ethnicity, bioethics, and prenatal diagnosis: the

..Ethnicity, Bioethics, and PrenatalDiagnosis: The Amniocentesis Decisionsof Mexican-Origin Women and TheirPartnersC. H. Browner, PhD, MPH, H. Mabel Preloran, PhD, and Simon J. Cox, BA

Prenatal diagnosis is one of the oldest ofthe "new" reproductive technologies. From thefirst amniocentesis tests performed in theUnited States in the early 1970s, prenatal diag-nosis provoked intense controversy andexposed a need for ethical standards that mightgovern its use. In the early 1980s, former sur-geon general C. Everett Koop denouncedsome forms of prenatal testing as "search anddestroy" missions.'(P") The pioneers ofprena-tal diagnosis took shelter from such criticismbehind an ethical principle that has prevailedever since: parents should be free to opt forprenatal diagnosis and to use the informationit provides as they see fit.

This ethic of parental autonomy andself-determination was institutionalized inthe form of "nondirective" genetic counsel-ing.2 Nondirective genetic counseling ismeant to avoid the paternalism and prescrip-tivism that typifies the doctor-patient rela-tionship. In counseling a pregnant woman orcouple about prenatal diagnosis, the intendedrole of the medical professional is to provideinformation to "help individuals-couplesunderstand their options and the present stateof medical knowledge so they can makeinformed decisions."3(p549)

At the time these guidelines and ethicalstandards were being established, amniocen-tesis was the principal diagnostic techniquein use. In the process of detecting untreat-able birth defects that can be averted only byabortion, amniocentesis exposes a fetus tothe risk of miscarriage. Because of its risksand costs, amniocentesis was first offeredonly to women at high risk for fetal defects,the preponderance of whom were White,middle-class, highly educated women whohad postponed pregnancy to pursue careers.The risk of bearing a fetus with some formof chromosomal abnormality rises from 1 in450 for a 30-year-old woman to 1 in 65 for awoman aged 40 years.4(pp3-5) To a largeextent, the first clients for prenatal diagnosiswere a self-selected group who had activelysought out the new prenatal diagnostictechnologies and who were prepared to ter-minate the pregnancy if an anomoly wasdiagnosed.

The advent of maternal serum ax-feto-protein (AFP) screening in the mid-1980s haschanged this situation. Unlike amniocentesis,

AFP screening is noninvasive and inexpen-sive, but it is sensitive to some of the mostfrequent birth defects, including Down syn-drome and spina bifida. AFP screening there-fore soon came to be recommended as thestandard of care for all pregnant women.5 Yet,although a positiveAFP result can indicate anincreased likelihood of an anomoly, amnio-centesis is required to confirm the problemand, if possible, identify it. Consequently, awoman who tests AFP positive is typicallyreferred to a genetic counselor and offeredamniocentesis.

Widespread AFP testing has thus intro-duced amniocentesis to new constituenciesof women, women who had not previouslyconsidered genetic services. Among thesenew constituencies are women from lower-and working-class ethnic minority back-grounds, who tend to complete their child-bearing at younger ages than White upper-middle-class women.4'6 There have been fewsystematic studies of how ethnicity mayinfluence the experience of prenatal diagno-sis and genetic counseling.7-9 But there issome evidence that ethnicity matters in pre-natal diagnosis decisions. Between 1995and 1997, the California maternal serumAFP screening program reported that 66.1%of Hispanic women and 70.5% of Blackwomen accepted amniocentesis after testingAFP positive, compared with 75.8% ofWhite women.10 These data are consistentwith the anecdotal reports of other healthcare providers, which suggest that Hispanicwomen are less likely than women of otherethnic groups to accept amniocentesis.

Most inquiries into decision makingabout amniocentesis are bioethical, not empir-

C. H. Browner and H. Mabel Preloran are with theCenter for the Study of Health and Culture, Depart-ment of Psychiatry and Biobehavioral Sciences,University of California, Los Angeles (UCLA).They are also members of the Socio-BehavioralGroup of the Mental Retardation Research Center atUCLA. Simon J. Cox is with the Russell SageFoundation, New York, NY.

Requests for reprints should be sent to C. H.Browner, PhD, MPH, Center for the Study of Healthand Culture, Department of Psychiatry and Bio-behavioral Sciences, UCLA, Los Angeles, CA90024-1759 (e-mail: [email protected]).

November 1999, Vol. 89, No. 11

Public Health Matters

ical. Bioethics proceeds by deduction fromabstract principles. Insofar as it has empiricalreferences, these tend to be singular casesdrawn from the courts or the media, which are,by definition, unrepresentative of the generalpopulation of cases. The ethical calculus thatemerges from these deliberations bears littlerelation to how real women, faced with actualamniocentesis decisions, make their choices.Bioethics is, therefore, poorly equipped to dealwith the question ofhow women from ethnicminority backgrounds make their amniocente-sis decisions.'1(pi9)

Few women experience prenatal diagno-sis as an occasion for the exercise of moralautonomy. Prenatal diagnosis puts more infor-mation in the hands of women, but it givesthem no way of acting on that informationsave for what is, for some, the unpalatableoption of abortion. It is also possible to acceptor refuse amniocentesis without experiencingit as a "decision" at all. A woman may simplyaccept every test that is offered, includingamniocentesis, without truly registering whatshe is accepting. Or she may simply miss pre-natal care appointments without truly register-ing what she is forgoing.'2tp5l)

If bioethics tells us little about the waywomen in general make decisions aboutprenatal diagnosis, it tells us even lessabout the new constituencies of womenfaced with these decisions as a result ofwidespread AFP testing. Large numbers ofwomen who screen positive now are fromcultures that may not share mainstream USviews about the role ofmedicine and prena-tal care in childbearing, the meaning of dis-ability, or the appropriate way to assess andrespond to scientific risks and uncertain-ties. In such groups, gender ideologies andthe balance of domestic power may alsobe quite different from the US norm. Butbioethics has yet to fully recognize that thisnew diversity of clients may have profoundimplications for the way prenatal diagnosisis implemented.

What is needed, then, is empiricallygrounded research into the questions thatbioethics neglects: How are women makingamniocentesis decisions? What considera-tions, ethical or otherwise, are important tothem? And, in the wake of the routinizationofAFP testing, who are these women? Howdo their diverse ethnic and cultural back-grounds shape their experience of prenatalcare and genetic services?

This article is intended as a contributionto this body of research. Our aim was toinvestigate how women of Mexican originand their male partners decide whether toundergo amniocentesis (after the woman hasscreened AFP positive) and how they experi-ence prenatal diagnosis and the genetic coun-

seling that accompanies it. We focused on theMexican-origin population because it islarge, young, and rapidly growing, with oneof the country's highest birth rates.13 Womenof Mexican origin (and other Hispanicwomen) are also at higher risk for neural tubedefects than women in any other US group.'4

What methodologies are appropriate forresearch of this kind? Given that the clients ofprenatal diagnosis are now highly diverse, anyempirical work needs careful measures ofhowfar its findings are representative of the popu-lation in question. As noted above, bioethicshas argued on the basis of singular, potentiallyunrepresentative cases. But as prenatal testinghas become routine, the subject population hasbecome large enough to allow (and require)useful quantitative work that draws statisticallysignificant inferences from relatively largesamples of cases.' l(ppl-l5)

At the same time, it is necessary tolearn something about the way decisionsare made and the considerations the clientsthemselves consider important. These arevariables that are inherently phenomeno-logical and are discernible only to theresearch subject. We therefore need open-ended questions that allow women andtheir partners to list the considerations thatare important to them rather than have torespond to answers imposed by researchers.To some extent, such self-report data maybe readily amenable to aggregation andquantitative analysis. But the data will alsocontain valuable insights that may not bequantifiable or directly commensurableacross cases. Interpretation of these resultswill necessarily remain case-specific, withthe unit of analysis staying at the level ofthe individual woman or her partner, andany generalizations should be made withgreat caution.

We attempted to meet these diversemethodological requirements by adopting amultimethod approach. We combined the sta-tistical analysis of data from a relatively largesample ofpatient charts with a content analy-sis ofsemistructured interviews and extendedstudies ofparticular cases.

From a public health perspective, thequestions addressed here are significant notleast because of their immediate implicationsfor the counseling of pregnant women. Newclients pose new challenges for those whocounsel pregnant women. The nondirectiveprinciple they have relied on since the 1970stook hold at a time when genetic counseling'sclient population was small, homogeneous,and well informed about genetic services.How well do the principles of nondirectivecounseling translate to the new constituenciesofwomen brought into genetic counseling bya positive AFP result?

The California PrenatalScreening Program

This research was conducted amongwomen and their male partners enrolled inCalifornia's state-administered program forprenatal diagnosis. Since 1986, Californiahas mandated that all pregnant women whobegin prenatal care prior to their 20th weekof pregnancy be offered maternal serumAFP screening. In mid-1995, total humanchorionic gonadotropin and unconjugatedestriol were added toAFP to improve screen-ing for Down syndrome and trisomy 18. Thescreening program is financed like an insur-ance pool, with everyone paying a single fee($115 since 1995) that covers both the costof the AFP screening, offered to all women,and the cost of subsequent services, includ-ing genetic counseling and amniocentesis,offered to the minority of women who testAFP positive. The fee is paid for by mostprivate insurers and by California's Medic-aid program, MediCal; therefore, the costs ofprenatal diagnosis were rarely a considera-tion for the women in our sample.

Approximately three quarters of thewomen who are offered AFP screeningaccept, and 7% to 13% test positive.10 15-17These women are offered a session lastingapproximately 45 minutes with a licensedgenetic counselor who is trained to the levelof a master's degree in genetics and counsel-ing techniques. The counseling session isintended to be nondirective2: counselors areexpected to provide information and answerquestions but not to make recommendations.

A high-resolution, level 2 ultrasoundreveals a benign explanation (such as amisdated pregnancy) for about half of theAFP-positive cases as well as detecting ahandful of severe birth defects, such asanencephaly and severe spina bifida. In theremaining cases, amniocentesis is the onlyway to confirm and identify any prenatallydetectable problems. Complications fromamniocentesis are uncommon but includecramping, bleeding, infection, and, onoccasion, fetal injury or miscarriage. Thelarge majority of women who undergoamniocentesis are found to have no fetaldefects. When fetal defects are detected, thewomen are informed about any availabletreatments. They are also offered an abor-tion, which in California may be performedthrough the 24th week ofpregnancy.

Research Questions

The principal research question motivat-ing this study was, How do women of Mexi-can origin who have screened AFP positive

American Journal of Public Health 1659November 1999, Vol. 89, No. I I


make decisions about amniocentesis? Welooked for attitudes and sociodemographiccharacteristics that distinguished betweenwomen who accepted amniocentesis andthose who refused. Reports from health careproviders suggested that among the most sig-nificant factors are the couple's level of edu-cation, the degree to which each is accultur-ated to the United States, their religiousaffiliation and observance, and their attitudestoward abortion. We hypothesized that themore educated or acculturated a woman, themore likely she would be to accept amnio-centesis. In addition, given that the religiousgroups most Mexican-origin women belongto (most are Catholics, Mormons, or mem-bers of fundamentalist Protestant sects)strictly prohibit abortion, we hypothesizedthat the more religiously observant a woman,the less likely she would be to accept amnio-centesis. The degree to which a woman andher partner opposed abortion, independent oftheir religious convictions, was also expectedto correlate with refusal of amniocentesis.

Beyond these background sociodemo-graphic and attitudinal characteristics, weinvestigated the considerations the womenand their male partners felt were mostimportant in their decision making. Weexplored the attitudes of Mexican-originwomen toward medicine in general andamniocentesis in particular. Such womenmay be less wedded than others to the ideathat medical interventions are essential for ahealthy birth. Specifically, we conjecturedthat women who felt confident in their abil-ity to bear a healthy child would be morelikely to refuse amniocentesis.

In a third area of study, we examined therole the male partner played in a woman'samniocentesis decisions and the influence ofthird parties other than doctors and counselors,such as family members and friends. Healthcare providers regularly explain the high rateof amniocentesis refusals among Hispanicwomen by citing the culture of machismo,whereby Hispanic men assume a high degreeof authority over their wives and may refuseamniocentesis on their wives' behalf.

Finally, we investigated how the womenexperienced genetic counseling. Women inthe interview sample and their partners wereasked to rate their satisfaction with thegenetic counseling they received. We alsogauged whether study participants perceivedtheir counselors as neutral or in favor ofacceptance or refusal.

Methods

Data for this investigation were drawnfrom 2 sources: patient charts and face-to-

face interviews with a sample ofwomen andtheir male partners.

Patient Chart Sample

We reviewed the patient charts of4 genet-ics clinics in southem California between Jan-uary 1 and December 31, 1996. For eachof the 379 Spanish-surnamed women whoscreened AFP positive, we obtained data onsociodemographic characteristics, reproduc-tive and family history, AFP screening results,and the amniocentesis decision. The patientchart sample was relatively large, and becausethe data were collected anonymously, the sam-ple was not biased toward patients who werewilling to give informed consent. Properlycoded, the data on each ofthese variables weresuitable for calculating a range of descriptivestatistics and significance tests.

The patient charts allowed us to relateeach patient's sociodemographic characteris-tics and reproductive history to the result ofher amniocentesis decision. The data tell usnothing, however, about the decision-makingprocess. To explain the decision-makingprocess, self-report data are necessary.

Interview Sample

Our semistructured interviews providedthis self-report data. Between 1995 and 1997,we conducted semistructured, face-to-faceinterviews with 2 waves of eligible women ofMexican origin. Participants were recruitedfrom 6 southern California state-approvedprenatal diagnosis centers. For the pilotphase, we recruited an opportunistic sampleof 25 couples. In the main phase of theresearch, 991 potential participants werescreened, and the 122 women who fit thestudy's criteria (Mexican-origin women withHispanic partners who had been offeredamniocentesis after a positive AFP result)were included in the interview sample. Thuswe used interview data from 147 women and120 men (interviews with the remaining27 men could not be scheduled for variousreasons). The interviews gave us informationon a wide range ofvariables mentioned by thewomen themselves, including phenomenolog-ical variables that had a bearing on the process,not merely the result, of their decision making.

Questions explored in the pilot studywere evaluated and the most reliable, compre-hensible, and informative questions wereincorporated into a semistructured interviewguide. The order ofquestions remained largelythe same in each interview, but interviewersfollowed up on topics that the respondentsthemselves raised. Where necessary, standard-ized probes were used to seek firther informa-tion or clarification.

The interviews were conducted after thewomen had made their amniocentesis deci-sions, but some were still awaiting theiramniocentesis results. Interviewing thewoman and her male partner together provedworthwhile in the pilot survey, andjoint inter-views were conducted with the 49% of cou-ples who requested them. Sixty-nine percentof the participants chose to be interviewed inSpanish and 31% in English. Interviews weregenerally conducted in participants' homes;some follow-up interviews were conductedby telephone.

Many of the interview questions werecoded quantitatively to allow us to make sta-tistically significant inferences about phe-nomenological variables that are difficult tomeasure. The interview results reported hereare based on the combined responses of thepilot and main samples. Many of the ques-tions were also coded qualitatively, with allresponses to a single question or topic ana-lyzed for content and patterns.'8

We have included several summaries ofthe interviews of individual women. Not onlydo these case summaries provide texture,nuance, and detail, but they also change theunit of analysis from the particular interviewquestion to the individual interviewee. Theytherefore address some of the blind spots ofquantitative coding.

Our qualitative content analysis focuseson the considerations that women found piv-otal in choosing whether or not to acceptamniocentesis. This analysis focuses less onthe turmoil and conflict some women felt thanon how their conflict was resolved. Thewomen were asked to account for their deci-sions, not to articulate their dilemmas. Thecase summaries provide examples of the wayinternal conflicts were played out.

As previous research has shown,'2" 9 notall women experience their amniocentesischoice as a dilemma or even as a consciousdecision. Content analysis considers the fac-tors that entered into a decision in isolationfrom the decision maker. Consequently, con-tent analysis does not distinguish effectivelybetween women who wrestled with a numberof competing factors in making their decisionand women for whom the choice was straight-forward. The individual case summaries illus-trate their differences more clearly.

Results

Some sociodemographic characteristicsof the interview sample are shown in Table 1.With regard to reproductive history, thewomen in the sample reported an average of2.01 previous pregnancies (range = 0-8;SD = 1.94). Forty-six (31.3%) of these

1660 American Journal of Public Health November 1999, Vol. 89, No. I I


women reported having had miscarriages;34 (73.9%) reported only 1 miscarriage andthe rest 2 to 4. Twenty-five (17.9%) of140 women reported having had 1 or more

induced abortions; 21 (84.0%) reported1 such procedure and the others 2 or 3. Fif-teen (10.3%) of 145 said they had had chil-dren who died (13 reported 1 such child and2 reported 2), and 10 (6.9%) of 144 said theyhad children who were born with anomalies(n's < 147 owing to missing data). There were

no statistically significant differences inreproductive history between the interviewsample and the more randomly drawn chartsample.

Sixty percent of the women in the inter-view sample accepted amniocentesis and40% declined, a refusal rate significantlyhigher than in most other US populations.'0All but 2 ofthe amniocenteses were negative.Both women who tested positive opted forabortion. In addition, the level 2 ultrasoundsof 2 women who refused amniocentesis indi-cated the likelihood of problems. One mis-carried; the other gave birth to a baby withmultiple anomalies.

The sociodemographic variables andreproductive histories abstracted from thepatient charts were found to have littleexplanatory power. Few sociodemographicfactors differentiated those who acceptedamniocentesis from those who declined. Wefound no significant differences in age or

educational background for either sex, andno significant differences in householdincome. Both groups of women had similarreproductive histories, and there was no dif-ference in family histories ofbirth defects.

We had hypothesized that men andwomen who were less acculturated to theUnited States would be more likely to declineamniocentesis. This hypothesis received par-tial support. Women born in Mexico were

significantly more likely than women born inthe United States to refuse amniocentesis(X2 = 4.67, P = .031), but neither scores on astandardized acculturation instrument20 nor

length oftime in the United States, for immi-grants, was predictive. For men, neither birth-place nor acculturation score was associatedwith amniocentesis decisions.

We had anticipated that more religiouslyobservant Catholics would be more likely todecline amniocentesis because ofthe Church'sstrict opposition to abortion. The sample wasoverwhelmingly (83%) Catholic, but neitherthose who accepted nor those who declinedwere particularly observant. Only 16%reported attending Mass every Sunday, andjust 13% regularly went to confession andtook communion. Neither women's nor men'sreligious background and practice were asso-

ciated with amniocentesis decision.

Although women who declined amnio-centesis were significantly more likely thanthose who accepted to describe themselves as

strictly opposed to abortion (x2 = 13.99, P =

.007), there was significant overlap betweenthe 2 groups of women in their views on

abortion. Just over half (53.4%) ofthose whoaccepted amniocentesis said they would ever

consider having an abortion, as did 26% ofthose who declined. Moreover, fully 42% ofthose who agreed to undergo amniocentesissaid they felt that abortion was personallyunacceptable under all circumstances. (Twothirds of women who refused also held thisview.) There was no association between themale partners' views on abortion and thewomen's decisions about amniocentesis.

During the pilot phase of the research,we asked interview subjects to describe all ofthe considerations that contributed to theirdecision; we probed until we were satis-fied that no additional factors had beenconsidered. This exercise resulted in a listof 30 reported reasons for refusing or accept-ing amniocentesis. We then asked partici-pants in the main study to rate, on a 4-pointscale (0-3), the importance of each of thesefactors in their own decision. We performedt tests and found statistically significant dif-ferences (P<.001) between those whoaccepted and those who declined for twothirds of the reasons (Table 2). Spouses

reported similar but not identical reasons fortheir decisions.

Some of the strongest differencesbetween those who accepted and those whorefused were found in their attitudes towarddoctors, medicine, and science. Those whoaccepted were more likely to indicate thatphysicians' recommendations were an impor-tant factor in their decision. They were alsomore likely to believe that a negative amnio-centesis result would resolve uncertainty andprovide reassurance, while a positive resultwould provide doctors with helpful informa-tion. By contrast, those who refused were

more skeptical about the accuracy and valueof scientific information and reported ahigher degree of discomfort with technology,machines, and needles. They were muchmore likely than those who accepted tobelieve that their fetus was healthy, despitethe positive AFP test result. The 2 groupswere similar in their belief that their decisionwas best for the baby and that God wouldhelp them regardless ofthe outcome.

It is common for any woman and herpartner to be anxious about the prospect ofamniocentesis, but these anxieties seemedespecially strong in our study population.Many found the idea of amniocentesis fright-ening. Even after genetic counseling, severalconcluded that amniocentesis is "like a smalloperation." Sixty-two percent described the

American Joumal of Public Health 1661

TABLE 1-Characteristics of Mexican-Origin Women and Their Male PartnersInterviewed About Their Amniocentesis Decisions: SouthernCalifornia, 1995-1997

Women Men(n =147) (n =120)

No. % No. %

EthnicityaMexican American 45 30.6 34 28.3Mexican immigrant 102 69.4 76 63.3Other Hispanic 0 0 10 8.3

Education<6y 37 25.3 31 26.77-9 y 72 49.3 60 51.7>10 y 37 25.3 25 21.6

Household income, $/y<10000 49 34.5 34 28.810 001-20000 42 29.6 43 36.4>20000 39 27.5 34 28.8Don't know 12 8.5 7 5.9

ReligionCatholic 125 85.0 96 80.0Other 15 10.2 10 8.3None 7 4.8 14 11.7

Note. Numbers add up to fewer than 147 (women) and 120 (men) because of missingdata.

a"Mexican American" was defined as having been born in the United States or havingimmigrated before completing primary school, "Mexican immigranf' as having immigratedafter completing primary school, and "other' as from a Hispanic background other thanMexican.

November 1999, Vol. 89, No. I I


procedure as "rather" or "extremely" risky forthe fetus; 22% described it as "rather" or

"extremely" risky for the mother. Several menand women refused in part because theyfeared that the woman could die or suffersevere complications. These negative views ofamniocentesis were far stronger among thosewho refused than among those who accepted.Those who declined rated the procedure as

significantly more risky than those whoaccepted (t= 9.28, P<.001) and said that fearofmiscarriage was a significant factor in theirdecision to decline (t= 8.77, P<.001).

Male Partners 'Role in AmniocentesisDecision Making

On the basis of reports by health care

providers, we had hypothesized that thesewomen's male partners would play an impor-tant role in their amniocentesis decisions,particularly among less acculturated women.This hypothesis was not supported by our

findings. Of the 120 couples interviewed,51% of both women and men said that the

woman had made the decision; 24% saidit was a joint decision. Only 14% of thewomen indicated that their partner's opinioncounted more than their own. The remaining11% said that someone else (e.g., parents, in-laws, siblings, a clinician) made the decision.As expected, the least acculturated men andwomen said that the man's opinion hadcounted most, but this pattern was not verystrong (NS).

Moreover, even when the decision was

a joint one, the woman's opinion generallycarried more weight. While many men were

actively involved in their wives' pregnancies(e.g., attending prenatal consultations, antici-pating being present at the delivery), bothsexes regarded the man's role as that of a

supportive helper in pregnancy rather than as

an equal partner. They agreed that the womanshould have the final word in the amniocen-tesis decision because it was her body thatwould be tested, or because her previousreproductive experiences and knowledgemade her more qualified to decide. Althoughthese findings contradict common assump-

tions about the culture of machismo, thisview ofmen's supporting rather than control-ling women's reproductive behavior is con-

sistent with the "separate spheres" model ofgender organization that is fundamental toHispanic societies: women are consideredresponsible for all family-related matters,men for public and economic activities.

Genetic Counseling Experience

For the most part, participants reportedvery favorable reactions to the genetic coun-

seling they received; the mean satisfactionrating was 8.5 out of 10 for women and 8.6for male partners who attended the counsel-ing sessions. Not surprisingly, perhaps, thosewho accepted amniocentesis gave the coun-

seling higher ratings than those who declined(t=-2.94, P= .004). Only 12% of the entiresample said they felt pressured by their coun-selors to accept amniocentesis, while 32%said they would have liked their counselors tobe more directive, or at least more explicitabout their own opinion of the value of fetaldiagnostic technologies.

At the same time, probing revealed thatmany participants were uncomfortable withwhat they felt to be the counselor's implicitpressure to agree to amniocentesis. Forinstance, several said they would have pre-ferred that the counselor terminate the ses-

sion as soon as they declined the amniocen-tesis offer. Instead, the counselors continueduntil they had completed their protocols,often asking the women several more timesif they were sure they did not want the pro-cedure. This type of counseling sometimesheightened the women's ambivalence andled them into a protracted period of indeci-sion before they settled on a final course ofaction.

Case Summaries

The case summaries illustrate how theindividual factors analyzed above can playout in different ways during the process ofdecision making. Laura and Dolores bothrefused amniocentesis, but they went throughquite different decision-making processes.The same is true for Maricarmen and Clara,both of whom accepted amniocentesis.Moreover, women who ultimately came todifferent decisions about the procedure, suchas Dolores and Clara, expressed similar fears,concerns, and attitudes.

Maricarmen: accepted, without diffi-culty. Maricarmen was 34 when she was

interviewed and was pregnant with her fourthchild. Her reproductive history had been

uneventful. She was born and raised in

Hidalgo, Mexico, where she received a pri-

1662 American Journal of Public Health

TABLE 2-Importance of Various Factors in Mexican-Origin Women's Decisionto Refuse or Accept Amniocentesis: Southern California, 1995-1997

MeanRefusers Acceptors t P

Reassurance 0.54 2.58 -12.51 0.000Doctor's advice 0.33 1.77 -9.04 0.000Fear of miscarriage 2.52 0.94 9.05 0.000Amniocentesis against God's will 0.83 0.18 4.04 0.000Faith in God's help 2.11 1.52 2.85 0.005To be prepared 0.40 2.50 -13.92 0.000"For the good of the baby" 2.35 2.64 -1.98 0.051Will rely on prayer if there is a problem 0.60 1.22 -3.57 0.001Difficulty in conceiving 0.58 0.20 2.54 0.013Previous experience with amniocentesis 0.07 0.22 -1.52 0.132"I feel the baby'sfine" 1.44 0.55 4.57 0.000Last or only chance to be a mother 0.72 0.40 1.79 0.077Child is wanted regardless of health 2.21 0.88 6.74 0.000"Pro-life" position 1.37 0.45 4.25 0.000Amniocentesis suggested, not recommended 0.46 0.39 0.45 0.657Fear of machines or needles 1.41 0.80 3.64 0.000"Don't want to knoW' 0.96 0.33 3.57 0.001Fear of spouse's insistence on abortion 0.22 0.26 -1.20 0.234

if there is a problemTest cannot cure a problem if detected 1.42 0.60 4.58 0.000Test may be wrong 1.39 0.52 4.89 0.000"We're in good health" 1.36 0.66 3.36 0.001Wish to abort if there is a problem 0.15 1.59 -7.82 0.000Amniocentesis can harm the mother 0.81 0.35 3.73 0.000Amniocentesis can harm the fetus 2.42 0.89 9.30 0.000Will feel guilty if test results in miscarriage 1.91 0.64 6.69 0.000So doctors can help the fetus 0.45 1.31 -5.29 0.000Knowledge of a problem may result in 0.36 0.69 -1.49 0.138

temptation to abortFear of inherited illness 0.16 0.44 -0.92 0.362First child 0.51 0.62 0.58 0.564Amniocentesis recommended by 0.42 0.47 0.43 0.670a respected person

almportance was rated on a 4-point scale: 0= not at all important; 1 = a little important;2 = quite important; 3 = very important.

November 1999, Vol. 89, No. I I


mary school education. In 1978, a sister whowas living and working in California per-suaded Maricarmen to come to the UnitedStates. Her husband, Manuel, joined her2 years later. He was from the same small vil-lage as she, and once in the United Statesfound intermittent work in construction.Maricarmen had worked in a restaurant untilshe began having children; since then she hadworked as an occasional domestic laborer.

Both Maricarmen and Manuel expressedmarked deference toward medical authority.Maricarmen believed that medical profession-als would not suggest a test without reason ormake a mistake. "Not in this," she explained,"because it's with blood and machines....That's what the machines are for." She wouldeven accept the doctors' advice if they sug-gested abortion. When asked ifshe had agreedto amniocentesis so that she would not feelguilty if something was, in fact, wrong withthe baby, she said, "Yes, I'm going to do every-thing the doctor says. I'm going to ask thatthey do all the analyses."

Maricarmen was motivated to haveamniocentesis by her fear of giving birth to achild with Down syndrome. "This is what Iworry about most, that it will be born mon-goloid." Maricarmen had strong attitudesabout specific disabilities. She was not con-cerned about heart disease or deafness, butshe was greatly concerned about blindness,mental problems, physical disfigurement,paralysis, epilepsy, and cleft palate. She saidshe would abort if she knew her child wouldbe sterile: "What's a boy who can't do thisgoing to be? [Such a] person is incomplete."

Maricarmen had already had an inducedabortion for financial reasons, and she saidshe would have another if the amniocentesisdetected a problem. She said she was not reli-gious and did not necessarily hold to theCatholic doctrine that life begins before birth;however, she did believe that praying canhelp change a person's luck. Maricarmen hadfound it difficult to wait 2 weeks for the testresults, and her husband had not been verysupportive during that time because he waspreoccupied with his chronically uncertainemployment situation.

Laura: refused, with conviction. Laurawas 29 years old and had come to the UnitedStates from Mexico with her parents whenshe was 6. They had been peasant farmers inrural Jalisco and continued to do agriculturalwork in the United States. Laura was workingin a grocery store at the time ofthe study; herhusband, who was born in the United States,worked as a parking lot attendant. This wasLaura's third pregnancy; her first 2 childrenhad been born without problems.

Laura was skeptical about her positiveAFP result, because a coworker had previously

suffered a false-positive result: "She was veryworried and crying ... thinking the baby wasgoing to be born sick, and the baby was bornjust fine." Laura reasoned that since she wouldhave the baby in any event, there was no pointin going through the anxiety of amniocentesis.She believed in the value ofknowledge for itsown sake, but in this case, she said, the knowl-edge was not worth the potential cost. "What ifmy baby was fine, but I'm one of those thatloses it [as a result of the amniocentesis]? No,no, no. I would feel guilty for what I did." Sheknew ofno hereditary disorders in her own orher husband's family and was confident thatthe baby would be healthy.

Laura said that the genetic counselingsession met her needs but that it was, tosome extent, redundant. "The truth is that Ididn't even want to go because in any event Iwas going to have the baby. I had alreadyread some booklets, the ones they offer inthe event thatAFP comes out high or low. SoI had in my mind decided not to do it [havethe amniocentesis]." She said she was notopposed to abortion for other women, in thecase of rape, for instance, or even for herselfif it were necessary. But in this case, "I wantto have a baby, and if it's fine or if it's sickI'm going to have it."

Laura described the decision as hers."Nearly always in things like this, I'm the onethat makes the decisions because it's my body."She was separated from her husband and didnot give him any details about the test. Herboss and her sisters gave her advice and sup-port. She also talked with her 12-year-old son:"I told him there was the possibility that thechild would be born sick and he said, 'It's bet-ter to have it.' And I said, 'It wouldn't botheryou if it was born sick; you'd love it?' And hesaid, 'Of course. It's going to be my brother.'And this gave me more security."

Although the couple was separated,Laura's husband, Xavier, supported her deci-sion. He was opposed to abortion; if therehad been a problem with the baby, he wouldhave "accepted what God had sent." Whenasked why they refused the amniocentesis, hesaid, "Because she didn't want it, and Ibelieve that everything will be fine becausewe are healthy."

Clara: accepted, with difficulty. Clarawas born in 1977 and had been in the UnitedStates 10 years. She worked as a hairdresser.Mario, her husband, came to the UnitedStates when he was 20 hoping to find stableemployment; although his father was aschoolteacher, Mario did not finish highschool. Instead he had apprenticed himself toan auto mechanic. Clara had miscarried twicein the past. She particularly wanted this childbecause problems in her marriage meant thismight be the only child she would have.

Clara accepted the AFP test hoping forreassurance: "I wanted to do it. I thought theywould be able to tell me that everything wasfine and then I could relax and finish the preg-nancy." The AFP test results were positive andthe ultrasound was inconclusive, requiringamniocentesis to determine whether there wasa problem.

Work prevented Clara's husband fromattending the genetic counseling session. Thediscussion of possible defects at the sessionleft a powerful impression on Clara. "Theysaid the baby could come with an enormoushead, that it could come with dead legs.... Iwas terrified to think about all these things."Clara was afraid both of the defects theamniocentesis might detect and ofthe amnio-centesis procedure itself. The amniocentesiscould have been performed that day, butClara needed time to think through her deci-sion. She told the hospital she needed to talkit over with her husband. "I didn't want [theamniocentesis] but on the other side I wantedto.... I was afraid ofeverything they told me[about] if I didn't do it ... and I was afraid todo it.... I wanted to go and think before say-ing yes or no."

Clara thought that if she refused amnio-centesis she might offend her physician, whorecommended the procedure in hopes ofunderstanding what had caused her previousmiscarriages. "I didn't want her to think Ididn't have faith in her. I have a lot of faith inmy doctor." She was less satisfied with thegenetic consultation. She detected implicitpressure in favor of the test. "They say onecan say yes or no but I believe that they wantyou to do it.... First I said no, I don't wantto do it, but [the counselor] said, 'There isvery little risk, very little pain, it's like hav-ing an injection.' . . . It seems to me that if Ihad refused the test they would have said,'Oh, this woman is ignorant, she knowsnothing."'

When asked what made her decide inthe end to accept the amniocentesis, shereplied, "To tell you the truth, I think it wasthe fear. I think that the people there put afear into you. I was so afraid, I thought betterto end this.... I wanted to know because notknowing is terrifying." She also wanted tofeel that she was doing everything she couldto help ensure a healthy birth. "If it came[with severe birth defects] and I hadn't donethe tests I would think, well, I didn't doeverything that they told me and for notdoing everything, now I have to pay."

Clara's husband said that he wanted herto have the test and he would want her to havean abortion if the fetus were found to havebirth defects. "In my situation now it'simpossible to raise a sick child, and if I can't,no one else will. Abortion might be painful



but you can't bring a child into the world tosuffer." Clara would not condemn abortionfor others-she saw it as a "very personalthing"-but she did not anticipate abortingthis pregnancy. "I wanted to know but not toend the pregnancy, nothing like that. Thispregnancy cost me a lot. Because before Ilost 2 . . . I said, 'My God, let it be bornhealthy, let all be perfect. If it's sick, I'll alsoaccept it . .. because I want to accept whatyou send me."'

Dolores: refused, on the advice ofathirdparty. Dolores, aged 25, had been in theUnited States just 2 years. Her husband, Mar-tin, had immigrated several years earlier andfinally felt secure enough economically tosend for her and their daughters. Althoughher father was a brick maker and her motherwas a laundress, Dolores had gone to secre-tarial school and was more educated thanmany women in the sample. This was herfourth pregnancy. She had 2 daughters,aged 6 and 3 years, and had recently suffereda miscarriage.

Dolores admitted, somewhat reluc-tantly, that she had refused the amniocente-sis on the advice of a fortuneteller, a friendof the family, who had told her that she hadtested AFP positive because she was beingbewitched by one of her husband's formergirlfriends. The fortuneteller told her not toaccept any invasive medical procedures andto minimize the threat to the pregnancy byresting, remaining in the house, and notaccepting visitors. Dolores was also afraidof the risk of miscarriage associated withamniocentesis, especially because she hadmiscarried before. "They themselves saidyou could lose your baby, and it was veryhard for me to get pregnant. I lost onealready and I don't want to lose this onetoo." Martin would have liked to know forsure whether there were problems with thefetus, but he deferred to the advice of thefortuneteller. "If she's being bewitched, thenit's better not to do the test."

Martin could not afford to miss work toaccompany Dolores to the genetic counsel-ing session. Dolores believed the counselorwanted her to accept amniocentesis, but thecounselor had little credibility for Dolores,because she did not believe that Dolores hadbeen bewitched. Dolores even turned downthe noninvasive ultrasound, because "if I'malready in the testing room, later they'll wantme to do the other one and I won't be able toget out of there." When asked why Doloreswent to the hospital for the counseling ses-sion at all, Martin explained that she went"to comply with what the doctors ordered,because once the baby is born she will gothere. It's better not to miss appointments."Martin was concerned that the doctors would

think they were ignorant for turning downthe test.

Dolores saw the value of foreknowl-edge. "Sometimes you can change luck,other times no, it's fixed. But it's clearlybetter to know, even if you can't changeit." But she also recognized that worryingabout the uncertain prognoses of medicalscience takes a toll. When asked if shesaw any benefit in preparing for possibleproblems, she replied, "No. In fact, I thinkit's the reverse. I see that you can be bitterall the time and what for? . . Because Ihave heard of a thousand cases of [thedoctors] saying one thing and anotherthing comes out."

Martin and Dolores were both againstabortion in most circumstances. "In mycountry they'll put you in jail ifyou do some-thing like that," Dolores said. Both couldenvision circumstances in which abortionmight be unavoidable, for example, if thepregnancy endangered the life of a motherwho already had a child she must care for. Inthe case of a fetus with serious birth defects,Dolores said, "I believe that if the mother isstrong-I, for example, I could care for it, Icould have it. But there are mothers who areweak. In that case [an abortion] would be forthe good of the baby. Better not to bring itinto the world to be tossed about from oneside to another."

Discussion

The case summaries illustrate some ofthe general limitations of our more quanti-tative patient chart data. The patient chartsallowed us to test, inter alia, whether repro-ductive history helps determine a woman'samniocentesis decision. But the case sum-maries show that the facts of a woman'sreproductive history can weigh in on eitherside of the amniocentesis dilemma. Awoman who has miscarried in the past mayaccept amniocentesis to find out whethersimilar problems will occur with her cur-rent pregnancy, or she may be so concernedabout losing this pregnancy that she refusesto run the risks associated with amniocen-tesis. Compare the cases of Clara, whosedoctor felt amniocentesis would providereassurance after her 2 previous miscar-riages, and Dolores, who refused amnio-centesis partly out of fear that it wouldcause a repeat of her earlier miscarriage.These 2 cases illustrate that it is not alwaysthe bald facts that count in an amniocente-sis decision, but the way those facts areapprehended by the woman. Hence self-report data, whatever their limitations, areindispensable.

The Burden of Unwanted DecisionMaking

Prenatal testing is offered to women inobservance of their right to choose. How-ever, women do not universally welcome thechoice prenatal diagnosis offers them.According to some commentators, the offerof prenatal diagnosis confers on women "theburden of unwanted decision making."3(p542)Write Kolker and Burke,

Even those who refuse testing-for example,women who would not consider aborting adefective fetus and those who would rathernot subject their pregnancies to what theydeem hazardous and unnecessary inter-ventions-must confront the decision. Theymust make their reasons for declining explicitand convincing to themselves, their partners,their health care givers, and sometimes tostrangers.4(p5)

This argument holds, a fortiori, for thenew constituencies of women introduced toamniocentesis through routinized AFP test-ing. The original clients ofprenatal diagnosis,who were at risk because of their advancedmaternal age, embraced the tests, seekingreassurance. Many of the women who nowtest AFP positive, however, were fully confi-dent of their ability to bear a healthy childbefore they received the AFP result. For thesewomen, the AFP test and the offer of amnio-centesis are as much a cause of doubt andanxiety as a source ofreassurance.

Unprepared for the offer of amniocente-sis, many women in our sample were unsurehow to respond. Ultimately, accepting theamniocentesis may be the only way to resolvethe worry and doubt that the offer of amnio-centesis has itself precipitated. Clara, forexample, explained, "I realized I had to do[the amniocentesis] because if not, I wouldhave gone crazy with the terrible things I hadin my head.... I wanted to know because notknowing is terrifying." Among the women inour sample who refused, there were those whohad never wanted amniocentesis but hadnonetheless felt pushed into a protracted,wrenching dilemma by genetic counselorswho did not take the first no for an answer.Worried by the AFP result, but having refusedamniocentesis, many actively pursued alterna-tive courses of action, such as praying, takingtraditional remedies, and avoiding stress, toassuage their fears and help them feel asthough they were doing everything they couldto ensure a healthy pregnancy (C. H. Brownerand H. M. Preloran, unpublished observa-tions, 1995-1997).

The "burden ofunwanted decision mak-ing" includes the burden of taking responsi-bility for whatever one decides. An importantpart of accounting for a decision is justifyingit to others and avoiding blame for oneself: as

1664 American Journal of Public Health November 1999, Vol. 89, No. I I


well as being "explicit and convincing," one'sreasons must be beyond reproach. Fear ofbeing blamed for a course of action can inter-fere with a strictly rational judgment. Themore risky but blameless option can becomepreferable to the less risky option to whichblame can be attached. This thought processenters into both refusals and acceptances.Laura rejected amniocentesis because shefeared feeling guilty if it caused a miscar-riage; Clara accepted because she feared shewould be culpable if the baby were born witha defect.

Genetic Counseling ofMexican-OriginWomen and Their Partners

Does genetic counseling help womenmanage the burdens of amniocentesis deci-sion making? How well does the nondirectiveprinciple, originally developed for use amonga class ofwell-prepared, self-selected women,translate to new groups of women who areunprepared to make amniocentesis decisions?

Genetic counseling received high markson average from the women in our sample,but it is clear that it was not wholly successfulin bridging the imposing gulf "between sci-ence and social experience.,,22(p"07) Both thosewho accepted and those who refused showedsigns of misunderstanding prenatal diagnos-tic technologies and the birth defects theydetect. As noted above, those who refusedtended to mischaracterize amniocentesis as"a small operation" and exaggerate its risks.Sixty-two percent described the procedure as"rather" or "extremely" risky for the fetus. Infact, miscarriage rates at the hospitals partici-pating in the study ranged from 1 per 500amniocentesis procedures to 1 per 200 (C. H.Browner and H. M. Preloran, unpublisheddata, 1996).

If the risks of amniocentesis were over-stated, our case summaries show that therisks ofbirth defects were also exaggerated inthe minds ofmany women. There are a largenumber of possible abnormalities that couldtrigger a positive AFP result, but the proba-bility of occurrence is relatively small. In oursample of 147 women, all of whom hadtested AFP positive, only 4 (2.7%) actuallyhad fetuses with birth defects. It may be thatwomen register the sheer number of possibledefects, some of which can be disturbing tocontemplate, without registering how smallthe chances of occurrence are. Asked if sheremembered the defects discussed in thegenetic counseling session, Clara seemedoverwhelmed by the number of problemsmentioned. "They said the baby could comewith an enormous head, that it could comewith dead legs .... I was terrified to thinkabout all these things."

A common misunderstanding evidentamong those who accepted amniocentesiswas the assumption that only one of theoptions available to them-acceptance oftheprocedure-was sanctioned by medical sci-ence. People are accustomed to paternalismin their relationships with medical profes-sionals; they expect "doctor's orders," pre-scriptions, and reassurances. Nondirectivecounseling is alien to most people's experi-ence of medicine. A large proportion (32%)ofthe sample said they would have liked theircounselor to be more directive, as ifthe coun-selor could have told them what would bebest for them.

When, against expectations, the coun-selors refrained from prescribing a course ofaction, many women scrutinized their coun-selors, looking for implicit cues as to whatthey should do. These cues were often forth-coming. As Clara explained, "First I said no, Idon't want to do it, but [the counselor] said,'There is very little risk, very little pain, it'slike having an injection."' And it is not onlythe doctor's (or the counselor's) medicalopinion that counts. Among the more power-ful reasons for considering amniocentesisreported by many in the sample, includingClara and Martin (Dolores' husband), werethe desire to maintain good relations with themedical staff, who would ultimately helpthem through childbirth, and the wish toavoid being thought ignorant.

These implicit cues, whether real orperceived, explain why many cited theirdoctor's opinion as decisive, although only12% reported any pressure from medicalpersonnel. Indeed, among the women whostrictly opposed abortion, a principalexplanatory factor distinguishing thosewho accepted from those who declined wasthe degree to which they deferred to med-ical authority. Contrary to the principles ofnondirective counseling, the "locus of con-trol" for these women lay with medicalprofessionals.

Conclusion

The bioethical debate on prenatal diag-nosis, animated by abstract principles anduntamed by fact, oscillates between extremes-between those who vilify the techniques andthose who support them uncritically. Bothsides of the debate translate their convictionsabout abortion into convictions about prena-tal diagnosis. Most of those who support awoman's right to choose abortion argue fromthe same principle that a woman has the rightto choose prenatal diagnosis. Those whooppose abortion, or who oppose terminationofa wanted pregnancy for genetic reasons, see

prenatal diagnosis simply as the handmaidenof abortion and the harbinger of an imminenteugenics.

This investigation, relying not on bioeth-ical principles but on empirical research,shows that neither of these positions is repre-sentative of Mexican-origin women facedwith actual amniocentesis decisions. Thesewomen's attitudes toward abortion are not thedeciding factors. Their prime concerns areto do what is best to ensure a healthy preg-nancy and to resolve uncertainty, but, as ourcase summaries show, these concerns canlead women to opposite decisions on amnio-centesis. Prenatal diagnosis offers no rightanswers.

What conclusions can be drawn fromthis study for prenatal diagnosis programsand health care providers? Two dangers areevident.

Health care providers have a tendency toassume that Hispanic womens' decisions aregoverned by deep-rooted, cultural givens-such as religion, machismo, or opposition toabortion-that are nonnegotiable and beyondthe influence of doctors, counselors, or med-ical norms. A nondirective counseling ses-sion, in which a woman is supposed to weighher options in the light ofmedical knowledgeand come to an autonomous, rational choice,would seem redundant for women whosedecisions have already been made for themby their culture. As we have shown, thisassumption is a misreading of such womens'motives. Religion, machismo, and attitudestoward abortion are far from being the decid-ing factors in their amniocentesis decisions.In fact, our findings suggest that the mostimportant factors include the women's under-standing of the risks of amniocentesis, theirfear of birth defects, their faith in medicine,and their relationships with their doctors.There is the danger that health care profes-sionals, blinded by a complacent belief inethnic determinism, will neglect these moresubtle influences on Hispanic womens'amniocentesis decisions. As a consequencehealth care professionals will remain insensi-tive to the way Hispanic womens' decisionscan and do fall prey to misunderstandings,exaggerated fears or faith, and a misplacedconcern about maintaining good relationswith medical professionals.

Prenatal diagnosis programs such asCalifornia's universal AFP screening pro-gram operate on the principle that a woman'sright to make informed decisions about herpregnancy entitles her to the information thatfetal testing provides. Some women in oursample wanted this information and wereprepared to act on it. But there were others onwhom this knowledge and the expectation ofself-determination imposed a heavy psycho-



logical cost. There is a danger that as AFPtesting becomes routine, women of bothtypes will accept it automatically and thepotential psychological costs of prenataldiagnosis will not be considered until afterthe event. Hence, most of the women in thisstudy, who were not well informed aboutgenetic services, found themselves plungedunprepared into a fraught decision about thefuture of their pregnancy. At some point, theright not to know-the right to trust in one'sown ability to bear a healthy child-was sur-rendered, but the moment of surrender wasnot made clear at the time. El

AcknowledgmentsThis research was supported in part by UC-MEXUS,the UCLA Center for the Study of Women, and theNational Center for Human Genome Research(lRO1 HG001384-01), and by the Russell SageFoundation, where Dr Browner was a visiting fellowin 1998 and 1999. The support of each of these insti-tutions is gratefully acknowledged.

Maria Christina Casado, Irma Herrera, Car-olina Izquirerdo, Nancy Monterrosa, Jesus Montoya,Jeffrey McNairy, and Ricardo Rivera helped withthe interviews. Casado, Monterrosa, and Rivera alsoassisted with preliminary data analysis. Ann Walker,director of genetic counseling at the University ofCalifornia, Irvine, and Elizabeth Billings ofthe Indi-ana University Prenatal Diagnosis Center generouslymade unpublished data available. Permission wasobtained from the institutional review board of eachparticipating genetics clinic, and interviewees signedconsent forms.

References1. Caplan AL. Foreword. In: Kolker A, Burke BM.

Prenatal Testing: A Sociological Perspective.Westport, Conn: Bergin& Garvey; 1994:xmii-xvii.

2. Harper PS. Practical Genetic Counseling. 4thed. Oxford, England: Butterworth-Heinemann;1993.

3. Fletcher JC, Wertz DC. Feminist criticism ofprenatal diagnosis: a response. Clin ObstetGynecol. 1993;36:541-567.

4. Kolker A, Burke BM. Prenatal Testing: A Soci-ological Perspective. Westport, Conn: Bergin &Garvey; 1994.

5. Professional liability implications ofAFP tests.In: DPLAlert. Washington, DC: Dept ofProfes-sional Liability, American College of Obstetri-cians and Gynecologists; May 1985.

6. Rothenberg KH, Thomson EJ, eds. Women andPrenatal Testing: Facing the Challenges ofGenetic Technology. Columbus: Ohio StateUniversity Press; 1994.

7. Nsiah-Jefferson L. Access to rproductive geneticservices for low-income women and women ofcolor. FetalDiagn Ther. 1993;8(suppl 1):107-127.

8. Press N, Browner CH. Characteristics ofwomen who refuse an offer of prenatal diagno-sis: data from the California maternal serumalpha fetoprotein blood test experience. Am JMed Genet. 1998;78:433-445.

9. Rapp R. Sociocultural differences in the impactof amniocentesis: an anthropological researchreport. Fetal Diagn Ther. 1993;8(suppl 1):90-96.

10. Cunningham GC. An Evaluation ofCaliforniasPublic-Private Maternal Serum Screening Pro-gram. Paper presented at: Ninth InternationalConference on Prenatal Diagnosis and Therapy;June 11, 1998; Los Angeles, Calif.

11. Rothman BK. Beyond "Choice ": Toward Pro-creative Ethics for the Millennium. New York,NY: The Ford Foundation; nd.

12. Rothman BK. The Tentative Pregnancy: Prena-tal Diagnosis and the Future ofMotherhood.New York, NY: Viking; 1987.

13. Vital Statistics of California, 1994. Sacramento:California Department ofHealth Services; 1995.

14. Stierman L. Birth Defects in California: 1983-1990. Emeryville: California Birth DefectsMonitoring Program; 1994.

15. Burton BK, Dillard RG, Clark EN. The psycho-logical impact of false positive elevations ofmaternal serum alpha-fetoprotein. Am J ObstetGynecol. 1985;151:77-82.

16. Evans MI, Belsky RL, Clementino NA, et al.Establishment of a collaborative university-commercial maternal serum alpha-fetoproteinscreening program: a model for tertiary centeroutreach. Am J Obstet Gynecol. 1987;156:1441-1449.

17. Greenberg F. The impact of MSAFP screeningon genetic services, 1984-1986. Am J MedGenet. 1988;31:223-230.

18. Patton MQ. Qualitative Evaluation andResearch Methods. Newbury Park, Calif: SagePublications; 1990.

19. Browner CH, Preloran HM. Latinas, amniocen-tesis, and the discourse of choice. Cult MedPsychiatry. In press.

20. Marin G, Sabogal F, Marin B, Otero-Sabogal R,Perez-Stable EJ. Development of a short accul-turation scale for Hispanics. Hispanic J BehavSci. 1987;9:183-205.

21. Hirsch JS. En el norte la mujer manda: gender,generation, and geography in a Mexican trans-national community. Am Behav Sci. 1999;42:1332-1349.

22. Rapp R. Moral pioneers: women, men andfetuses on a frontier of reproductive technology.Women Health. 1987;13:101-116.

1666 American Journal of Public Health November 1999, Vol. 89, No. 11

ethnicity, bioethics, and prenatal diagnosis: the

Documents