Expert surgical consensus for prenatal counseling using the Delphi method

Journal of Pediatric Surgery xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

Journal of Pediatric Surgery journal homepage: www.elsevier.com/locate/jpedsurg

Expert surgical consensus for prenatal counseling using the Delphi method☆,☆☆ Loren Berman a,b,⁎, Jordan Jackson c, Kristen Miller d, Rebecca Kowalski e, Paul Kolm e, Francois I. Luks f,g a

Sidney Kimmel Medical College of Thomas Jefferson University, 1025 Walnut St. #100, Philadelphia, PA 19107, USA Nemours/Alfred I. duPont Hospital for Children, P.O. Box 269, Wilmington, DE 19899, USA University of California San Francisco—East Bay, 1411 East 31st St., QIC 22134, Oakland, CA 94602, USA d National Center for Human Factors in Healthcare, MedStar Health 3007 Tilden St NW, Washington, DC 20008, USA e Value Institute at Christiana Care Health System and Christiana Care Emergency Department, 4755 Ogletown-Stanton Rd., Newark, DE 19718, USA f Alpert Medical School of Brown University, Box G-A1, Providence, RI 02912, USA g Hasbro Children's Hospital, 593 Eddy St., Providence, RI 02903, USA b c

a r t i c l e

i n f o

Article history: Received 12 May 2017 Received in revised form 27 September 2017 Accepted 21 November 2017 Available online xxxx Key words: Congenital diaphragmatic hernia Congenital pulmonary airway malformation Prenatal counseling Delphi method Expert opinion

a b s t r a c t Background: Pediatric surgeons frequently offer prenatal consultation for congenital pulmonary airway malformation (CPAM) and congenital diaphragmatic hernia (CDH); however, there is no evidence-based consensus to guide prenatal decision making and counseling for these conditions. Eliciting feedback from experts is integral to defining best practice regarding prenatal counseling and intervention. Methods: A Delphi consensus process was undertaken using a panel of pediatric surgeons identified as experts in fetal therapy to address current limitations. Areas of discrepancy in the literature on CPAM and CDH were identified and used to generate a list of content and intervention questions. Experts were invited to participate in an online Delphi survey. Items that did not reach first-round consensus were broken down into additional questions, and consensus was achieved in the second round. Results: Fifty-four surgeons (69%) responded to at least one of the two survey rounds. During round one, consensus was reached on 54 of 89 survey questions (61%), and 45 new questions were developed. During round two, consensus was reached on 53 of 60 survey questions (88%). Conclusions: We determined expert consensus to establish guidelines regarding perinatal management of CPAM and CDH. Our results can help educate pediatric surgeons participating in perinatal care of these patients. Level of Evidence: V. © 2017 Elsevier Inc. All rights reserved.

Birth defects are one of the leading causes of infant mortality, accounting for more than 20% of all infant deaths, and creating an ongoing source of morbidity for many afflicted children who survive infancy [1]. Prenatal counseling plays an increasingly important role in influencing decision making during pregnancy. For surgical anomalies, referral to appropriate surgical specialists in addition to maternal–fetal medicine consultants is critical. There are several reports in the literature of

☆ Conflicts of interest: None. ☆☆ This work was supported by an Institutional Development Award (IDeA) from the National Institute of General Medical Sciences of the National Institutes of Health under grant number U54-GM104941 (PI: Binder-Macleod). ⁎ Corresponding author at: Department of Surgery, Nemours/Alfred I. duPont Hospital for Children, P.O. Box 269, Wilmington, DE 19899, USA. Tel.: +1 302 651 5888; fax: +1 302 651 5990. E-mail addresses: [email protected] (L. Berman), [email protected] (J. Jackson), [email protected] (K. Miller), [email protected] (R. Kowalski), [email protected] (P. Kolm), [email protected] (F.I. Luks).

prenatal consultation for certain congenital anomalies leading to increased anxiety or unrealistic expectations. In some cases, decisions regarding termination of pregnancy may be based on erroneous information [2,3]. Congenital diaphragmatic hernia (CDH) and congenital pulmonary airway formation (CPAM) are two diagnoses for which prenatal consultation can have an important impact on decision making, ranging from prenatal intervention to location and mode of delivery [4,5]. Technical advances in prenatal diagnosis have led to increased accuracy and the ability to detect anomalies earlier in pregnancy, and many of these conditions are surgically correctable [6]. Pediatric surgeons, along with maternal fetal medicine specialists and neonatologists, play an essential role in interpreting these findings, conducting prenatal consultation, and participating in perinatal decision making [7]. This can lead to important decisions, ranging from where and how the baby will be delivered to considering an in utero intervention or terminating the pregnancy [1,8]. Although there are some data to drive this decisionmaking process, there is lack of clear consensus in the literature on

https://doi.org/10.1016/j.jpedsurg.2017.11.056 0022-3468/© 2017 Elsevier Inc. All rights reserved.

Please cite this article as: Berman L, et al, Expert surgical consensus for prenatal counseling using the Delphi method, J Pediatr Surg (2017), https:// doi.org/10.1016/j.jpedsurg.2017.11.056

2

L. Berman et al. / Journal of Pediatric Surgery xxx (2017) xxx–xxx

many aspects of these decisions. Care of these patients is therefore often guided by expert opinion and experience. It is critical that pediatric surgeons, upon completion of their training, be able to provide evidencebased information in a way that will allow future parents to make decisions consistent with their values, and set up realistic expectations as to how the anomaly will impact the life of their child. There are several obstacles to accomplishing this important goal. First, there are limited long-term outcome data on children with congenital anomalies; therefore, much of what may be shared is the anecdotal experience of the surgeon. Second, a recent survey of pediatric surgery fellowship graduates showed that 47% of pediatric surgeons felt underprepared to perform prenatal consults, citing that there was inadequate exposure to prenatal consultation during fellowship (with 54% of them participating in five or fewer prenatal consultations during their training), and a lack of clear resources to guide prenatal decision making and counseling [9]. Eliciting feedback from pediatric surgical experts is integral to defining best practice and standardizing fetal counseling content and intervention. The Delphi method is a validated scientific approach to solving a complex problem through expert consensus [10,11]. It builds consensus by submitting a series of questionnaires to a panel of identified experts, permitting involvement of geographically distant and informed individuals [12,13]. Anonymity is inherent to the method, designed to offset conventional means of pooling opinions and avoid the introduction of bias through a particular vocal or dominant individual or group of individuals. In contrast to other data-gathering and analysis techniques, the Delphi method employs multiple iterations in the feedback process, allowing and encouraging the selected experts to reassess initial judgments about the information provided in previous iterations. Thus, in a Delphi study, the results of previous iterations regarding specific statements and/or items can change or be modified by individual panel members in later versions based on their ability to review and assess the comments and feedback provided by the other Delphi panelists. A Delphi method expert guides the scientific process of defining expert consensus. The aim of this study was to create and facilitate a model of online Delphi consensus survey on perinatal management, using CDH and CPAM as examples. We identified areas of discrepancy and controversy in the literature and assembled a panel of experts to complete a Delphi consensus survey to develop consensus-based recommendations regarding terminology, diagnostic work-up, delivery plans, postnatal management, and content that should be discussed during prenatal consultation for CDH and CPAM. 1. Methods 1.1. Study population Members of the American Pediatric Surgical Association (APSA) Fetal Diagnosis and Therapy Committee, the North American Fetal Therapy Network (NAFTNet) Steering Committee and Executive Board, the Society for Maternal and Fetal Medicine, and the American Academy of Pediatrics (AAP) Committee on the Fetus and Newborn were contacted via email. Each of these individuals was asked to name other individuals who they would consider experts in prenatal consultation. Only pediatric surgeons (total of 78 surgeons) were included in the final expert list. We chose to limit participation to pediatric surgeons since our goal in conducting this survey was to define content to be discussed during pediatric surgical consultation. 1.2. Survey development and administration The current literature on the perinatal surgical care of CPAM and CDH was reviewed. Areas of discussion regarding prenatal imaging practice, interpretation of imaging findings, indication for fetal intervention, and critical points of prenatal counseling, were identified. A

questionnaire was created to address these questions with the help of an expert in prenatal consultation and an expert in the Delphi method. Questions included binary (yes/no) responses followed by scaled measures to assess a participant's strength of agreement. Select multiplechoice questions evaluated optimal time frames for diagnosis and treatment; cutoffs for test results; and optimal choice of diagnostic tests, intervention, and treatment (Table 1). The survey was then pilot-tested by two additional pediatric surgeons who were experts in prenatal consultation. The survey was administered online using Research Electronic Data Capture (REDCap), a secure web application for building and managing online surveys and databases [14]. An invitation email was sent followed by a series of weekly reminder emails. After the initial round of the survey, items that achieved consensus were summarized and included in the second round, allowing respondents to agree or disagree. Items that did not reach consensus in the first round were broken down into additional questions based on comments from first-round respondents. Final consensus was defined as 80% agreement for binary responses and greater than 50% agreement for multiple-choice questions. After round two, it was determined that consensus had been reached, or that two answers to a question had equal weight (divided expert opinion), and no additional rounds were administered. Open-ended questions were included to explore experts' opinions when they did not agree with any of the answer choices offered. All research procedures were approved by the institutional review board.

1.3. Data analysis The statistics used in Delphi studies are measures of central tendency (means, median, and mode) and level of dispersion (standard deviation and inter-quartile range) to present information concerning the collective judgments of respondents [15]. Consensus on a topic was determined based on the percentage of responses within a prescribed range and by measuring the stability of subjects' responses in successive iterations. There were certain items for which we did not reach the 80% threshold or N 50% agreement, but dissenters voiced their reasons in the open-ended response section included for each survey item, and these views are represented in the final summary of findings. For those items that lacked consensus, multivariable logistic regression (consensus—yes/no) was used to assess whether there were demographic characteristics that were associated with lack of consensus. Demographic characteristics of the respondents included age, gender, race, Table 1 Delphi survey content. • What imaging should be obtained prenatally? How often? • How should prenatal imaging be used to stratify severity? • When should fetuses be referred to fetal treatment center? Where should fetuses be delivered? Scheduled vs spontaneous? How does this vary based on stratification of severity? • What should be discussed during prenatal consultation in terms of long-term outcomes? • Should minimally invasive surgery (MIS) repair be considered? CPAM • What is the proper terminology for discussing congenital lung lesions? • What is the natural history in terms of growth, plateau, and regression prenatally? • What imaging should be obtained prenatally? How often? • How should severity be stratified? • What is the threshold for referral to a fetal treatment center? For maternal steroid administration? For fetal intervention for solid and cystic lesions? • Where and when should the mother deliver? Scheduled vs spontaneous? How does this vary based on stratification of severity? • Should observation without plan for resection be offered in asymptomatic patients? • Should segmentectomy rather than lobectomy be performed if technically feasible? CDH

CDH, congenital diaphragmatic hernia; CPAM, congenital pulmonary airway formation.

Please cite this article as: Berman L, et al, Expert surgical consensus for prenatal counseling using the Delphi method, J Pediatr Surg (2017), https:// doi.org/10.1016/j.jpedsurg.2017.11.056

L. Berman et al. / Journal of Pediatric Surgery xxx (2017) xxx–xxx

year training completed, and consultation volume. A p value b 0.05 for any demographic covariate was considered statistically significant.

Table 2 Participant characteristics.

2. Results A total of 54 out of 78 surgeons responded to either round one (38, 49% response rate) or round two (40, 51% response rate) of the survey, for a 69% overall response rate (Fig. 1). After two rounds of the survey, consensus was reached on 53 out of 60 survey items (88%). For the remaining seven survey items, there was a 50/50 split in surgeon opinion, and recommendations based on these responses were developed after review of open-ended comments. Demographic characteristics, year of pediatric surgery fellowship completion, and practice experience are presented in Table 2. The average number of prenatal consultations performed per year among all respondents was eight for CDH (range 0–25) and nine for CPAM (range 3–30). 2.1. Congenital pulmonary airway formation Survey questions addressed seven domains of CPAM: terminology, natural history, prenatal imaging (MRI), severity stratification, prenatal management, delivery plan, and postnatal management. We asked a total of 80 questions related to CPAM, using branching logic to display follow-up questions based on participant answers. For example, if a participant responded “yes” to the question, “Should presence of hydrops be used to stratify severity?” the survey presented a follow-up question, “How strongly do you believe this measure should be used to stratify

3

Participants

Round 1

Round 2

Overall

Total participants, n (% of total) Age (years), average, range Race, n (%) Asian Black White More than 1 race Unknown Gender, n (%) Female Male Year pediatric surgery fellowship completed (median, range) Practice, n (%) Academic Private practice Private/academic mix Unknown CPAM consultations Practice average/year, range Individual average/year, range CDH consultations Practice average/year, range Individual average/year, range

39 (50) 50, 37–72

40 (51) 49, 37–72

54 (69) 50, 37–72

6 (15) 1 (2.5) 31 (80) 1 (2.5) 0 (0)

4 (10) 0 (0) 32 (80) 1 (2.5) 3 (7.5)

7 (13) 1 (2) 43 (79) 1 (2) 2 (4)

8 (20) 31 (80) 2001, 1972–2013

10 (25) 30 (75) 2002, 1978–2015

11 (20) 43 (80) 2001, 1972–2015

33 (85) 1 (3) 5 (12) 0 (0)

33 (82.5) 1 (2.5) 4 (10) 2 (5)

44 (81) 1 (2) 6 (11) 3 (6)

19, 6–40 9, 3–30

21, 5–100 9, 3–30

20, 5–100 9, 3–30

16, 7–40 8, 0–25

17, 5–60 8, 0–25

16, 5–60 8, 0–25

CDH, congenital diaphragmatic hernia; CPAM, congenital pulmonary airway formation.

The Delphi Method Round 1 survey (89 questions) distributed to 78 surgeons

Surgeon

Surgeon

Surgeon

Surgeon

Surgeon

Surgeon

Surgeon

Surgeon

39/78 (50%) participants responded. Consensus was reached on 54/89 (61%) questions.

Surgeon

Surgeon

Surgeon

45 new questions developed to address questions that did not obtain a consensus. 15 questions developed to reassess and confirm areas of consensus.

Round 2 survey (60 questions) distributed to 78 surgeons

Surgeon

Surgeon

Surgeon

Surgeon

Surgeon

Surgeon

Surgeon

Surgeon

Surgeon

Surgeon

Surgeon

40/78 (51%) participants responded. Consensus was reached on 53/60 (88%) questions. Fig. 1. Delphi method for two survey rounds.

Please cite this article as: Berman L, et al, Expert surgical consensus for prenatal counseling using the Delphi method, J Pediatr Surg (2017), https:// doi.org/10.1016/j.jpedsurg.2017.11.056

4

L. Berman et al. / Journal of Pediatric Surgery xxx (2017) xxx–xxx

risk?” These questions addressed a total of 57 items (27 binary questions), and consensus was reached on 34 items. To determine consensus on the 23 remaining items, an additional 21 survey questions were developed. In the second round survey, there were a total of 44 questions (30 binary questions), and consensus was reached on 28 of the 30 binary questions (16 reached 90% or greater consensus and 12 reached 80% or greater). Open-ended responses were used to define consensus, or conclude that there were multiple reasonable options for the remaining questions. For example, 60% of respondents felt that scheduled delivery was indicated for high risk CPAMs, while 40% did not. Those who did not explained in open-ended responses that it was important for fetuses with high-risk CPAM to deliver as close to term as possible. Those who did favor scheduled delivery argued that outcomes are best if the multidisciplinary team is assembled and ready to care for the infant, and that this is more challenging for an unscheduled delivery. We summarized these results by stating that both options are acceptable depending on local resources. Similarly, 60% of respondents would consider a segmental resection as opposed to lobectomy for resection of CPAM, but 40% explained that they think segmental resection is rarely, if ever, technically feasible. We therefore summarized the results by saying that segmental resection is an option if technically feasible at time of surgery. The following sections (2.1.1–2.1.7) summarize consensus on CPAM as established by 54 experts in the online Delphi consensus survey. 2.1.1. Cystic adenomatoid malformation general terminology and definitions Experts were asked to establish the definition of CPAM and clarify use of terminology. The term “CPAM” encompasses congenital cystic adenomatoid malformation (CCAM) and bronchopulmonary sequestration. Bronchogenic cysts, or foregut cysts, should not be considered CPAMs. They can, however, be mistaken for CPAMs on prenatal imaging and should be included in the differential diagnosis in some cases. Sequestration is suggested by presence of a feeding vessel, but this is not always a reliable finding. Regardless of the exact etiology, CPAM is a disease spectrum, and characterizing these lesions precisely is not particularly helpful, especially not before birth. Bronchial atresia and congenital lobar emphysema should also be considered to be somewhere on this spectrum. 2.1.2. Cystic adenomatoid malformation natural history The natural history of CPAM can be quite variable. Experts came to consensus regarding the typical gestational age for growth (18–28 weeks), plateau (26–30 weeks), and regression or decrease in size (after 30 weeks). Lesions may regress completely. Ultrasound should be performed every 2 weeks during growth and plateau and every 4 weeks once the lesion starts to regress (frequency may vary slightly depending on severity of the lesion and potential need for steroids or intervention). 2.1.3. Cystic adenomatoid malformation prenatal imaging Experts were asked about imaging techniques, frequency, and optimal gestational age. Magnetic resonance imaging is not necessary for every CPAM but can be helpful in establishing diagnosis (for example, to differentiate from CDH) or to identify a feeding vessel. Ideal timing for MRI depends on the indication: at the time of referral (usually 18 to 20 weeks) if termination is being considered, 22 to 26 weeks if prenatal intervention is being considered, or after 26 weeks (after growth phase) if information is needed to predict postnatal outcome and to plan postnatal management. 2.1.4. Cystic adenomatoid malformation severity stratification Cyst–volume ratio (CVR), presence of hydrops, mediastinal shift, reversal of flow within the umbilical vein, and abnormal cardiac echocardiography should be used to stratify severity of CPAM. Traditionally, maternal–fetal medicine specialists define hydrops as the abnormal

presence of fluid in two or more body compartments. Experts identified the following findings on prenatal imaging for the definition of hydrops in a fetus with CPAM: skin edema, pericardial effusion, ascites, and pleural effusion. Inversion of diaphragm, polyhydramnios, and mediastinal shift should not count toward diagnosis of hydrops. 2.1.5. Prenatal management and intervention for cystic adenomatoid malformation Experts were asked about indications for referral to a fetal treatment center, administration of maternal steroids, and fetal intervention. Indications for referral to a fetal treatment center include the following: hydrops, CVR greater than 1.6, CVR 1.0 to 1.5 during growth or plateau phase, mediastinal shift during growth or plateau phase, and inversion of diaphragm during growth or plateau phase. Some surgeons felt that all fetuses with CPAM should be referred, or that patients should only be referred after a trial of maternal steroids. Indications for maternal steroid administration include the following: hydrops, CVR greater than 1.6, and CVR 1.0 to 1.5 during growth or plateau phase. In an open-ended response, one respondent contributed, “Steroids are not useful in macrocystic lesions, [or] if the hydrops is caused by bronchial atresia with lobar or lung hypoplasia […].” Hydrops is an indication for intervention for cystic lesions. Both cyst aspiration and thoracoamniotic shunting are reasonable options for intervention. The threshold for intervention should be higher for solid lesions, as it is rare to have a fetus with a large solid lesion that does not get better after maternal steroids. Fetal surgery for CPAM is rarely indicated or advisable. In an open-ended response, one participant contributed, “If the lesion is a CCAM that is true. However, we have seen occasional cases where solid (microcystic) lesions do not respond to multiple courses of steroids. [Furthermore], lobar bronchial atresias can cause hydrops and do not respond to steroids. Finally, there is the occasional sequestration that can cause hydrops and will not respond to steroids.” 2.1.6. Delivery plan for cystic adenomatoid malformation Low/moderate-risk CPAMs should not have scheduled delivery. Scheduled delivery at term should be considered for high-risk CPAMs (hydrops, CVR greater than 1.6), but cesarean section is not required for nonobstetric reasons. Experts felt that there was benefit in allowing patients to deliver as close to term as possible. The following important caveat was stated: “If delivery is not scheduled, then true readiness is important–and few centers truly are ready 24/7.” Because of the complexity of decisions regarding delivery plans, participants were presented with different case scenarios. Case scenario 1: A 35-week gestational age fetus with large a CPAM, CVR of 5.0, mediastinal shift, inverted diaphragm, strong concerns about pulmonary hypoplasia, no response to maternal steroid administration. What do you do? Scheduled delivery with an operating room standby was the most favored option. If necessary, resection in the delivery room with ex utero intrapartum (EXIT) to thoracotomy may be considered. The role of the EXIT procedure is controversial, as illustrated in the following open-ended responses: • “I do not think there is a role for EXIT.” • “The EXIT procedure can be very useful for CPAMs as large as the one described. When you have a solid lesion with minimal or no visualization of lung parenchyma an EXIT should be performed. The degree of pulmonary hypoplasia cannot be predicted in these patients. ECMO [is likely necessary] but is not a certainty.” • “[I] would have had prenatal discussions with mother regarding the option of comfort care of an extremely distressed neonate in the delivery room, and provided the option of consultation with palliative care.” Case scenario: A fetus with a small CPAM that is no longer visible on ultrasound at 32 weeks. Mother would prefer to deliver at her local

Please cite this article as: Berman L, et al, Expert surgical consensus for prenatal counseling using the Delphi method, J Pediatr Surg (2017), https:// doi.org/10.1016/j.jpedsurg.2017.11.056

L. Berman et al. / Journal of Pediatric Surgery xxx (2017) xxx–xxx

community hospital, which has a level 3 neonatal intensive care unit (NICU) but no immediate access to pediatric surgery. Would you be comfortable with this plan? There was consensus that delivery of this fetus in a community hospital is acceptable. 2.1.7. Postnatal management of infants with asymptomatic cystic adenomatoid malformation There was consensus that elective resection is the best option. Observation may be considered for asymptomatic CPAMs, but risks of infection, malignancy, and pneumothorax should be discussed. If the option of observation is chosen, the imaging modality and frequency of follow-up vary depending on the lesion: • “In general after perinatal [chest radiograph (CXR)], we obtain a [computed tomography (CT)] of the chest at 4-6 months to guide further counseling, For observed lesions, we are trying to move to follow-up MRI scanning after the first postnatal CT.” • “Would do one CT, then follow by annual CXR for a few years.” Most experts strongly recommend resection even if they do discuss the option of observation with families. They voiced concern that surgery is more difficult after a lesion becomes infected because of scarring and adhesions, and argued that the morbidity of long-term radiation exposure and possible malignancy makes resection the best option: • “Frequency and type of radiographic follow-up isn't well defined. The plan for observation requires some radiographic monitoring, which often entails additional radiation exposure.” • “There is no good imaging modality to monitor asymptomatic CPAMs. [Primary pulmonary blastoma] cannot be reliably distinguished from CPAM by imaging. Asymptomatic CPAMs can be thoracoscopically resected without significant morbidity and a two-day hospitalization. Risk/reward [ratio] favors removing the lesion.” If an operation is performed, it should be done during the first year of life. If it is technically feasible to perform a segmental resection rather than lobectomy for a patient with an intralobar sequestration, segmental resection should be considered. For a CCAM in the same scenario, it may be more advisable to perform a lobectomy. Many participants did not think it technically feasible to perform a segmental resection for a CCAM, and were concerned about margin status. 2.2. Congenital diaphragmatic hernia Survey questions addressed six domains of CDH: prenatal imaging (MRI), severity stratification, delivery plan, long-term outcome consultation topics, long-term outcome incidence rates, and minimally invasive repair. We asked a total of 46 questions related to CDH, again using branching logic to display follow-up questions specific to a certain response. These questions addressed a total of 32 items (14 binary questions), and consensus was reached on 20 items. We developed 24 additional questions to determine consensus on the remaining 12 items, and included a total of 33 questions in round two of the survey. Consensus was reached on 19 of the 30 binary survey questions (90% consensus on 10 items and 80% consensus on nine additional items). For the questions that did not reach consensus, consensus was determined from review of the dissenter's comments. There was lack of consensus on whether echocardiographic parameters should be used to stratify the severity of CDH, with 45% of respondents in favor and 55% against using these parameters. Open-ended comments showed that the only important echocardiographic parameter was the presence of a congenital cardiac anomaly, which is believed to worsen prognosis. In the follow-up questionnaire, there was consensus that other echocardiographic parameters, such as cardiac vector and pulmonary artery diameter, were not important. As with CPAM, there was a lack of consensus regarding scheduled delivery for CDH, but we reached a

5

similar conclusion through the open-ended comments that spontaneous delivery is acceptable as long as a multidisciplinary team can be assembled to care for the baby. The following sections (2.2.1–2.2.5) summarize consensus on CDH as established by 54 experts in the online Delphi consensus survey. 2.2.1. Prenatal imaging for fetuses with congenital diaphragmatic hernia Experts were asked questions regarding prenatal evaluation (ultrasound, MRI, and genetic testing) in CDH. An MRI may be helpful in the prenatal evaluation of CDH if the diagnosis is in question. Optimal gestational age for MRI is around 24 to 28 weeks, but later studies may yield more accurate results. It should be done early if confirmation of the diagnosis may lead to a decision to terminate the pregnancy. Ultrasound is helpful and should be performed every 2 to 4 weeks. Genetic testing, including karyotype and microarray, should be performed. 2.2.2. Metrics for predicting severity of congenital diaphragmatic hernia Lung-to-head ratio (LHR), percent predicted lung volume (PPLV) (or observed:expected (O/E) total lung volume [TLV]), and liver position are helpful in stratifying severity of CDH. Severe cardiac anomalies associated with CDH worsen the overall prognosis. Cardiac echocardiographic parameters that are suggestive of pulmonary hypertension, such as pulmonary artery diameter, may have a role in predicting severity. Most experts stated that right-sided diaphragmatic hernia had the same or worse prognosis as a left-sided defect with similar parameters, with the exception of liver position (the liver is almost always up with a right-sided defect). Respondents shared the following: “Literature is not definitive on the assumption that right CDH has higher risk of adverse outcome (adjusted for usual risk predictors) than Left.” and “Right CDH is just less predictable to analyze prenatally. MRI lung volumes are most helpful, but infants often do better than expected. The amount of liver herniation is a factor, too.” 2.2.3. Delivery plan for congenital diaphragmatic hernia All CDH should be considered high-risk, regardless of prenatal imaging parameters. Scheduled delivery at term should be considered. Many, but not all, specialists believe that scheduled delivery is necessary, and that early delivery should be balanced with the benefit of optimal lung maturity. Similar to high-risk CPAM, true readiness is important if delivery is not scheduled. Mandatory cesarean section is not indicated for fetuses with CDH. Delivery should occur in a level 3 NICU with access to pediatric surgery and extracorporeal membrane oxygenation (ECMO). Because of the complexity of decision-making regarding delivery plans, experts were presented with different case scenarios. Case scenario: The mother of a fetus with CDH would like to deliver in a community hospital with a level 3 NICU, within transfer distance of a pediatric ECMO center. Would you feel comfortable with this plan in each of the following scenarios? (Experts were presented with mild CDH (LHR N 1.4, PPLV N40%, O/E TLV N 50%, O/E LHR N 35%), moderate CDH (LHR 1–1.4, PPLV 20%–30%, O/E TLV 30%–50%, O/E LHR 25%– 35%), and severe CDH (LHR b 1.0, PPLV b 20%, O/E TLV b 30%, O/E LHR b 25%). For all three levels of CDH severity, experts thought it preferable for delivery to occur at an ECMO center. For a mild CDH, it may be appropriate to make an exception. Respondents shared the following: • “Pulmonary hypertension, i.e. the reason for needing ECMO does not completely correlate with prenatal prognostic parameters. Any postnatal transport of a CDH baby potentially needing ECMO is not acceptable if prenatal diagnosis was made.” • “Centers without ECMO should be able to demonstrate ability to care for CDH patients. I do not believe that because a center has ECMO, they are more capable of caring for CDH babies, particularly when these centers have ECMO rates as high as 30% (indicating over-utilization of a very high-risk therapy).”

Please cite this article as: Berman L, et al, Expert surgical consensus for prenatal counseling using the Delphi method, J Pediatr Surg (2017), https:// doi.org/10.1016/j.jpedsurg.2017.11.056

6

L. Berman et al. / Journal of Pediatric Surgery xxx (2017) xxx–xxx

• “In patients with possibility of severe pulmonary hypertension, transportation can exacerbate the hypertension. Ideally, the patient should be kept calm with minimal stimulation and there are too many things that can go wrong with transfer.” • “Would strongly prefer that delivery occur at facility with neonatology support familiar with the care of CDH. But the reality is that the mother is the patient.”

2.2.4. Long-term outcome discussion topics for congenital diaphragmatic hernia The following long-term outcomes should definitely be discussed during prenatal consultation for a fetus with CDH: recurrence of hernia, death, long-term oxygen supplementation, ventilator dependency, neurodevelopmental issues, and gastroesophageal reflux disease (GERD). One should consider discussing the following long-term outcomes: asthma, reactive airway disease, hearing loss, gastrostomy tube dependence, failure to thrive, chest wall deformities, and intestinal obstruction. For each outcome, participants were asked what complication/incidence rates they report, if any, during prenatal consultations. These questions were asked for CDH with and without the need for ECMO (Table 3). In open-ended responses, experts shared the following: “We do not counsel patients based on ECMO or non-ECMO. We counsel based on prenatal estimates of severity,” and “I like to discuss the issues but do not use percentages. I rather explain that these issues are more common than for the general population and some are rare but others more likely. If they ask for numbers I provide approximates to the best of my knowledge. I do discuss non-ECMO v. ECMO and higher risks of all complications and prolonged stay with ECMO.”

2.2.5. Option of minimally invasive repair of congenital diaphragmatic hernia The majority of experts (78%) felt that minimally invasive repair of CDH should be discussed if the baby was an appropriate candidate for this approach, but many raised concerns about increased perioperative morbidity and recurrence rate. Participants shared the following: “I do not recommend [minimally invasive repair] for any baby with any degree of pulmonary hypertension. Risk of inducing need for ECMO, higher rate of recurrence, does not reduce hospital stay,” and “Current evidence clearly shows a higher recurrence rate with minimally invasive repair. Choosing the repair that has higher recurrence rates just because of cosmesis and improved pain scores does not seem appropriate.”

2.3. Associations between participant characteristics and opinions The items lacking consensus for CPAM included natural history, ideal gestational age for MRI to establish diagnosis, and ideal gestational age for MRI to identify a potential feeding vessel. Items not reaching consensus for CDH included scheduled delivery, prenatal consultation topics (other long-term pulmonary issues, long-term hearing loss, gastroesophageal tube dependence, failure to thrive, chest wall deformities, intestinal obstruction), and minimally invasive repair. There were no statistically significant associations between demographic characteristics and likelihood of agreeing with consensus. 3. Discussion A Delphi survey of experts in prenatal consultation and management of CDH and CPAM led to consensus on topics that are often questioned or unresolved in the literature. The Delphi method is often used in medicine and surgery to address complex problems that cannot be answered from direct literature review. This method was used to define quality indicators for colorectal surgery patients [16] and ovarian cancer patients [17], and in pediatrics to create guidelines for prioritization of supportive care topics in childhood cancer [18]. The results of this Delphi survey of experts in fetal management and consultation fill in important gaps where there is controversy in the literature. The following three main controversies will be further developed: risk stratification for CPAM, the option of observation for asymptomatic CPAMs, and the prenatal evaluation and delivery planning for infants with CDH. Congenital pulmonary airway malformation encompasses a spectrum of developmental abnormalities, which includes CCAM, bronchopulmonary sequestration, and congenital lobar emphysema [19]. These diagnoses may exist on a spectrum rather than falling into a discrete category. Congenital pulmonary airway malformation severity stratification is extremely important in the prenatal period, as it has implications for prenatal management and intervention, as well as delivery planning. There are numerous factors that are taken into account to determine severity including CVR, presence of hydrops or mediastinal shift, reversal of flow within the umbilical vein, and an abnormal cardiac echocardiogram. Cyst volume ratio is a measurement used to estimate the size of the lesion by multiplying cyst length × height × weight × 0.52 divided by head circumference. The size of the fetal lung mass has well-known correlation with outcome; however, the appropriate CVR cutoff to use in parent counseling and decision-making around prenatal intervention and delivery planning is not clear from the literature. Crombleholme et al. [20] in a study reporting initial use of the CVR, found that a cutoff of CVR greater than 1.6 was significant for determining the risk of development of hydrops, and thus correlated

Table 3 Complication rates for CDH requiring ECMO vs not requiring ECMO. Outcome

Rate for ECMO CDH (%)

Rate for non-ECMO CDH (%)

Prolonged ventilator dependence (N1 month) Long term supplemental oxygen need (N6 months) Asthma or reactive airway disease as a school-aged child Neurologic impairment: learning difficulties Neurologic impairment: unable to attend regular school Long-term hearing loss: mild loss without need for medical assistance Long-term hearing loss: requiring hearing aid Long-term hearing loss: hearing loss impacting speech development GERD: managed with medication GERD: managed with fundoplication Failure to thrive resulting in need for gastrostomy placement Intestinal obstruction Chest wall deformities: mild Chest wall deformities: requiring surgery Scoliosis CDH recurrence after repair Death

25–75 25–75 25–75 50 10–20 Up To 50 5–10 5–10 75 20 Up to 50 5–10 10–20 Up to 10 Up to 20 10–50 20–50

Up to 10 Up to 10 10–50 5–20 Up to 5 Up to 5 1 1 Up to 75 Up to 10 Up to 10 5–10 10–20 Up to 5 Up to 20 5–20 5–20

CDH, congenital diaphragmatic hernia; CPAM, congenital pulmonary airway formation; ECMO, extracorporeal membrane oxygenation; GERD, gastroesophageal reflux disease.

Please cite this article as: Berman L, et al, Expert surgical consensus for prenatal counseling using the Delphi method, J Pediatr Surg (2017), https:// doi.org/10.1016/j.jpedsurg.2017.11.056

L. Berman et al. / Journal of Pediatric Surgery xxx (2017) xxx–xxx

with a higher severity and need for fetal intervention. They reported that those with a CVR less than or equal to 1.6, and without a dominant cyst, had a less than 3% chance of developing hydrops. Of the fetuses with CVR greater than 1.6, 75% developed hydrops. Cass et al.[4] showed that a CVR threshold of 2.0 was most significant, with only 2% of fetuses with CVR less than 2.0 having hydrops, and an overall survival rate of 98%. Kunisaki et al. [21] reported that those with an initial CVR greater than 1.6 have a perinatal survival rate of 63%, compared with 100% survival for the four fetuses with smaller lesions. It is difficult to ascertain from review of the literature how CVR should be used to stratify severity. Analysis of our Delphi survey concluded that a CVR of greater than 1.6 should be used as the cutoff for stratification into a high-risk category. Therefore, a CVR of greater than 1.6 qualifies the patient for referral to a fetal treatment center, maternal steroid administration, and scheduled delivery. Other consensus-based indications for referral to a fetal treatment center included presence of hydrops, mediastinal shift during growth or plateau phase, and inversion of the diaphragm during growth or plateau phase. The management of asymptomatic patients with CPAM remains somewhat controversial with respect to ideal timing of the operation and whether an operation should be done at all. Some surgeons argue that delaying surgery until symptom onset may have higher morbidity, but the incidence of symptom development as a whole has not been well studied, and without a well-understood risk of development of such symptoms, patients may unnecessarily be subjected to surgical risks [22–28]. Kapralik et al. conducted a meta-analysis of studies comparing elective surgery versus observation of CPAMs. In their series, two thirds of patients in the observation cohort developed symptoms and therefore required surgery. These symptomatic patients were more than four times as likely to experience postoperative complications. The authors concluded that early resection is safe and prevents risks associated with symptom development, and that surgery after development of symptoms may result in worse outcomes [29]. Stanton et al. [30] also compared elective surgical resection with conservative management. They found that elective surgery patients had fewer complications; however, only 3.2% of patients who were managed conservatively developed symptoms. According to the APSA Outcomes Committee review on treatment of CPAMs, there is no literature-based consensus on the question of observation or elective resection for asymptomatic CPAM. There is also no consensus regarding how to properly monitor asymptomatic lesions [31]. Our Delphi survey results achieved consensus that elective resection during the first year of life is the best option for asymptomatic CPAM. Congenital diaphragmatic hernia is a serious and potentially lifethreatening condition, with survival rates of about 50%. Several studies have attempted to identify whether there is an association between delivery planning (location and mode of delivery) and outcome in CDH. Nasr and associates [32], in a retrospective cohort study, compared mortality for “inborn” and “outborn” babies with CDH. Inborn births occurred in a hospital with a NICU or connected to a hospital with a NICU. Outborn births were those that required transfer by ambulance or flight to a hospital with a NICU. They found that outborn birth was a significant independent predictor of mortality, and thus advocated for delivery of fetuses with prenatally diagnosed CDH in hospitals with access to a NICU. Safavi et al. [33] aimed to identify the delivery mode associated with the most favorable outcomes. In comparing vaginal delivery versus cesarean section, they found that the delivery route did not influence outcome and therefore concluded that there was no one specific delivery plan or route that was superior for fetuses with prenatally diagnosed CDH. Frenckner et al. [34], in a retrospective cohort study, also aimed to compare the outcome of CDH infants by delivery method. They also showed no significant differences in survival for cesarean section, induced vaginal delivery, or spontaneous vaginal delivery. They concluded that there is no clear evidence that mode of delivery affects the

7

outcome, but that there is little doubt that the fetus should be delivered at a center with pediatric surgery and access to a NICU [34]. None of these studies assessed prenatal risk stratification of the CDH to guide the decision-making around delivery planning. Our Delphi survey achieved consensus that all CDH fetuses and deliveries should be considered high risk, regardless of prenatal imaging parameters. Therefore, scheduled delivery at full-term should be considered. Many, but certainly not all, specialists believe that scheduled delivery is necessary, but should be balanced with the benefit of better optimization of lung maturity. Mandatory cesarean section was not found to be indicated for fetuses with CDH, but there was a consensus reached that delivery should occur in a level 3 NICU with access to pediatric surgery and ECMO. It is important to note that there were some experts who felt that for a mild CDH (LHR N 1.4, PPLV N40%, O/E TLV N 50%, O/E LHR N 35%), it might not be necessary to deliver in a tertiary center. This study has several limitations that are important to acknowledge. The Delphi technique has a strong track record as a legitimate and valid research technique [35,36]. Methodological guidance for conducting electronic Delphi technique in health education is limited, however, with best-practice implementation strategies yet to emerge. Traditional Delphi technique limitations exist including design sensitivity, recruitment, time commitments, attrition, and identifying consensus measurements. More specifically, since many of the experts we surveyed practice in high-volume centers, there is likely to be some bias toward recommending referral of patients to these centers. We attempted to address this bias by analyzing associations between center volume and likelihood of agreement with consensus statements, and found no significant associations. The findings of our Delphi survey have provided resolution to several areas of controversy in the literature regarding the prenatal counseling and perinatal care of neonates with CPAM and CDH. Our findings represent a body of knowledge that can be incorporated into fellowship training curriculums and educational materials for those performing prenatal consultation. There are several quaternary care centers throughout the country that specialize in fetal consultation and provide team-based, protocoldriven care; however, a significant proportion of the United States population does not have ready access to one of these centers [37]. Most prenatal consultations for surgically-correctable anomalies are performed by surgeons practicing outside of a major regional center. We have developed a method to render expert opinion for congenital anomalies, gathered using a validated, scientific process. This expert opinion can be used to define and ultimately standardize the content that should be discussed during prenatal consultation, and promote decision-making that will optimize outcomes for these patients. Furthermore, we have demonstrated successful use of the Delphi method to achieve expert consensus where evidence is lacking, and we believe that this approach should be applied to other controversial topics in pediatric surgery to facilitate creation of consensus-based guidelines.

References [1] Aite L, Trucchi A, Nahom A, et al. Multidisciplinary management of fetal surgical anomalies: the impact on maternal anxiety. Eur J Pediatr Surg 2002;12:90–4. [2] Aite L, Zaccara A, Trucchi A, et al. Is counselling for CCAM that difficult? Learning from parental experience. J Prenat Med 2011;5:65–8. [3] Aite L, Zaccara A, Trucchi A, et al. When uncertainty generates more anxiety than severity: the prenatal experience with cystic adenomatoid malformation of the lung. J Perinat Med 2009;37:539–42. https://doi.org/10.1515/JPM.2009.098. [4] Cass DL, Olutoye OO, Cassady CI, et al. Prenatal diagnosis and outcome of fetal lung masses. J Pediatr Surg 2011;46:292–8. https://doi.org/10.1016/j.jpedsurg.2010.11. 004. [5] Lazar DA, Cass DL, Rodriguez MA, et al. Impact of prenatal evaluation and protocolbased perinatal management on congenital diaphragmatic hernia outcomes. J Pediatr Surg 2011;46:808–13. https://doi.org/10.1016/j.jpedsurg.2011.02.009. [6] Pooh RK, Kurjak A. 3D/4D sonography moved prenatal diagnosis of fetal anomalies from the second to the first trimester of pregnancy. J Matern Fetal Neonatal Med 2012;25:433–55. https://doi.org/10.3109/14767058.2011.636107.

Please cite this article as: Berman L, et al, Expert surgical consensus for prenatal counseling using the Delphi method, J Pediatr Surg (2017), https:// doi.org/10.1016/j.jpedsurg.2017.11.056

8

L. Berman et al. / Journal of Pediatric Surgery xxx (2017) xxx–xxx

[7] Brown SD, Ecker JL, Ward JR, et al. Prenatally diagnosed fetal conditions in the age of fetal care: does who counsels matter? Am J Obstet Gynecol 2012;206:409.e401–11. https://doi.org/10.1016/j.ajog.2012.01.026. [8] Crombleholme TM, D'Alton M, Cendron M, et al. Prenatal diagnosis and the pediatric surgeon: the impact of prenatal consultation on perinatal management. J Pediatr Surg 1996;31(1):156–62 [discussion 162–53]. [9] Berman L, Kabre R, Kazak A, et al. Exposure to prenatal consultation during pediatric surgery residency: implications for training. J Pediatr Surg 2016;51:131–6. https:// doi.org/10.1016/j.jpedsurg.2015.10.030. [10] Brook RH. The RAND/UCLA appropriateness method. Agency for Healthcare Policy and Research. Rockville, MD: Conference proceedings; 1995. [11] Sinha IP, Smyth RL, Williamson PR. Using the Delphi technique to determine which outcomes to measure in clinical trials: recommendations for the future based on a systematic review of existing studies. PLoS Med 2011;8:e1000393. https://doi.org/ 10.1371/journal.pmed.1000393. [12] Young SJ, Jamieson LM. Delivery methodology of the Delphi: a comparison of two approaches. J Park Recreation Admin 2001;19:42–58. [13] Dalkey NC, Helmer O. An experimental application of the Delphi method to the use of experts. Manag Sci 1963;9:458–67. [14] Harris PA, Taylor R, Thielke R, et al. Research electronic data capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform 2009;42:377–81. https://doi.org/10. 1016/j.jbi.2008.08.010. [15] Linestone HA, Turoff M. The Delphi method: techniques and applications. Reading, MA: Addison-Wesley Pub Co, Inc; 1975. [16] McGory ML, Shekelle PG, Ko CY. Development of quality indicators for patients undergoing colorectal cancer surgery. J Natl Cancer Inst 2006;98:1623–33. [17] Gagliardi AR, Fung MF, Langer B, et al. Development of ovarian cancer surgery quality indicators using a modified Delphi approach. Gynecol Oncol 2005;97:446–56. [18] Loeffen EA, Mulder RL, Kremer LC, et al. Development of clinical practice guidelines for supportive care in childhood cancer—prioritization of topics using a Delphi approach. Support Care Cancer 2015;23:1987–95. https://doi.org/10.1007/s00520014-2559-7. [19] Martin RJ, Fanaroff AA, Walsh MC. Neonatal respiratory disorders. Fanaroff and Martin's neonatal–perinatal medicine: diseases of the fetus and infant. Saunders, Philadelphia: Elsevier; 2015. p. 1113–36. [20] Crombleholme TM, Coleman B, Hedrick H, et al. Cystic adenomatoid malformation volume ratio predicts outcome in prenatally diagnosed cystic adenomatoid malformation of the lung. J Pediatr Surg 2002;37:331–8. [21] Kunisaki SM, Barnewolt CE, Estroff JA, et al. Large fetal congenital cystic adenomatoid malformations: growth trends and patient survival. J Pediatr Surg 2007;42:404–10.

[22] Aspirot A, Puligandla PS, Bouchard S, et al. A contemporary evaluation of surgical outcome in neonates and infants undergoing lung resection. J Pediatr Surg 2008; 43:508–12. [23] Conforti A, Aloi I, Trucchi A, et al. Asymptomatic congenital cystic adenomatoid malformation of the lung: is it time to operate? J Thorac Cardiovasc Surg 2009;138: 826–30. https://doi.org/10.1016/j.jtcvs.2009.01.014. [24] dell'Agnola C, Tadini B, Mosca F, et al. Advantages of prenatal diagnosis and early surgery for congenital cystic disease of the lung. J Perinat Med 1996;24:621–31. [25] Marshall KW, Blane CE, Teitelbaum DH, et al. Congenital cystic adenomatoid malformation: impact of prenatal diagnosis and changing strategies in the treatment of the asymptomatic patient. AJR Am J Roentgenol 2000;175:1551–4. [26] Sueyoshi R, Okazaki T, Urushihara N, et al. Managing prenatally diagnosed asymptomatic congenital cystic adenomatoid malformation. Pediatr Surg Int 2008;24:1111–5. [27] Laje P, Liechty KW. Postnatal management and outcome of prenatally diagnosed lung lesions. Prenat Diagn 2008;28:612–8. https://doi.org/10.1002/pd.1966. [28] Wong A, Vieten D, Singh S. Long-term outcome of asymptomatic patients with congenital cystic adenomatoid malformation. Pediatr Surg Int 2009;25:479–85. https:// doi.org/10.1007/s00383-009-2371-5. [29] Kapralik J, Wayne C, Chan E, Nasr A. Surgical versus conservative management of congenital pulmonary airway malformation in children: A systematic review and meta-analysis. Journal of pediatric surgery 2016;51(3):508–12. [30] Stanton M, Njere I, Ade-Ajayi N, et al. Systematic review and meta-analysis of the postnatal management of congenital cystic lung lesions. J Pediatr Surg 2009;44:1027–33. [31] Downard CD, Calkins CM, Williams RF, et al. Treatment of congenital pulmonary airway malformations: a systematic review from the APSA outcomes and evidence based practice committee. Pediatr Surg Int 2017;33(9):939–53. https://doi.org/10. 1007/s00383-017-4098-z. [32] Nasr A, Langer JC, Canadian Pediatric Surgery Network. Influence of location of delivery on outcome in neonates with congenital diaphragmatic hernia. J Pediatr Surg 2011;46:814–6. https://doi.org/10.1016/j.jpedsurg.2011.02.007. [33] Safavi A, Lin Y, Skarsgard ED. Perinatal management of congenital diaphragmatic hernia: when and how should babies be delivered? Results from the Canadian Pediatric Surgery Network. J Pediatr Surg 2010;45:2334–9. https://doi.org/10.1016/j. jpedsurg.2010.08.026. [34] Frenckner BP, Lally PA, Hintz SR, et al. Prenatal diagnosis of congential diaphragmatic hernia: how should the babies be delivered? J Pediatr Surg 2007;42:1533–8. [35] de Meyrick J. The Delphi method and health research. Health Educ 2003;103:7–16. [36] Landeta J. Current validity of the Delphi method in social sciences. Technol Forecast Soc Change 2006;73:467–82. [37] Brantley MD, Davis NL, Goodman DA, et al. Perinatal regionalization: a geospatial view of perinatal critical care, United States, 2010–2013. Am J Obstet Gynecol 2017;216:185.e1–185.e10. https://doi.org/10.1016/j.ajog.2016.10.011.

Please cite this article as: Berman L, et al, Expert surgical consensus for prenatal counseling using the Delphi method, J Pediatr Surg (2017), https:// doi.org/10.1016/j.jpedsurg.2017.11.056