| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH |
From the 1 Departments of Obstetrics, Gynecology, & Reproductive Sciences, 2 Epidemiology & Biostatistics, and 3 Medicine and the 4 Medical Effectiveness Research Center for Diverse Populations, School of Medicine, University of California, San Francisco, California; 5 Express Scripts, Inc., and Department of Medicine, Washington University School of Medicine, St. Louis, Missouri; and 6 Department of Obstetrics and Gynecology, Kaiser Permanente, San Francisco, California.
 |
ABSTRACT |
|---|
TOP
 ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
METHODS: This was a prospective cohort study of 827 English-, Spanish-, or Chinese-speaking pregnant women presenting for care by 20 weeks of gestation at 1 of 23 San Francisco Bay–area obstetrics clinics and practices. Our primary outcome measure for women aged less than 35 years was any prenatal genetic testing use compared with none, and for women aged 35 years or older, prenatal testing strategy (no testing, screening test first, straight to invasive diagnostic testing). Baseline questionnaires were completed before any prenatal test use; test use was assessed after 30 gestational weeks.
RESULTS: Among women aged less than 35 years, no racial or ethnic differences in test use emerged. Multivariable analyses yielded three testing predictors: prenatal care site (P = .024), inclination to terminate pregnancy of a Down-syndrome–affected fetus (odds ratio 2.94, P = .002) and belief that modern medicine interferes too much in pregnancy (odds ratio .85, P = .036). Among women aged 35 years or older, observed racial or ethnic and socioeconomic differences in testing strategy were mediated by faith and fatalism, value of testing information, and perceived miscarriage risk. Multivariable predictors of testing strategy included these 3 mediators (P = .035, P < .001, P = .037, respectively) and health care system distrust (P = .045). A total of 29.5% of screen-positive women declined amniocentesis; 6.6% of women screening negative underwent amniocentesis.
CONCLUSION: Racial or ethnic and socioeconomic differences in prenatal testing strategy are mediated by risk perception and attitudes. Screening is not the best choice for many women. Optimal prenatal testing counseling requires clarification of risks and consideration of key attitudes and preferences regarding the possible sequence of events after testing decisions.
LEVEL OF EVIDENCE: II-2
Substantial attention has been paid to racial or ethnic and socioeconomic differences in prenatal test uptake. Where differences have been observed, African Americans and Latinas have demonstrated less inclination to undergo diagnostic testing than white and East Asian women. Failure to facilitate informed choices has been found to explain the lower use of prenatal screening for Down syndrome by women from minority ethnic groups and lower socioeconomic status in the United Kingdom.8 Elsewhere, acculturation, specifically language skill, has been found to moderate the effect of these factors.9,10 To further investigate the extent to which sociodemographic variation in prenatal test use can be explained by understanding of testing risks and benefits and preferences, values, and attitudes, we conducted a prospective study of a diverse cohort of pregnant women presenting for prenatal care by their 20th gestational week.
|
MATERIALS AND METHODS |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
All women were asked to complete a paper-and-pencil questionnaire; 386 women also participated in a series of computerized preference-elicitation exercises. Study materials were translated by native Spanish and Chinese speakers and then back-translated to ensure accuracy of content. Fluent bilingual interviewers conducted interviews in subjects' homes or at other convenient locations and administered questionnaires orally if preferred by the study participant. Prenatal test use was assessed by telephone interview or chart review after 30 gestational weeks.
Our primary outcome was prenatal testing strategy, representing the overall pattern and sequence of testing decisions, not just whether specific individual tests were used. Because testing guidelines and insurance coverage are different for women aged younger than 35 years compared with 35 years or older, we defined these outcomes differently for each of these 2 groups. For women aged younger than 35 years, who are generally offered only screening as the initial test, we defined a dichotomous testing outcome: undergoing or forgoing prenatal testing (screening or diagnostic) for chromosomal disorders. For women aged 35 years or older, who typically are offered a choice between invasive diagnostic testing and screening, we employed a 3-level testing strategy outcome: no testing at all, an initial screening test (regardless of whether this was followed by a diagnostic test), or going straight to invasive diagnostic testing. Secondary outcomes focused on subsequent invasive diagnostic testing decisions for women who first underwent a screening test.
Our survey instrument included questions about sociodemographics, reproductive history, and familiarity with Down syndrome. Knowledge about testing and Down syndrome was assessed using 3 statements ("A normal amniocentesis result means my baby will not have any birth defects," "Children who have Down syndrome are more likely to have heart problems than other children," and "If the expanded alpha-fetoprotein blood test is abnormal it means that my baby definitely will have a birth defect"). Perceived risk of carrying an affected fetus and experiencing a procedure-related miscarriage (if opting for invasive testing) were assessed by asking participants to put an "X" on a 10-cm visual analog scale ranging from "no risk" to "very high risk." Two scales entitled "Perceived Understanding of Testing" and "Decision Uncertainty" measured testing awareness using 5 adapted items from the "Factors Contributing to Uncertainty" subscale and 2 items from the "Decision Uncertainty" subscales of the Decisional Conflict Scale.13 Both scales were modified to focus on prenatal genetic testing. Attitude scales included a 2-item "Value of Testing Information" scale assessing the desirability of prenatal testing information; a 4-item "Faith or Fatalism" scale measuring cultural, religious, and fatalistic attitudes toward prenatal testing and birth outcomes; and a 2-item "Health Care System Distrust" scale addressing general distrust of the health care system and experience of unfair treatment when seeking medical care. Attitudes toward having a child with Down syndrome and medicalization of pregnancy were measured using single items in which respondents were asked to indicate the extent to which they agreed or disagreed with the statements "Modern medicine interferes too much with my pregnancy" and "I would rather have a child who has Down syndrome than no child." Two items were included to assess attitude toward pregnancy termination in the event of an affected fetus ("Would you have an abortion if your fetus were found to have Down syndrome?") and views on autonomy regarding prenatal testing ("When it comes to special testing in this pregnancy, like amniocentesis, who should decide?"). We used the time tradeoff metric to measure patient preferences (utilities) for outcomes of decisions to accept or forgo prenatal testing,14 and we calculated 3 preference difference scores comparing the utility for a procedure-related miscarriage to that of having a child with Down syndrome; the value of the reassurance brought about by negative results; and the strength of preference for a definitive diagnosis compared with risk-refining information.
We assumed 80% power, 2-tailed 
= .05, and a set of highly intercorrelated covariates and estimated that among the 475 women aged younger than 35 years, the minimal detectable odds ratio equaled 1.66. For the 344 women aged older than 35 years, there were 2 primary comparisons among levels of the test choice outcome, for which we estimated minimal detectable odds ratios of 1.59 and 1.98, respectively.
We used logistic regression models to estimate the effects of study explanatory variables on the 2- and 3-level testing strategy outcomes for the younger and older age groups, respectively. For women aged younger than 35 years, binary logistic regression models were fit. For women aged 35 years or older, multinomial logistic regression modeled the 3-level outcome variable,15,16 which simultaneously modeled 2 comparisons among the outcome categories: invasive testing first compared with no testing and invasive testing compared with screening first. Custom tests also considered the outcome comparing screening with no testing. Technically, the custom tests provided no added information about model effects; they were provided for descriptive purposes. Explanatory variables included sociodemographics, knowledge and risk comprehension, attitudes and influences. Additional univariable logistic models, which included preference difference scores, were fit for the subset of the sample from whom preference data were collected. Within each age stratum, both binary and multivariable regression models were fit. For each multivariable model, the set of explanatory variables was reduced through a backward elimination procedure with the P value for retention set to .20.
To determine whether any effects of race or ethnicity and socioeconomic status in the population 35 years or older were mediated by knowledge or attitude variables, all demographic indicators (race or ethnicity, maternal age, income, education, occupation, interview language, parity) were included as exogenous variables in a mediation analysis. Potential mediators included the knowledge and attitude variables that were retained in the corresponding final multivariable logistic regression model. We concluded by conducting analyses of predictors of subsequent invasive test use among women of all ages who had a screening test first and received positive or negative results and of going straight to invasive testing among women aged younger than 35 years.
Because the data contained missing values, each substantive model was fit to 20 multiply imputed data sets created with SAS PROC MI (SAS OnlineDoc 9.1.3, SAS Institute Inc., Cary, NC). Imputations of outcome and explanatory variables were performed separately by age stratum. Imputed values for binary and categorical variables were rounded and truncated to the nearest category.16,17 All parameter estimates and significance tests were calculated by combining results across the imputed 20 data sets.18,19
|
RESULTS |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
A somewhat different picture emerged from our analyses of women aged 35 years or older. Univariable logistic regression analyses yielded race or ethnicity as a significant predictor of testing strategy (Table 3). Generally, compared with the other groups, African-American women had lower odds of undergoing prenatal testing. All socioeconomic status and acculturation variables also were predictive of testing strategy, as were prenatal care site, 3 of the 5 knowledge, risk comprehension, and awareness variables, 6 of the 7 attitudes, and all 3 preference difference scores. The multivariable model yielded 4 independent predictors of testing strategy: women with higher perceived risk of procedure-related miscarriage were less likely to go straight to invasive testing (P = .037), whereas women with higher value of testing information were more likely to undergo invasive testing (P < .001), and women with greater health care system distrust (P = .045) or higher levels of faith or fatalism (P = .035) were less likely to use any testing.
|
|
Separate linear models regressed each candidate mediator onto the sociodemographic indictors; the resulting parameter estimates, B, follow. Significant effects of being African American were found for value placed on testing (B = –0.66, P = .019), and faith or fatalism (B = .99, P < .001; Figure 1), with African-American women scoring lower on the value of testing and higher on the faith or fatalism scales. Income was significantly related to perceived risk of procedure-related miscarriage (B = –.02, P = .008) and faith or fatalism (B = –.16, P = .001), with higher-income women having lower perceived miscarriage risk and faith or fatalism scores. No sociodemographic indicator was related to health care system distrust. Finally, the prenatal testing strategy outcome was regressed onto the sociodemographic indicators and the candidate mediators. The effects of race or ethnicity and income were nonsignificant (P = .519 and .141, respectively), suggesting complete mediation of these effects.
|
Forty-four women who started with a screening test received positive results (midtrimester risk for trisomy 21 > 1:190). Twenty-three of these women were aged younger than 35 years old, and 21 were aged 35 years or older. A total of 29.5% did not go on to have amniocentesis. Among these women, only high value of testing information emerged as a predictor of amniocentesis use in multivariable analysis (OR = 1.55, P = .033). Additionally, 6.6% of the women who screened negative elected to undergo amniocentesis. Significant predictors included higher maternal age (OR = 1.29, P < .001), greater value of testing information (OR = 1.55, P = .033), and being nulliparous (OR = 3.08, P = .026). Finally, 4.6% of the women who were aged younger than 35 years went straight to invasive testing. Significant multivariable predictors included higher maternal age (OR = 1.09, P = .017), occupational class (OR = .61, P = .039), and value of testing information (OR = 1.45, P = .049).
|
DISCUSSION |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Among women aged 35 years or older, a different set of attitude and belief factors came into play: value placed on testing information, health care system distrust, and accepting God's will or having a more fatalistic attitude toward Down syndrome. Perceived risk of procedure-related miscarriage also was found to predict prenatal testing strategy. These findings suggest that, in this age group, prenatal testing decisions incorporate factors typically included in formal approaches to decision making (the value of information obtained weighed against the "cost" of obtaining that information). However, these considerations are tempered by or overlaid with other beliefs (attitudes about faith, and trust—or lack thereof—in the health care system).
Importantly, associations between prenatal testing strategies among women aged 35 years or older and sociodemographic variables were explained by mediating factors. We found that lower test use by African-American women in univariable analysis was explained by greater faith or fatalism and lower value for testing information and that increased test use among higher income women was mediated by lower faith or fatalism and perceived procedure-related miscarriage risk.
Finally, the fact that 29.5% of women who screened positive in our study declined to have invasive testing suggests that efforts to get all women to undergo screening may result in unwanted or unnecessary testing, with the attendant risk of anxiety engendered by false-positive results. They may also put many women screening positive in an untenable situation—anxious about a positive result but unwilling to incur the risks of diagnostic testing to obtain the definitive results that would be required to prepare for the birth of an affected infant or terminate the pregnancy. The risk cutoffs used in the newer screening approaches maintain a 5% false-positive rate20; it is therefore likely that there will remain a significant proportion of screen-positive women who will be unwilling to pursue invasive testing. Moreover, 6.6% of women who screened negative in our study subsequently underwent amniocentesis, suggesting that despite the reassurance that most women who screen negative experience, some will prefer the certainty of diagnostic confirmation of their fetus' karyotype. For such women, undergoing screening may be of limited value. Whether improved sensitivity of screening tests would diminish the percentage of women screening negative who undergo amniocentesis is unknown. It is likely, however, that some women—regardless of age—will continue to prefer the definitive information that only invasive testing can provide.
Our study has several limitations. First, it was conducted before the dissemination of first-trimester screening. Nonetheless, we believe that our observations on the use of noninvasive screening tests are generalizable to the formulation of practices regarding optimal use of newer first-trimester methods. The question of whether to undergo or forgo prenatal testing altogether, as well as whether to have a screening test first or go straight to invasive testing, remains important. Moreover, there are still many people for whom first-trimester screening is not an option. In the State of California, for example, expanded alpha-fetoprotein remains the only serum screening test covered by Medi-Cal. Future studies will shed light on whether patient attitudes change once first-trimester screening becomes standard of care for the majority of patients.
In addition, although we were successful in recruiting a diverse population of women, our sample was limited to a single geographic area whose demographic and attitudinal characteristics and testing options may not be representative of the rest of the United States. Our study also was limited in scope. We did not collect data on the extent of provider counseling before the interview or on access to and coverage for invasive testing, factors which provide potential explanations for the differences we found by site of care. Finally, our knowledge and attitude scales covered selected areas of hypothesized importance, but were not designed to be comprehensive. Importantly, we were able to identify key drivers of the racial or ethnic and socioeconomic differences that have been noted repeatedly, furthering our understanding of the apparent disparities in prenatal test use.
We believe our findings have implications for improving patient counseling about prenatal testing. The key roles of risk perception and perceived understanding of testing highlight the importance of educating patients about testing. Choice of prenatal testing strategy is also dependent on values that vary among women of all ages and risk levels, underscoring the need to present 3 reasonable options (no testing, screening first, and diagnostic testing first) and the possible outcomes of each, including the need for further testing to get definitive results if positive screening results are obtained. The right strategy for the individual woman should be consistent with her values and desire for information about fetal status, and informed by the implications of both negative and positive screening and diagnosis results.
Our study also contributes to the mounting evidence that risk-based testing practices are at odds with preference-based approaches, and supports the view that the latter approach is preferable. Although reliance on risk-based thresholds has been questioned, and concern about routinizing testing has been raised,26 investigators have suggested implementing a "screening first" policy for women of all ages, effectively eliminating the option to go straight to invasive testing at any age.27,28 Such recommendations are likely to increase in popularity in view of recent findings regarding the increased accuracy of first-trimester screening strategies.4 They also may be easier to implement and be more equitable from a public health perspective by limiting invasive procedures to women who have clearly defined elevated risk. However, routinizing "screening first" could diminish the opportunity for at least some women to make choices that might be more reflective of their preferences and values.
Findings from our study also have broad implications for research addressing health care disparities. Studies exploring sociodemographic patterns in test use have progressed from simple descriptive analyses highlighting the problem to more analytical investigations, which are beginning to elucidate underlying causes. Our study underscores the importance of rigorously exploring factors beyond race or ethnicity and socioeconomic status, to benefit patients, providers, and policy makers in their decision making.
|
Footnotes |
|---|
Corresponding author: Miriam Kuppermann, PhD, MPH, Department of Obstetrics, Gynecology, & Reproductive Sciences, University of California, San Francisco, 3333 California Street, Suite 335, San Francisco, CA, 94143-0856; e-mail: kuppermannm{at}obgyn.ucsf.edu.
doi:10.1097/01.AOG.0000214953.90248.db
|
REFERENCES |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2. Wapner R, Thom E, Simpson JL, Pergament E, Silver R, Filkins K, et al. First-trimester screening for trisomies 21 and 18. N Engl J Med 2003;349:1405–13.
3. Wald NJ, Rodeck C, Hackshaw AK, Walters J, Chitty L, Mackinson AM. First and second trimester antenatal screening for Down's syndrome: the results of the Serum, Urine and Ultrasound Screening Study (SURUSS). J Med Screen 2003;10:56–104.[Medline]
4. Malone FD, Canick JA, Ball RH, Nyberg DA, Comstock CH, Bukowski R, et al. First-trimester or second-trimester screening, or both, for Down's syndrome. N Engl J Med 2005;353:2001–11.
5. Marion JP, Kassam G, Fernhoff PM, Brantley KE, Carroll L, Zacharias J, et al. Acceptance of amniocentesis by low-income patients in an urban hospital. Am J Obstet Gynecol 1980;138:11–5.[Medline]
6. Knott PD, Penketh RJ, Lucas MK. Uptake of amniocentesis in women aged 38 years or more by the time of the expected date of delivery: a two-year retrospective study. Br J Obstet Gynaecol 1986;93:1246–50.[Medline]
7. Kuppermann M, Gates E, Washington AE. Racial-ethnic differences in prenatal diagnostic test use and outcomes: preferences, socioeconomics, or patient knowledge? Obstet Gynecol 1996;87:675–82.[Abstract]
8. Dormandy E, Michie S, Hooper R, Marteau TM. Low uptake of prenatal screening for Down syndrome in minority ethnic groups and socially deprived groups: a reflection of women's attitudes or a failure to facilitate informed choices? Int J Epidemiol 2005;34:346–52.
9. Press N, Browner CH. Characteristics of women who refuse an offer of prenatal diagnosis: data from the California maternal serum alpha fetoprotein blood test experience. Am J Med Genet 1998;78:433–45.[Medline]
10. Browner CH, Mabel Preloran H, Casado MC, Bass HN, Walker AP. Genetic counseling gone awry: miscommunication between prenatal genetic service providers and Mexican-origin clients. Soc Sci Med 2003;56:1933–46.[Medline]
11. Kuppermann M, Nease RF, Learman LA, Gates E, Blumberg B, Washington AE. Procedure-related miscarriages and Down syndrome-affected births: implications for prenatal testing based on women's preferences. Obstet Gynecol 2000;96:511–6.
12. Learman LA, Kuppermann M, Gates E, Nease RF Jr, Gildengorin V, Washington AE. Social and familial context of prenatal genetic testing decisions: Are there racial/ethnic differences? Am J Med Genet C Semin Med Genet 2003;119:19–26.[Medline]
13. O'Connor AM. Validation of a decisional conflict scale. Med Decis Making 1995;15:25–30.[Medline]
14. Kuppermann M, Nease RF Jr, Gates E, Learman LA, Blumberg B, Gildengorin V, et al. How do women of diverse backgrounds value prenatal testing outcomes? Prenat Diagn 2004;24:424–9.[Medline]
15. Agresti A. Categorical data analysis. 2nd ed. New York (NY): John Wiley-Interscience; 2002.
16. Schafer JL. Analysis of incomplete multivariate data. New York (NY): Chapman & Hall; 1997.
17. Allison PD. Missing Data. Thousand Oaks (CA): Sage Publications; 2002.
18. Rubin DB. Multiple imputation for nonresponse in surveys. New York (NY): Wiley; 1987.
19. Meng X-L, Rubin DB. Performing likelihood ratio tests with multiply-imputed data sets. Biometrika. 1992;79:103–11.
20. Malone FD, Canick JA, Ball RH, Nyberg DA, Comstock CH, Bukowski R, et al. First-trimester or second-trimester screening, or both, for Down's Syndrome. N Engl J Med 2005;353:2001–11.
21. Pauker SP, Pauker SG. Prenatal diagnosis—why is 35 a magic number? N Engl J Med 1994;330:1151–2.
22. Hook EB Genetic counseling and prenatal cytogenic services: policy implications and detailed cost-benefit analyses of programs for the prevention of Down syndrome. In: Porter IH, Hook EB, editors. Service and education in medical genetics. New York (NY): Academic Press; 1979. p. 29–54.
23. Kuppermann M, Nease RF, Learman LA, Gates E, Blumberg B, Washington AE. Procedure-related miscarriages and Down syndrome-affected births: implications for prenatal testing based on women's preferences. Obstet Gynecol 2000;96:511–6.
24. Druzin ML, Chervenak F, McCullough LB, Blatman RN, Neidich JA. Should all pregnant patients be offered prenatal diagnosis regardless of age? Obstet Gynecol 1993;81:615–8.
25. Harris RA, Washington AE, Nease RF Jr, Kuppermann M. Cost utility of prenatal diagnosis and the risk-based threshold. Lancet 2004;363:276–82.[Medline]
26. Press N, Browner CH. Why women say yes to prenatal diagnosis. Soc Sci Med 1997;45:979–89.[Medline]
27. Haddow JE, Palomaki GE, Knight GJ, Cunningham GC, Lustig LS, Boyd PA. Reducing the need for amniocentesis in women 35 years of age or older with serum markers for screening. N Engl J Med 1994;330:1114–8.
28. Egan JF, Benn P, Borgida AF, Rodis JF, Campbell WA, Vintzileos AM. Efficacy of screening for fetal Down syndrome in the United States from 1974 to 1997. Obstet Gynecol 2000;96:979–85.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
.05. 
P
P