|Home | About | Journals | Submit | Contact Us | Français|
The majority of women prescribed category D or X medications may not receive adequate contraceptive counseling or a reliable contraceptive method. Physicians who prescribe potentially teratogenic medications have a responsibility to provide women with contraceptive counseling, a method of highly-effective contraception, or both.
Investigate the knowledge, beliefs and barriers of primary care physicians toward providing adequate contraception to women taking potential teratogens.
Self-administered confidential survey of primary care internal medicine physicians at an urban, academic medical center.
Knowledge of potential teratogenic medications and contraceptive failure rates was assessed. Participants’ beliefs about adequacy of their medical education, practice limitations and attitudes toward improving provision of contraception to women on potential teratogens were assessed.
One hundred and ten physicians responded (57.3%). Nearly two-thirds (62.3%) of participants had cared for reproductive aged women taking category D or X medications in the past year. The mean percent of correctly identified category D or X medications was 58.4% (SD 22.1%). The mean percent correct for knowledge of published contraceptive failure rates was 64.6% (SD 23.1%). Most respondents (87.6%) felt it is the responsibility of primary care physicians to provide contraception. Time constraints were reported to be somewhat or very limiting by 61.3% and over half felt medical school (63.2%) or residency (58.1%) inadequately prepared them to prescribe or counsel about contraceptives.
Primary care physicians commonly encounter reproductive age women taking category D or X medications, but may lack sufficient knowledge and time to counsel about potential teratogens and contraception to prevent fetal exposure to these drugs.
The online version of this article (doi:10.1007/s11606-009-1215-2) contains supplementary material, which is available to authorized users.
Birth defects are a major cause of infant mortality and childhood morbidity.1,2 Ten percent of congenital anomalies have been associated with exposure to a confirmed teratogen, a chemical that can cause structural and/or functional birth defects.3 This could increase the risk of birth defects as much as three-fold.3 With nearly half of pregnancies in the US unplanned,4 family planning efforts for women taking potentially teratogenic medications must be vigilant. Studies of large administrative healthcare databases indicate that women taking potentially teratogenic medications, classified D or X by the US Food and Drug administration (Table 1), are no more likely to be using a highly-effective method of contraception than those who are taking medications that are lower risk of teratogenicity, category A, B, or C medications.5,6 Approximately 11.7 million prescriptions for category D or X drugs are written for women of reproductive age each year.5,6 It is estimated that 5.8% of US pregnancies may be exposed to category D and X medications.7,8
Pregnancy for women taking potentially teratogenic medication carries more risk from both fetal exposure to potential teratogens and the effect of pregnancy on the underlying medical condition requiring treatment with such medications. Physicians who prescribe potentially teratogenic medications have a responsibility to provide contraceptive counseling, a method of highly-effective contraception, or both to all women who will be receiving those medications.5–7,9 Authors of prior studies discuss the importance of physician behavior in ensuring that women of childbearing age are adequately counseled about the teratogenic potential of medications and that women are provided highly-effective methods of contraception.5–7,9 To our knowledge, no published research has studied physician behavior.
To explore physicians’ roles in ensuring adequate contraceptive protection among women for whom pregnancy represents the highest risk, we conducted a survey of the primary care physicians (PCPs) at a single, urban, academic medical center and its affiliated community practices. The goals of the survey were to investigate the barriers PCPs encounter in provision of contraception to women on category D or X medications and their knowledge of both contraceptive efficacies and classification of potential teratogens. We hoped to assess whether PCPs know which commonly prescribed medications are classified as D or X, measure PCPs knowledge of contraceptive effectiveness (failure rates), evaluate if PCPs believe their medical education regarding potential teratogens and contraception was adequate, and determine limitations to contraceptive counseling and prescribing within the practice of primary care medicine.
A confidential self-administered survey instrument (see appendix available online at www.jgim.org) was developed with experts in survey design and methodology. Specialists in the fields of contraception, pharmacology, teratology and primary care internal medicine assisted with the content and assessed the survey for face validity. It was then pretested among members of the internal medicine department at our institution, modified to reflect suggested changes from these reviewers, and then it was transferred to an electronic format using a web-based survey service (www.surveymonkey.com). The electronic and written versions of the survey, data collection and entry techniques were then pilot tested by having the authors and selected faculty of the gynecology department complete both written and web-based versions. Based on our pilot testing, the survey took less than ten minutes to complete, regardless of whether it was done on paper or via the web-based format.
To maximize response, the survey was provided in both written and electronic formats with matched unique identifiers to prevent duplication.10 An email invitation with an embedded html link to the web-based survey was sent to all potential participants with two subsequent reminder emails sent at two and five weeks to non-responders. The invitation included an endorsement from the chief of staff of the hospital and the chief of general internal medicine. A written version of the survey was mailed on the same day as the initial email notification inviting participants to use the web-based format. The paper mailing included a copy of the survey, a self-addressed stamped envelope and a $5.00 Starbucks™ gift card as a thank you for the subject’s participation. Corresponding with the email notifications, reminder postcards were sent via traditional mail two weeks later after the initial mailing. A second full copy of the survey was mailed to all non-responders five weeks after the first solicitation. The initial invitation to participate in the survey was sent in early November, 2008 with subsequent reminders through the middle of December, 2008. Any response received before the end of January, 2009 was included for analysis.
Information about the physician’s medical school and residency training and practice setting were collected. Participants were asked to categorize medications as potentially teratogenic—classified as D or X—versus those that are lower risk of teratogenicity—category A, B, or C. The list included 33 of the most commonly prescribed medications in the US for 200711 with 42% (14/33) classified as D or X.12 Medications listed as either category C or D depending on reference source or timing of exposure were considered potential teratogens. Contraceptive failure rates are published as either typical use or perfect use statistics.13 Though the survey asked participants to identify “typical failure rates” for birth control methods, an answer that was consistent with either typical or perfect use was accepted as correct. Questions rated on a four-point Likert-type scale assessed physicians’ beliefs about their medical education regarding prescribing and counseling about contraceptives and teratogenic medications. Lastly, participants were asked to rate limitations and potential interventions they felt would be most useful in improving their ability to regularly provide contraception to women taking potential teratogens. There were no instructions about whether respondents could use reference materials during completion of the survey.
All 194 PCPs associated with a large, academic hospital in Chicago, Illinois, were identified by the medical staff office of the hospital which provided contact, demographic and educational information about the physicians. Inclusion criteria for potential participants were that the individual is a licensed MD or DO who participates in the hospital referral and admitting network specializing in primary care internal medicine. Of note, all PCPs associated with the study institution specialize in internal medicine. The subjects self-selected for inclusion by choosing to participate and complete the survey. This study was approved by the institutional review board of Northwestern University with a waiver of informed consent.
All data obtained from written surveys were manually entered into the web-based format by a single research assistant and checked for entry errors. The survey data were downloaded from the web-based application and combined with the demographic data obtained from the medical staff office. Data management and analysis were completed using SAS statistical software (SAS version 9.2, Cary, NC). Questions regarding attitudes and beliefs with four-point scale Likert-type scales were dichotomized (strongly disagree/disagree and agree/strongly agree) as were those questions regarding practice limitations (not at all limiting/not too limiting and. somewhat limiting/very limiting). Physician respondents and nonrespondents were compared using the Student’s t-test for continuous data and chi-square test for categorical variables to determine potential response biases. Respondents who completed the survey by mail were compared with those who completed the web-based version. Descriptive statistics for all data are reported.
Of the 194 PCPs identified, two were excluded because they had participated extensively in the survey development and pretesting process leaving 192 eligible respondents. A total of 110 completed a survey (response rate of 57.3%). Characteristics of participants are listed in Table 2. Compared with nonresponders, physicians who responded were in practice for fewer years (10.9 vs. 13.5 years, p=0.04). The majority of respondents (66/110) chose paper over the web-based format, but younger respondents (41.5 vs. 45.6 years, p=0.02) and those who were in practice for less time (12.0 vs. 9.0 years, p=0.04) were significantly more likely to respond via the web.14,15 When asked to describe their current practice environments, 12% selected more than one practice setting. Of the remaining respondents who chose only one practice type, 26.3% reported they are part of a multi-specialty office, 22.1% work in an academic/university practice, 16.9% are in private practice, 11.6% in a single-specialty office, 9.5% work in a hospital owned practice, 9.5% are in solo practice, 3.2% work in a community clinic, and the final 1.1% in a hospital based clinic.
Most participants (62.3%) cared for a reproductive aged woman who was taking a category D or X medication in the past year. The majority of respondents reported they have refrained from prescribing (70.4%), changed (68.5%) or discontinued (63.0%) category D or X medications because of concern for possible fetal exposures. Respondents were asked to identify what references they use when determining pregnancy risk classification for potential teratogens. A single respondent reported using only paper based references such as textbooks, whereas 46.3% use only electronic references such as websites or handheld computer reference software. The vast majority of respondents, 81.5%, report using at least two sources. Almost 64% of respondents agreed that handheld medication reference tools are important. Greater than half, 56.1% of respondents report they are using an electronic health record (EHR) for documenting patient visits. Most respondents (72.2%) report seeing three or four patients per hour. Only 4% of respondents have prescribed an isotretinoin medication, such as Accutane®, and thus participated in an FDA mandated contraceptive compliance program.
Of the 107 respondents who completed the teratogen knowledge exercise, the overall mean percent correctly identified was 58.4% (SD 22.1%). Table 3 shows that accuracy varied across all FDA pregnancy risk categories. The most common errors reflected respondents’ beliefs that medications in category A, B or C carried greater risk and thus incorrectly rated them as D or X. Physicians were less likely to correctly identify medications designated category A, B or C (mean 51.7%, SD 24.1%) than category C/D, D or X medications (71.4.%, SD 24.2%).
The mean percent correct for responses about contraceptive failure rates was 64.6% (SD 23.1%). Overestimating the effectiveness of the specific contraceptive method was the most common error, regardless of type. Physicians were less likely to correctly identify failure rates for less reliable contraception, such as natural family planning and barrier methods (mean 41.4%, SD 25%) than for more reliable methods, such as hormonal contraception and sterilization (mean 87.1%, SD 27.4%). The mean percent correct for each contraceptive method is listed in Table 4.
Most (87.6%) respondents reported they feel it is the responsibility of the PCP to provide family planning and contraceptive services. When asked to rank what they believed were the most important factors in counseling patients about possible adverse effects of potentially teratogenic medications, 98.1% felt physician counseling was most important followed by information provided by a pharmacist. When asked about their likelihood to prescribe various contraceptive methods to women on category D or X medications, most of the respondents reported they were somewhat or very likely to prescribe hormonal contraception or refer to another provider for a subdermal implant, intrauterine contraception or sterilization. As detailed in Table 5, the majority of respondents felt their medical school or residency training was inadequate in preparing them to counsel or prescribe contraception. However, more than half (57.5%) agreed that their internal medicine residency adequately prepared them to counsel about potential teratogens.
When asked to identify how limited participants felt by a selection of potential barriers to providing adequate contraception to women on potentially teratogenic medications, time constraints was the only one the majority (61.3%) of respondents identified. Other items that the majority of participants did not feel limited by are listed in Table 5. Although 69.1% of respondents have billed for either contraceptive counseling or management, the majority did not feel that poor reimbursement was an issue. Most respondents (88.6%) agreed they are interested in obtaining more information. Specifically, the vast majority agreed they would benefit from continuing medical education on counseling patients about potential teratogens (88.6%) and contraception (80.0%). Nearly 85% of respondents agreed that a laminated pocket reference regarding teratogenic medications sorted by therapeutic category would be useful, whereas nearly all (99.1%) reported that an EHR alert would be worthwhile. Roughly 63% of participants felt all category D or X medications should have FDA mandated contraceptive compliance programs such as those that have been used for isotretinoin compounds (e.g. Accutane®). Approximately two-thirds of respondents agreed a referral or telephone consultation services for assistance in providing appropriate contraception for women on potential teratogens would be useful. Nearly all (97.1%) agreed there is a need for information about potential teratogens designed for patient understanding, and 98.1% agreed that an evidence-based pregnancy risk classification system would be useful.
Nearly two-thirds of the respondents reported they encountered a reproductive-aged woman on a category D or X medication in the past year. PCPs who discuss a potential teratogenic therapy with patients should review the need to avoid pregnancy and methods of contraception. Schwarz et al, noted the importance of clinicians’ knowledge of highly-effective contraceptive methods in the prevention of pregnancy when discussing the potential harm from teratogenic medications.6 The majority of respondents identified time constraints as a limitation in providing adequate counseling. Although the majority of respondents did not feel limited by their knowledge of potential teratogens or contraceptive methods to prevent fetal exposures, this study indicates physicians’ actual knowledge may be lower than they believe it to be. If physicians are unaware of their limitations, they may not perceive them as barriers. Interventions aimed at reducing fetal exposures to potential teratogens should include a campaign to raise awareness of these issues. Such a campaign should aim to improve the knowledge of prescribers, pharmacists, women and their partners regarding the need for women taking potential teratogens to be on a highly-effective contraceptive. This could decrease exposures and potentially lower the rate of teratogenic induced congenital anomalies. PCPs frequently discuss preventive health initiatives with their patients. The prevention of teratogenic exposure by provision of highly-effective contraception decreases potential harm to both mother and fetus. Contraceptive counseling can significantly increase the odds of contraceptive use and unintended pregnancy decreases with provision of contraception.13,16
To our knowledge, this was the first study to evaluate physicians’ knowledge, beliefs and barriers to providing contraception for women taking potentially teratogenic medications. To effectively implement improvements, it is important to understand physician behavior and what barriers they perceive in their practice. Although our response rate of 57.3% is higher than the typical range from 20–47% for surveys of physicians, because this is a cross-sectional survey, the findings apply only to the current day practice of the physicians who participated.10,14,15 Other limitations in the generalizability of our findings are related to the relative homogeneity of the respondents and a bias toward younger physicians participating. However, previous studies indirectly agree with our results that prescribers of potential teratogens may face significant barriers to ensuring women taking potential teratogens are protected from unintended pregnancy. These studies demonstrated that women prescribed potential teratogens are no more likely to be using a highly-effective method of contraception than those who are taking medications that have a lower risk of teratogenicity.5,6 Published estimates report that 5.8% of US pregnancies may be exposed to category D and X medications.7,8 Another potential limitation is that it is unclear if participants made use of reference materials to answer knowledge questions. However, given the level of knowledge reported among PCPs, it is unlikely this was common. Though these findings may not be applicable to other practice settings, we believe that because our participants are associated with an urban, academic medical center their self-reported behaviors and knowledge may be better than average community PCPs.
PCPs were selected as the best cohort of physicians to study because they prescribe more contraception and the largest proportion of category D and X medication than any other single specialty.6 Given the breadth of patients that PCPs see, it is unrealistic to expect these physicians to have knowledge about teratogens and contraception always at hand. Furthermore, patients on category D or X medications have underlying medical issues to discuss leaving less time for the clinician to adequately counsel or provide contraception. Most respondents felt that handheld, bedside tools would be useful. Many of the respondents in this study felt their medical school and/or residency training in internal medicine was inadequate in preparing them to provide contraception or to deal with discussions of potential teratogenic medications. As such, medical school and residency curricula should be strengthened to include more information on such topics and to ensure exposure to clinical scenarios that would provide training in this area. A continuing medical education program could be created to emphasize the need to recognize opportunities to prevent teratogenic exposure. Conversely, those PCPs who have less familiarity with contraception should refer patients taking potentially teratogenic drugs to another clinician such as a gynecologist or other women’s health specialist.
Many of the same principles being studied to reduce medical errors across healthcare settings and to incorporate electronic health record (EHR) based tools17–19 could be applied to preventing inadvertent exposure to potential teratogens. The EHR could be a powerful tool to alert physicians of potential teratogenic risk in reproductive aged women, prompting a discussion about contraception or an alternate therapy. With the current efforts to implement EHR programs18,19 and update the FDA pregnancy risk classification system,20 this is an ideal time to develop standards of practice within the EHR that tie into the new FDA regulations. An EHR that is used in the exam room or at the bedside can be programmed to recognize a prescription being written for a potential teratogen to a reproductive aged female and prompt the physician to proceed with contraceptive counseling, prescribing or referral to another clinician if necessary. This will help to eliminate the need for physicians to recognize all medications which are considered potentially teratogenic. A similar alert could be used within the pharmacy software system. Implementation of such an EHR-based safeguard could be an effective tool, but will require significant upkeep as new medications are approved and new evidence for and against fetal risk is published. Given the public health burdens of teratogen exposure resulting in birth defects and the subsequent infant mortality and childhood morbidity, there is significant benefit to employing such an initiative.1,2,6,21,22
Assessment of physician behaviors is complicated and a single institution survey is unlikely to adequately quantify all the factors that affect prescribing practices. These results could inform the design of a larger study of prescribers across geographic regions and medical specialties to better define the barriers to providing contraception to women on potentially teratogenic medications. Future studies should also evaluate the impact of pharmacist and patient behaviors in reducing risk. Though the underlying issues associated with prevention of fetal exposure to potential teratogens are complex, it is imperative to make efforts to reduce such occurrences.
Below is the link to the electronic supplementary material.
Appendix: Written version of survey (PDF 164 kb)
Funded by a grant supplied by a generous anonymous donor. This research was presented at the Fellowship in Family Planning National Conference in Chicago, May, 2009. We would like to thank the faculty and staff of the Research and Education for Academic Achievement (REACH) Network and the Northwestern University Clinical and Translational Sciences Institute (NUCATS) for their support during this project (grant UL1RR025741).
This research could not have been completed without the assistance of Charlie Zei and Jason Thompson, research assistant and programmer, respectively.
Conflicts of Interest David Eisenberg has received honoraria and consulting fees from Bayer Schering Pharma and Haymarket Media. Catherine Stika has received a grant from Boehringer Ingelheim and has a grant pending from Biosante Pharma.
Dr. Eisenberg was a fellow at Northwestern University when this research was conducted, but he is now on faculty at the Washington University School of Medicine