What factors influence health professionals to use decision aids for Down syndrome prenatal screening?

This qualitative study was embedded in a sequential exploratory mixed-methods study aiming to implement the use of a PtDA to foster shared decision-making in the context of prenatal screening for DS. The overall study is part of the PEGASUS project (PErsonalized Genomics for prenatal Aneuploidy Screening USing maternal blood, Canada). Ethics approval was obtained from the research ethics boards of the Centre de Santé et de Services Sociaux de la Vieille-Capitale (#2013-2014-29) and the CHU de Québec (#B14-02-1929).

Prenatal care in the province of Quebec, Canada, is offered by obstetrician-gynecologists (about 51 % of pregnancies), family physicians (about 46 %) or midwives (about 3 %) [22]. We wanted to maximize the diversity of perspectives by drawing from clinical sites representing different team approaches to prenatal care and different clienteles. According to Godin et al., for this kind of qualitative study, a minimal sample of 25 to 30 participants is recommended to identify all relevant opinions [23]. We thus aimed to recruit a convenience sample of 45 participants with 15 participants in each professional category. We recruited health professionals from three health centers in the Quebec City area, Canada: obstetrician-gynecologists in a university hospital, family physicians in a family practice teaching unit and midwives in a birthing center.

We included health professionals who were: a) involved in prenatal care; b) family physicians, midwives, obstetrician-gynecologists or interns in these professions; and c) working in any of the three health centers targeted. We excluded health professionals who were on parental or sick leave.

The project coordinator and a research assistant first met the family physicians, the midwives and the obstetrician-gynecologists during a regular professional meeting to explain the study process and secure their collaboration. A research assistant was then assigned to each site to generate a list of health professionals likely to be eligible based on their affiliations, functions and employment status. The research assistant then approached these health professionals, explained the project briefly, assured them that data would be anonymous and confidential, confirmed their eligibility, and invited them to participate. Each participant then signed a consent form and agreed to an interview appointment over the following ten business days.

Participants were interviewed either at their health center or on the phone, depending on their preferences and availability. Before the interview, participants watched a 10-min video showing two simulated consecutive prenatal follow-up consultations between a health professional and a pregnant women and her partner. During the first visit, the health professional gives a PtDA to the couple, explains its purpose, reviews its content and explains the risks and benefits of doing or not doing the prenatal testing. At this point the video presents a close-up of the content of the PtDA page by page and item by item. However, the interviewed health professionals were not offered a printed copy of the decision aid to look at. The PtDA presented in the video had been developed earlier by the team based on an existing decision support tool [24]. Then the health professional invites the couple to look over the PtDA at home. The health professional clarifies that the pregnant women is facing the decision to undergo prenatal screening for DS and that more information can be found in the PtDA that will help the decision-making process. In the second visit (4 weeks later), the pregnant woman, still accompanied by her partner, discusses the decision with the health professional and says she has made her choice. However, the viewer does not find out what she has chosen so that participants would not be influenced by her choice. Immediately after watching the video, to identify factors influencing the use of a PtDA for DS screening during a clinical pregnancy follow-up, semi-structured interviews took place consisting of 11 open-ended questions based on the Theoretical Domains Framework: 1) knowledge/awareness, 2) advantages, 3) disadvantages, 4) incentives, 5) emotions, 6) anticipated regret, 7) social approval, 8) social disapproval, 9) appraisal/evaluation, 10) facilitators and 11) barriers [25].

The Theoretical Domains Framework is comprised of 12 theoretical domains relevant to behavioral change [25‐27]. It postulates that factors influencing behavior change can be mapped onto these theoretical domains and can be used to design effective theory-based implementation interventions. In this study, the behavior of interest was defined as follows: use (action) of a PtDA about prenatal screening test for DS (target) in the context of a prenatal care follow-up with a health professional (context) [28]. Time was not specified, as the behavior was hypothetical. At the end of the interview, we also assessed participants’ socio-demographic characteristics. Interviews were audio-recorded and transcribed verbatim. Each recording had an identification code and was uploaded to our database.

To identify participants’ salient beliefs (henceforth referred to as “opinions”), the second author (MELP) analyzed the full transcripts using a content analysis approach. Then, using N-Vivo v.10 software (QSR International, Melbourne, Australia), similar opinions were grouped into themes (henceforth referred to as “influential factors”) which were mapped onto the Theoretical Domains Framework constructs and then onto 10 of its 12 domains: 1) beliefs about consequences, 2) environmental context and resources, 3) social influences, 4) social/professional role and identity, 5) knowledge, 6) emotions, 7) beliefs about capabilities, 8) motivation and goals, 9) skills, and 10) behavioral regulation. No influential factors mapped onto the two remaining theoretical domains, which are 11) memory, attention and decision processes, and 12) nature of the behaviors. These domains were less relevant to our participants, probably because they had hardly ever used PtDAs in the clinic before [25]. The first author (JL) independently read all the transcripts and double-checked the tree-node. She suggested the addition of relevant categories or themes and the removal of less relevant ones, and discussed a final coding scheme with the second author. Any discrepancies were resolved through discussion with team members. The same procedure was undertaken with all identified influential factors. The number of participants who identified the same influential factor, along with the percentage rate, was calculated for each health professional category (n, %, Table 2) and for all professional categories together (N, %, Table 2). Modal beliefs were also identified (superscripted "a" in Table 2), i.e. the most frequently reported influential factors for each construct (up to 75 %), to give a sense of the relative importance of this factor (some factors, for example, were mentioned more than once by a single participant) [23]. The frequencies of modal beliefs were calculated by summing the total number of quotes for each influential factor (Q) as well as the percentage of times that the same opinion was repeated per construct (%QC) (Table 2).

Out of 73 eligible health professionals, 42 were approached between July 25, 2014 and February 28, 2015, and 36 agreed to complete interviews (86 % overall participation rate), including 15 family physicians (79 %), 12 midwives (92 %) and nine obstetrician-gynecologists (90 %). The remaining 31 eligible family physicians were not approached because we stopped recruitment once we had reached the expected sample size for this category (n = 15, Fig. 1). Twenty-seven women (75 %) and nine men (25 %) were recruited; four participants were interns (11 %, Table 1); and two obstetrician-gynecologists were interviewed by phone (6 %) instead of in person (n = 34, 94 %, data not shown). The mean age was 42.1 ± 11.6 years old, mean practice experience was 12.6 ± 12.0 years, and mean number of prenatal visits per week was 17.4 ± 21.5. The mean length of interviews including video viewing (10 min) was 28 ± 6 min (Table 1).

A total of 64 influential factors (see Additional file 1) were reported by participating health professionals regarding the use of a PtDA for deciding about prenatal screening for DS during a pregnancy follow-up visit, including 35 factors that were reported by 20 % or more participants in any health professional category (Table 2).

This domain has four theoretical constructs: advantages, disadvantages, anticipated regret and appraisal. The two most frequently reported advantages of using a PtDA for decisions about prenatal screening for DS were: 1) it helps patients to think about the decision [N = 18(50 %); Q = 25(44 %)^a]; and 2) its visual content is helpful for patients [N = 11(31 %); Q = 14(25 %)^a]. Obstetrician-gynecologists reported these two advantages less frequently (n = 3, 33 % and n = 2, 22 %, respectively) than did the other two professionals. Five midwives (n = 5, 42 %) said that an advantage of the PtDA was that it enables patients to express their preferences, while only one obstetrician-gynecologist (n = 1, 11 %) and two family-physicians (n = 2, 13 %) reported this. Four midwives (n = 4, 33 %) and three obstetrician-gynecologists (n = 3, 33 %) thought that another advantage was that PtDAs promote decision making, but only one family physician thought this (n = 1, 7 %, Table 2).

The most frequently reported disadvantage was that PtDA use did not fit in with the timing of the current DS prenatal screening program covered by the government of the province of Quebec [N = 13(36 %); Q = 15(45 %)^a]. While the video proposes 4 weeks for thinking about the decision between the first prenatal visit and the second, pregnant women may actually have less than 3 weeks between the first visit, generally between 10 and 12 weeks of pregnancy, and the first blood sample for DS screening, which must be between 10 and 13 weeks of pregnancy. While only one family physician mentioned this (n = 1, 7 %), five midwives (n = 5, 42 %) and seven obstetrician-gynecologists (n = 7, 78 %) mentioned this factor. The latter also said more frequently that the PtDA content is incomplete (n = 4, 44 %, Table 2).

Overall, for the anticipated regret construct, more health professionals said they would regret it if they did not use a PtDA [N = 16(44 %); Q = 16(62 %)^a] compared to those who said they would not regret it [N = 10(28 %); Q = 10(38 %)^a]. Almost all midwives mentioned they would regret not using it (n = 8, 67 %), while family physicians (n = 5, 33 %) and obstetrician-gynecologists (n = 3, 33 %) were less likely to regret not using it (Table 2).

For the appraisal construct, the majority of health professionals appraised the use of a PtDA positively, i.e. thought its use would be good, useful, beneficial, or excellent [N = 29(81 %); Q = 33(92 %)^a], especially midwives (n = 12, 100 %) (Table 2).

This domain has two constructs: facilitators and barriers. The most frequently reported facilitator was its availability in the office [N = 27(75 %); Q = 47(67 %)^a]. Midwives mentioned comprehensibility for patients more frequently than the others (n = 5, 42 %) and a quarter mentioned having enough time to present it as a facilitator (n = 3, 25 %, Table 2).

The two most frequently reported barriers were: 1) not having enough time to present it during a clinical encounter [N = 26(72 %); Q = 48(58 %)^a]; and 2) if it is too complex for patients [N = 16(44 %); Q = 21(25 %)^a]. Some family physicians considered PtDAs in print form only as a barrier (n = 3, 20 %) while some midwives considered unbalanced/biased PtDAs as a barrier (n = 3, 25 %, Table 2).

This domain has two opposite constructs: people who would approve and people who would disapprove. Health professionals most frequently reported their colleagues as being both the people who would approve of their using a PtDA [N = 27(75 %); Q = 29(64 %)^a], and to a lesser extent, the people who would disapprove of it [N = 11(31 %); Q = 11(58 %)^a]. No obstetrician-gynecologists reported their family as a social influence (Table 2), and only a few health professionals said that their patients/clients would approve [N = 5(14 %); Q = 5(11 %), see Additional file 1] or disapprove of PtDA use [N = 2(6 %); Q = 2(11 %), see Additional file 1].

This domain has one construct: moral norms. The most frequently reported factor relating to moral norms was that seven health professionals felt it was their duty to present the PtDA to their patients [N = 7(19 %); Q = 8(67 %)^a] (Table 2).

The majority of health professionals said they did not know of any PtDAs [N = 23(64 %); Q = 23(59 %)^a] although some said they knew about the government pamphlet (not a decision aid [21]) on the prenatal screening program covered by the province of Quebec [N = 10(28 %); Q = 10(26 %)^a] [29]. Some participants reported that they knew of PtDAs for other decisions (prostate cancer, breast cancer, and respiratory tract infections) [N = 6(17 %); Q = 6(15 %)]. More midwives than other health professionals were aware of the existence of PtDAs (only four said they did not know of any PtDAs, n = 4, 33 %) and of the government pamphlet (n = 9, 75 %) about prenatal screening for DS. Family physicians appeared to be more familiar with PtDAs for other decisions (n = 4, 27 %, Table 2).

The most frequently reported emotion regarding the use of a PtDA was that its use would give satisfaction [N = 8(22 %); Q = 9(53 %)^a], especially among obstetrician-gynecologists (n = 4, 44 %, Table 2).

This domain has one construct: self-efficacy. The most frequently reported factor about self-efficacy was that health professionals would feel comfortable using it [N = 11(31 %); Q = 13(68 %)^a], especially midwives (n = 7, 58 %, Table 2).

This domain has one construct: incentives. The most frequently reported incentive was that health professionals would use the PtDA because it is a relevant source of information [N = 24(67 %); Q = 38(75 %)^a]. However, fewer obstetrician-gynecologists thought so (n = 4, 44 %). Compared to the others, fewer family physicians mentioned the incentive of having to decide with their patients (n = 1, 7 %, Table 2).

Health professionals reported only one influential factor in the skills development construct. Overall, a quarter of the health professionals said they would need prior training to use the PtDA properly during a clinical encounter [N = 9(25 %); Q = 11(100 %)^a], while fewer obstetrician-gynecologists thought so (n = 1, 11 %, Table 2).

This domain has one construct: action planning. The two most frequently reported preliminary steps were that the PtDA could be explained by a nurse before the consultation [N = 11(31 %); Q = 17(57 %)^a], followed by that the PtDA could be given to patients before the consultation [N = 7(19 %); Q = 13(43 %)^a]. None of the midwives reported these two factors (Table 2).

Table 3 presents examples of opinions (translated from the French originals) by participants that illustrate each of the 35 influential factors (IF) listed in Table 2.