In 2002, 105,4581 abortions were performed in Canada, with an abortion ratio of 32.1 per 100 live births.2 Despite the high prevalence of abortion, access is limited by geographic factors, political influences on funding, antiabortion harassment, and lack of providers. In the United States, studies have shown a decrease in the number of providers, partially due to factors such as retirement of older providers and insufficient training of new providers.3 In Canada, organizations have brought attention to similar concerns about a provider shortage.4 A minority of obstetrics and gynecology (ob-gyn) programs include routine abortion training in the United States,5–8 and there are no data about Canadian ob-gyn residency abortion training. A 1993 survey of Canadian residents focusing on their attitudes about abortion found that 66% of residents were willing to participate in abortion training, and 52% planned to provide abortion services after residency.
Canadian residency programs, like U.S. programs, follow the educational objectives for training set out by the Council on Residency Education in Obstetrics and Gynecology (CREOG). In the area of family planning, CREOG recommends that graduating residents understand the indications, contraindications, and principles of uterine evacuation for pregnancy termination, incomplete abortion, and fetal death. The Council further recommends that residents be able to describe pregnancy termination techniques, including suction curettage and dilation and evacuation (D&E).9
Given no data about abortion training and limited data about resident’s attitudes toward training in Canadian ob-gyn programs, we conducted a survey to assess quality of training in abortion care, resident participation in training, and intention to become abortion providers.
MATERIALS AND METHODS
In 2004, an anonymous bilingual questionnaire was sent by mail to all postgraduate year (PGY)-4 and PGY-5 ob-gyn residents (n = 130) and residency program directors (n = 16) in Canada. The questionnaires inquired about individual and program demographic information and details of abortion training. Information about personal religiosity was requested from residents by asking, “Do you consider yourself religious?” on a Likert scale based on a previous study.10 Residents were also surveyed regarding their participation in training, comfort with various methods, and intentions to provide abortions before and after residency.
The survey contained no personal identifying information except province, which we did not report because of potential loss of anonymity. A postcard identifying the recipient by code was sent with the first mailing, and each respondent was asked to mail it at the same time as the anonymous survey. After one month, a second questionnaire was mailed to those who had not returned the postcard. The resident questionnaire included a monetary compensation of 5 Canadian dollars. The results were entered by two investigators using Excel spreadsheets with rules and limits to minimize errors. All analyses were performed with STATA 8.0 (StataCorp, College Station, TX). Most analyses were univariable comparisons of proportions with χ2 tests. The multivariable analyses were logistic regression models and included variables that correlated with the outcome in univariable analyses at P < .1. The study was approved by the University of California, San Francisco, and Berkeley institutional review boards and funded by an anonymous foundation.
A total of 92 of 130 residents (71%) and 15 of 16 program directors (94%) returned the questionnaires. The majority of responding residents were women (75%) and Caucasian (74%), with a mean age of 32 years. Forty-three percent described themselves as religious, the most common religion being Catholicism (37%), 74% were married or had a long-term partner, and 47% had children.
All residency programs are university-based and without religious affiliation, and the majority (67%) train 4–6 residents per year. According to both residents and program directors, at least 97% of programs include abortion training, and the training is considered routine in approximately half of programs and elective in half (Table 1). Approximately one fourth of programs teach manual vacuum aspiration, and almost all teach electric vacuum aspiration and second-trimester induction terminations. Although all program directors reported training in D&E, only 67% of residents reported training. Similarly, 87% of directors and 67% of residents reported training in medical abortion. Most training takes place within the hospital. The majority of first-trimester training occurs in the first 2 years of residency, whereas second-trimester training varies by program. Twenty-seven percent of responding residents and 60% of directors reported that the department arranges training.
According to residents, the majority of program directors (72%) and department chairs (59%) are supportive of abortion training. Program directors reported that 14% of the primary faculty and 25% of the affiliated attendings perform first-trimester abortions, and 7% and 16%, respectively, perform second-trimester procedures. Programs with routine training are more likely to have a supportive program director (P = .001) or departmental chair (P = .03). Training at programs in which it is routine is more likely to occur within the hospital than in a clinic (P =.01) and to be arranged by the programs instead of the residents (P < .001).
According to the program directors, programs with routine training have higher rates of resident participation for both first-trimester (62% versus 12%, P = .02) and second-trimester training (78% versus 18%, P = .01) than programs with elective training. Three of the 15 program directors (20%) reported encountering resistance to abortion training; two reported residents as the source of resistance, and one reported multiple sources, including nursing, anesthesia, ob-gyn colleagues, residents, and medical students.
The majority of responding residents (71%) participated in abortion training (Table 2). More than half felt competent after training to perform first-trimester aspiration (73%) and second-trimester induction (70%). Thirty-nine percent of residents felt competent in first-trimester medical abortion and 26% in D&E. The average maximum gestation to which residents were trained to do abortions was 19 weeks (range 12–24 weeks), and 60% were trained to do only inductions beyond 17 weeks of gestation.
Before starting residency, 50% of residents planned to provide some type of elective abortion, and 50% planned to not provide or were undecided. Of the 46 residents who planned to provide abortions when beginning residency, 42 (91%) participated in abortion training, and conversely, of 46 who planned not to provide abortions when beginning residency, 24 (52%) participated.
After residency, 46% planned to provide elective, first-trimester surgical abortion, and 38% planned to provide first-trimester medical abortion (Table 3). Five respondents who did not plan to do abortions before starting residency stated that they planned to provide them after graduating. At the beginning of residency training, 33% of residents planned to provide elective second-trimester abortions. After residency training, 21% said they would perform elective second-trimester abortions, and 76% stated they would perform therapeutic second trimester abortions in the cases of fetal anomaly, previable preterm premature rupture of membranes (PPROM), or danger to the woman’s life. Of those who were planning to provide second-trimester abortions for any indication, 41% planned to use D&E up to a certain period of gestation (average: 16 weeks) and then induction for later gestational durations. Four percent would perform D&Es for all gestational ages, and about 45% would always use induction of labor in the second trimester.
Table 4 presents the unadjusted and adjusted odds ratios (ORs) of variables correlating with participation in training and intention to provide abortion. The independent correlates of participation in training were prior intention to provide abortions (OR 5.9, 95% confidence interval [CI] 1.5–22.5) and the program arranging training instead of the residents (OR 10.1, 95% CI 1.1–94.2). Therefore, even residents who were religious or who did not plan to provide abortions before residency were significantly more likely to participate if the program arranged training (P=.04). Intention to provide elective abortions after graduating correlated with the following variables in univariable analyses: female gender (P=.08), personal low religiosity (P≤.001), not planning to enter a subspecialty (P=.05), having planned to provide abortions when starting residency (P≤.001), having participated in training (P≤.001), and routine training status (P=.002). In multivariable analyses, the independent correlates of provision were intention to do abortions before residency (OR 216.8, 95% CI 16.2–2,905.8), low religiosity (OR 18.5, 95% CI 1.5–234.5), and routine status of training during residency (OR 26.4, 955 CI 1.7–405.0).
Our study had a high response rate (71% of residents and 94% of program directors), but due to the small number of ob-gyn programs in Canada (16), the sample size was small. The study design lends itself to selection bias, particularly of resident respondents, because those most motivated about these issues are more likely to return the survey. Because we do not have data about nonresponders, we cannot confirm or deny this possibility. These deficiencies may have led to biases about participation in training and intention to provide. However, the high (94%) response rate by program directors and the general agreement between residents and program directors lead us to believe that we collected accurate information about program characteristics. Further, the small sample size limited our ability to create multivariable models to identify independent predictors of plans to perform abortions or participation in training and may have led us to conclude no associations between some predictors and the outcomes. Still, we were able to identify a few personal and training variables that were significantly, independently associated with these outcomes despite the large confidence intervals.
Most Canadian ob-gyn departments provide abortion services to their patients in the first and second trimesters of pregnancy and offer abortion training to residents. However, in approximately half of programs, it is not integrated routinely into the curriculum, and in these programs competence relies on residents’ interest in training. The strongest correlate of routine training was a supportive program director, emphasizing that program directors may play an important role in integrating training into the curriculum.
In our study a higher proportion of residents (71%) reported participation in abortion training than in a 1993 survey (60%).11 Participation in training is important for reasons other than future abortion provision. For example, ob-gyns who have been exposed to abortions are more likely to include pregnancy counseling for women and to be accepting of abortion.12,13 We found that residents were more likely to participate in training if they perceived that the program arranged the training for them, even when controlling for personal religiosity and routine training status. The difference in proportions of residents (27%) and directors (60%) who described the training as arranged by the program may reflect a difference in perception, an overestimate by program directors, or an indication of response bias by residents who had less integrated training. Regardless, it seems that, for some residents, their decision to participate may be affected by departmental culture, integration of training, and their perception about training arrangement.
At the end of residency, the majority felt competent in first-trimester surgical abortions, but only 39% felt competent in first-trimester medical abortion (Table 2). Because medical abortion has been only recently introduced, it is likely that most programs are not able to train residents to competence. Most residents feel competent providing induction of labor for second-trimester pregnancy terminations. The number who feel competent performing D&Es (26%) is approximately half of the proportion who reported competence in a 1993 study of resident attitudes about abortion (42%),11 which is possibly due to a decrease in faculty expertise in D&E. Because D&E has been shown to be safer than induction,14 improving competence in D&E may increase safety for patients. Notably, for both medical abortion and D&E, the procedures with the lowest level of reported competency, we found a discrepancy between the perspectives of program directors and residents about training availability. This may reflect misunderstandings about training by either group or response bias of residents.
For second-trimester abortions, fewer residents planned to provide elective abortions after (21%) than before (33%) residency. However, the majority of residents planned to provide therapeutic abortions in the second trimester (76%) for previable PPROM, anomaly, or threat to the health of the woman. We did not ask whether they had planned to provide second-trimester therapeutic abortions at the beginning of their residencies and therefore do not know if this proportion increased or decreased. The determinant of willingness to provide second-trimester terminations seems to be personal—the indication for termination, rather than the level of training. It may be that many residents trained in second-trimester terminations are only exposed to women with therapeutic indications and thus are unfamiliar with patients with other reasons for seeking a later abortion. It also may reflect the beliefs of the general public about acceptable indications for abortion. Regardless, in terms of improving access to second-trimester abortion services for women, a majority of residents feel comfortable providing services for at least some indications.
Residents were more likely to intend to provide abortions after graduation if they were not religious or planned to do abortions before residency. In addition, even controlling for these personal variables and participation in training, they were more likely to provide if the training was considered routine in their program. Although our results suggest that personal beliefs and attitudes before residency are important predictors of abortion provision, our data also reflect the importance of providing adequate, integrated training to increase the odds of future provision.
In summary, abortion training in Canadian ob-gyn residency programs is adequate but could be more integrated and arranged by the programs to increase participation and competence, in particular in medical abortion and D&E. One option for achieving this improvement is to adopt a similar policy to that in the United States, in which residency programs are required to offer abortion training but allow residents to opt out and allow training to be off-site if necessary. Furthermore, the variables that correlated with participation in training and abortion provision are likely to be generalizable to countries with similar training structures and political climates, such as the United States. All ob-gyn programs should consider how their training might be improved to increase participation and postresidency abortion provision. Future studies are warranted to address, not only intention to provide abortion, but also the more important outcome of actual service provision.