To develop a patient decision aid (PtDA) for mode of delivery after caesarean section that integrates personalised prediction of vaginal birth after caesarean (VBAC) with the elicitation of patient preferences and evidence-based information.
A PtDA was developed and pilot tested using the International Patients Decision Aid Standards (IPDAS) criteria.
Obstetric health care in the Netherlands.
A multidisciplinary steering group, an expert panel, and 25 future users of the PtDA, i.e. women with a previous caesarean section.
The development consisted of a construction phase (defini-tion of scope and purpose, and selection of content, framework, and format) and a pilot testing phase by interview. The process was supervised by a multidisciplinary steering group.
Main outcome measures
Usability, clarity, and relevance.
The construction phase resulted in a booklet including unbiased balanced information on mode of birth after caesarean section, a preference elicitation exercise, and tailored risk information, including a prediction model for successful VBAC. During pilot testing, visualisation of risks and clarity formed the main basis for revisions. Pilot testing showed the availability of tailored structured information to be the main factor involving women in decision-making. The PtDA meets 39 out of 50 IPDAS criteria (78%): 23 out of 23 criteria for content (100%) and 16 out of 20 criteria for the development process (80%). Criteria for effectiveness (n = 7) were not evaluated.
An evidence-based PtDA was developed, with the probability of successful VBAC and the availability of structured information as key items. It is likely that the PtDA enhances the quality of decision-making on mode of birth after caesarean section.
Decision-making in pregnancy is particularly complex because women must consider not only their own health but also that of their unborn babies. In the decision on mode of delivery after caesarean section, the importance of involving and supporting pregnant women has been highlighted by various international guidelines.[1-4] When no contraindications are present, women should be provided with a choice between an elective repeat caesarean section (ERCS) and an intended vaginal birth after caesarean (VBAC; i.e. a trial of labour).[1-4] It is recommended that this decision is based on a discussion of the risks, benefits, and implications of both options.[1-5] Mean reported intended VBAC success rates are as high as 72–76%, and a successful VBAC is associated with fewer major complications, a shorter recovery period, and higher perceived maternal satisfaction scores.[8, 9] In addition, an intended VBAC is in general more cost-effective than ERCS; however the drawback of intended VBAC is possible failure leading to an emergency caesarean section, which is associated with a higher risk of major complications, including hysterectomy and operative injury, and lower levels of maternal satisfaction,[8, 9] than with either ERCS or successful VBAC. The existence of major international differences in (intended) VBAC rates suggests that choices for mode of delivery after caesarean section do not always reflect women's preferences, but may result from non-medical factors like hospital policy, concerns about medical liability, or provider preference. The question is how guideline recommendations on mode of birth after caesarean section can be implemented in practice. Because of the complex nature of the decision-making process, and because it concerns a preference-sensitive decision, mode of delivery should be discussed within a shared decision-making setting. Shared decision-making can be enhanced by the use of patient decision aids (PtDAs), which provide unbiased information in a clear and comprehensive way. PtDAs have proven to be effective tools for facilitating communication between patient and healthcare provider, and for increasing the knowledge and involvement of patients. The individualised prediction of successful intended VBAC may contribute to the personalisation of the decision-making process, and could place potential risks into context. Although there are several prediction models,[6, 12, 13] and there are two PtDAs on mode of birth after caesarean section,[14, 15] as yet there has been no published integration of a prediction model with a PtDA. In this study we aim to develop and pilot test a PtDA for mode of delivery after previous caesarean section that integrates a personalised prediction of successful VBAC with the clarification and elicitation of patient preferences and evidence-based information.
Construction of the PtDA
The PtDA was developed using the systematic process illustrated in Figure 1.
Scope and purpose
The purpose of the PtDA was to improve decision-making on mode of birth after previous caesarean section, as part of a guideline implementation strategy for the improvement of current Dutch care for women after caesarean section. The target audience, i.e. future users of the PtDA, were defined as pregnant women with a history of one previous caesarean section without a contraindication for an intended VBAC. Contraindications were determined in accordance with current international guidelines on pregnancy and childbirth,[1-5] and included previous uterine rupture or other relevant uterine scar, or a contraindication for vaginal labour such as malpresentation or placenta praevia.[1-4] It was decided not to include information in the PtDA on twin or breech pregnancies, or in cases when more than one caesarean section had already been performed, as we considered counselling in these cases to be essentially different.
Based on current international guidelines on pregnancy and childbirth after caesarean section,[1-5] literature, and expert opinion, we identified eight domains important for decision-making on mode of birth after prior caesarean section: (1) risks and (2) benefits of intended VBAC and ERCS; (3) induction of labour; (4) the consequences of the decision for future pregnancies; (5) predictors of successful intended VBAC; (6) the woman's preference for a vaginal or caesarean section delivery; (7) the possibility of a birth plan (i.e. a document containing the woman's preferences when giving birth); and (8) previous birth experiences. Current guideline recommendations dictated the level of detail of the medical contents of domains 1–5.[1-5] Current data on risk estimates were included, based upon the principle of best available evidence. For domain 5, a prediction model was included that estimates the probability of successful intended VBAC. This prediction model was developed by our research group in a western European cohort of women with a high rate of intended VBAC. As prescribed by current methodological guidelines for prediction research, we selected potential predictors based on clinical reasoning and evidence from the literature.[18, 19] The final set of predictors included in the model are: estimated fetal weight ≥90th percentile; previous caesarean section for failure to progress; previous vaginal delivery; induction of labour; pre-pregnancy body mass index; and ethnicity. The development and internal validation of this prediction model is described elsewhere. Domains 6, 7, and 8 were elaborated based on expert opinion and a review of the literature.
Selection of framework and format
Based on the content selection, a practical framework was derived that consisted of the aforementioned eight domains. A prototype of the PtDA was developed in accordance with the international quality criteria of the International Patients Decision Aid Standards (IPDAS). More information regarding the selection process for the framework and the format can be found in Appendix S1.
Pilot testing and revision
Pilot testing (i.e. alpha testing) comprised several rounds of review and revision of the prototype, based on information gathered during interviewing (ref. no. 12-4-034). We recruited women in 2012 at an obstetric outpatient clinic of a Dutch university teaching hospital. Inclusion criteria were: pregnancy with a history of one caesarean section and a current singleton cephalic pregnancy; and a minimum age of 18 years. Women with a contraindication for an intended VBAC or who could not read Dutch were excluded. The primary outcome measures were usability, clarity, relevance of the content, and relevance of the PtDA in general. After informed consent was obtained, women participated in a 30–minute face-to-face audio-recorded interview. The first part of the interview was unstructured according to the ‘think aloud method’. This method allowed us to obtain data using observation and without interrupting participants' thinking processes. The second part of the interview was conducted using a questionnaire. Issues of particular interest to us were the opinions of women on the stage at which the PtDA was introduced, opinions about the prediction model, whether there was a minimum outcome of the prediction model that women considered acceptable when deciding to try for a VBAC, unwanted adverse effects of the PtDA (such as increased anxiety), and opinions on the way in which risk was communicated and the preference elicitation tool.
The interviews were transcribed verbatim and translated into a thematic framework. Interviews were analysed by two independent researchers (E.S. and B.A. or M.K.). Data collection and revision of the PtDA was continued over several rounds until data saturation was achieved.
Construction of the PtDA
The practical framework, i.e. the outline of the PtDA, was constructed on the basis of the selected medical content and our eight domains. With the availability of an evidence-based report on birth after caesarean section by Guise et al., the quantification of risk was mainly based on data from meta-analyses. We included four additional articles because they provided data that were more applicable for the target audience of the PtDA, based on population characteristics, were recently published, and therefore not included in Guise et al.'s report, or were included in that report but not in the form of a meta-analysis. The risk of uterine rupture was shown to range from 2.1 to 15 per 1000; we used 8 per 1000 (for spontaneous onset of labour) as a mean, as this was more appropriate for the Dutch setting. This was based on data from a Dutch prospective cohort study, which reported a higher incidence (8–15 per 1000) of uterine rupture in women who attempt VBAC than was reported by the meta-analysis (4.3 per 1000), and in a recent large cohort study by Fitzpatrick et al. (2.1 per 1000). Our final selection of the risks of intended VBAC and ERCS is shown in Table 1.
Table 1. Risks of intended vaginal birth after caesarean and elective repeat caesarean section: estimates that were selected for the contents of the patient decision aid based on literature research
Composite of uterine rupture, hysterectomy, and surgical injury.
Composite of cysthotomy, bowel injury, uretheral injury, and bowel obstruction.
No evidence for a significant difference in the long term
The PtDA was subdivided into a general introduction, seven steps (chapters) that guide the decision-making process, and a supplement that clarifies the induction of labour after previous caesarean section. The introduction to the PtDA describes its aim, explains the choice that has to be made regarding mode of delivery, and lists the options.
Step one of the PtDA consists of a question regarding any existing preference for either VBAC or ERCS before reading the PtDA. This enables the woman to identify preliminary thoughts, beliefs, and decisions, but this preference can also be discussed with the provider.
Step two includes two questions that can be used to evaluate previous birth experiences, fear of vaginal delivery, and other external influences that the woman regards as relevant.
In step three the woman is shown a fact sheet laying out the risks and benefits of both modes of delivery. Risks are illustrated by using pictographs on a scale of per 1000 women. This level of detail with regard to the risks led us to give careful and extensive consideration to the method of risk visualisation. We derived several formats of the pictographs and elaborated these during pilot testing. Step three also includes the prediction model for successful intended VBAC. This element of the PtDA is shown in Figure 2. Providers enter variables into the prediction formula using a computerised calculator and the outcome is noted on the corresponding (empty) pictograph within the PtDA. In order to provide context to the outcome of the prediction model, a mean probability of 70–75% of successful VBAC is included.[1, 23] All risks and benefits are explained using plain language, pictographs, and summaries, in a balanced, unbiased manner.
Step four involves a preference elicitation tool, which was designed as a list of statements that women (and their partners) can consider to clarify their values and thoughts around several aspects of the decision. The statements can be scored on a four-point Likert scale to show level of agreement with the statements. Statements were derived from the entire contents of the PtDA and were completed with data from qualitative studies of women's views on deciding about mode of birth after caesarean section.[27, 28] The preference elicitation tool can be viewed in Appendix S2.
Step five of the PtDA explains the option of making agreements on the delivery (birth plan), such as an epidural analgesia. This step facilitates discussion of a birth plan between the woman and healthcare provider.
Step six requires a preliminary decision.
Step seven is aimed to encourage the woman and healthcare provider to agree on follow-up.
In the supplement, induction of labour after caesarean section is considered: risks and methods of induction are explained, and this encourages a recalculation of the probability of successful intended VBAC.
Pilot testing and revision
In total, 16 eligible women with a history of one previous caesarean section were included in the first pilot. The baseline characteristics are described in Table 2. The mean age of the participants was 33 years: all were native Dutch speakers and 14 out of 16 (87%) had the median socio-economic status (SES). Pilot 1 consisted of three rounds in which interviews were conducted and the PtDA was revised. The majority of the revisions concerned wording and clarity. Many women stated that they experienced a lack of information within the current care set-up. With regard to timing of exposure to the PtDA, a range of 12–36 weeks of gestation was considered appropriate. Furthermore, 15 out of 16 participants thought that the outcome of the prediction model for successful intended VBAC helped their decision-making process. Two women would choose an ERCS if their predicted probability of successful VBAC was 50% or less, and one woman set the cut-off value at 60–65%. Thirteen women noted that they could not define a minimum outcome of the prediction model that they considered acceptable for trying a VBAC.
Table 2. Baseline characteristics of pregnant women with a history of one caesarean section who participated in pilot testing of the patient decision aid
Pilot 1 (n =16)
Pilot 2 (n =9)
Derived from post-code related socio-economic status scores, based upon data on income, educational level, and profession.
In general, all women thought that the PtDA was clear and informative. The PtDA was considered relevant by all participants. The stepwise approach of the decision-making process was repeatedly cited as a positive point. Also, most women felt that the PtDA was complete, balanced, and objective. An aspect that was repeatedly noted as requiring improvement was the pictographs of the health outcomes. In rounds 1 and 2, many women stated that one or more of the pictographs did not clarify the risks. They thought that the pictographs were distracting and added nothing to the decision-making process. In the third round, women thought that the presented figures were clear, but two out of seven women still thought that the pictures could be omitted.
In the third round of pilot 1, no new information was obtained. Subsequently, we presented the PtDA to an expert panel for critical review and comments. Eight out of 11 experts approved this version; however, three out of 11 experts expressed concerns regarding the presentation of risks. A new outline was suggested for the pictographs that involved using cubes instead of pregnant women and means instead of ranges. Based on this feedback, the PtDA was revised and this new version was tested in pilot 2.
Nine eligible women were included. The baseline characteristics are summarized in Table 2. The mean age of the participants was 36 years; eight out of nine (89%) women were native Dutch speakers and four out of nine (44%) women had the median SES. Pilot 2 resulted in minor additional changes regarding the wording and clarity of the PtDA. The revised pictographs were considered to be clear, and seven out of nine women thought that the pictographs supported the information in the tables. Eight out of nine women thought that the PtDA was relevant. After pilot 2, no major revisions were required. After final adjustments the PtDA was presented to the expert panel, and was approved by all of the members.
We used the checklist of the IPDAS collaboration to estimate the quality of the developed PtDA.[21, 29] Of the 64 items on the checklist, 50 quality criteria were applicable to our study based on the scope of the PtDA. The final version of the PtDA met 39 out of the 50 applicable IPDAS criteria (78%).[21, 29] With regard to content, all 23 criteria were met. For the domain development process, we met 16 out of 20 criteria. The four criteria that we did not meet were whether the PtDA is written according to a validated readability score (10.4), and whether the PtDA is understood by those with limited reading skills (10.6). Although we pilot tested the PtDA, we did not actually field test (i.e. beta test) the PtDA in the current study (1.4, 1.5). The effectiveness of the PtDA was not evaluated in the current development study; therefore we have not yet met the seven criteria (12.1–12.8) concerning effectiveness.
The main findings
The current study provides insight into the systematic process of developing a decision-support tool by means of interdisciplinary cooperation of healthcare professionals and the critical input of future users, i.e. women and healthcare providers. The agreed format is a booklet with a stepwise approach for decision-making. Opinions on the stage at which women should be introduced to the PtDA varied widely between individual women and providers; therefore, we agreed on introduction at any gestational age before 36 weeks based on the needs of the woman concerned. The PtDA is evidence based and its content is in agreement with current recommendations and guidelines. Hence, this PtDA could be seen as a guideline implementation tool. The quantification of risks was mainly based on data from meta-analyses. Although no studies on this topic with a grade–A level of evidence were available, many large cohort studies provide insight into relevant risks. The need for structured information was confirmed, as women indicated that clear and relevant information is currently hard to find. Women emphasised the clarity of the step-wise approach to decision-making and of the preference elicitation tool. As risks had to be shown on a scale per 1000 women, several versions of the pictographs had to be pilot tested. Thanks to the critical input of women and healthcare experts, we established a format that was usable and clear. The pilot study confirmed that women value the use of a prediction model for decision-making, although many women could not define a minimum outcome of the prediction model that they considered acceptable for trying a VBAC. However, our results cannot directly show the impact that the application of the prediction model will have on women's choices. The actual application of the prediction model requires the woman and the gynaecologist to interpret the outcome of the prediction model together, and this is necessary in order to measure its effect. Based on the results of pilot testing, it is likely that this PtDA supports the decision-making process, but results on the effectiveness of the strategy must be further explored.
Strengths and weaknesses
We developed the PtDA according to a thorough systematic approach. Quality control was performed using critical input and review of the expert panel. The opinions of 25 women amongst the target audience were taken into account during development; however, readers may question the representativeness of the pilot study because it was performed in just one tertiary hospital and only included a small number of women with low SES. We acknowledge that results should be confirmed in other settings with a larger study population. Other drawbacks of our design are that we did not specifically test the PtDA embedded within the routine clinical pathway (field testing). Also, we did not target women with limited reading skills, and this may have influenced attitudes about the pictographs, for example. Therefore, the results of this pilot study should be interpreted with caution and should be confirmed and complemented by the results of a larger, prospective trial.
Two other PtDAs on mode of birth after caesarean section have been published. The PtDA of Shorten et al. was published in 2004 and comprised a booklet with a value clarification exercise. A systematic review conducted by Vlemmix et al. estimated the agreement of published obstetrics PtDAs with the IPDAS criteria. According to this review, the PtDA of Shorten et al. meets 34 of 50 IPDAS criteria. Another PtDA, developed by Emmett et al. and published in 2007, comprised a computerised information programme with a decision analysis tool. This PtDA meets 29 of 50 IPDAS criteria. In the current study, we were able to meet 39 out of 50 applicable IPDAS criteria, which compares well with other PtDAs in obstetrics, especially when taking into account that effectiveness criteria (n = 7) will be further explored in continuing research. Our PtDA has similarities to other available PtDAs on delivery after caesarean section in the choice of the key issues that are addressed;[14, 15] however, our PtDA also provides a new decision-making element, namely the integration of an individualised estimation of intended VBAC success with evidence-based information. We consider the prediction model to be applicable to other western European populations with comparable VBAC rates and ethnicities. With regard to other specific information on risk estimates, the PtDA should be evaluated along local guidelines and literature, as differences in obstetric policy and case mix may impair the validity of the data. Hence, the PtDA can be adapted for different populations.
We expect that the introduction of our PtDA in current care will enhance the involvement of women in decision-making on mode of delivery after caesarean section. Subsequently, it is hoped that this will reduce decisional conflicts and increase patient satisfaction. The application of the prediction model might alter birth preferences and reduce the rate of emergency caesarean section; if this effect were to be observed, we would expect a concomitant reduction in major maternal and neonatal morbidity. Before this PtDA can be incorporated into current care, its effectiveness needs to be ascertained in a controlled prospective trial. Currently, a prospective trial amongst 12 Dutch hospitals is being carried out in order to determine its effectiveness in terms of caesarean section rates, adherence to guidelines, decisional conflict, patient satisfaction, feasibility, and costs (SIMPLE TRIAL: NCT00800384).
This article provides insight into the systematic development of a PtDA that aims to enhance the involvement of women in decision-making on mode of delivery after caesarean section. It has uncovered the main issues around the involvement of women in decision-making, namely the medical and non-medical items that women and experts consider relevant for decision-making, a method of communicating risks, the additive value of using a prediction model for successful intended VBAC, and the stage at which women are introduced to the PtDA. The need for a healthcare strategy for mode of delivery after caesarean section was confirmed. In the current format the PtDA is applicable for western European countries, and specific information with regard to risk estimates may be adapted for different populations. The pilot study suggests that the PtDA and the included prediction model support the decision-making process. Its effectiveness in terms of caesarean section rates, adherence to guidelines, decisional conflict, patient satisfaction, feasibility, and costs will be further examined in a prospective controlled trial.
Disclosure of interests
The authors declare that they have no conflicts of interest to declare.
Contribution to authorship
RH and HS obtained the funding for the trial. All authors contributed to the protocol and design of this study. ES, BA, and MK performed and analysed the interviews. ES, TW, RH, and HS designed the first draft of the PtDA. EV, CD, SK, AK, SM, JN, RA, KB, TH, BM, MN, MP, JR, FR, RV, and MW critically reviewed the decision aid and contributed to its contents. ES drafted the article, with input, critical review, and editing from all of authors. All of the authors accept full responsibility for the overall content of this article.
Details of ethics approval
The Medical Ethical Committee (CMO) of Maastricht University (azM/UM) declared, on 26 March 2012, that no ethical approval was required for this study protocol (ref. no. 12-4-034).
This study was funded by ZonMW (grant no. 17100.3006): the Netherlands Organisation for Health Research and Development.
The authors would like to thank the following: the Netherlands Organisation for Health Research and Development (ZonMw), who funded this project (grant no. 17100.3006), V. Stégehuis for her contribution to data collection, the research staff for recruiting participants, and all of the women who participated in the pilot study.