Materials and Methods
In 1997, we conducted a multicenter, cross-sectional survey of 1362 postpartum women from seven hospitals in four states, North Carolina, Florida, New York, and Connecticut. Women from central North Carolina delivered in one of four hospitals, Duke University Medical Center (2085 annual deliveries) and Durham Regional Hospital (2785 annual deliveries) in Durham, North Carolina, the University of North Carolina Hospitals (2075 annual deliveries) in Chapel Hill, North Carolina, and Wake Medical Center (3350 annual deliveries) in Raleigh, North Carolina. One hospital was included from each of the other states, Jackson Memorial Hospital (6130 annual deliveries) in Miami, Florida, University Hospital of Brooklyn (2300 annual deliveries) in Brooklyn, New York, and Yale New Haven Hospital (5200 annual deliveries) in New Haven, Connecticut.
Nurses at all sites identified women not to be approached because of illness or fetal death. In North Carolina, interviewers used the delivery log to ascertain all vaginal deliveries occurring in the prior 24 hours and cesarean births in the prior 48 hours. Women were approached sequentially by birth order and interviewed unless there were more births than average. Among 1759 births during the study period, we interviewed 754 (36.5%). Among 1038 women approached, 134 (12.9%) refused, most because there was a visitor in the room.
In Brooklyn, the study staff approached and interviewed every other woman identified on the nurses' daily log on the postpartum ward. An estimated 61.3% of women who gave birth on study days were interviewed. The refusal proportion was 11.1%. The Miami site followed the same procedure except they approached every third woman, interviewing an estimated 41.8% of women who gave birth on study days. The refusal proportion was 0.8%. In New Haven, we used the delivery log to identify women who delivered in the previous 48 hours. All eligible women were invited to participate. We interviewed about 34.6% of women at that site who had given birth during the study interval. The refusal proportion was 10.3%.
The average number of days spent interviewing at each location was 50.8 (range 29–63 days). On average, we interviewed 42.7% (range 23.2–61.3%) of the expected births on those days. The overall refusal proportion among women approached was 10.5%. At each hospital, interviewers used labor and delivery logs to record all vaginal deliveries within the past 24 hours and cesarean deliveries within the past 48 hours.
Trained interviewers collected informed consent and administered standardized interviews in the participants' hospital rooms. Interviews lasted approximately 15 minutes and were conducted in English, Spanish, or Haitian Creole according to patients' primary language or preference. The protocol, consent forms, and interview instruments were approved by institutional review boards at each site and at the Centers for Disease Control and Prevention (CDC).
Demographic information collected included maternal age, ethnicity, race, marital status, education level, annual household income, and source of payment for prenatal care, and whether women received prenatal care. The interviewer asked whether information on HIV and AIDS was given in the prenatal clinic, whether an HIV test was offered during pregnancy, and whether women were tested for HIV during their pregnancies. The survey also measured women's knowledge, attitudes, and beliefs in three separate domains, including attitudes towards perinatal HIV testing, knowledge about perinatal HIV transmission, and beliefs about trust in the government and the scientific community.
Proportions were calculated for categoric data, including demographics; counseling and testing rates; and individual responses to knowledge, attitude, and belief items. We calculated 95% confidence intervals on proportions using binomial distributions. We stratified all analyses by location. Differences for categoric data were assessed using χ2 analysis.
We constructed a knowledge score (0 to 5) based on accuracy of responses to five questions about perinatal HIV transmission, with each correct response accorded one point. A dichotomous version was created based on the mean and median knowledge scores (four or fewer correct versus more than four correct). A variable for mistrust of health authorities was constructed from two questions, one on the belief that the government was keeping a cure for HIV from the public and one on the belief that HIV had been invented in a laboratory to hurt people. Mistrust included answering affirmative to either or both questions.
Crude analyses used location, age, race or ethnicity, income, education, marital status, and source of payment for prenatal care for knowledge and mistrust. Three separate multivariable models were subsequently constructed. The first model analyzed factors associated (P < .10) with the dichotomous knowledge score in the crude analyses. A second model analyzed factors associated with mistrust. A third multivariable analysis used all of the associated demographic variables and level of mistrust as determinants for reported receipt of a prenatal HIV test. Multivariable modeling yielded odds ratios (ORs) adjusted for all other factors. Analysis and model diagnostics were done using Epi Info 6 (CDC, Atlanta, Georgia) and JMP 3.2 (SAS Institute Inc., Cary, NC).
We interviewed 1362 women, which was a substantial proportion of those who gave birth at each location on the days we interviewed. The women differed across the four study locations in age, ethnicity, race, marital status, level of education, amount of household income, and source of payment for prenatal care (Table 1). Only ten women reported they received no prenatal care. Nearly 90% of women reported receiving information on HIV and AIDS at the prenatal clinic (Table 2). Almost 88% of women across all locations reported being offered HIV tests during pregnancy, three-fourths of whom reported accepting them. Overall, regardless of whether women were offered HIV tests, over two-thirds (68%) reported they were tested for HIV during a prenatal care visit. Most women (84%) thought that all pregnant women should be tested for HIV, of whom 70% indicated there should be a law that all pregnant women be tested. Overall, 60% of women thought it should be a law. In general, the proportion of women who supported prenatal testing of all women for HIV during pregnancy was higher than the proportion of women who supported testing of all newborns for HIV. Ninety-five percent of women from across all locations responded affirmatively when asked whether they would return for their infant's postnatal visit at which they would learn the infant's HIV test result, if there were a law that mandated newborn HIV testing. Counseling and testing rates and attitudes toward testing varied across the study locations (Table 2).
Women's overall knowledge of perinatal HIV transmission was evaluated according to their responses to five questions (Table 3). The median knowledge score was 4 and the mean score was 3.5. The proportions of women responding correctly to the questions were five correct responses (20%), four correct responses (31%), three correct responses (30%), two correct responses (15%), one correct response (3%), and no correct responses (under 1%). Only 5% of women were unaware that a woman with HIV could pass infection to her infant. Over 20% of women responded incorrectly that all infants born to women with HIV are also infected. Over 25% of women did not know that if an infant tests positive for HIV when it is born, the mother is infected as well. Nearly 40% of women did not know or were unsure whether a woman with HIV could transmit it to her infant during breast-feeding. Almost half of the women did not know or were unsure whether there was a medicine that pregnant women could take to keep from passing HIV to their infants.
Crude analysis showed several demographic factors to be associated with a knowledge score of 4 or more, including study location (P < .001), non-Hispanic ethnicity (P < .001), being white (P = .040), having greater than a high school education (P < .001), and having an annual household income more than $20,000 (P = .002). In a multivariable model that controlled for location and included Hispanic ethnicity, race, education, and income, only a higher educational level was associated with a significantly higher perinatal transmission knowledge score (Table 4).
Over 30% of women believed (or were unsure) that there was a cure for AIDS hidden by the government, and over 20% believed (or were unsure) that HIV was invented in a laboratory to hurt people (Table 3). Exactly 20% of women held one or both beliefs that suggested mistrust of the government or scientific community.
In the crude analysis, demographic factors significantly associated with mistrust included study location (P < .001), being 25 years old or younger (P = .018), being of Hispanic ethnicity (P = .006), being nonwhite, being unmarried, having a high school education or less or having an annual household income less than or equal to $20,000, and having a government source of payment for prenatal care (all factors, P < .001). In a multivariable model that controlled for location-related differences, being nonwhite, having a high school education or less, and having a public source of payment for prenatal care were all associated with more mistrust of health authorities (Table 4).
Having a knowledge score of 4 or more was unrelated to receipt of an HIV test during pregnancy. Women who expressed mistrust were more likely than others to report having been tested for HIV during pregnancy (77% versus 65%, respectively). A multivariable model was constructed to evaluate and identify predictors of those who received HIV tests. According to the model, women who were mistrustful were somewhat, but not significantly, more likely to accept prenatal HIV tests (OR 1.3, 95% confidence interval [CI] 0.9, 1.9) after adjustment for location and sociodemographic factors (Table 4).
Our survey showed positive attitudes about HIV testing. Despite the 68% prenatal HIV test rate in our study, 84% of women thought that all pregnant women should be tested for HIV infection, and most thought testing should be legally mandated. The recent report issued by the Institute of Medicine Committee on Perinatal Transmission of HIV recommended adoption of universal HIV testing, with patient notification, as a routine component of prenatal care.13 Our study suggests that that recommendation should be well accepted by women.
Despite a high acceptance of routine testing, our results show that a substantial proportion of new mothers lack rudimentary knowledge about perinatal transmission of HIV. In particular, women were most likely not to know about ways to prevent perinatal transmission. Knowledge about breast milk transmission was reported correctly by only 60% of women, and knowledge about a medication to help prevent perinatal transmission was reported correctly by only half. However, our 1997 study suggested that women's knowledge of medication to reduce perinatal transmission has improved in comparison with two previous studies, one of which was conducted in 1995, which found that only 20–24% of pregnant women were aware of such medication.9,11 Although other studies have linked knowledge,14 specifically knowledge of a medicine to reduce perinatal transmission,9 to HIV test acceptance during pregnancy, our results do not support those findings.
In addition to insufficient knowledge about perinatal HIV transmission, clinicians should also be aware of our troubling finding that many women have considerable distrust of government and the scientific community regarding the HIV-AIDS epidemic. Previous work documented high levels of mistrust among blacks with respect to the AIDS epidemic.15 The Tuskegee syphilis study historically marked their legitimate discontent with the public health system.16 In a multivariate model, being nonwhite, having a lower educational level, having a public source of payment for prenatal care, and being from a location other than North Carolina were all related to reported mistrust. Regional differences in the degree of mistrust also might indicate the influence of local media. Prenatal HIV testing was more common among women who mistrusted than those who trusted, suggesting that the decision to be tested during pregnancy was based on other more influential factors, such as regional testing recommendations, local clinic testing practices, and providers recommending testing.
The strength of this study is its inclusion of women from geographically diverse sites. The study sample can be considered representative of women who gave birth at those locations because we selected women nonsystematically and information was gathered on HIV testing for the entire prenatal period. Inherent in that design were also some limitations. Study locations might not be representative of the nation as a whole. The data relied on accurate recall of events that took place several months earlier. The interviews were conducted 24–48 hours postpartum, when women's thoughts and feelings are typically focused on their new infants and on recovering from the birthing experience. The interviews were limited to about 10 minutes of focused questions and did not include questions to assess knowledge about HIV and more sensitive questions about risk factors for infection or perceived risk of infection. We did not correlate women's self-report of HIV testing with medical record documentation of testing. That will be the topic of a future report for a subset of women interviewed in North Carolina.
This study underscores the continued need for education about prevention of perinatal HIV transmission. Although we disproved our hypothesis of a link between knowledge, trust, and a willingness to be tested, lack of knowledge and presence of mistrust might hamper other important prevention components, such as care-seeking behavior for women who are HIV infected. Lack of knowledge and mistrust also might slow other public health efforts to control the epidemic that rely on public understanding and support. Our data suggest that women who have not completed high school are particularly in need of measures to increase knowledge and build a partnership of trust with public health authorities. Given that this sector of the population is at a greater risk of acquiring HIV, those measures are urgent.
1. Centers for Disease Control and Prevention. Unexplained immunodeficiency and opportunistic infections in infants—New York, New Jersey, California. MMWR Morb Mortal Wkly Rep 1982;31:665–7.
2. Centers for Disease Control and Prevention. AIDS among children—United States, 1996. MMWR Morb Mortal Wkly Rep 1996; 45:1005–10.
3. Mofeson LM. Interaction between timing of perinatal human immunodeficiency virus infection and the design of preventive and therapeutic interventions. Acta Paediatr 1997;421(Suppl):1–9.
4. Connor EM, Sperling RS, Gelber R, Kiseley P, Scott G, O'Sullivan M, et al. Reduction of maternal infant transmission of human immunodeficiency virus type 1 with zidovudine treatment. N Engl J Med 1994;331:1173–80.
5. Centers for Disease Control and Prevention. AIDS Clinical Trials Group Protocol 076 Study Group. U.S. Public Health Service recommendations for human immunodeficiency virus counseling and voluntary testing for pregnant women. MMWR Morb Mortal Wkly Rep 1995;43:1–15.
6. Koniak-Griffin D, Brecht M. AIDS risk behaviors, knowledge, and attitudes among pregnant and young mothers. Health Educ Behav 1997;24:613–24.
7. McNicol LB, Hadersbeck RE, Dickens DR, Brown JE. AIDS and pregnancy: Survey of knowledge, attitudes and beliefs, and self-identification of risk. J Obstet Gynecol Neonatal Nurs 1991;20:65–72.
8. Rapkin AJ, Erickson PI. Differences in knowledge of and risk factors for AIDs between Hispanic and non-Hispanic women attending an urban family planning clinic. AIDS 1990;4:889–99.
9. Curusi D, Learman LA, Posner SF. Human immunodeficiency virus test refusal in pregnancy: A challenge to voluntary testing. Obstet Gynecol 1998;91:540–5.
10. Silverman NS, Rohner DM, Turner BJ. Attitudes towards healthcare, HIV infection, and perinatal transmission in a cohort of inner-city, pregnant women. Am J Perinatol 1997;14:341–6.
11. Kass NE, Faden RR, O'Campo P, Gielen AC. Policy options for prenatal screening programs for HIV: The preferences of inner-city pregnant women. AIDS Public Policy J 1992;7:225–33.
12. Royce RA, Walter EB, Fernandez MI, Wilson TE, Ickovics JR, Simonds RJ. Barriers to universal prenatal HIV testing in four US locations in 1997. Am J Public Health. In press.
13. Institute of Medicine (IOM). Reducing the odds: Preventing perinatal HIV in the United States. Washington, DC: National Academy Press, 1999.
14. Messiah A, Rey D, Obadia Y, Rotily M, Moatti J. HIV testing, knowledge, attitudes, beliefs, and practices among minorities: Pregnant women of North African origin in Southeastern France. J Natl Med Assoc 1998;90:87–92.
15. Quinn SC. AIDS and the African-American woman: The triple burden of race, class, and gender. Health Educ Q 1993,20:305–20.
16. Thomas SB, Quinn SC. The Tuskegee syphilis study, 1932 to 1972: Implications for HIV education and AIDS risk education programs in the black community. Am J Public Health 1991;81:1498–504.