Engagement of Stakeholders, Researchers, and Funders

A total of 13 stakeholders (2 Federal, 11 non-Federal) representing the perspective of patient advocacy groups, academic researchers, obstetricians and gynecologists, nursing and nurse-midwifery professional organizations, the payor perspective, and national foundations and societies agreed to participate in one or more of the stages of ranking and prioritization. The group includes five Key Informants/Technical Expert Panel members from the draft CER. Throughout the teleconferences and surveys, stakeholder participation varied from 69 percent to 84 percent. Figure 4 presents stakeholder recruitment and participation results.

Figure 4

Stakeholder recruitment and participation results.

Identification of Evidence Gaps

In Phase 1, the EPC investigators started with the 17 evidence gaps identified in the draft review and translated these gaps into a list of 12 sample research questions and 12 sample methodologic recommendations that have potential utility for reducing cesarean birth. We invited the stakeholder panel to participate in a teleconference and then a web-based survey to make the list broader and more comprehensive and to suggest topics we may have omitted.

During the teleconference and snowballing survey, stakeholders expanded the initial list to 47 research questions and 17 methodologic recommendations. These items encompassed a wide range of topics and are presented in Tables 3–4. Figure 5 shows the snowballing process and results for Phase 1.

Table 3

Snowballed list of research questions.

Table 4

Snowballed list of methodologic recommendations.

Figure 5

Snowballing results. CER = Comparative Effectiveness Review; EPC = Evidence-based Practice Center

Prioritization

During initial voting, stakeholders were asked to respond to a Web-based survey in which they distributed 47 points among the research gaps and 17 points among the methodologic issues (Figure 6). Once the surveys were completed, we totaled points for each item, ranked items based on number of points assigned, and eliminated the lowest one-third responses. (Appendix D). The top five research questions were the following:

Figure 6

Prioritization results. PiCMe = Prioritization Criteria Method

1. What factors drive a patient's decision to undergo a primary cesarean during labor, e.g. prior cesarean, general fears, fear of future pelvic floor disorders?
2. Why do some patients prefer to undergo elective cesarean?
3. Do different staffing models such as use of hospitalists and integration of midwives reduce the number of cesarean births?
4. What physician factors contribute to the use of cesarean during labor, e.g. residency training, attitude toward cesarean, practice setting, practice size, shift/time of day, use of hospitalists, personal birth experience?
5. Does use of the Consortium for Safe Labor labor curves reduce use of cesarean?

The top two methodologic recommendations were:

1. Capture all categories of birth outcomes (cesarean, emergent cesarean, assisted vaginal, and spontaneous vaginal births) and related complications in order to assess if reductions in cesarean occur at the cost of increased use of other interventions or increased complications.
2. Future studies should include a full range of practice settings including community hospitals, birth centers, and health systems.

After the initial voting, we held a second teleconference to discuss results. During the call, stakeholders agreed to eliminate the lowest one-third of items, and combined similar questions to reduce redundancies. This process resulted in a list of 26 research questions and 10 methodologic recommendations.

Next, stakeholders completed an electronic survey and ranked the remaining 26 research questions and ten methodologic improvements from 1 (highest priority) to 26 (lowest priority) and from 1 to 10, respectively. We totaled points for these items, ranked them from fewest points (highest priority) to most points and eliminated the bottom one-third items, leaving 16 research questions and seven methodologic recommendations.

After the initial ranking, the top five research questions were:

1. Can tighter standards for induction (indicated or elective) among primiparous patients reduce use of cesarean?
2. Do different staffing models, e.g. models that use hospitalists or midwives, reduce the number of cesarean births?
3. How does implementing uniform definitions for arrest of labor and its management influence use of cesarean?
4. Would changing the timeframes for normal progress in latent and active labor reduce primary cesareans?
5. What physician factors contribute to the use of cesarean during labor, e.g. residency training, attitude toward cesarean, practice setting, practice size, shift/time of day, use of hospitalists, personal birth experience?

The top two methodologic recommendations were:

1. Capture all categories of birth outcomes (cesarean, emergent cesarean, assisted vaginal, and spontaneous vaginal births) and related complications in order to assess if reductions in cesarean occur at the cost of increased use of other interventions or increased complications.
2. Conduct studies that allow stratification on patient characteristics such as nulliparity and multiparity and have adequate power to detect differences across strata.

Research Needs

For the final prioritization step, we scored research questions and methodologic recommendations based on total points assigned to seven AHRQ potential value criteria. We established tiers for top-, middle-, and lower-ranked items and created cutoff points where natural breaks in total points assigned occurred. In Tables 5–6, we present research questions and methodologic recommendations by tier. We present the top five research questions and top two methodologic recommendations for each AHRQ criterion in Appendix D, Tables D7–D8.

Table 5

Highest priority research questions.

Table 6

Highest priority methodologic recommendations.

The top-tier research questions reflect a focus on standardization strategies for induction and arrest of labor (three of five), systems-strategies (one of five) and staffing models (one of five). For strategies that standardize induction and arrest of labor, we recommend cluster randomized controlled trials with randomization of entire labor and delivery units. For trials of systems-level strategies and staffing models, we recommend studies multisite studies to improve power and generalizability.

Identification of Current and Ongoing Studies

Few studies explicitly set out to reduce cesarean births. We identified six ongoing randomized controlled trials using the search criteria from the original review. The RCTs addressed pregnancy/fetal monitoring (1), prolonged pregnancy (1), maternal obesity/weight gain (3), and labor dystocia (1). The RCTs that address fetal monitoring and labor dystocia may address the research needs we identified in this project.