Article Text

Download PDFPDF

Obstetrician-gynaecologists' opinions about conscientious refusal of a request for abortion: results from a national vignette experiment
  1. Kenneth A Rasinski1,
  2. John D Yoon1,
  3. Youssef G Kalad2,
  4. Farr A Curlin3
  1. 1Department of Medicine, The University of Chicago, Chicago, Illinois, USA
  2. 2Undergraduate Division, The University of Chicago, Chicago, Illinois, USA
  3. 3Department of Medicine, MacLean Center for Clinical Medical Ethics, The University of Chicago, Chicago, Illinois, USA
  1. Correspondence to Kenneth A Rasinski, Research Associate, Department of Medicine, The University of Chicago, 5841 South Maryland Avenue, Chicago, IL 60637, USA; krasinsk{at}


Background and objectives Conscientious refusal of abortion has been discussed widely by medical ethicists but little information on practitioners' opinions exists. The American College of Obstetricians and Gynecologists (ACOG) issued recommendations about conscientious refusal. We used a vignette experiment to examine obstetrician-gynecologists' (OB/GYN) support for the recommendations.

Design A national survey of OB/GYN physicians contained a vignette experiment in which an OB/GYN doctor refused a requested elective abortion. The vignette varied two issues recently addressed by the ACOG ethics committee—whether the doctor referred and whether the doctor disclosed their objection to the abortion.

Participants and setting 1800 OB/GYN randomly selected physicians were asked to complete a mail survey containing the vignette. The response rate was 66% (n=1154) after excluding 40 ineligible cases.

Measurement Physicians indicated their approval for the vignette doctor's decision.

Main results Overall, 43% of OB/GYN physicians responded that the conscientious refusal exercised by the vignette physician was appropriate. 70% rated the vignette doctor as acting appropriately when a referral was made. This dropped to 51% when the doctor disclosed objections to the patient, and to 12% when the doctor disclosed objections and refused to make a referral. Consistent with previous research, males were more likely to support disclosure and refusal to refer. Highly religious physicians supported non-referral but not disclosure.

Conclusion OB/GYN physicians are less likely to support conscientious refusal of abortion if physicians disclose their objections to patients. This is at odds with ACOG recommendations and with some models of the doctor–patient relationship.

  • Right to refuse treatment
  • moral and religious aspects
  • demographic surveys/attitudes
View Full Text

Statistics from


The 1973 Supreme Court case, Roe v Wade, affirmed a woman's legal right to seek elective abortion during her first trimester of pregnancy. Soon afterwards, Congress passed ‘conscience clause’ legislation establishing the legal right of medical care professionals to refuse to provide abortion and sterilisation.1–3 The concept of conscientious refusal has stirred controversy and debate within medicine—addressing issues of physician autonomy and rights, professional obligations, patient autonomy, patient wellbeing, and the doctor–patient relationship.2 4–7

In 2007 the American College of Obstetricians and Gynecologists (ACOG) addressed the issue of conscientious refusal in their Committee Opinion 385.8 The opinion states that ‘Physicians and other healthcare professionals have the duty to refer patients in a timely manner to other providers if they do not feel that they can in conscience provide the standard reproductive services that their patients request’ (p5) and ‘Where conscience implores physicians to deviate from standard practices, including abortion, sterilization, and provision of contraceptives, they must provide potential patients with accurate and prior notice of their personal moral commitments’ (p5).

Ethical opinions on the topic of conscientious refusal abound but there are few empirical studies of practicing physicians' opinions and practices with regard to the topic as it applies to specific reproductive health services. While a recent national survey of physicians found that 71% agreed with the general statement that physicians who object to a requested procedure have an obligation to refer the patient, and 63% agreed that it would be ethical for a physician to describe plainly to the patient the moral, ethical or religious reasons for their objection to a requested procedure, no specific procedure was mentioned in these questions.7

We designed an experiment, implemented in a national survey of obstetrician-gynecologists' (OB/GYN) physicians, which focused specifically on conscientious refusal in the case of abortion. Our goal was to understand the importance that OB/GYN physicians place on referral and disclosure. The experiment was designed to measure whether a respondent approved of a vignette OB/GYN doctor who refuses to perform an elective abortion. Two variations in the experiment related directly to the ACOG recommendations—whether the vignette doctor made a referral after exercising conscientious refusal and whether the doctor disclosed the reason for objection to abortion. We also manipulated the vignette doctor's sex and the type and source of objection. Moreover, we examined whether responses to the vignette differed by selected characteristics of physician respondents.


Sample design and data collection

From October 2008 to June 2009 we mailed a confidential, self-administered, 12-page questionnaire to a national sample of 1800 OB/GYN physicians 65 years of age or younger practicing in the USA. This sample was drawn from the American Medical Association's physician Master File—a database designed to include all practicing US physicians. In order to increase the number of respondents in non-Christian religious groups (important for other goals of our research but not addressed in this study), we stratified the sample by ethnic names (Arabic, South Asian, Jewish, other) using validated surname lists9–11 and oversampled within these groups. A case weight was constructed to compensate for the oversampling from ethnic name groups and for higher non-response rates among graduates of foreign medical schools and among the Arabic and South Asian ethnic name strata.12 The use of this weight in analyses adjusts the sample data to the US population of practicing OB/GYN physicians aged 65 and under.

Physicians received an advance letter and up to three separate mailings of the questionnaire. The first questionnaire mailing included a US$20 bill and the third offered an additional US$30 for participation. We received completed questionnaires from 1154 of the eligible 1760 physicians (40 were no longer practicing, had left the country or could not be located after mailing to two different addresses) yielding a 66% response rate. All data were double keyed, cross compared and corrected against the original questionnaires in order to minimise error from data entry. The study was approved by The University of Chicago Institutional Review Board.

Experimental design

One of the survey questions described a clinical scenario in which a 23-year-old single graduate student presented to her OB/GYN and was found to be 8 weeks pregnant. She requested an abortion but the doctor denied her request. Table 1 gives the exact wording of the vignette along with the variables that were experimentally manipulated to create the different vignette versions.

Table 1

Design of vignette experiment

Five manipulated vignette variables, each with two levels, yielded 32 unique vignette versions. The 32 versions were randomly distributed across sampled physicians. After reading the vignette the physician was asked: ‘In your judgement, how appropriate are the physician's actions in this case?’ To facilitate analysis we collapsed the four-level response category into two categories—those who answered ‘somewhat’ or ‘very’ inappropriate were coded as 0 and those who answered ‘somewhat’ or ‘very’ appropriate were coded as 1. This served as the outcome variable in all of the analyses. The experiment was analysed using the logistic regression procedure in the STATA V.11 statistical software with the survey design feature implemented to account for the effects of sample stratification and weighting.

Respondent physicians' sex and self-reported level of religiosity were also analysed to investigate their influence in moderating any differences in appropriateness judgements that were due to different versions of key experimental variables. Physician sex was available from the AMA Master file. Self-reported religiosity was measured by responses to the following survey question ‘How important would you say religion is in your own life?’ Response categories were ‘the most important’, ‘very important’, ‘fairly important’ and ‘not very important’. The first two categories were combined to define highly religious physicians and the latter two were combined to define physicians who place less importance on religion in their lives.


Demographic and background characteristics of the sample are presented in table 2. Physicians' sex and age are related such that male physicians (m=51) are significantly older than females (m=44) (t(1152)=13.04, p<01). Because of this, age was statistically controlled in analyses that examined physicians' sex differences. We analysed the experiment in two phases. In the first phase we examined only the effects of the scenario features that we experimentally manipulated. We focused primarily on two features—referral: either referring or not referring the patient to a provider who would perform the abortion, and disclosure: either telling or not telling about the objection. We expected that more OB/GYNS would show support for a doctor who made a referral than one who refused to make a referral. In addition, consistent with the ACOG recommendation, we expected to find greater OB/GYN support for the vignette doctor who disclosed his or her objection to abortion.

Table 2

Respondent characteristics: frequencies and weighted percents

Across all versions of the experiment, 43% of OB/GYN physicians responded that the conscientious refusal exercised by the vignette physician was very or somewhat appropriate. Support for conscientious refusal varied substantially for different versions of the referral and disclosure scenarios. Results are shown in table 3. The scenario doctor who referred after making a conscientious refusal was favoured by 56 percentage points over the one who did not (p<0.001). The scenario doctor who did not disclose their objection to abortion to the patient was favoured by 21 percentage points over the doctor who did disclose (p<0.001). In addition, within each disclosure condition referral was favoured over non-referral. However, referral received much more approval when the reason for the objection was not disclosed than when it was (significant interaction effect, p<0.001). In summary, table 3 shows that the physician who did not disclose his or her objections but referred the patient to an abortion provider was judged by this national sample of OB/GYN physicians to be acting more appropriately than the physician who disclosed and referred.

Table 3

Percentage of OB/GYN physicians approving the actions of the vignette physician, by referral and disclosure status

Interactions of OB/GYN individual differences with experimental factors

Next we examined the interaction of respondent physician sex and level of religiosity with key experimental variables. Based on the results of previous research,7 we expected that male OB/GYNS (compared to females) and those high in religiosity (compared to those low in religiosity) would be more supportive of the decision not to refer and more supportive of disclosing objections to the patient. As expected, OB/GYN sex and referral interacted significantly (p<0.01) such that a higher percentage of male OB/GYN physicians (18%) compared to females (9%, p<0.001) approved the vignette doctor's decision not to refer. Similarly, OB/GYN sex and disclosure interacted significantly (p<0.05) such that a higher percentage of male OB/GYN physicians (38%) compared to female physicians (26%, p<0.05) approved of the vignette doctor's decision to disclose the reason for objection.

We also examined the interaction between OB/GYN physicians' level of religiosity, on the one hand, and referral and disclosure, on the other. Religious physicians were more supportive of the vignette doctor's decision not to refer (21%) than were non-religious physicians (7%, p<0.001). Religious and non-religious physicians' support of the vignette doctor did not differ depending on disclosure.


We surveyed a national sample of OB/GYN physicians to assess their support for conscientious refusal to perform an elective abortion and to understand the extent to which different conditions of the refusal moderate that support. The vignette format promoted realism,13 14 and the use of a randomised experimental design allowed us to make independent tests of moderators of physicians' support. Overall, fewer than half of OB/GYN physicians supported conscientious refusal.

The findings suggest OB/GYNS resonate with some parts of the ACOG recommendations, but not with other parts. In support of the ACOG recommendation, 70% of the sample approved of the vignette doctor who, though refusing to perform the abortion based on ethical or moral objections, made a referral. This level of support compares favourably with the 71% agreement that Curlin and his associates found in support of the general statement that physicians who morally object to a procedure should be willing to make a referral.7 In contrast, only 33% of OB/GYNS approved of the vignette physician telling the patient the reasons for his or her objection to abortion. This is just over half of the support that Curlin et al7 found with the general question asking if it would be ethical for physicians to describe plainly to the patient the moral, ethical or religious reasons for their objection to a requested procedure. (Of note, the present study specified abortion whereas the former study did not specify any particular practice, and whereas the present study includes only OB/GYN physicians, the former study included physicians from all specialties.)

Only one in three OB/GYNS supported a physician disclosing his or her moral objections to abortion to a patient, and when the vignette physician explained why he or she would not provide an abortion, that led OB/GYNS to be less, not more, supportive of the refusal. These findings complicate the ACOG recommendation that the conscientiously refusing physician ‘…must provide potential patients with accurate and prior notice of their personal moral commitments.’ Perhaps many OB/GYNS do not believe they have the authority or the expertise to properly explain their objections to patients. Or OB/GYNS might support the sort of anticipatory disclosure that the ACOG opinion piece seems to envision more than they support disclosing one's objections to a patient who is actively seeking an abortion. To provide anticipatory disclosure, OB/GYNS could simply tell patients up front that they do not provide certain types of services without addressing morality or values. Davis, however, suggests that this sort of disclosure is likely to result in confusion and would not serve the patient well. He argues, ‘…doctors have a duty to reveal their moral reasons for refusal because, if they do not, patients may mistakenly conclude that there are medical reasons for the refusal’.4

OB/GYNS' attitudes about disclosure also suggest that they do not think Emanuel and Emanuel's ‘deliberative’ model of the doctor/patient relationship15 fits a scenario in which a patient requests an abortion. In the deliberative model, ‘The physician's objectives include suggesting why certain health-related values are more worthy and should be aspired to’. This model would ‘…shift the publicly assumed conception of patient autonomy that shapes both the physician's and the patient's expectations from patient control to moral development’. Given OB/GYNS' objections to moral disclosure regarding abortion, future research might explore how open physicians' are in other clinical contexts to the idea of seeking patients' moral development through the doctor–patient relationship.

Other scholars encourage physicians to disclose their reasons for objecting to abortion. Pellegrino16 argues that by explaining themselves to patients, physicians actually enhance patient autonomy. He argues

Respect for the physician's moral autonomy does not sanction proselytisation to promote the physician's values, which takes advantage of a patient's vulnerability. Nonetheless, in the case of abortion, for example, the pro-life physician should respectfully state her moral reasons for refusal. Moralistic condemnation or personal vilification of the woman who chooses an abortion is not justifiable. On the other hand, explaining one's serious moral objections to abortion or assisted suicide … enhances autonomy by providing the full spectrum of choices and risks pertinent in such a decision.

Davis makes a similar point4:

The doctor should communicate his objection in a way that does not diminish the patient's ability to exercise her own moral judgement. This does not mean that a doctor may not persuade a patient to revoke her request for the procedure, but, if he does he must do so by helping the patient see another point of view and not by overwhelming and effectively coercing the patient.

Our data do not make clear whether OB/GYNS believe that disclosing one's objections effectively coerces a patient or if they believe that disclosing objections is impermissible for other reasons.

The results are problematic for OB/GYNS who endorse conscientious refusal to the point of not providing a referral. Pellegrino states that:

Respect for the patient's autonomy does not include referral to a physician who will carry out the procedure if that procedure involves an act the physician deems intrinsically and seriously wrong. For a conscientious physician, this would be an inadmissible degree of formal cooperation. It would be an additional violation of his moral integrity to cooperate in this way.

However the two conditions that displayed physician autonomy—moral disclosure and non-referral—were least supported by OB/GYNS. Survey research on what physicians do or do not consider a breech of their autonomy may be quite useful to inform this component of the debates about conscientious refusal.

There are limitations to this research. As noted above, our experiment set up a situation that does not precisely mirror the ACOG committee opinion that, ‘Where conscience implores physicians to deviate from standard practices, including abortion, sterilization, and provision of contraceptives, they must provide potential [italics ours] patients with accurate and prior notice of their personal moral commitments’.8 That opinion would seem to require physicians to disclose to women those treatments the physician will not provide before the women become patients, or at least to discuss the limitations at an early visit. It may not be feasible for physicians to meet this standard with respect to every doctor/patient encounter.

In addition, while our response rate is excellent for national physician surveys there is no guarantee that those who did not respond would have done so in the same way as the responders. And, although clinical vignettes have proven useful in physician surveys our respondents were still reacting to a hypothetical situation and their responses may not be consistent with their behaviour in an actual encounter. Future studies should ask OB/GYNS to explain why they do not support physicians disclosing the reasons for refusing abortion. Studies should also examine whether physicians support disclosure more at some stages of the doctor–patient relationship than at others—for example, during a new patient visit versus after a patient has already requested an abortion. Nonetheless, this study combines the broad scope of a carefully conducted national survey, the realism of clinical vignettes, and the inferential capability of a randomised experiment. It therefore provides a powerful tool for assessing how physicians might respond to policy recommendations regarding conscientious refusals.17

To conclude, our results suggest that at this point in time the majority of OB/GYNS support physicians conscientiously refusing to perform an abortion, as long as the physician refers the patient to an accommodating provider. They are somewhat less supportive of conscientious refusal if the physician discloses her moral or ethical objections to the patient and are much less supportive of conscientious refusal if the physician refuses to refer.


View Abstract


  • Funding This project was funded by grants from the Greenwald Foundation and the Templeton Foundation both awarded to FAC.

  • Competing interests None.

  • Ethics approval This study was conducted with the approval of The University of Chicago.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.