Facilitating autonomous, confident and satisfying choices: a mixed-method study of women’s choice-making in prenatal screening for common aneuploidies

We conduct a pilot study to investigate women’s choice-making experiences in population-based prenatal screening within the public maternal service system at the Department of Obstetrics and Gynecology, Helsinki and Uusimaa Hospital District (HUS) Finland, where aforementioned non-invasive screening tests (FTS and STS) are offered free of charge to all women. Finnish community-based maternity care unit (Neuvola), a public sector activity led by general practitioners and providing free maternal care service to all pregnant women, is responsible for informing pregnant women about prenatal screening and testing program and helps to arrange the tests. Every pregnant woman receives a booklet (Additional file 1) of prenatal screening and testing from the nearest maternity care unit. This study was approved by the Ethics Committee for Gynecology and Obstetrics, Pediatrics and Psychiatry, Helsinki University Hospital (permission number: 220/13/03/03/2015).

We applied a mixed-method approach using qualitative data and quantitative data in one study. This could lead to greater validity and rigor and provide a better understanding of the research problem than either research approach alone [27]. Many researchers have agreed that the mixed methods approach can be particularly useful in healthcare research that requires a broader range of perspectives to view the complexity in this field [28‐30]. We followed the sequential exploratory design [31], in which qualitative data are collected and analyzed first, followed by quantitative study [32, 33]. This design is useful for exploring the unknown variables, developing new instruments based on an in-depth qualitative analysis, and confirming or generalizing qualitative findings by quantitative study. This is suitable to our case, where the complex phenomenon of women’s choice-making experiences needs structuralization and operationalization. We explore how women experience the choice-making process in prenatal screening (qualitative research question), and identify what are the experiential factors significantly influencing women’s choice satisfaction (quantitative research question).

The details of women’s choice-making experiences were explored by semi-structured interviews. Storytelling or narrative inquiry as a way of investigating the subjective experience and delving into individual’s “lifeworld” [34‐36] produced remarkable contributions for the qualitative study. The inquired modes of experience were quite broad, including individual’s feeling, seeing, hearing, reflecting, expecting, acting, and all other kinds of emotional, mental and physical activities. Interviews, that were conducted after the women had gone through the screening, started with open-ended questions about how the woman made a choice between different screening tests. The concrete questions were developed from Theory of Planned Behavior (TPB) [37, 38], which helps to understand the choice-making process from women’s perspective with three main constructs: behavioral beliefs, subjective norms and perceived behavioral control. “Behavioral beliefs” refers to personal evaluation of a behavior (e.g. choosing the FTS) and by this we explored women’s reflections on the advantages and disadvantages of the choice consequences (e.g. the consequences of choosing the FTS). “Subjective norms” refers to social influence and by this we investigated how women’s choices were socially affected by other people. “Perceived behavioral control” refers to the resources and opportunities available to person to make a certain behavior (choice) possible and in this study it helped to evaluate the ease and difficulty of making the choice. Midwives from the Screening Unit and Fetomaternal Medical Center (FMC) in Helsinki helped to recruit the interviewees who participated in prenatal screening by presenting the information letter explaining our study to women and inviting them to participate in the study on a voluntary basis. All the informants voluntarily participating in the interviews signed a written informed consent. Qualitative data collection were collected in a period of December 2015–May 2016. Interviews were conducted in Finnish (author HT conducted) or English (AC conducted). Interviews were tape recorded, transcribed in English and imported into Atlas.ti qualitative data analysis software. The data were processed by thematic analysis. Author AC and HT independently coded the data, identifying the items that could delineate women’s choice-making processes and experiences. Then the two authors discussed, reached an agreement on the items and categorized them into the themes. We enhanced the validity of our findings by organizing discussions in the research team, consulting experts from the Screening Unit, and getting feedbacks from professionals. In total, we interviewed 28 women.

Following the guidelines of scale development provided by DeVellis [39], we developed a framework of measurements for women’s choice-making experiences in prenatal screening and designed a cross-sectional self-completion survey on a retrospective basis to assess women’s experiences in choice-making processes. The generation of measurements for choice-making experience was based on the themes of choice-making process and experience uncovered by the qualitative study as well as some current validated measurements. In the questionnaire, we also measure women’s satisfaction on the choices they made as the primary outcome measurement. Items for measurements were made in a statement format with 5-point Likert scale from “strongly disagree” to “strongly agree”. We asked participants to provide demographics information, including maternal age, education level, marital status, occupation, gravidity and previous experiences of fetal aneuploidy. The questionnaire was critically reviewed and approved by the expert in fetomaternal medicine (VS) who was responsible for the organization of prenatal screening at HUS at the time of commencing this study. In order to test and improve the quality of the measurements, we invited three women to fill in the questionnaire and asked them to report their understandings, confusions and difficulties when responding to each item. Questionnaire was made in English and then translated into Finnish and Swedish. Translation service agency, researchers and medical professionals co-worked in translation and back-translation. Our questionnaires were distributed to the women who participated in prenatal screening. Midwives from delivery hospitals, in which certified midwives perform screenings for pregnant women, helped to disseminate the anonymous questionnaires in the Screening Unit where screening tests and ultrasound are performed. The survey lasted 10 months, from June 2016 to March 2017. Factor analysis was employed to further validate the measurements. Factor analysis empirically determined the number of dimensions/constructs underlying a set of items and refined the inclusion of items for each dimension [39]. Each variable of choice-making experience was produced by summating its measurement items and assessed by one-sample t test, testing whether the variable mean was equal to a neutral value. Multivariable linear regression analysis was used to detect the effects of these variables on choice satisfaction. Quantitative data was managed and analyzed by STATA 13. A p-value of < 0.05 was used to establish statistical significance.

We interviewed 28 women who participated in prenatal screening, of which 22 (78.6%) chose FTS and 6 (21,4%) participated in STS. Table 1 displays the six dimensions of women’s choice-making processes and experiences in prenatal screening: activeness (attitude and behavior), informedness (information needs and sources), preferences/considerations, social influence (perceived social support and pressure), difficulty (difficulty in making choice) and emotional status. For each item under these dimensions, we counted the number of women mentioning it to evaluate its importance, and pictured the choice-making differences between the two groups of women participating in different screening tests. Sample quotations are shown in Table 1.

Starting from June 2016, 940 questionnaires were distributed, 185 were returned by March 2017 with an overall response rate of 19.7% and 183 were valid. In our sample, 163 (89.1%) women participated in the FTS, 3 (1.6%) participated in the STS and 17 (10.4%) did not know which screening test they participated in. The demographic characteristics of the sample are shown in Table 2.

The design of the choice-making experience measurements was based on the results of the qualitative study, group discussions and validated measurements including Decisional Conflict Scale [40, 41] (Additional file 2) and six-item short-form of the state scale of the Spielberger State-Trait Anxiety Inventory [42] (Additional file 3). Additional file 4 summarizes the initially developed constructs (7 constructs for choice-making experience and one for choice satisfaction) with the meanings, measurement items and score ranges. The constructs created by factor analysis (Additional file 5) for all the items measuring the choice-making experience were similar with the original conceptual ones, which means that the original categories have a certain level of validity to account for covariation among the items. One adjustment based on factor analysis was that the item “someone else made the choices for me” that was supposed to measure activeness of choice-making grouped with the measurement of social pressure. The second adjustment was that the items that were planned to measure preference consistency and the items that were planned to measure choice support grouped together as one factor, which was named with “choice-making confidence”. The third one was that the items reflecting positive emotional status and those reflecting negative emotional status did not belong to one emotion category, and they were divided into two factors. Thus, based on the factor analysis, a more structured and validated set of instruments measuring choice-making experience (Additional file 6) were generated, which included “active choice”, “informedness”, “confidence”, “social pressure”, “difficulty”, “positive emotion” and “negative emotion”.

Table 3 quantitatively outlines women’s choice-making experiences in prenatal screening. In general, the survey respondents did not have negative experiences when making choices for screening tests: they were well informed (mean: 22.41; score range: 7–35); they felt confident (mean: 16.83; score range: 5–25); they did not perceive social pressure (mean: 2.94; score range: 2–10); the perceived difficulty of making choices was very low (mean: 5.50; score range: 3–15); they had positive emotion status when making choices (mean: 9.93; score range: 3–15) and the anxiety level was low (mean: 17.62; score range: 9–45); but they did not actively make choices in prenatal screening (mean: 8.66; score range: 4–20). The choice satisfaction with the mean of 8.86 (score range: 2–10) was high.

Table 4 presents the result of the multiple linear regression identifying the experiential factors that significantly influence women’s choice satisfaction in prenatal screening. Demographic factors and screening results (normal, abnormal or do not know yet) at the point of filling in the questionnaire were input as control variables in the regression model. Regression revealed that education (β = − 0.136; p = 0.047), gravidity (β = − 0.182; p = 0.008), activeness in choice-making (p = 0.023), confidence (β = 0.388; p < 0.001), perceived social pressure (β = − 0.306; p < 0.001) and perceived difficulty (β = − 0.274; p < 0.001) significantly influenced women’s choice satisfaction in prenatal screening. Women with higher levels of education were more likely to have lower choice satisfaction. Women had lower choice satisfaction as the pregnancy times increased. If women made more active choices, choice satisfaction would increase. While choice-making confidence had strong positive effects on choice satisfaction, social pressure and perceived difficulty were negatively related to choice satisfaction.

By integrating qualitative and quantitative analysis, one of our principal observations is that when there is a population risk of aneuploidy but the risk has not been individually revealed, women’s activeness in choice-making is not very high. A document with clear information on chromosomal screening, the offered test options and the possible results of the tests may be sufficient for women to read and be informed. Some features of the different methods, including the testing coverage, timing and accuracy should be clearly presented. Another important observation is that although they were offered the booklet shown in Additional file 1, the women in our study were not well informed and many of them were not aware of the STS possibility. The possible explanations are that they regarded the firstly presented option as the best option, they relied solely on public health nurses’ advice or they thought that in their situation it was not necessary to read the information extensively. The regression result indicates that information and knowledge is not the decisive factor for choice satisfaction in prenatal screening. Gourounti et al. [43] have made a similar observation that knowledge component was not decisive for the calculation of informed choice. In general, women do not spend a lot of time considering the screening options. Making an easy and quick choice is what they appear to expect. Recommendation with backup options given by health professionals may work well in this context.

Our study indicates that the recommendation and advice from health professionals (such as the public health nurse of maternity care unit) is highly important for women to make decisions. It is in line with the finding by Aune and Möller [25]: although women knew the health professionals were neutral in their statements and provided ‘balanced’ information and non-directive counseling, there was still a request for more advice and suggestions. According to Porter and Macintyre [44], women tend to assume that whatever the service system offers is well thought and likely the best option. However, this creates a danger of viewing everything the system suggests as routine procedures that do not require a decision [26]. Therefore, even though in previous studies it has been suggested that women did not see such system recommendations as being incompatible with making autonomous choices, there is a need for clarification of the nature of the discussion and counseling. Guidelines on how to support women’s choice-making based on their capability, willingness and preferences should be clearly defined [8].

A large body of research indicates that pregnant women want to be part of the decision-making process in prenatal screening and testing and make autonomous choices [25, 45, 46]. In our study, women’s activeness in choice-making highly influences choice satisfaction. Increased active choice-making leads to higher choice satisfaction, whereas the women in our study do not make active choices in general. This may reflect that the current choice environment in Finland for prenatal screening program does not provide women with enough opportunities or motivations to make active choices. According to women’s accounts, they appreciated the recommendation from maternity care units, but they would have wanted to know the different options. Offered with recommendation together with other options, women would be given better chances to think and make choices by themselves.

Our result of women’s preference on the screening tests is consistent with many other studies: women have a strong preference for the combination of first trimester maternal serum screening and ultrasound measurement of the nuchal translucency and fetal crown-rump length, over STS, because of the earlier estimation, higher accuracy and providing more information about the baby [47‐50]. However, some women have only the option of STS due to special situations such as traveling abroad occurring at the window of the FTS. Our regression analysis showed that choice-making confidence highly influences women’s choice satisfaction and preference consistency partially boosts the confidence. In order to facilitate preference-consistent choices in prenatal screening, women should be offered with opportunities to discuss their wishes and personal situations and clarify their preferences [9].

In general, women did not feel much pressure to participate in the screening tests and most of them reported they received some support from others, mostly from public health nurses of maternity care units and husbands/partners. This empirically supports Lawson and Pierson’s [51] conclusion that the women identified their physicians and their husbands as the most important individuals from whom they sought support. We can see from the quantitative study that while social pressure decreases choice satisfaction, confidence in choice-making that embraces social support can increase choice satisfaction. This is in tune with Lawson and Pierson’s [51] argument that women who are making decisions in a socially supportive context have higher decisional competence and lower decisional conflict. Therefore, the service provider should build a socially supportive choice environment and encourage women to seek support and advice not only from healthcare professionals but also from others (e.g. husband/partner and friends).

In our study, education level negatively influences women’s choice satisfaction, which means that women with higher education level have lower choice satisfaction. It is possible that more highly educated women have higher expectations and more requirements on choice environment. As the number of women with higher education will continue to increase, improving choice environment for prenatal screening demands swift actions. In addition, gravidity (i.e. the number of times that a woman has been pregnant, including current pregnancy) also strongly affect women’s choice satisfaction: the more pregnancies, the lower the choice satisfaction. We can conceive that more experienced women have more critical views about the choices and decisions, which should be taken into account by the choice architects.

In the qualitative study, reliability was secured by the fact that two researchers carefully planned the interview protocol together, and they consulted medical and language professionals concerning the translations of the interview questions and terms. The two researchers also analyzed the interview data first separately and then together, and a strong mutual agreement on coding and categorization was reached. In addition, using the computerized data analysis package Atlas.ti enhanced the reliability [52]. For the quantitative study, Cronbach’s alpha was used to assess the measurements reliability, testing how well the measurement items of one variable measure the same construct or idea [52, 53]. Table 3 includes Cronbach’s alpha of each variable and the figures are acceptable, indicating reliable measurements. In this study, deriving the measurements from the interviews established the construct validity that concerns how well the measurements reflect the investigated constructs [54]. The similarity between the findings from the interviews and the survey confirmed a certain level of construct validity [55]. Specifically, the acceptable results of confirmatory factor analysis (CFA) [56] indicate convergent and discriminant validity of this study (Additional file 7).

One limitation in our study was the low response rate of the survey. This was probably due to asking the respondents to return the survey by mail (with prepaid envelopes) instead of collecting the survey answers at the Screening Unit. Furthermore, we did not set any deadlines for returning the questionnaires and this could lead to many respondents forgetting to return the survey. The other considered issue was that some results could be biased by the generally high education level of women in our sample, which applied to both the interviews and the survey. The generalization of the results may suffer somewhat due to this, as education level has been found to be a significant variable predictive for the choice of tests [13]. It can also be suspected that the beliefs and values of the women with different education levels differ to a certain extent [11]. However, this sample bias should be considered in the nationwide context. According to reports of Statistics Finland, the women in the district of Helsinki and Uusimaa, i.e. in the capital area, are the most highly educated women in whole of Finland [57]. This demographic composition naturally influences the likelihood of getting more answers from women with higher education. There is also a tendency of people who are academically inclined to be more willing to participate in surveys and other studies.

Despite the rather low response rate and slight education sample bias, we consider that the study sample can well represent the maternity population in the hospital district of Helsinki and Uusimaa. According to the Finnish perinatal statistics report for 2015, the mean age of our sample (32) is very close to the mean age of the parturients in Helsinki and Uusimaa region (31,4) [58]. And in terms of the family status, the study group (Married: 55.19%; Co-habitation: 38.80%) is close to the parturient population of the whole country in 2015 [58].

For further study, we plan to expand the sample to include other groups of women (e.g. women who take invasive tests, or women have a lower education status) or cover other regions or other countries. We will conduct the research in other medical fields (e.g. cancer screening) to test the generalization of our conclusions. Meanwhile, we could further the study by addressing the possible influence of personal beliefs and social values on choice-making experience and choice satisfaction.