OBJECTIVES: The ethics of maternal-fetal surgery involves weighing the importance of potential benefits, risks, and other consequences involving the pregnant woman, fetus, and other family members. We assessed clinicians’ ratings of the importance of 9 considerations relevant to maternal-fetal surgery.
METHODS: This study was a discrete choice experiment contained within a 2015 national mail-based survey of 1200 neonatologists, pediatric surgeons, and maternal-fetal medicine physicians, with latent class analysis subsequently used to identify groups of physicians with similar ratings.
RESULTS: Of 1176 eligible participants, 660 (56%) completed the discrete choice experiment. The highest-ranked consideration was of neonatal benefits, which was followed by consideration of the risk of maternal complications. By using latent class analysis, we identified 4 attitudinal groups with similar patterns of prioritization: “fetocentric” (n = 232), risk-sensitive (n = 197), maternal autonomy (n = 167), and family impact and social support (n = 64). Neonatologists were more likely to be in the fetocentric group, whereas surgeons were more likely to be in the risk-sensitive group, and maternal-fetal medicine physicians made up the largest percentage of the family impact and social support group.
CONCLUSIONS: Physicians vary in how they weigh the importance of social and ethical considerations regarding maternal-fetal surgery. Understanding these differences may help prevent or mitigate disagreements or tensions that may arise in the management of these patients.
- IQR —
- interquartile range
- MFM —
- maternal-fetal medicine specialist
What’s Known on This Subject:
Decision-making for fetal surgery represents a complex calculus whereby one must weigh multiple potential benefits, risks, and other consequences involving the pregnant woman, fetus, and family. Questions remain about how clinicians vary in their prioritization of these considerations.
What This Study Adds:
Physicians vary in how they weigh the importance of social and ethical considerations regarding maternal-fetal surgery; in this study, we provide insight regarding the potential disagreements or tensions that may arise in the care of these patients.
The ethics of maternal-fetal surgery combines the typical considerations encountered in the care of individual patients with the added complexity of potentially having to balance heightened risk of morbidity for the pregnant woman against the likelihood of benefit for the future child.1–5 How clinicians think about and manage this calculus is not well characterized.
Compared with most pediatric or maternal interventions, maternal-fetal surgery poses additional or heightened collateral trade-offs all in an effort to benefit the fetus. Many women who elect to undergo open maternal-fetal surgery relocate closer to a fetal care center and voluntarily incur significant inconveniences associated with bed rest and the attendant stresses on partners, other children, and other caregivers that these choices bring.6–10 Although common in pediatrics and other areas of specialty medical care, little is known about how clinicians weigh these collateral effects relative to the direct effects of maternal-fetal surgery on the pregnant woman and fetus.
In this study, we sought to assess the relative importance of various social and ethical considerations regarding maternal-fetal surgery among maternal-fetal medicine specialists (MFMs), neonatologists, and pediatric surgery specialists. We hypothesized that we would be able to identify distinct groups of physicians whose ratings in response to randomly assigned pairings of hypothetical social and ethical considerations in maternal-fetal surgery would cluster attitudinally in a manner that would explain a substantial proportion of the variability in importance judgments among the clinicians, beyond that attributable to specialty-based differences.
Human Subjects Protections
This study was reviewed and deemed exempt by the Mayo Clinic Institutional Review Board.
Study Participants and Data Collection
In the spring of 2015, we mailed a confidential, self-administered, 11-page, 32-item paper questionnaire titled “Emerging Issues in Maternal-Fetal Surgery” to a random sample of 1200 practicing US physicians. The sample included 400 MFMs, 400 neonatologists, and 400 pediatric surgeons. These physicians were randomly selected from member lists of the Society for Maternal-Fetal Medicine, the American Academy of Pediatrics Section on Perinatal Pediatrics, and the American Academy of Pediatrics Section on Surgery, respectively. In accordance with the tailored design method,11 up to 3 separate mailings were sent; the first mailing included a $20 bill.
Survey Development: Social and Ethical Considerations
After reviewing the literature, consulting with content experts, and conducting physician interviews of MFMs, surgeons, and neonatologists at 5 maternal-fetal centers, we developed a set of social and ethical considerations that we expected to inform physician assessment of the potential risks and benefits of maternal-fetal surgery. We then pilot-tested the considerations with MFMs, surgeons, and neonatologists and revised the considerations on the basis of feedback from cognitive interviews. The final 9 considerations used in the current study were as follows: (1) neonatal benefits: the child benefits from the operation; (2) maternal autonomy: the pregnant woman’s right to make decisions; (3) risk of prematurity: increased chance that the child will be delivered before term; (4) risk of fetal death: increased chance the fetus dies or the infant dies soon after birth; (5) risk of maternal complications: increased chance the woman’s health suffers; (6) risk to future reproductive health, such as the risk of uterine rupture with future pregnancies necessitating cesarean deliveries; (7) maternal social support: ensuring that the pregnant woman has physical and emotional support; (8) impact on other family members: the impact on fathers, siblings, or grandparents; and (9) maternal psychological benefit: increased maternal satisfaction or self-esteem.
Discrete Choice Experiment
We used a discrete choice experiment to enable physicians to select the social and ethical considerations that they deemed to be the most and least important for women contemplating prenatal surgery. In contrast to Likert-type rating scales, the discrete choice experiment creates a series of balanced subsets of the considerations, requires respondents to compare the considerations in a given subset and choose the most important consideration, and then combines the choices across all the subsets to produce a rank ordering and rating of all considerations (9, in this study) that is psychometrically superior and avoids ceiling effects.12,13 This discrete choice experiment methodology has been used in other experiments; in one experiment, for example, it was used to elucidate what attributes parents perceive to be most important when caring for a seriously ill child.14
In the present experiment, respondents were taken through a series of subsets of the 9 social and ethical considerations. Respondents were instructed to answer each set on the basis of their initial reaction. There were a total of 8 unique sets that each included 4 different considerations, designed so that each consideration was shown a relatively equal number of times and in a balanced set of combinations and permutations with the other considerations. Respondents were instructed to select the 1 least important and 1 most important consideration in each unique set. This resulted in a total of 16 choices. On the basis of these choices, the respondent’s rank ordering of the considerations was derived. We then used maximum difference scaling to calculate a point value for each consideration. The scaling was set so that the total number of points for all considerations would equal 100, and the values were assigned to reflect the relative importance of a specific consideration in relation to all of the others.15 Therefore, the scores are interpretable relative to each other as a ratio (for example, a consideration with a score of 50 is twice as important as a consideration with a score of 25).
Other Study Measures
We also examined physician demographic characteristics (age, sex, region, specialty, and years in practice) and practice setting types, including affiliation with a fetal care center.
Data Management and Analysis
Responses were double-entered and imported into Stata version 13.1 (StataCorp, College Station, TX). We used the American Association for Public Opinion Research RR2 response rate definition.16 Maximum difference scaling and latent class analysis were performed by using Sawtooth software (Light House Studio, Version 9.0.1, Orem, UT).
We determined the physician ratings and rankings of the 9 social and ethical considerations by using MaxDiff (Light House Studio, Orem, UT), which applies multinomial logistic regression to estimate the probability of choosing each consideration (as most important or least important) given the considerations shown in the set, transforming raw scores to a 0-to-100 scale, wherein the total of all transformed scores of all considerations equals 100. This results in a relative scaling of importance, as described above.
We then performed latent class analysis of the transformed consideration scores to identify an unobserved categorical latent class variable. We examined solutions with 2 to 6 distinct classes and replicated each latent class solution 5 times, beginning at random starting values. We considered the best latent class solution to be that with both conceptual meaning and the best fit (as indicated by the lowest Bayesian information criterion and adjusted Bayesian information criterion).17–19 The Bayesian information criterion improved only slightly from groups 2 to 6. We selected the 4-group model for ease of interpretability. We assigned physicians to the latent class group for which they had the highest probability of membership.
Descriptive statistics for physician characteristics and practice settings are presented for the full sample as well as by latent class. Descriptive statistics for categorical variables are reported as frequency and percentage, and continuous variables are reported as mean and SD. Categorical variables were compared between latent classes by using the χ2 test. Continuous variables were compared between latent classes by using one-way analysis of variance. Statistical inferences were based on 2-tailed tests with significance set at P < .05.
Of the 1200 potential respondents, 24 (2%) could not be contacted. Of the remaining 1176 eligible participants, 670 returned surveys, yielding a response rate of 57%. Six hundred and sixty respondents completed the discrete choice experiment (effective response rate: 56%). Effective response rates did not differ significantly by specialty (MFMs: 54%; neonatologists: 57%; pediatric surgeons: 60%; P = .2). The characteristics of respondents are shown in Table 1.
Mean Rankings and Ratings of Social and Ethical Considerations
Among the 9 social and ethical considerations included in the discrete choice experiment (Fig 1), participants ranked neonatal benefits (mean rating [SD]: 24.8 [5.3]) and risk of maternal complications (mean rating [SD]: 23.3 [5.9]) as the 2 most important considerations. These were followed by the risk of fetal death (mean rating [SD]: 17.2 [9.3]) and maternal autonomy (mean rating [SD]: 13.6 [10.5]). In contrast, participants rated maternal psychological benefit (mean rating [SD]: 1.5 [3.0]), maternal social support (mean rating [SD]: 0.8 [2.1]), and impact on other family members (mean rating [SD]: 0.6 [2.6]) as the least important considerations. These later considerations were much less likely to be chosen as the most important, compared with neonatal benefits (mean ratings of 1.5, 0.8, or 0.6 vs 24.8).
Individual Variation of Rankings and Ratings of Social and Ethical Considerations
In addition to assessing the average relative importance of these 9 considerations across all respondents, we also examined individual variations in ratings. There was a broad distribution for many of the considerations (Fig 2). For the 2 most important considerations, neonatal benefits and risk of maternal complications, the interquartile range (IQR) of ratings was relatively tight despite outliers (median: 25.6, IQR: 23.8–27.5 and median: 24.9, IQR: 21.7–26.6, respectively). This was also true for the 3 least important considerations, which included maternal psychological benefit, maternal social support, and impact on other family members (median: 0.4, IQR: 0.2–1.4; median: 0.1, IQR: 0.05–0.5; and median: 0.01, IQR: 0.003–0.09, respectively). In contrast, the 4 middle considerations (risk of fetal death, maternal autonomy, risk of prematurity, and risk to future reproductive health) had the largest distribution. For example, risk of fetal death received a median rating of 21.1, yet had an IQR of ratings from 8.7 to 24.8. Maternal autonomy received a median rating of 14.8, yet had an IQR of ratings from 2.5 to 22.9. And risk to future reproductive health received a median rating of 5.0, yet had an IQR of ratings from 1.8 to 13.3.
Latent Class Groupings
We identified 4 physician groups on the basis of the latent class analysis described above. on the basis of the underlying content of the clustered item groupings, we labeled the 4 groups as follows: "fetocentric” (n = 232), risk-sensitive (n = 197), maternal autonomy (n = 167), and family impact and social support (n = 64). For each group, the ranked prioritization (Fig 3) was used to define and differentiate each group and can also be viewed in a complementary manner, in regard to how each group rated the relative importance of each consideration (Fig 4). All 4 groups gave high priority to neonatal benefits and risk of maternal complications (these considerations were in the top 3 for all 4 groups); however, the other rankings differed substantially among groups. The first group, which we designated as risk-sensitive, gave prioritization to both risks for the fetus and for the pregnant woman. In contrast, the fetocentric group prioritized the risk of fetal death and the risk of prematurity more than any other group; however, this group did not prioritize risk to future reproductive health. In addition, this group gave the lowest rating for risk of maternal complications (mean 19.6 vs 23.4–24.4, respectively). The maternal autonomy group rated autonomy significantly higher than the other groups (mean: 23 vs 6.7–16.2) and also gave some consideration to maternal psychological benefit (mean: 3.6). Finally, the family impact and social support group also gave some consideration to maternal psychological benefit (mean: 3.7) but was the only group to give any consideration to impact on other family members (mean: 5.1) or maternal social support (mean: 3.3).
We examined whether the 4 groups were associated with respondent characteristics (Table 1). Surgeons were more likely to be in the risk-sensitive group, whereas neonatologists were more likely to be in the fetocentric group, and MFMs made up the largest percentage of the family impact and social support group.
This study expands our understanding of how physicians weigh the importance of potential risks and benefits of maternal-fetal surgery. Their rating of 9 unique social and ethical considerations in the discrete choice experimental design described here revealed both clear consensus among priority considerations and diversity among several secondary considerations. Neonatal benefits and the risk of maternal complications were the clear dominant considerations. Almost half of the importance points were given to these 2 considerations (combined mean rating: 48.1). In contrast, 4 distinct physician groups also emerged for secondary considerations, including fetocentric, risk-sensitive, maternal autonomy, and family impact and social support considerations. Combined, these data reveal a nuanced portrait of the multiple dimensions of physicians’ considerations of the risks and benefits of maternal-fetal surgery.
Although we did not find significant associations with respondents’ group membership and their demographic characteristics or practice settings, respondents’ specialty was significantly associated. Pediatric surgeons were more likely to be in the risk-sensitive group (43% of pediatric surgeons). Surgeons’ risk aversion may be explained in part by an omission bias, a natural human tendency to favor potentially harmful omissions (not operating) over equally or less potentially harmful commissions (performing an operation).20,21 Physician risk tolerance has been shown to predict the use of health care resources,22–24 and in the case of maternal-fetal surgery, may influence which operations are offered and to whom. Pediatric surgeons were also least likely to be in the family impact and social support group (6% of pediatric surgeons). Neonatologists were more likely to be in the fetocentric group (44% of neonatologists), possibly reflecting the fact that their work is most impacted by the effects of surgery on the fetus. The MFMs were more evenly spread among the 4 groups, compared with their pediatric subspecialty colleagues, with the largest group of MFMs belonging to the fetocentric group (37% of MFMs) and the smallest group belonging to the impact on family and social support group (13% of MFMs).
Specialists of each type were present in all latent class groups, tempering our finding of specialty-based associations, with no single specialty constituting a majority within any of the latent class groups. Physicians within these latent classes likely evaluated the risks and benefits of the surgery differently. These differences may account in part for practice variation regarding the management of pregnant women and fetuses.25,26 Furthermore, understanding how physicians weigh these considerations differently can, if disagreements or tensions arise in the care of these patients, help to focus discussions about maternal-fetal surgery and patient management.
This study has 5 principal limitations to consider. First, the 9 considerations that we used in the discrete choice experiment include the most widely discussed considerations in the literature, but there may be other unmeasured considerations that we and others have not adequately considered. Second, physicians’ ratings of considerations are likely to vary somewhat, depending on the specific condition and procedure under consideration. For example, a physician may prioritize different considerations when contemplating a potentially fatal diagnosis (such as fetal hydrops secondary to a sacrococcygeal teratoma) versus a nonlethal diagnosis (such as spina bifida).27 Different considerations may also be assigned a different level of priority when a physician is contemplating the invasiveness of a specific procedure (such as fetoscopic surgery versus open uterine surgery). Third, the latent class procedure could have yielded a different number of groups; we selected the 4 groups on the basis of statistical fit for the 4-group solution and for conceptual clarity. Fourth, although the survey instrument was pilot-tested with members of each specialty, the discrete choice experiment methodology has not been used extensively in examining medical decision-making, and some respondents may not have understood the wording or the discrete choice task. Furthermore, respondents’ choices on the survey may not reflect how they would weight different considerations in clinical practice. Finally, even with a robust response rate, the answers of nonrespondents may have differed from those who did respond.
Although specialists associated with maternal-fetal surgery largely agree that neonatal benefits and the risk of material complications are among the leading considerations when deciding whether to perform surgery, significant diversity exists regarding how physicians weigh the 7 other considerations regarding the risks and benefits presented by maternal-fetal surgery. This diversity may be related to each physician’s values, personal experience, specialty, and interpretation of the literature. Understanding that physicians vary in how they weigh the importance of these considerations may help clarify and manage differences of opinion regarding how to best care for these patients.
- Accepted September 19, 2017.
- Address correspondence to Chris Feudtner, MD, PhD, MPH, The Children's Hospital of Philadelphia, 3535 Market St, Philadelphia, PA 191044. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Funded by a grant from the Greenwall Foundation.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Smajdor A
- Johnson MP,
- Bennett KA,
- Rand L, et al; Management of Myelomeningocele Study Investigators
- Verkerk MA,
- Lindemann H,
- McLaughlin J, et al
- Lindemann H
- Dillman D,
- Smyth J,
- Christian L
- Cohen S
- Cohen S,
- Orme B
- Louviere J,
- Woodworth G
- American Association for Public Opinion Research
- Brown SD,
- Feudtner C,
- Truog RD
- Copyright © 2017 by the American Academy of Pediatrics