Women's interpretation of an abnormal result on measurement of fetal nuchal translucency and maternal serum screening for prenatal testing of Down syndrome
To assess the effects of sociodemographic and health-provider factors on women's understanding of abnormal results on measurement of nuchal translucency (NT) and maternal serum screening (MSS), 18 months after the implementation of a policy aimed at increasing women's awareness regarding MSS.
A representative sample of women (n = 734) who gave birth in Parisian maternity units in 1999 were asked about their understanding of an abnormal result on MSS and NT. We assessed the effects of sociodemographic and health-provider factors on the probability of women interpreting an abnormal result correctly, misinterpreting it as a definitive diagnosis, or declaring that they did not know how to interpret the result. Response rate was 92% and the analyses included multinomial models.
For both MSS and NT measurement, the majority of women interpreted an abnormal result correctly. However, there were substantial sociodemographic differences in the probability of women interpreting an abnormal result correctly, and more so in the probability of their declaring not to know how to interpret the result. The probability of correct interpretations was substantially higher, and that of declaring not to know how to interpret the result substantially lower, for MSS than NT measurement. However, for several sociodemographic groups, the proportion of women who misinterpreted an abnormal result on screening as indicative of a definitive diagnosis was also higher for MSS as compared with NT measurement.
These findings underscore the need for additional efforts, along with alternative strategies, to inform women about the implications of prenatal screening, particularly in the case of measurement of NT. Copyright © 2006 ISUOG. Published by John Wiley & Sons, Ltd.
Along with maternal serum screening (MSS) during the first and second trimesters of pregnancy1, 2, measurement of nuchal translucency (NT) has come to play an increasingly important role in prenatal testing for chromosomal anomalies3–5. Beginning in 1996, measurement of NT at weeks 11–13 of gestation has become an integral part of prenatal screening for aneuploidy in France6. The main aim of prenatal screening with measurement of NT and MSS is to provide a more informed and rational basis for deciding whether to resort to invasive prenatal diagnosis with amniocentesis or chorionic villus sampling.
Given the goal of informed decision-making for prenatal testing7, an evaluation of prenatal testing programs needs to include women's understanding of the test results. However, few studies have examined women's knowledge about maternal serum screening8–10, and none has examined the factors that might influence women's knowledge about measurement of NT, using representative population-based data.
Determinants of women's knowledge of MSS may differ from those for NT measurement, as the provision of services for these two modalities of prenatal screening differ in important ways. This is particularly the case in France where, in 1997, a specific policy was implemented with the aim of increasing access and informed decision-making with regard to serum screening. Health-care providers are required by a government decree to inform women of the purpose of serum screening and the implications of its results9. Health professionals are further required to provide an official statement to confirm that information was provided, and women must give their informed consent for serum screening to be carried out. On the other hand, measurement of NT is done systematically as part of the general procedure for ultrasound examinations during pregnancy and women may not always be aware of the measurement being taken or its specific contribution to prenatal diagnosis of Down syndrome.
The objective of this study was to assess the effects of sociodemographic and health-provider factors on the probabilities of: (1) correctly interpreting an abnormal result for MSS and measurement of NT; (2) declaring that one does not know how to interpret an abnormal result; and (3) misinterpreting an abnormal result on prenatal screening as being indicative of a definitive (abnormal) diagnosis.
Patients and Methods
We conducted a survey of women giving birth in 15 maternity units in Paris in 1999. The survey was conducted 18 months after the implementation of a policy aimed at increasing women's information regarding MSS. The objectives of the survey included an assessment of women's access to and knowledge of the techniques available for prenatal screening and diagnosis of Down syndrome. Details of the survey have been described elsewhere11. Briefly, the survey was conducted in each of the participating maternity units on two different days separated by an interval of 10 days. These units accounted for 75% of the 38 835 births in Paris in 1999. Women who had recently given birth were interviewed by midwives or other medical personnel in the units. Women who had had a termination of pregnancy, a child with a congenital malformation or other neonatal disease were excluded from the survey.
795 women were present in the maternity units during the study period. 734 (92.3%) women responded to the survey. The main reasons for non-response (n = 61, (7.7%)) included difficulties with understanding French when no interpreter was available (n = 40, (5.0%)) and refusals (n = 13, (1.6%)). Based on data from the French National Institute for Statistics and Economic Studies (INSEE) and the French National Perinatal Survey of 199812, our study population in 1999 was representative of women who gave birth in Paris in 1998 in terms of maternal age, place of residence, country of birth, level of education, and distribution of births between public and private units.
For both MSS and measurement of NT the women were asked for the correct interpretation of an abnormal result on prenatal testing. Specifically, interviewers asked women: ‘In your opinion, if the result of maternal serum screening (measurement of NT) is abnormal, is the fetus affected with Down syndrome?’ Four choices were given for the response in the following order: ‘Yes, certainly’; ‘Maybe’; ‘No’; and ‘Don't know’. Women were asked to choose one response only, which was recorded by the interviewer. Responses ‘Maybe’ and ‘No’ were considered correct in that they both recognize the uncertain (probabilistic) nature of screening results.
Of the 734 women included in the survey, the question for NT measurement was asked of 670 (91.3%) women who reported having had an ultrasound examination in the first trimester of their pregnancy; a response was missing for 49 (7.3%) women who were excluded from analyses. For MSS the question was asked of 587 (80.0%) women who reported that serum screening had been offered to them; a response was missing for 17 (2.9%) women who were excluded. The study population thus comprised 621 women who had responded to the question for NT measurement and 570 women who had responded to the question on MSS. We assessed the effects of sociodemographic and health-provider factors on the probability of each of the three response categories (‘Yes, certainly’, ‘Maybe/No’, and ‘Don't know’). Sociodemographic variables in the study included maternal age (categorized as <30, 30–34, 35–37 or ≥38 years), education (categorized as ≤12 or >12 years), geographic origin/country of birth (categorized as French, African or other) and parity (categorized as 0, 1–2 or ≥3). Health-provider factors included type of follow-up clinic for antenatal care (public/private) and specialty of the health provider (midwife, gynecologist, obstetrician, other) during the first trimester of pregnancy.
We used the Chi-square test for analysis of the statistical significance of associations between the distribution of responses and sociodemographic and health-provider factors. We also tested the statistical significance of differences in individual proportions for each response category (e.g. responding correctly with ‘Maybe/No’) for NT measurement vs. serum screening. For multivariate analysis we used multinomial logit models13. The outcome reference category for comparing the probabilities of responses was ‘Don't know’ (acknowledged absence of knowledge to interpret an abnormal result on screening). This was done in order to assess the relative probabilities of both correct and incorrect responses with the probabilities of not knowing the correct response and saying so (i.e. responding with ‘Don't know’).
We report the multinomial results in terms of relative risk (RR) ratios, which may be interpreted as follows. For example, when considering education, an RR ratio of 5.4 for the response ‘Maybe/No’ as compared with ‘Don't know’ for women with >12 years of education implies that the likelihood of the correct response as compared with ‘Don't know’ was more than five times higher for women with >12 years of education as compared with women with ≤12 years of education. All statistical analyses were done using the STATA statistical software14.
Table 1 shows the frequency of responses to the question: ‘In your opinion, if the result of MSS (measurement of NT) is abnormal, is the fetus affected with Down syndrome?’ For both tests, the majority of women responded correctly (i.e. with ‘Maybe/No’); 68.8% (95% CI, 65.0–72.4%) for NT measurement and a higher percentage of 83.7% (95% CI, 80.4–86.6%) for MSS. The percentage of women who did not know how to interpret an abnormal result and said so was about three-fold greater for NT (26.6% (95% CI, 23.1–30.2%)) as compared with serum screening (8.6% (95% CI, 6.4–11.2%)). Comparable proportions of women (5–8%) incorrectly interpreted an abnormal screening result as indicative of a certain diagnosis, i.e. responded with ‘Yes, certainly’.
Table 1. Women's understanding of an abnormal result on measurement of nuchal translucency (NT) and maternal serum screening (MSS) in Parisian maternity units, 1999. For each test the women were asked: ‘In your opinion, if the result of NT or MSS is abnormal, is the fetus affected with Down syndrome?’ Four choices were given for the response in the following order: ‘Yes, certainly’; ‘Maybe’; ‘No’; and ‘Don't know’, and the women were asked to choose one response only
|Yes, certainly||29 (4.7)||3.1–6.6||44 (7.7)||5.7–10.2|
|Maybe/No†||427 (68.8)||65.0–72.4||477 (83.7)||80.4–86.6|
|Don't know||165 (26.6)||23.1–30.2||49 (8.6)||6.4–11.2|
|All||621 (100)|| ||570 (100)|| |
For both MSS and measurement of NT, there were significant sociodemographic differences in the frequency of various response categories (Table 2). A substantially higher proportion of women with higher levels of education and those born in France interpreted an abnormal result correctly and a much lower proportion declared that they did not know how to interpret an abnormal result on screening (P < 0.05).
Table 2. The relationship between socioeconomic and health-provider factors and the understanding of an abnormal result on prenatal screening. For each test the women were asked: ‘In your opinion, if the result of maternal serum screening (MSS) or measurement of nuchal translucency (NT) is abnormal, is the fetus affected with Down syndrome?’ Four choices were given for the response in the following order: ‘Yes, certainly’; ‘Maybe’; ‘No’; and ‘Don't know’, and the women were asked to choose one response only
|Maternal age (years)||0.11|| || || || ||0.04|
| < 30||262||3.1||64.9||32.1|| ||257||7.0||80.9||12.1|| |
| 30–34||212||6.6||71.7||21.7|| ||208||7.2||86.1||6.7|| |
| 35–37||94||5.3||73.4||21.3|| ||86||10.5||88.4||1.2|| |
| ≥ 38||53||3.8||67.9||28.3|| ||19||10.5||73.7||15.8|| |
|Maternal education (years)||< 0.001|| || || || ||< 0.001|
| ≤ 12||197||5.1||53.3||41.6|| ||188||12.8||68.1||19.2|| |
| > 12||421||4.3||76.1||19.7|| ||379||5.3||91.8||2.9|| |
|Geographic origin||< 0.001|| || || || ||< 0.001|
| French||457||5.3||75.7||19.0|| ||414||6.3||88.9||4.8|| |
| African||73||2.7||37.0||60.3|| ||66||10.6||60.6||28.8|| |
| Other||89||3.4||59.6||37.1|| ||88||12.5||76.1||11.4|| |
|Parity||0.34|| || || || ||0.37|
| 0||253||4.7||72.7||22.5|| ||235||8.9||83.8||7.2|| |
| 1–2||296||5.1||65.2||29.7|| ||276||7.2||84.1||8.7|| |
| ≥ 3||72||2.8||69.4||27.8|| ||58||3.4||82.8||13.8|| |
|Follow-up clinic‡||0.68|| || || || ||0.001|
| Public||222||5.0||66.7||28.4|| ||210||8.1||77.6||14.3|| |
| Private||397||4.3||70.0||25.7|| ||357||7.6||87.1||5.3|| |
|Health provider‡||0.19|| || || || ||0.03|
| Midwife||84||7.1||61.9||31.0|| ||91||11.0||72.5||16.5|| |
| Gynecologist||223||4.5||65.0||30.4|| ||194||8.2||87.1||4.6|| |
| Obstetrician||266||3.8||72.2||24.0|| ||236||6.8||84.3||8.9|| |
| Other||46||4.3||80.4||15.2|| ||46||4.3||87.0||8.7|| |
Maternal age had a significant effect on the distribution of responses for MSS (P = 0.04) but not for NT measurement (P = 0.11). Older (≥ 38 years) women had the lowest probability of interpreting an abnormal result on serum screening correctly. Women aged 35–37 years had the lowest probability of declaring that they did not know how to interpret the result, but also a higher probability of interpreting an abnormal result as indicative of a definitive diagnosis as compared with younger women (P < 0.01).
Health-provider factors, namely type of follow-up clinic (public/private) and specialty of the health provider, did not have a statistically significant effect on the distribution of responses for NT measurement (P = 0.68 and 0.19, respectively). For maternal serum screening however, both the type of follow-up clinic and the specialty of the health provider were significantly associated with the responses (P = 0.001 and P = 0.03, respectively). In particular, women in private clinics had a higher probability of correctly interpreting an abnormal result on serum screening.
For most sociodemographic and health-provider groups, the probability of a correct interpretation of an abnormal result on screening was higher, and that of declaring not to know how to interpret the result lower, for maternal serum screening as compared with NT measurement (Table 2). On the other hand, in general, and especially for women with lower levels of education and those born outside France, the probability of misinterpreting an abnormal result as indicative of a definitive diagnosis was greater for serum screening than for NT measurement (P < 0.05).
Table 3 shows the results of adjusted multinomial logit analyses. Results from both the unadjusted (not shown) and adjusted multinomial logit models suggested that for both screening tests, women with higher levels of education and those of French origin were more likely to respond correctly than they were to declare not to know how to interpret an abnormal result. Women aged 35–37 years were substantially more likely to interpret an abnormal result on serum screening correctly as compared with women aged < 30 years. However, women aged 35–37 years were also much more likely to misinterpret an abnormal result as indicative of a definitive diagnosis than responding with ‘Don't know’.
Table 3. Multinomial logit analyses of the relationship between socioeconomic and health-provider factors and the understanding of an abnormal result on prenatal screening—measurement of nuchal translucency (NT) and maternal serum screening (MSS)—adjusted estimates*. For each test the women were asked: ‘In your opinion, if the result of NT or MSS is abnormal, is the fetus affected with Down syndrome?’ Four choices were given for the response in the following order: ‘Yes, certainly’; ‘Maybe’; ‘No’; and ‘Don't know’, and the women were asked to choose one response only
|Maternal age (years)||0.15|| || || || ||0.13|
| < 30||1.0||Ref||1.0||Ref|| ||1.0||Ref||1.0||Ref|| |
| 30–34||3.3||1.2–9.0||1.6||1.0–2.6|| ||2.2||0.8–6.3||2.0||0.9–4.4|| |
| 35–37||3.0||0.8–11.9||2.0||1.0–3.9|| ||21.1||2.3–193.9||11.7||1.5–93.2|| |
| ≥ 38||2.5||0.4–14.0||1.7||0.8–3.7|| ||2.0||0.2–17.0||1.0||0.2–5.0|| |
|Maternal education (years)||< 0.001|| || || || ||< 0.001|
| ≤ 12||1.0||Ref||1.0||Ref|| ||1.0||Ref||1.0||Ref|| |
| > 12||1.3||0.5–3.2||2.2||1.5–3.4|| ||1.9||0.7–5.1||5.4||2.5–11.9|| |
|Geographic origin||< 0.001|| || || || ||0.007|
| French||1.0||Ref||1.0||Ref|| ||1.0||Ref||1.0||Ref|| |
| African||0.2||0.0–1.0||0.2||0.1–4.0|| ||0.4||0.1–1.4||0.2||0.1–0.5|| |
| Other||0.3||0.1–1.1||0.4||0.2–0.7|| ||1.2||0.4–3.6||0.7||0.3–1.6|| |
|Parity||0.16|| || || || ||0.26|
| 0||1.0||Ref||1.0||Ref|| ||1.0||Ref||1.0||Ref|| |
| 1–2||0.5||0.2–1.3||0.6||0.4–0.9|| ||0.6||0.2–1.5||0.9||0.4–2.0|| |
| ≥ 3||0.3||0.1–1.9||0.6||0.3–1.3|| ||0.1||0.0–0.9||0.5||0.1–1.5|| |
|Follow-up clinic§||0.94|| || || || ||0.36|
| Public||1.0||Ref||1.0||Ref|| ||1.0||Ref||1.0||Ref|| |
| Private||1.2||0.4–4.2||1.0||0.6–1.8|| ||2.1||0.6–6.9||1.8||0.8–4.5|| |
|Health provider§||0.10|| || || || ||0.93|
| Midwife||2.6||0.7–10.4||1.0||0.5–2.1|| ||1.6||0.5–5.5||1.0||0.4–2.4|| |
| Gynecologist||0.9||0.3–2.6||0.6||0.4–1.0|| ||1.5||0.4–5.0||1.2||0.4–3.2|| |
| Obstetrician||1.0||Ref||1.0||Ref|| ||1.0||Ref||1.0||Ref|| |
| Other||2.5||0.4–15.4||2.4||0.9–6.5|| ||0.7||0.1–5.1||1.0||0.3–4.2|| |
For women with higher levels of education and those aged 35–37 years, the relative probability of a correct response as compared with acknowledged absence of knowledge (i.e. ‘Don't know’) was substantially greater for serum screening than for NT measurement. After controlling for other sociodemographic and health-provider factors, women with > 12 years of education were more than five times (RR ratio 5.4, 95% CI 2.5–11.9), and those aged 35–37 years almost 12 times (RR ratio 11.7, 95% CI 1.5–93.2) more likely to interpret an abnormal result on serum screening correctly as compared to declaring not to know how to interpret the result; the corresponding estimates for NT measurement were about two-fold greater for both maternal education > 12 years (RR ratio 2.2, 95% CI 1.5–3.4) and for women aged 35–37 years (RR ratio 2.0, 95% CI 1.0–3.9). However, the relative probability of misinterpreting an abnormal result as indicative of a definitive diagnosis as compared with declared lack of knowledge (‘Don't know’) was also greater for MSS than for NT measurement for both women aged 35–37 years and for those with higher levels of education (Table 3).
Our results suggest that substantial disparities exist in women's understanding of prenatal screening results. Women's understanding of an abnormal result on MSS appeared to be more accurate than that for NT measurement. This finding was consistent for most sociodemographic and health-provider groups and might in part be related to the implementation of a specific policy aimed at increasing access and informed decision-making for serum screening in France. We also found, however, that disparities in the understanding of prenatal screening results persist for both serum screening and measurement of NT. Moreover, in a study of a nationally representative sample of women in France15, a substantial proportion of women (1/6, and more so for those with a lower socioeconomic status) did not know whether they had been offered serum screening.
Our results further suggest that the additional information conveyed about MSS is at times misunderstood. For most sociodemographic and health-provider groups, we found an increase in the probability of misinterpreting an abnormal result as indicative of a definitive diagnosis in the case of serum screening as compared with measurement of NT. This was in particular true for women aged 35–37 years. The latter are perhaps the group of women in France who are most concerned with prenatal screening results, as older (≥ 38 years) women in France can have the cost of amniocentesis reimbursed, and a majority choose to have amniocentesis with or without prenatal screening16.
In a comprehensive and systematic review of previous studies of the offer and uptake of prenatal testing in the United Kingdom17, Rowe et al. found some evidence of ethnic inequalities in access to prenatal testing for hemoglobin disorders and Down syndrome. These inequalities were related to the offer and/or uptake of prenatal testing. A previous study in France15 also found that a substantially lower proportion of women from lower socioeconomic groups knew whether or not serum screening had been offered to them as part of prenatal surveillance. However, few studies have specifically examined women's understanding of prenatal testing results8–10. Our study results and the limited information available in the previous literature regarding women's knowledge of prenatal testing for Down syndrome10, 17 suggest that lack of adequate information constitutes an important socioeconomic barrier for informed decision-making in prenatal testing, particularly in the case of measurement of NT.
Several caveats and limitations must be considered in interpreting our data. The fact that women were surveyed after childbirth does introduce the possibility of ‘hindsight’. For example, women who had tested positive on screening and went on to have an unaffected child or a normal result on amniocentesis might have given a different response if they had been questioned immediately after the screening or at some other point in the course of their pregnancy. However, false-positive results on screening represent a small minority of screening results and hence are unlikely to have had a major effect on the main conclusions of our study. Nevertheless, it would have been interesting to know whether and how women may have changed their interpretations of screening results in the course of their pregnancy.
Another potential limitation relates to the difference in the study populations for MSS vs. NT measurement. Specifically, women who were asked about the interpretation of an abnormal result on serum screening included only those to whom such screening had been proposed (or those who were aware of the fact that serum screening had been offered). For the question on NT measurement, all the women who had had an ultrasound scan during the first trimester were included. It is possible, therefore, that differences in selection mechanisms of the study population for MSS and NT measurement might have favored a higher proportion of correct responses for the former modality.
We evaluated the potential impact of such selection bias by conducting a sensitivity analysis in which the same set of analyses as those presented here were conducted for the subgroup of women (n = 505) who had responded to both questions. Results of this analysis (detailed results available from the authors) were comparable with those we present here in terms of differences between MSS and NT measurement. In addition, the proportion of women who misinterpreted an abnormal result as indicative of a definitive diagnosis was higher for MSS. This latter finding suggests that favorable selection is unlikely to explain the differences observed between MSS and NT measurement.
We did not inquire explicitly about the interpretation of a normal (i.e. negative) result. Nevertheless, understanding a normal result is as important as that of an abnormal result if women are to make a fully informed decision about prenatal testing. We intend to include this question in a follow-up of the present study, as well as questions that aim at assessing women's quantitative understanding of prenatal screening results18, including their evaluations of the probabilities of false-negative and false-positive results19. Another caveat regarding our findings is that, while our results are probably representative of pregnant women in Paris based on our study population characteristics, the effects of the French policy in other regions may not be entirely reflected in our data.
Our results have several implications for issues related to the evaluation of prenatal testing programs and, in particular, questions related to disparities in effective access and informed decision-making about prenatal testing. It appears that the specific policy aimed at increasing the information available to women about MSS might have improved women's understanding of screening results, as their interpretation of abnormal MSS results was in general, more accurate than that for NT measurement. Therefore, a similar policy of including an information sheet as part of all prenatal screening tests, including measurement of NT, may be helpful in increasing the opportunity for informed decision-making for prenatal testing.
However, our results also suggest that while the information provided as part of the policy on serum screening benefited women from essentially all sociodemographic groups, disparities in women's understanding of prenatal screening results persist for both serum screening and measurement of NT. This may in part be related to the limited time available for health providers to inform women of the implications of prenatal screening. In addition, the level of detail and complexity of the information provided in the current information sheet for serum screening may be a factor20. These findings underscore the need for organizational changes that would allow additional time and effort for health providers to communicate the information regarding prenatal screening. This may be particularly necessary for measurement of NT, given our finding of women's lower level of understanding of an abnormal result for this parameter. In addition, alternative21 and less detailed formulations22–26 of the currently provided information for serum screening may be more effective in communicating the essential concepts needed for understanding screening results.
Women with higher levels of socioeconomic status appeared to be more aware of the uncertain nature of an abnormal result on screening. This may in part explain the finding in a previous study that suggests that women with higher levels of education are more likely to opt for amniocentesis irrespective of serum screening16. In addition, older (≥ 38 years) women were least likely to understand an abnormal result on prenatal screening, which can limit the utility of this modality for reducing the use of invasive prenatal diagnostic techniques for older women27, 28. More broadly, however, decisions regarding the use of invasive prenatal diagnostic techniques depend on women's preferences vis-à-vis the outcomes of prenatal diagnosis29, 30, in particular women's preferences regarding the trade-off between diagnostic certainty and risk of fetal loss due to invasive diagnostic techniques.
Our findings underscore the need for additional efforts, along with alternative strategies, for informing women with regard to the implications of prenatal testing, particularly in the case of measurement of nuchal translucency thickness. Our results also suggest that women aged 35–37 years and those with a higher level of education may be more aware of, or more concerned about, the uncertain nature of prenatal screening. This may explain in part why they opt more often for invasive prenatal diagnosis16, 31–34 which allows a definitive diagnosis of Down syndrome.
This study was supported by the INSERM, CNRS, and MIRE-DREES program of Biomedical sciences, Health and Society (SHS-2002) for which we express our gratitude. Babak Khoshnood also received funding from a grant by the Fondation pour la Recherche Médicale (FRM).
We thank the staff of the Paris maternity units for their participation in the collection of data used for this analysis. The Paris Registry received financial support from INSERM (Institut National de la Santé et de la Recherche Médicale), DGS (Direction Générale de la Santé) and InVS (Institut de Veille Sanitaire).