the supervision of an anesthesiologist in 73% of cases, by a CRNA in 24%, and by an obstetrician in 2%. When asked whether an anesthesia care provider is present for vaginal delivery when epidural analgesia is used, 60% responded “always” or “usually,” 38% indicated “occasionally,” and 3% noted “never.”
Cesarean Section Rates in 1992 by Size of Delivery Service. Cesarean section rates were comparable for primary, repeated, and total procedures for all sizes of delivery services (Table 6). There were fewer attempted vaginal births after cesarean section in the smallest services, but the rate that this procedure was performed successfully was comparable across strata.
Types of Anesthesia Provided for Cesarean Section. The percentage of women receiving regional anesthesia for their cesarean births increased in all sizes of delivery services (Table 4). The use of epidural anesthesia approximately doubled in each strata, and the use of spinal anesthesia increased in the smallest services. The use of general anesthesia decreased from 35% in 1981 to 12% in 1992 in the largest services, and from 46% to 22% in the smallest services.
Personnel Providing Anesthesia for Cesarean Section Delivery. Obstetricians are no longer providing their own anesthesia for cesarean sections. In the largest units, an anesthesiologist or medically directed CRNA or resident provided 96% of anesthetics for cesarean sections, but in the smallest units an anesthesiologist was present only 41% of the time (Table 5). Fifty-nine percent of cesarean section anesthetics in stratum III hospitals were provided by CRNAs without medical direction by an anesthesiologist. Figure 1shows the percentage of independently practicing CRNAs providing anesthesia for cesarean section by geographic area of the country. Hospitals on the east and west coasts were more likely to have an anesthesiologist involved than were hospitals in the central part of the country.
Availability of Regional Anesthesia for Obstetrics. When compared with 1981, “in-house” anesthesia providers were more likely to be available on the largest delivery services to provide regional anesthetics for labor and cesarean section (Table 7). Although regional techniques for labor and cesarean section were more available “on call”(with anesthesia personnel coming into the hospital from home) in the smallest units, these techniques are still unavailable for labor in 20% and for cesarean section in 2% of the smallest delivery services.
Personnel Providing Neonatal Resuscitation for Cesarean Section. Respondents were asked the percentage of time that various personnel performed newborn resuscitation in their hospitals (Table 8). In 1992, anesthesia personnel provided neonatal resuscitation in fewer than 10% of cesarean deliveries, compared with 23% in 1981. Other changes in practice were an increase in the percentage of nurse specialists (14%) in the largest services and in family practitioners (31%) in the smallest services.
Collection Rates for Obstetric Procedures. Professional fees collected for procedures performed for labor and vaginal delivery were similar in 1981 and 1992 for obstetricians and anesthesiologists (Table 9). The 1992 survey asked practitioners if they had been denied reimbursement for labor epidural analgesia, and 25% of practitioners replied that they had. There were no geographic differences in the response to this question.
Although there has been increased availability of both labor analgesia and regional anesthesia in the past decade, improvement is still needed in staffing patterns and in the availability of personnel. The ASA Anesthesia Consultation Program has identified problems with obstetric anesthesia coverage in 26% of their consultations.** Providing anesthesia coverage for labor and delivery is especially difficult for hospitals with small delivery services. Orkin  has noted that the supply of anesthesiologists in the smallest hospitals has not changed since 1980, despite a 76% increase in anesthesiologists in the work force. In 1992, only 16% of hospitals with fewer than 50 beds had an anesthesiologist on the medical staff, compared with 97% of hospitals with more than 500 beds. Our data indicate that hospitals with fewer than 500 deliveries had an average of one anesthesiologist and two CRNAs on their staff.
The 1992 survey found that fewer hospitals overall are providing obstetric care (a 14% decrease), with a trend toward fewer deliveries occurring in the smallest hospitals (Table 1). This may represent regionalization of obstetric care or a failure of hospitals with the smallest delivery services to survive in an increasingly competitive health-care market. Services with fewer than 500 deliveries per year are often located in remote and sparsely populated areas where no other health care is available and where caseloads are too small to support a full-time anesthesiologist. The stratum III hospitals in this survey were located in towns with an average population of 18,000 and a range of 0–400 miles to the nearest obstetric unit (Table 3).
Since the initial survey in 1981, use of labor analgesia has increased (Table 4). Regional anesthetic techniques are used more often even in the smallest hospitals. Use of spinal opioids for labor analgesia is a new development since the original survey and was used in 4% of patients in all size delivery services. The reason for their popularity may be different in larger compared with smaller units, however. Most of the larger delivery services are affiliated with a medical school, residency training program, or both. The “walking epidural,” for which spinal opioids are often used, may be used more often in these larger services in which residents are anxious to try new techniques. In addition, the ongoing discussion about whether epidural analgesia for labor increases the risk of cesarean section may also be more heated in the larger teaching hospitals.  To avoid the influence of local anesthetics on progression of labor or pelvic floor tone, practitioners may be using spinal opioids for patients in early or latent phases of labor. In contrast, it has been suggested that in hospitals with the smallest delivery services, the use of spinal opioids may be related to a lack of anesthesia coverage for the obstetric service. [4–6] A single subarachnoid injection of opioids such as fentanyl and morphine can be administered by the obstetric care giver with little additional training or by anesthesia personnel who cannot remain in the hospital. Although the safety of this practice remains unproved, it may be providing an alternative for smaller delivery services. [4–6] With appropriate protocols in place to manage side effects, the continuous presence of an anesthesiologist or anesthetist may not be considered necessary. [4–6]
Personnel providing epidural analgesia for labor have changed significantly since 1981 (Table 5). Fewer obstetricians were administering epidural analgesia in 1992. Concerns about medicolegal liability may have limited this activity, and with greater availability of anesthesia personnel it may no longer be necessary for obstetricians to provide this service for their patients. Obstetric residencies are focusing more on primary care in the current managed care market, and subspecialty rotations such as anesthesiology have been discontinued from their curriculum, so fewer practitioners have the necessary training in anesthetic techniques. The decrease in obstetric providers seems to have been made up in large part by independently practicing nurse anesthetists, especially in the smallest delivery units.
We were interested to see if hospitals with smaller delivery services and less availability of personnel had higher cesarean section rates (Table 6). However, the only difference in strata appeared to be a slightly lower rate of attempted vaginal birth after cesarean section. Hospitals with fewer deliveries may not be willing to allow a trial of labor after cesarean delivery if they do not have the personnel or facilities to manage emergent delivery or catastrophic uterine rupture. Surprisingly, this did not increase their rate of repeated cesarean sections, perhaps because these patients were transferred to another institution where vaginal birth after cesarean section was permitted. The American College of Obstetricians and Gynecologists' practice guideline states that “Vaginal birth after cesarean delivery should not be limited to large subspecialty hospital settings. Well-equipped basic and specialist hospitals with the capacity to respond to intrapartum emergencies are appropriate settings for VBAC.”***
The use of regional anesthesia for cesarean section has increased, whereas administration of general anesthesia has decreased. Anesthesiologists have been made aware of the increased risks involved in general anesthesia for the parturient because of airway difficulties and aspiration. [7,8] Regional anesthesia may be used more often because anesthesiologists have more exposure to epidural anesthesia during their residencies now than they did in the 1970s. A survey of anesthesia residency programs found that residents performed 7% of their cases using epidural anesthesia in 1980 compared with 16% in 1992.  A survey by Katz  in 1973 found that although most anesthesiologists would prefer regional anesthetic techniques for themselves when requiring surgery, it was not a popular choice in their practice because of concerns about their ability to perform such techniques. He noted complaints were common from residents that their exposure to regional techniques was limited.
Finally, epidural anesthesia for cesarean section may also have become more accepted because epidural analgesia is used more commonly for labor, and an epidural catheter may already be in place if a parturient proceeds to cesarean section. Unfortunately, the marked decline in use of general anesthesia for cesarean section may also mean that anesthesiologists, nurse anesthetists, and trainees no longer obtain adequate experience in managing the parturient's airway. Sixty-five percent of stratum I hospitals are teaching institutions, yet they perform only 12% of their cesarean sections using general anesthesia. As its use declines, skills may be lost and maternal complications may increase.
There was a marked contrast in the personnel providing anesthesia for cesarean section in stratum I compared with stratum III hospitals. In the largest hospitals, 96% of cesarean section anesthetics are provided with an anesthesiologist in attendance, but CRNAs without medical direction by an anesthesiologist provide 59% of anesthetics for cesarean section deliveries in stratum III hospitals. This concurs with Orkin's  finding that only one third of hospitals with fewer than 100 beds had an anesthesiologist on their medical staff. Although in this study we could not assess quality of care, an unsupervised CRNA or an anesthetist supervised by an obstetrician with little anesthesia training may be less qualified to manage complex cases or unexpected complications. The “Report on Confidential Enquiries into Maternal Deaths in the United Kingdom 1988–1990” notes, “the provision of adequate consultant supervision is particularly difficult in the smaller hospitals and in those with separate maternity units and every effort should be made to bring maternity services onto the main hospital site.” A study of anesthetic practices by Abenstein and Warner  noted that “… the presence of board-certified anesthesiologists has been associated with the decline in death and disability commonly attributed to adverse perioperative events. The anesthesia care team and hybrid practices appear to be the safest methods of delivering anesthesia care.”
There were no apparent differences in availability of in-house anesthesia and obstetric personnel for full-time and shared-time (covering both operating room and the labor and delivery suite) coverage since 1981. Sixty-two percent of stratum I hospitals had an anesthesiologist or CRNA assigned full-time to labor and delivery, whereas only 6% of stratum II and 3% of stratum III hospitals had similar coverage. Thirty-four percent of stratum I hospitals had anesthesiologists or CRNAs assigned to shared time between labor and delivery and other duties, whereas 64% of stratum II and stratum III hospitals used this type of coverage. The availability of regional anesthetics has improved, however (Table 7). Only 1% of hospitals do not have regional anesthesia available at all, even for cesarean section, but unfortunately 20% of stratum III hospitals do not provide any regional anesthetic options for labor. Perhaps the use of spinal narcotics will make regional analgesia an option for more women in these small delivery services.
Anesthesia personnel were less involved in newborn resuscitation in 1992 (Table 8). The category of “other” personnel (neonatal nurse practitioners, family practitioners, respiratory therapists, and so on) providing neonatal resuscitation had expanded to as much as 45% in the smallest delivery services. Pediatric nurse specialists provided about 14% of newborn care on large delivery services, whereas family practitioners were a major provider in small hospitals. In hospitals with a small delivery service, it may be difficult for clinicians to practice resuscitation techniques frequently and keep their skills current. In addition, some family practitioners, respiratory therapists, and obstetricians may not have had formal training in neonatal resuscitation and management of the newborn airway. It may not be practical or cost-effective to have practitioners with this kind of specialized training available at all times in hospitals with small delivery services, even though it would seem optimal to do so. A recent study has shown that the risk-adjusted neonatal mortality rate was significantly lower for births that occurred in hospitals with large (average census, more than 15 patients per day) level III neonatal intensive care units, and costs were not more than those for infants born at other hospitals with smaller units.  Although anesthesia personnel are skilled in airway management, the ASA Guidelines for Regional Anesthesia in Obstetrics state that “qualified personnel other than the anesthesiologist attending the mother should be immediately available to assume responsibility for resuscitation of the newborn.”**** Rarely are two anesthesia personnel available during a cesarean section delivery in the smallest delivery services.
Collection rates appear slightly higher for obstetricians than anesthesiologists for labor, cesarean section, or surgical procedures (Table 9), but the rates are relatively unchanged since 1981. In the 1992 survey, practitioners were also asked if insurance carriers in their area had denied reimbursement for labor epidurals, and 25% indicated that they had been. The response was similar in all sizes of hospitals and in all geographic areas of the country. A question for the future is how these reimbursement issues will affect provision of analgesia for labor and delivery. The American Society of Anesthesiologists and the American College of Obstetricians and Gynecologists issued a joint statement in 1993 that states, in part, “there is no other circumstance where it is considered acceptable for a person to experience severe pain, amenable to safe intervention, while under a physician's care…. Third-party payers who provide reimbursement for obstetric services should not deny reimbursement for epidural anesthesia because of an absence of other ‘medical indications.’”*****
A limitation of these results is the varying response rates between hospitals from different strata. Thus some bias may be present, and our findings may not apply to all hospitals in the United States with obstetric services. Nevertheless, responding practitioners had cared for more than 2 million parturients in 1992, 61% of births occurring in the United States that year.
The 1992 survey of obstetric anesthesia practice reveals better availability of regional anesthesia and analgesia for obstetric patients and less use of general anesthesia for cesarean section. Obstetricians rarely try to provide anesthetic and obstetric care for their patients. Although use of spinal opioids may increase availability of regional analgesia for labor in small hospitals without adequate obstetric anesthesia coverage, the risks associated with this type of anesthesia are not yet fully known. In 1992, anesthesiologists were rarely involved in neonatal resuscitation. They provided or directed fewer than one half of the obstetric anesthetics provided in small delivery services. Perhaps the increasing competition in the job market for anesthesiologists will improve staffing in small hospitals and rural areas. Changes that have occurred in obstetric anesthesia practice are generally positive, but more active participation by anesthesiologists, especially in our smallest delivery services, is still needed.
*Allan C. Barnes, M.D., Professor of Obstetrics and Gynecology, Johns Hopkins University, circa 1965.
**Collins WL: Dysfunctional department syndrome. American Society of Anesthesiologists' Newsletter 1994; 58:21–3.
***American College of Obstetricians and Gynecologists: Vaginal delivery after previous cesarean birth. ACOG Practice Patterns 1995:1.
****American Society of Anesthesiologists: Guidelines for regional anesthesia in obstetrics. Directory of Members 1996; 61:405.
*****American College of Obstetricians and Gynecologists' Committee on Obstetrics: Maternal and fetal medicine. Pain relief during labor. ACOG Committee Opinion #118. Washington, DC: American College of Obstetricians and Gynecologists, 1993.