Objective
The objective of the study was to evaluate attitudes and practice patterns of obstetricians related to screening for group B streptococcal colonization and providing intrapartum antibiotic prophylaxis against early-onset neonatal infections with group B streptococcus.
Study Design
We mailed a survey to 546 members of the American College of Obstetricians and Gynecologists, including members of the Collaborative Ambulatory Research Network and non-Collaborative Ambulatory Research Network members. Stratified random selection was used to generate samples from both of these groups.
Results
The survey response rate was 60% for Collaborative Ambulatory Research Network members and 42% for non-Collaborative Ambulatory Research Network members. Of the 206 respondents who reported providing prenatal care, 97% collect screening samples at 35–37 weeks’ gestational age. Anatomic sites used to collect samples were more variable: 62% include lower vagina and rectum, 26% include lower vagina and perianal skin but not rectum, and 5% include neither the perianal skin nor the rectum. First-line agents for intrapartum antibiotic prophylaxis were penicillin (71%), ampicillin (27%), and cefazolin (2%). For patients reporting a nonanaphylactic penicillin allergy, drugs used for intrapartum antibiotic prophylaxis were more varied: cefazolin (51%), clindamycin (36%), vancomycin (8%), and erythromycin (5%). For patients undergoing a labor induction starting with a cervical ripening agent, less than 40% typically give the first dose of intrapartum antibiotic prophylaxis before or at the time of cervical ripening agent administration, and 15% wait until the patient reaches the active phase of labor.
Conclusion
Gaps in knowledge and reported practice related to the prevention of early-onset neonatal group B streptococcus infections were similar to gaps in implementation of guidelines demonstrated in past studies. New approaches to improve implementation are warranted.
About a quarter of pregnant women are vaginally and/or rectally colonized with Streptococcus agalactiae , commonly known as group B streptococcus (GBS). In the absence of intervention, this colonization results in early-onset (first week of life) neonatal infection in 1–2 per 1000 live births in the general obstetric population.
A few decades ago, intravenous intrapartum antibiotic prophylaxis was shown to substantially reduce the incidence of early-onset neonatal GBS infection. Since that time, there has been widespread adoption of screening of pregnant women for colonization and use of intrapartum prophylaxis, as recommended by the most recent revision of the guidelines from the Centers for Disease Control and Prevention (CDC) and key partners that include the American College of Obstetricians and Gynecologists (ACOG) ; the ACOG’s GBS prevention statement is fully harmonized with that issued by the CDC.
This change in practice has resulted in an 80% reduction in early-onset GBS infection since the early 1990s. Currently there are approximately 950 cases per year in the United States, consistent with 0.24 per 1000 live births.
Despite this improvement in early-onset neonatal infection, a recent case series from the Active Bacterial Core surveillance system concluded that optimal implementation of the guidelines could reduce the rate by another 26–59%, with prenatal screening and intrapartum prophylaxis being the most common error types. We wanted to try to better understand the reasons for these errors. Therefore, the objective of this study was to evaluate the attitudes and practice patterns of obstetricians related to various aspects of screening for GBS colonization and providing intrapartum antibiotic prophylaxis against early-onset neonatal infections with GBS.
Materials and Methods
Survey questionnaires were sent to 546 ACOG fellows and junior fellows in practice between January and July 2014. Of these recipients, 246 were members of the Collaborative Ambulatory Research Network (CARN). CARN is composed of practicing obstetrician-gynecologists who volunteer to participate in survey research, and it was established to facilitate the assessment of practice patterns and development of professional education. The other 300 survey recipients did not belong to CARN (non-CARN); they were randomly selected by computer from ACOG fellows and junior fellows in practice.
The survey, accompanied by a cover letter explaining the purpose of the study and a postage-paid return envelope, was first distributed in January 2014. Four follow-up mailings were sent to nonresponders. The study was approved by the Institutional Review Board of the University of Alabama at Birmingham.
The survey included demographic questions including sex, age, years in practice, board certification, and practice characteristics (type, location, patient mix, proportion of time spent doing obstetrics). Only physicians who provided prenatal care were asked to complete the survey. To assess the attitudes and practice patterns related to GBS, respondents were asked about prenatal screening, intrapartum screening, intrapartum prophylaxis, a potential GBS vaccine, and reasons for their attitudes and practice patterns regarding these topics.
Data were entered into a computerized spreadsheet (Excel 2010; Microsoft Corp, Redmond, WA). All data management and analyses were done using SAS version 9.3 (SAS Institute, Inc, Cary, NC). For responses, frequency counts, proportions, and exact 95% confidence intervals about these proportions were calculated. If a respondent provided no answer to a specific question, their data for that question were excluded both from the numerator and the denominator. In planned secondary analyses, results were stratified by physician sex and by years in practice since residency. Although physician sex is not a variable by which such data frequently are stratified, others have done so, and some have found differences in care delivered by sex.
For pairwise comparisons, we utilized the χ 2 test of association or the Fisher exact test, as appropriate. For ordinal outcomes, we used the Cochran-Armitage trend test. Because we conducted multiple comparisons, statistical significance was evaluated at a level of P = .01.
Results
The survey response rate was 50.0% overall (273 of 546), with 60.2% (148 of 246) for CARN and 41.7% (125 of 300) for non-CARN. Of the 273 respondents, 206 reported providing prenatal care, 116 CARN and 90 non-CARN. Our study group was comprised of these 206 obstetricians ( Figure ). Of these obstetrician respondents, 91.6% reported their primary specialty as general obstetrics and gynecology, 6.4% as maternal-fetal medicine, and 2.0% as obstetrics only or other. Additional demographic data regarding these physicians are presented in Table 1 . Reported descriptions of the patient populations these physicians serve are shown in Table 2 .
| Variable | Statistic |
|---|---|
| Age, y | 48.3 ± 11.9 |
| Years in practice after residency | 16.4 ± 12.4 |
| Sex | |
| Female | 66.0% |
| Male | 34.0% |
| Board certified a | 82.5% |
| Type of practice a | |
| Solo private practice | 14.2% |
| Obstetrician-gynecologist partnership/group | 47.6% |
| Multispecialty group | 13.7% |
| Military/government | 1.0% |
| University full-time faculty and practice | 15.2% |
| HMO/staff model | 3.9% |
| Other | 4.4% |
| Practice location a | |
| Urban-inner city | 19.0% |
| Urban-non–inner city | 26.3% |
| Suburban | 33.2% |
| Midsized town | 16.1% |
| Rural | 5.4% |
| Percentage of clinical activities | |
| Obstetrics | 50% (40-60) |
| Gynecology | 50% (35-60) |
| Percentage of time spent in various areas | |
| Clinical | 90% (80-95) |
| Teaching | 2% (0-10) |
| Research | 0% (0-0) |
| Administrative | 5% (0-10) |
a There were x missing values for each of these characteristics: board certified (n = 12), type of practice (n = 2), and practice location (n = 1).
| Variable | Statistic |
|---|---|
| Race | |
| White | 60% (40–75%) |
| Black | 10% (5–30%) |
| Hispanic | 10% (4–20%) |
| Asian | 2% (0–5%) |
| Native American | 0% (0–1%) |
| Other | 0% (0–2%) |
| Patient education a | |
| <12 y | 6.4% |
| High school degree | 48.8% |
| College degree or advanced degree | 27.6% |
| Unknown or population is too varied to respond accurately | 17.2% |
| Source of payment | |
| Private insurance | 60% (35–80%) |
| Medicaid | 23.5% (6–40%) |
| Medicare | 5% (1–10%) |
| Unfunded | 1% (0–75%) |
a There were 3 missing values for the variable, patient education.
When asked about routine timing for collection of samples, excluding urine cultures, to screen for GBS colonization, 97.1% of respondents (95% confidence interval [CI], 93.7–98.9%) reported collecting at 35-37 weeks, in compliance with the CDC guidelines. However, 4.9% of respondents (95% CI, 2.6–8.9%) also reported collecting samples at the first prenatal visit, and 0.5% (95% CI, less than 0.01% to 3.0%) reported collecting samples both at 35–37 weeks and later. There were 1.0% of respondents (95% CI, 0.4–3.7%) who reported routinely collecting samples earlier than 35 weeks and 1.5% (95% CI, 0.3–4.4%) who reported routinely collecting samples after 37 weeks. Of 204 respondents who answered the question, only 0.5% (95% CI, less than 0.01% to 3.0%) reported not screening for GBS colonization. For this and other questions, responses did not differ significantly between respondents who reported being board certified and those who did not (data not shown).
Anatomic sites routinely used to collect samples for prenatal screening were more variable: of the 204 respondents who answered the question, 62.3% (95% CI, 55.2–68.9) reported sampling from the lower vagina and rectum (in compliance with the CDC guidelines), and 25.5% (95% CI, 19.7–32.1) reported collecting from the lower vagina and perianal skin but not the rectum. There were 4.9% (95% CI, 2.4–8.8) of respondents who reported collecting from site(s) that included neither the perianal skin nor the rectum and 7.4% (95% CI, 4.2–11.8) who reported collecting from the rectum and/or perianal skin but not the vagina. Among respondents, 3.9% (95% CI, 1.7–7.6) included the cervix in the sites from which the sample is routinely collected.
Anatomic sites routinely used to collect samples for prenatal screening did not differ by physician sex ( P = .199) or whether in practice more or less than 15 years ( P = .232). The following factors related to decisions about sites for collection of samples were listed as important or somewhat important by a majority of respondents: compliance with guidelines (94.1%), hospital/group practice policy (75.0%), maximize identification of colonized patients (91.2%), and medicolegal concerns (59.8%). Tolerability/patient comfort was listed as important or somewhat important by 40.6% of respondents.
These responses were not significantly different when comparing respondents who reported routinely collecting samples from the lower vagina and rectum with those who reported collecting samples from the lower vagina and perianal skin but not the rectum, although answers related to maximizing identification of colonized patients (95.3% vs 82.7%; P = .014) and tolerability/patient comfort (35.7% vs 54.9%; P = .019) approached significance.
When asked about which test they used for prenatal screening for GBS colonization, 67.0% of respondents (95% CI, 60.0–73.5) reported using culture, 12.5% (95% CI, 8.3–17.9) reported using polymerase chain reaction (PCR), and 9.5% (95% CI, 5.8–14.4) reported using both culture and PCR. There were 10.0% (95% CI, 6.2–15.0) who were unsure of which test they used, and 1.0% (95% CI, 0–3.6) reported using different tests, depending on the patient’s insurance.
The CDC guidelines call for prenatal screening. However, intrapartum screening, using nucleic acid amplification techniques such as PCR, is available. Most respondents (67.2%; 95% CI, 61.4–72.5) reported that none of their intrapartum patients are screened for GBS colonization, and only 4.8% (95% CI, 2.2–8.9) reported screening more than half of their intrapartum patients for GBS colonization.
Most respondents (70.7%; 95% CI, 63.8–76.8) reported using penicillin as their first-line agent for intrapartum antibiotic prophylaxis. Ampicillin (27.4%; 95% CI, 21.3–34.1) and cefazolin (2.0%; 95% CI, 0.5–5.0) were the other drugs reported as being used as first-line agents for prophylaxis. Those who reported using ampicillin as first-line agents were significantly more likely to consider availability of the drug as important or somewhat important in their decision to use it than those who reported using penicillin as first line ( P = .006). Other factors were not rated differently between respondents in those 2 groups ( Table 3 ).
| Factor | Ampicillin (n = 55) | Penicillin (n = 142) | P value |
|---|---|---|---|
| Availability | 81.5% | 60.6% | .006 |
| GBS antibiotic resistance | 90.9% | 81.0% | .090 |
| Antibiotic resistance of other bacteria | 33.3% | 45.8% | .115 |
| Compliance with CDC guidelines a | 92.7% | 93.7% | .758 |
| Concerns about adverse events from other agents | 59.3% | 47.5% | .142 |
| Cost of antibiotic | 54.6% | 45.8% | .269 |
| Dosing interval | 46.3% | 36.2% | .195 |
| Hospital/group practice policy | 74.6% | 73.9% | .931 |
| Medicolegal concerns | 60.0% | 56.3% | .641 |
| Pediatrician’s approach to the care of the infant | 54.6% | 47.9% | .402 |
| Placental transfer | 58.2% | 43.6% | .066 |
| Tolerability/side effects | 76.4% | 61.0% | .042 |
For patients reporting a history of a nonurticarial rash when exposed to penicillin, the drugs that respondents reported using most often for intrapartum antibiotic prophylaxis were cefazolin (51.2%; 95% CI, 44.1–58.3), clindamycin (36.3%; 95% CI, 29.7–43.4), vancomycin (7.5%; 95% CI, 4.2–12.0), and erythromycin (4.5%; 95% CI, 2.1–8.3).
There were no significant differences between proportions of respondents who rated various factors as important or somewhat important regarding the choice of agent in this situation when stratifying results by those respondents who reported using cefazolin (the drug suggested by the CDC guidelines) and clindamycin (the next most common option and one that the guidelines recommend against) for intrapartum antibiotic prophylaxis in patients who report a history of a nonurticarial rash reaction to penicillin ( Table 4 ).
