Gray, Angela D.1,2; Carlson, Rebecca MS3; Morgan, Maria A. PhD2; Hawks, Debra MPH3; Schulkin, Jay PhD2
Human immunodeficiency virus (HIV) infection and the resulting acquired immune deficiency syndrome (AIDS) have become epidemic worldwide. In the United States, more than 500,000 people have died from HIV/AIDS, making it one of the leading causes of illness and death in the country.1 Presently, more than 1 million people are living with HIV/AIDS in the United States, and it is estimated that approximately 40,000 new HIV infections will occur this year.2
A growing proportion of HIV cases in the United States has been attributed to heterosexual transmission, with increased incidence especially among Hispanic and non-Hispanic black women.3 Over one fourth of newly diagnosed HIV cases are women.4 A major concern of infected women is mother-to-child HIV transmission. In an effort to reduce perinatal HIV transmission, the American College of Obstetricians and Gynecologists (ACOG), the American Academy of Pediatrics (AAP), the Centers for Disease Control and Prevention (CDC), and the Institute of Medicine (IOM) have recommended that all pregnant women be screened for HIV as a part of routine prenatal care.5–8 If women with HIV are diagnosed early, they can have access to prenatal, intrapartum and postpartum, and reproductive health treatment interventions that are important for their own health and the health of their newborns.5,6,9,10
Also, ACOG, AAP, CDC, and IOM have recommended that health care providers use an opt-out approach to HIV screening, wherein the patient is notified that she will be tested for HIV as part of the routine battery of prenatal blood tests unless she declines.5–8 The opt-out approach has been associated with higher rates of HIV screening than the opt-in approach, in which the patient must provide specific informed consent to an HIV test.11,12
Increasing acceptance of HIV testing is especially important because those who are infected, but unaware of their status, are at high risk for transmitting HIV to others.1 Obstetrician–gynecologists can be influential in the effort to encourage more women to be tested for HIV because obstetrician–gynecologists have become the primary care physicians for many women of reproductive age.13 As such, these physicians need to be prepared to discuss a range of medical issues, including HIV. The purpose of this study is to investigate the current knowledge and screening practices of obstetrician–gynecologists regarding HIV. We expect that most obstetrician–gynecologists will recommend HIV testing to all of their pregnant patients. We also expect that most obstetrician–gynecologists will report using the opt-in testing approach because most state regulations still require informed consent for HIV testing.
PARTICIPANTS AND METHODS
Survey questionnaires were sent to 1,032 ACOG Fellows and Junior Fellows in practice between June and July 2006. Six hundred eighty-seven of the recipients were members of the Collaborative Ambulatory Research Network (CARN). The Collaborative Ambulatory Research Network is a group of practicing obstetrician–gynecologists who volunteer to participate in survey research. The network was established to facilitate assessment of patterns in obstetric–gynecologic clinical practice and to help in the development of professional education. The other 345 recipients did not belong to CARN but were randomly selected by computer from ACOG Fellows and Junior Fellows in practice who had not received an ACOG questionnaire in the past 2 years (non-CARN). The survey was first distributed in June 2006 and then resent to nonresponders during two follow-up mailings.
This study included inquiries about physician characteristics, testing practices, and knowledge regarding HIV screening. The study was approved by the Georgetown University Institutional Review Board, and the survey questionnaire was accompanied by a cover letter explaining the purpose of the study. The survey packets sent to CARN members also included a questionnaire about HIV screening to be distributed to patients for completion; these findings will be described elsewhere. To view an online copy of the questionnaire, go to the Web site: http://www.acog.org/departments/dept_web.cfm?recno=16.
Data were analyzed with a personal computer–based software package (SPSS14.0; SPSS Inc, Chicago, IL). Descriptive statistics were computed for the measures used. Confidence intervals (95%) are provided for various proportions. Group differences in responses on continuous measures were assessed with independent sample t tests. Group differences on categorical measures were assessed with χ2 tests. Group differences on ordinal measures were assessed with the Mann-Whitney U or Kruskal Wallis tests. All analyses were tested for significance by using an alpha of 0.05.
A total of 582 surveys (56.4%) were returned. The response rate was 59.4% (408 of 687) for CARN participants and 50.4% for non-CARN participants (174 of 345). The demographics of the CARN participants were very similar to those of the non-CARN participants (Table 1). Because the CARN and non-CARN groups differed only in response rate and the use of the opt-in compared with the opt-out approach to testing, the data for the groups were collapsed during analysis.
Responses were received from obstetrician–gynecologists in every state in the United States except Wyoming, as well as from the District of Columbia, Puerto Rico, Canada, and military posts. Female obstetrician–gynecologists had a higher response rate than males to the survey (female 60.3%, 95% confidence interval [CI] 55.92–64.66%; male 52.1%, 95% CI 47.92–56.26%). The mean age of all participants was 47.55 years (standard deviation 10.24, range 30–86 years). More than one third of the respondents (37.9%) reported completing residency less than 11 years ago, 32.0% completed residency more than 20 years ago, and 30.1% completed residency 11–20 years ago. Almost three fourths (72.7%) of obstetrician–gynecologists reported that the majority of their patients were white, and four of five (80.2%) reported that the majority of their patients used private insurance.
The survey was divided into two sections. The first section was designed to be completed by obstetrician–gynecologists who treat pregnant patients. Those obstetrician–gynecologists who do not treat pregnant patients were instructed to skip to the section designed for all obstetrician–gynecologists to answer. Four fifths (80.4%, n=468) of the responding obstetrician–gynecologists reported that they treat pregnant women, and 19.6% (n=114) of the obstetrician–gynecologists reported that they do not treat pregnant women.
Almost all of the obstetrician–gynecologists (97%) reported that they recommend HIV testing to all of their pregnant patients. Of those who did not (n=14), the factor most commonly selected as influencing their decision to not recommend testing was having a low-risk patient population (n=12), followed by overly burdensome state pretest consent and counseling regulations (n=8). Overall, most obstetrician–gynecologists who see pregnant patients reported that they were at least moderately knowledgeable regarding HIV infection during pregnancy (83.3%, 95% CI 79.93–86.73%). The majority (92.0%) also reported that they maintain contact with an infectious disease specialist for consultation purposes. Of those who have managed pregnant HIV-positive patients, 67.0% reported that they always consult with or refer such patients to this specialist. Those obstetrician–gynecologists who completed residency fewer than 11 years ago rated themselves as more highly knowledgeable about HIV infection during pregnancy than did those who completed residency 11–20 years ago or more than 20 years ago (Kruskal Wallis χ2=24.96, P<.001).
Most (72.4%, 95% CI 68.34–76.48%) respondents considered 5% or fewer of their pregnant patients to be at high risk for HIV infection. Respondents estimated that a mean of 12.41% (95% CI 10.26–14.55%) of their patients initiate a request for HIV testing, with responses ranging from 0% (n=99) to 100% (n=4) and a median of 5%, indicating that most HIV testing is physician initiated. Respondents indicated that they would strongly recommend HIV testing for pregnant patients for several reasons (Table 2). Obstetrician–gynecologists reported that they use several methods to provide their pregnant patients with information about HIV testing, including the office nurse discussing HIV testing with the patient (56.2%), the physician discussing HIV testing with the patient (51.1%), and the office staff giving written information about HIV testing to the patient (37.4%). The majority of obstetrician–gynecologists reported that they screen their pregnant patients for HIV at about the same frequency as they did 5 years ago (66.8%, 95% CI 61.82–71.70%).
Obstetrician–gynecologists reported that patients gave several reasons for declining HIV testing, with patient perception of self as low risk as the most commonly selected (79.0%) (Table 3). If a patient arrived at labor and delivery with unknown or undocumented HIV status, the majority of obstetrician–gynecologists would recommend (56.3%, 95% CI 51.74–60.90%) rapid HIV testing, and almost one fifth (18.2%, 95% CI 14.62–21.74%) would not recommend rapid testing at that time. One fourth (25.5%) of respondents reported that rapid HIV testing at labor and delivery was not available at their hospitals.
Two-thirds (67.1%, 95% CI 68.34–76.48%) of obstetrician–gynecologists reported that they reoffer HIV testing to pregnant patients who decline testing early in pregnancy, with 46.8% reoffering it to all patients and 20.3% reoffering only to high-risk patients; 28.7% of respondents do not reoffer testing. Other steps that obstetrician–gynecologists take if a patient declines HIV testing include documenting instances where a pregnant patient declines HIV testing (75.6%, 95% CI 71.75–79.53%), asking why a patient declined testing (57.3%, 95% CI 52.78–61.74%), having the patient sign a refusal of testing form (37.0%, 95% CI 62.60–41.34), immediately repeating the recommendation for HIV testing (31.6%, 95% CI 27.41–35.83%), and doing nothing (3.6%). Respondents who indicated that their state requires recommending HIV testing were more likely (than those who said their state does not require recommending testing) to follow up a refusal by asking the patient why (61.2% compared with 49.7%, χ2=4.05; P<.05), having the patient sign a refusal of testing form (46.8% compared with 32.7%, χ2=6.33; P<.05), and reoffering testing later in pregnancy (74.6% compared with 61.8%. χ2=5.54; P<.05).
Just under one third (30.1%, 95% CI 25.91–34.27%) of respondents said their state has regulations that require recommending HIV testing during pregnancy, and 37.0% (95% CI 32.61–41.41%) said their state does not have such regulations. Almost one in three (32.9%, 95% CI 28.62–37.18%) obstetrician–gynecologists said they do not know if their state requires HIV testing during pregnancy. Of those obstetrician–gynecologists who said their state does have regulations, 7.2% did not know if the state required informed consent before such testing. Two approaches to testing pregnant women for HIV status are used in the United States: with the opt-in approach, the woman must specifically consent to an HIV test; with the opt-out approach, the pregnant woman is notified that an HIV test will be performed as part of the standard battery of prenatal tests, and she has the option to decline. Almost equal numbers of obstetrician–gynecologists reported testing their pregnant patients for HIV using the opt-in approach (52.0%) as reported using the opt-out approach (48.0%) (Table 4).14,15
During the survey period, June–July 2006, six states (Arkansas, Florida, Indiana, New Mexico, Oregon, Texas) had regulations requiring use of the opt-out approach when testing pregnant women for HIV, 12 states (Alaska, California, Connecticut, Delaware, Georgia, Illinois, Maryland, Michigan, New Jersey, Rhode Island, Tennessee, Virginia) required use of the opt-in approach, and the remainder had no pregnancy-specific regulations.14,15 For a Web site of specific state requirements for HIV testing, see the Box, “State by State HIV Testing Regulations for Pregnant Women.” In the absence of specific requirements, general HIV testing requirements usually apply to pregnant women. Testing strategies reported by obstetrician–gynecologists were not necessarily consistent with state regulations: 57% of obstetrician–gynecologists reported using the approach required in their state and 43% reported using an approach not consistent with their state's regulations. In the states with no specific requirement, the physician-reported testing approach was divided nearly equally between the opt-in and opt-out approaches.
One fourth of obstetrician–gynecologists who manage pregnant patients reported that they had never managed an HIV-positive pregnant patient (25.2%, 95% CI 21.28–29.14%). These physicians were more likely to report that 0% of their pregnant patients (20.3% compared with 6.4%, χ2=19.26, P<.001) and 5% or less of their nonpregnant patients (80.5% compared with 61.8%, χ2=13.68, P<.001) are at high risk for HIV infection. If a pregnant patient declines HIV testing, obstetrician–gynecologists who reported never having managed an HIV-positive patient were less likely than other respondents to ask the patient why she declined (48.3% compared with 60.3%, χ2=5.18, P<.05), reoffer HIV testing later in pregnancy (50.9% compared with 72.3%, χ2=17.15, P<.001), or document refusal in the medical record (64.4% compared with 79.4%, χ2=10.81, P<.001). These respondents were more likely to practice in suburban locations (55.9% compared with 36.7%, χ2=13.39, P<.001) and to have a patient population that was majority non-Hispanic white (84.6% compared with 65.2%, χ2=15.70, P<.001). They also did not rate themselves as highly knowledgeable regarding HIV infection during pregnancy as did other respondents (Mann-Whitney U=12,824.5; P < .05).
The results regarding HIV screening practices for nonpregnant patients include data from both those obstetrician–gynecologists who do and those who do not see pregnant patients (n=582). The majority of those responding (65.7%, 95% CI 61.78–69.64%) considered between 1% and 5% of their nonpregnant patients to be at high risk for HIV. Almost three fourths of the responding obstetrician–gynecologists (73.7%) reported that they provide pretest counseling before HIV testing, and 84.6% reported that they provide posttest counseling. During health screening, 72.0% of obstetrician–gynecologists indicated that they ask their patients if they have ever used injection drugs, and an additional 20.3% said the inquiry was patient-specific. More than half (58.5%) ask about the number of sexual partners the patient has had, with an additional 25.2% saying it was patient-specific. Fewer obstetrician–gynecologists inquire about the same behaviors in their patients' partners: just over one fourth (27.8%) ask whether their partners have ever used injection drugs, with an additional 32.7% saying it is patient-specific; only 16.3% inquire about the number of sexual partners that the patient's partners have had, with 29.5% saying it was patient-specific. Respondents indicated that they would strongly recommend HIV testing for nonpregnant patients for several reasons (Table 2). Only 4.2% of respondents said they consider a patient's race when deciding to recommend HIV testing.
The obstetrician–gynecologists reported that they used a variety of sources to stay informed about HIV screening practices and guidelines. The majority of obstetrician–gynecologists considered ACOG publications (93.9%), journals (67.3%), and non-ACOG practice guidelines (56.1%) to be important sources of information. Approximately half of the obstetrician–gynecologists considered continuing medical education programs (50.9%), federal guidelines (50.5%), and state health guidelines (50.1%) to be important sources of information. Obstetrician–gynecologists considered local jurisdiction health guidelines (47.0%), textbooks (54.4%), and literature searches (47.1%) to be minor sources of information.
Female respondents were more likely than male respondents to strongly recommend HIV testing for nonpregnant patients in four different cases: reports unprotected sex with multiple partners (strongly recommend: female 90.4%, male 81.1%, χ2=9.97, P<.01), history of blood transfusion between 1978 and 1985 (strongly recommend: female 84.3%, male 70.1%, χ2=16.44, P<.001), any history of blood transfusion (strongly recommend: female 53.9%, male 41.0%, χ2=11.30, P<.005), sexually active with no previous HIV testing (strongly recommend: female 41.1%, male 27.8%, χ2=12.93, P<.005).
Collaborative Ambulatory Research Network members were more likely to respond to the survey than non-CARN members (response rates were 59.4% and 50.4%, respectively). Collaborative Ambulatory Research Network members also were more likely to use the opt-out approach to HIV testing than non-CARN members (52.0% compared with 39.5%, respectively; χ2=6.28, P<.05).
As expected, most (97%) of the obstetrician–gynecologists in this study reported that they recommend HIV testing to all of their pregnant patients, a practice that is consistent with current national and federal recommendations.5–8 This finding also is consistent with the results of a study conducted by Troccoli et al16 that investigated HIV testing practices among North Carolina prenatal providers. Most (83%) obstetrician–gynecologists also reported that they were at least moderately knowledgeable about HIV infection. Major sources of information for these physicians included ACOG publications, journal articles, and non-ACOG practice guidelines.
Just over half (52%) of the responding obstetrician–gynecologists reported that they test pregnant women for HIV by using an opt-in approach to testing. Study results suggested that the obstetrician–gynecologists' approaches to HIV testing were not always consistent with the approach mandated by the regulations in the physicians' states. Although several states mandate informed consent before prenatal HIV testing (opt-in testing approach) and a few require providers to use the opt-out approach, many states have no specific requirements for prenatal HIV testing. In cases where the state does not have specific requirements, general HIV testing laws and regulations usually apply. In this study, in states that require the opt-in testing strategy and states that require the opt-out testing strategy, just over half (57%) of the respondents were using the approach mandated by their state. This finding suggests that, of those responding obstetrician–gynecologists who treat pregnant patients, a substantial proportion was not in compliance with their state regulations regarding approach to HIV testing in pregnant women. In states that did not have specific requirements for prenatal HIV testing, physicians were almost equally likely to use the opt-in approach (51%) as the opt-out approach (49%). This is in spite of the fact that in states without specific requirements, general HIV testing laws, which usually require informed consent, would apply.
Although ACOG, AAP, CDC, and IOM5–8 recommend that obstetric providers use an opt-out approach to prenatal HIV testing (a recommendation that is in conflict with many state regulations), the study results suggested that just under half (48%) of the obstetrician–gynecologists treating pregnant women test for HIV using the opt-out approach. Although few states mandate the use of the opt-out strategy, even in those that do, just over half (57%, n=38) of respondents report using that approach. Although the survey did not measure respondents' awareness of the national and federal guidelines for opt-out testing, these responses suggest that some obstetrician–gynecologists may not know that opt-out is the recommended testing approach. Thus, an increased awareness of state laws, federal guidelines, and professional organizations' recommendations may encourage obstetric providers to use the preferred opt-out approach to prenatal HIV testing where legally allowed.
Many obstetrician–gynecologists did not know whether their state had regulations that require recommending HIV testing during pregnancy (33%). The obstetrician–gynecologists' confusion could be due to the fact that, as previously mentioned, many states do not have HIV testing regulations that are specific to pregnant women. Additionally, in states that have recently revised their regulations to require the opt-out approach, a sufficient time interval may not have passed for obstetric providers to become familiar with and incorporate the new requirements into practice. It is important that obstetric providers have timely access to clear information regarding their state regulations for HIV testing during pregnancy.
Although the rate of recommending HIV testing is high among obstetrician–gynecologists, there are instances where some pregnant women decline to be tested for HIV. Obstetrician-gynecologists reported that the most common reason for a patient to refuse testing is the patient's perception of herself as low risk for infection. However, research has shown that using risk behaviors to determine need for testing actually may fail to identify many of those who are HIV infected.17,18 National and federal guidelines state that obstetric providers should recommend all pregnant women to be tested for HIV, regardless of patients' self-assessed risk.5–8
The results of this study also indicate that it is important to strengthen physician practices in instances where patients refuse HIV testing. Many obstetrician–gynecologists reported that they documented when a patient refused HIV testing, but just under half of the respondents reported that they reoffer HIV testing to all patients who initially declined. To effectively reduce the rate of mother-to-child HIV transmission, obstetric providers should discuss concerns with patients who initially decline testing and continue to encourage them to be tested.
Study results suggest that there may be a need for focused professional education on the use of rapid HIV testing at labor and delivery for women with undocumented HIV status. This is an opportunity to test women who may not have had prenatal care or entered into care late, and to further encourage women who previously declined testing. Although more than half of responding obstetrician–gynecologists reported recommending rapid HIV testing at labor and delivery, 43% either would not recommend testing at that time or said it was not available in their hospital.
Our study has limitations that should be acknowledged. As with all survey studies, our study may be subject to nonresponse bias, although we believe that our findings are reliable. There were 582 respondents and the response rate was 56%, which is at the high end of the 35–60% rate typical of our survey studies. In addition, responses were derived from diverse geographical locations and practice types. It is possible that those physicians more interested in HIV or knowledgeable about regulations regarding HIV screening were more likely to respond. However, questions addressing regulations regarding HIV screening were placed toward the end of the questionnaire. Further, a subset of our subject pool included CARN members who respond to several questionnaires a year covering a wide variety of topics; it is unlikely that HIV screening is a topic of greater interest to this group than to the group of randomly selected ACOG members, yet their response rate was 9 percentage points higher than non-CARN subjects and they differed significantly from them on only one nondemographic question. All of this suggests that response bias was minimized.
The results of this study suggest that obstetrician–gynecologists may benefit from additional information that could improve their knowledge and practice regarding HIV screening. Although most respondents report being at least moderately knowledgeable about HIV during pregnancy and report recommending HIV testing to all pregnant patients, they may need to be more aggressive in following up with patients who decline HIV testing. Additionally, there may be a need for focused education on the use of rapid HIV testing at labor and delivery. Strengthening obstetric providers' HIV screening practice patterns can lead to earlier diagnosis, enabling them to provide their patients who are infected with HIV better access to interventions that could improve their own health, the health of their newborns, and the health of future progeny.19
The results of this study suggest that many obstetrician–gynecologists do not know their own states' laws and regulations pertaining to HIV testing for pregnant women and point to the need for future research to examine obstetrician–gynecologists' knowledge about both national (ACOG, AAP) and federal (IOM, CDC) testing recommendations and the often contradictory state requirements under which they must practice. Further studies should focus on the reasons that many obstetrician–gynecologists are not compliant with their states' requirements and why some physicians may know the requirements but are not following them.