JOURNAL OF WOMEN’S HEALTH Volume 23, Number 8, 2014 ª Mary Ann Liebert, Inc. DOI: 10.1089/jwh.2013.4567

Contraceptive Counseling by General Internal Medicine Faculty and Residents Rachael R. Dirksen, MD,1 Benjamin Shulman, MS,2 Stephanie B. Teal, MD, MPH,3 and Amy G. Huebschmann, MD 4

Abstract

Background: Almost half of US pregnancies are unintended, resulting in many abortions and unwanted or mistimed births. Contraceptive counseling is an effective tool to increase patients’ use of contraception. Methods: Using an online 20-item questionnaire, we evaluated the frequency of contraceptive counseling provided to reproductive-age women during a prevention-focused visit by University of Colorado internal medicine resident and faculty providers. We also evaluated factors hypothesized to affect contraceptive counseling frequency. Results: Although more than 95% of the 146 medicine faculty and resident respondents agreed that contraceptive counseling is important, only one-quarter of providers reported providing contraceptive counseling ‘‘routinely’’ (defined as ‡ 80% of the time) to reproductive-age women during a prevention-focused visit. Providing contraceptive counseling routinely was strongly associated with taking an abbreviated sexual history ‘‘often’’/‘‘routinely’’ (odds ratio [OR] = 11.6 [3.3 to 40.0]) and with high self-efficacy to provide contraceptive counseling (OR = 6.5 [1.5 to 29.0]). However, fewer than two-thirds of providers reported taking an abbreviated sexual history ‘‘often’’/‘‘routinely.’’ More than 70% of providers reported inadequate knowledge of contraceptive methods as a contraceptive counseling barrier. However, providers’ perceived inadequate knowledge was not associated with traditional educational exposures, such as lectures and women’s health electives. Conclusions: In prevention-focused visits with reproductive-age women, a minority of internal medicine faculty and residents reported routine contraceptive counseling. Future efforts to increase contraceptive counseling among internists should include interventions that increase provider contraceptive counseling selfefficacy and ensure that providers obtain an abbreviated sexual history.

Introduction

A

lmost half of US pregnancies are unintended, resulting in more than one million induced abortions each year and many unwanted or mistimed births.1,2 Increasing the use of contraception and decreasing unintended pregnancy is an important public health goal designated in the Healthy People 2020 initiatives. Recent evidence demonstrates that contraceptive counseling by primary care providers does increase patients’ use of contraception at last intercourse.3 Despite our understanding that contraceptive counseling effectively increases patients’ use of contraceptive agents, survey data have shown that fewer than one-quarter of re-

productive-age women report receipt of either contraceptive counseling or a birth control prescription from a healthcare provider over a 12-month time period.4 A prior survey specifically evaluating internal medicine residents showed that residents infrequently assess the contraceptive needs of their outpatients.5 One reason for this is that practical training in contraceptive counseling and family planning is not consistently implemented for internists, even though the American Board of Internal Medicine has designated family planning training as a core competency for internal medicine residents since 1997.6,7 Contraceptive counseling is a vital part of the patient encounter for primary care providers, but internists appear particularly less likely to provide contraceptive

1

Department of Medicine, Division of General Internal Medicine, University of Iowa Carver College of Medicine, Iowa City, Iowa. Department of Biostatistics and Informatics, Colorado School of Public Health, University of Colorado School of Medicine, Aurora, Colorado. 3 Department of Obstetrics and Gynecology, University of Colorado School of Medicine, Aurora, Colorado. 4 Department of Medicine, Division of General Internal Medicine and Center for Women’s Health Research, University of Colorado School of Medicine, Aurora, Colorado. 2

707

708

counseling; in part, this may stem from insufficient training on the subject. Other factors previously shown to decrease the frequency of contraceptive counseling include the medical provider’s perceived lack of knowledge regarding contraceptive methods, the provider’s lack of time to address contraception owing to competing medical problems, a lack of routine sexual history taking during clinic visits, provider misconceptions regarding contraception, a low proportion of reproductive-age women in a provider’s practice, and provider gender.5,8,9 By understanding better the factors associated with the frequency of contraceptive counseling by internal medicine providers, future interventions may be developed to assist internists in providing comprehensive and safe contraceptive care to their reproductive-age female patients. The purpose of this study was to identify the frequency of contraceptive counseling provided to reproductive-age women during a prevention-focused visit by University of Colorado internal medicine residents and affiliated outpatient internist faculty. Also, we sought to identify modifiable factors that may predict a greater frequency of contraceptive counseling behavior. We hypothesized that the following factors would be associated with a higher frequency of contraceptive counseling: a higher percentage of reproductiveage women in one’s outpatient practice, a high frequency of abbreviated sexual history taking, high provider self-efficacy for contraceptive counseling, high perceived knowledge of contraceptive methods, low perceived lack of time, a culture of origin that does not oppose contraception, no contraception misconception, the presence of plans to pursue an outpatient-based career (residents only), and plans to continue working in outpatient internal medicine (faculty only). Finally, we assessed the association of provider knowledge of contraceptive agents with prior women’s health training in the form of residency lectures addressing contraceptive counseling (residents only), continuing medical education (CME) on contraceptive counseling (faculty only), and women’s health electives during residency. Materials and Methods Study design and participants

We sent an anonymous, 20-item SurveyMonkeyª10 questionnaire by electronic mail to University of Colorado internal medicine residents and outpatient general internal medicine faculty in March–April of 2012. Internal medicine clinic sites included two university clinics, three Denver County safetynet clinics, the Veteran’s Affairs Eastern Colorado clinic in Denver, and a private Denver clinic that serves uninsured and underinsured patients. Exclusion criteria for the study were (1) preliminary residents who complete only a single internship year in internal medicine, (2) reporting no reproductive-age women in one’s practice, and (3) coinvestigator of the current study. The Colorado Multiple Institutional Review Board approved the study as an exempt protocol. A total of 95 internal medicine outpatient faculty and 146 internal medicine residents were eligible and invited to participate. Survey development and measures

With permission of the author, we adapted a previously published survey assessing the frequency and barriers to

DIRKSEN ET AL.

contraceptive counseling among internal medicine residents.5 The previously published survey assessed contraceptive counseling factors, including sexual history, self-efficacy, outpatient medicine career plans, proportion of practice devoted to reproductive-age women, and provider gender. We adapted the survey to also assess the prevalence of barriers noted in prior published focus group data and observed by the study investigators (i.e., perceived lack of time, the presence of contraception misconceptions, cultural origin that opposes contraception). Separate surveys were created for residents and faculty, but the only questions that differed between surveys addressed factors that would vary naturally between study groups (e.g., postgraduate year for residents, years in clinical practice for faculty). The survey was pilot tested by five internal medicine chief residents, who were not participants in the study, to assess comprehension and responder burden. The demographics of survey participants were assessed with regard to age, gender, year in training (residents only), years practicing as attending (faculty only), location of clinic, and plans after residency (residents only) or plans to remain in outpatient internal medicine (faculty only). Study participants were asked to use a four-point Likert scale with verbal anchors to assess how often they address following medical history components during a prevention-focused visit with a woman aged 15–45 years: medications, immunizations, abbreviated sexual history, contraceptive counseling, seat-belt usage, and preconception counseling (rarely £ 20%, sometimes 21–49%, often 50–79%, routinely ‡ 80%). In the survey, we defined abbreviated sexual history taking as asking patients about such items as ‘‘current sexual activity, number of partners.’’ We defined contraceptive counseling as both assessing the ‘‘current contraceptive method’’ and providing contraceptive counseling ‘‘if needed.’’ Participants also estimated the proportion of their practice that consisted of reproductive-age women. Additional survey questions addressed potential factors hypothesized to affect contraceptive counseling based on prior studies.5,8 These factors included a low prevalence (defined as two categories of < 10% and < 20%) of reproductive-age women in their primary care practice, taking an abbreviated sexual history ‘‘rarely’’ or ‘‘sometimes’’ (defined as < 50% of prevention-focused visits with reproductive-age women), inadequate time to provide contraceptive counseling (defined as an answer of ‘‘strongly agree’’ or ‘‘somewhat agree’’ to the statement ‘‘I would provide contraceptive counseling to my patients more often if I had more time during an annual exam’’) and inadequate knowledge to provide contraceptive counseling (defined as an answer of ‘‘strongly agree’’ or ‘‘somewhat agree’’ to the statement ‘‘I would provide contraceptive counseling to my patients more often if I had more knowledge regarding contraceptive methods’’). Other factors hypothesized to affect contraceptive counseling frequency included the presence of religious or ethnic culture of origin that opposed contraception, lack of perceived importance of contraceptive counseling (defined as an answer of ‘‘somewhat disagree’’ or ‘‘strongly disagree’’ to the statement ‘‘It is important for me to know how to discuss different forms of contraception, including their effectiveness and potential adverse effects’’), low self-efficacy (defined as an answer of ‘‘strongly disagree’’ or ‘‘somewhat disagree’’ to the statement ‘‘I feel confident assessing my patients’ current contraceptive methods and discussing effective alternate

CONTRACEPTIVE COUNSELING BY INTERNISTS

methods when necessary’’), plans to pursue a nonoutpatient position (residents only) or plans to stop working in outpatient internal medicine within the next 5 years (faculty only), and contraception misconception presence (defined as an answer of ‘‘strongly disagree’’ or ‘‘somewhat disagree’’ to the statement ‘‘I would generally prescribe contraception to a woman who is interested in contraception, but has not had cervical cancer screening within the last 3 years’’). At the time of survey creation, the longest appropriate interval between cervical cancer screening tests was 3 years, based on US Preventive Services Task Force (USPSTF) recommendations. Requiring cervical cancer screening prior to contraception prescription is a known and documented barrier to obtaining contraception counseling. The survey was administered in March and April of 2012, and new USPSTF recommendations for potential longer intervals (5 years) were released in March 2012. As such, our results potentially underestimate the proportion of participants who would withhold contraception if cervical cancer screening were not up-to-date. Finally, participants answered questions regarding prior women’s health educational exposures, either through residency lectures on contraceptive counseling (residents only) or continuing medical education on contraceptive counseling (faculty only), or a women’s health elective during their training (all participants). Statistical analysis

Descriptive statistics of the study outcomes were reviewed to assess for statistical outliers. All study analyses were performed using SAS 9.2 (SAS Institute, Cary, NC). The primary outcome of self-reported frequency of contraceptive counseling among faculty and residents was analyzed using a chi-square test for association between the reported frequency and faculty/resident status. The relationship between the presence of barriers or facilitators and frequency of contraceptive counseling was analyzed using univariate logistic regression models to determine the odds ratio (OR) of providing contraceptive counseling routinely, as compared to often, sometimes, or rarely ( ‡ 80% vs. < 80%) given the presence or absence of each factor. Each barrier was analyzed in a separate logistic model. We also tested for association between contraceptive counseling and preconception counseling with Fisher’s exact test. Finally, chi-square tests were used to test for associations between three educational exposures (lectures, CME, and women’s health electives) and reporting adequate knowledge to provide contraceptive counseling. We attempted to develop a multivariate model to determine the strength of association of the more significant barriers with contraceptive counseling within the same statistical model. However, the frequency of abbreviated sexual history taking and high self-efficacy was sufficiently collinear with the frequency of contraceptive counseling that we could not build an appropriate multivariate model incorporating the key contraceptive counseling predictors. Results Study population characteristics

Of 95 outpatient internal medicine faculty members contacted, 66 (69.5%) responded; of 146 internal medicine residents contacted, 80 (54.8%) responded. The study population

709

consisted of resident and faculty respondents (n = 146) who reported caring for women 15–45 years of age in their outpatient clinic. The medicine resident respondents were almost 60% male, closely approximating the gender breakdown of the residents in the internal medicine residency program (59% male). Onethird of resident respondents were in their first year of postgraduate training, slightly fewer than one-third were in their second year, and slightly more than one-third were in their third year of training (Table 1). Of the faculty members who were analyzed, 47.0% were female, which is slightly lower than the percentage of women among the outpatient general internal medicine faculty who received the survey (51% female). Approximately half the faculty respondents were £ 10 years postresidency (Table 1). Factors affecting frequency of contraceptive counseling

Almost one-fifth of residents and one-third of faculty reported routine contraceptive counseling at prevention-focused visits (Table 2). Across all providers, routine contraceptive counseling rates of 25% were lower than reported routine medication-reconciliation rates (90%) and routine immunization-review rates (60%) but higher than routine preconceptioncounseling rates (7%) or routine seat-belt counseling rates (4%). Although there was not a significant association between status as faculty or resident and frequency of contraceptive counseling as a categorical variable ( p = 0.14, Table 2), faculty had a greater odds of performing routine contraceptive counseling than residents (OR 2.3 [1.1 to 4.9], unadjusted OR with 95% confidence interval [CI] (Table 3). Potential factors affecting contraceptive counseling for all providers were broken down into potential barriers and facilitators that may affect the frequency of contraceptive counseling (Table 2). The majority of faculty and residents reported that they perceive inadequate time (75.3%) and inadequate knowledge (74.0%) as reasons for not performing contraceptive counseling (Table 2). More than 95% of all providers reported that it was important to be able to discuss various forms of contraception, including their effectiveness and potential adverse effects. More than a third of respondents reported taking an abbreviated sexual history only ‘‘rarely’’ or ‘‘sometimes’’ during a prevention-focused visit.

Table 1. Study Population Demographics All respondents Residents Returned survey 146 (100.0%) and met survey criteria Female 64 (43.8%) PGY1 N/A PGY2 PGY3 Years N/A postresidency 0–5 6–10 11–15 > 15

Faculty

80 (54.8%) 66 (45.2%) 33 26 21 32

(41.3%) 31 (47.0%) (32.5%) N/A (26.3%) (40.0%) N/A

Missing data for these survey questions: < 5%. N/A, not applicable; PGY, post-graduate year.

14 19 9 21

(21.2%) (28.8%) (13.6%) (31.8%)

710

DIRKSEN ET AL.

Table 2. Frequency of Contraceptive Counseling and Assessment of Potential Factors Affecting Counseling All respondents n (%) Frequency of contraceptive counseling Routinely ( ‡ 80%) Often (50%–79%) Sometimes (21%–49%) Rarely ( £ 20%) Potential barriers Perceived inadequate time Perceived inadequate knowledge < 20% of practice are women 15–45 years of age < 10% of practice are women 15–45 years of age Obtain sexual history < 50% Presence of contraception misconception Low self-efficacy Presence of religious or ethnic culture of origin that opposes at least one form of contraception Plans as nonoutpatient (hospitalist, subspecialty, other) (residents only) Plan to stop working in outpatient in the next 5 years (faculty only) Low perceived importance Potential facilitators Prior women’s health elective (faculty and residents) Contraceptive-related CME in the past 5 years (faculty) Prior contraceptive counseling lecture (residents only)

37 48 43 14

(25.3%) (32.9%) (29.5%) (9.6%)

Medicine residents n (%) 15 27 28 9

(18.8%) (33.8%) (35.0%) (11.3%)

Outpatient GIM faculty n (%) 22 21 15 5

p-value

(33.3%) (31.8%) (22.7%) (7.6%)

0.1430

110 (75.3%) 108 (74.0%) 102 (69.9%)

68 (85.0%) 66 (82.5%) 57 (71.3%)

42 (63.6%) 42 (63.6%) 45 (68.2%)

0.0029 0.0097 0.6876

42 (28.8%)

20 (25.0%)

22 (33.3%)

0.2683

56 (38.4%) 46 (31.5%)

29 (36.3%) 25 (31.3%)

27 (40.9%) 21 (31.8%)

0.5645 0.9414

31 (21.2%) 15 (10.3%)

21 (26.3%) 12 (15.0%)

10 (15.2%) 3 (4.6%)

0.1027 0.0641

N/A

56 (70.0%)

N/A

N/A

N/A

N/A

2 (3.0%)

N/A

5 (3.4%)

3 (3.8%)

2 (3.0%)

1.000a

34 (23.3%)

10 (12.5%)

24 (36.4%)

0.0010

N/A

N/A

21 (31.8%)

N/A

N/A

63 (78.8%)

N/A

N/A

a Fisher’s exact test used to calculate this p-value, owing to small cell sizes; all other p-values calculated with chi-square test for association. CME, continuing medical education; GIM, general internal medicine.

Factors that were significantly associated with greater odds of providing contraceptive counseling routinely included female gender providers (vs. male), faculty providers (vs. residents), reported taking of a sexual history routinely or often (vs. sometimes or rarely), reported high self-efficacy for contraceptive counseling (vs. low self-efficacy), reported adequate knowledge regarding contraceptive methods, and reported adequate time (Table 3). Residents with outpatient internal medicine plans were significantly more likely to provide routine contraceptive counseling than residents with nonoutpatient plans (38% vs. 9%, p = 0.001); however, the OR could not be estimated reliably, because only five residents with nonoutpatient plans provided routine contraceptive counseling. The association of perceived importance of contraceptive counseling and contraceptive counseling frequency could also not be estimated, as none of the participants who rated importance as low provided contraceptive counseling routinely. Preconception counseling in preventive visits with women of childbearing age was reported by only 7% of all respondents, by 27% of respondents who reported routine contraceptive counseling, and by 0% of respondents who did not report routine contraceptive counseling ( p < 0.0001

for association of routine contraceptive counseling and routine preconception counseling). A prior women’s health elective was also significantly associated, with a 2.5 times greater odds of providing contraceptive counseling routinely (Table 3). However, there was not a significant association between provider belief of having inadequate knowledge and educational exposures in the form of women’s health electives, lectures, or CME in contraceptive counseling (Table 4). The majority of resident and faculty providers felt that they would provide contraceptive counseling more often if they had more knowledge regarding contraceptive methods. Of resident participants, 70.0% agreed that they would have preferred more training in contraceptive methods and counseling as part of their internal medicine residency. Of faculty participants, 72.7% agreed that they would like more CME regarding contraceptive methods and counseling. Discussion

Although we asked University of Colorado internal medicine providers about their frequency of provide contraceptive

CONTRACEPTIVE COUNSELING BY INTERNISTS

711

Table 3. Association of Hypothesized Factors with the Provision of Routine Contraceptive Counseling ORa Study population characteristics Female vs. male 4.3 Faculty vs. resident 2.3 Absence of potential barriers Perceived adequate time 3.1 vs. inadequate time Perceived adequate 3.9 knowledge vs. inadequate knowledge 2.4 ‡ 20% of practice devoted to women vs. < 20% 1.0 ‡ 10% of practice devoted to women vs. < 10% Routinely/often take 11.6 sexual history vs. not Misconception absent 1.5 vs. present Self-efficacy (high vs. low) 6.5 1.0 Religious or ethnic culture that opposes contraception absent vs. present N/Ab Plans to work in outpatient setting (residents only) vs. nonoutpatient setting N/Ac Plans to continue working in outpatient setting for at least the next 5 years (faculty only) vs. not Perceived importance N/Ad (dichotomize high vs. low) Potential facilitators Prior women’s health 2.5 elective (faculty and residents) yes vs. no 1.6 Contraceptive-related CME in the past 5 years (faculty only) yes vs. no N/Ae Prior contraceptive counseling lectures (residents only) yes vs. no a

Table 4. Association Between Educational Exposures and Reporting Inadequate Knowledge About Contraceptive Counseling

Lower Upper CI CI p-value 1.9 1.1

9.8 4.9

0.0004 0.0337

1.3

7.2

0.0090

1.7

8.9

0.0015

1.1

5.4

0.0271

0.4

2.2

0.8937

3.3

40.0

0.0001

0.6

3.3

0.3802

1.5 0.3

29.0 3.3

0.0135 0.9823

N/A

N/A

N/A

1.1

5.8

0.0276

0.5

4.6

0.4248

N/A

OR of providing contraceptive counseling routinely, as compared to often, sometimes, or never ( ‡ 80% vs. < 80%). b OR estimate not reliable: £ 5 residents with nonoutpatient plans reported routine counseling ( ‡ 80% of preventive visits). c OR not estimable: Only two faculty were not planning to continue practice in an outpatient setting. d OR not estimable: 0 providers who do not perceive contraceptive counseling as important (out of 4) reported routine counseling ( ‡ 80% of preventive visits). e OR estimate not reliable: < 5 participants with no prior contraceptive counseling lectures provided routine counseling ( ‡ 80% of preventive visits). CI, confidence interval; OR, odds ratio.

Inadequate knowledge Total n n (%) p-valuea Residents who had contraceptive counseling lectures Residents who did not have contraceptive counseling lectures Faculty who had CME regarding contraception Faculty who did not have CME regarding contraception Women’s health elective completed Women’s health elective not completed

63

53 (84.1%)

16

13 (81.3%)

21

12 (57.1%)

42

30 (71.4%)

34

26 (76.5%)

108

82 (75.9%)

0.78

0.26

0.95

a Chi-square test for association between educational exposure and reporting inadequate knowledge.

counseling in a fairly ideal circumstance—a preventionfocused visit for a woman of reproductive age—only 33.3% of faculty and 18.8% of residents reported providing contraceptive counseling routinely ( ‡ 80% of the time). In contrast to the low rates of counseling observed, more than 95% of respondents agreed that contraceptive counseling is important. These findings suggest that overcoming barriers to contraceptive counseling is important to increase contraceptive counseling frequency among internal medicine providers. High self-efficacy for contraceptive counseling and taking an abbreviated sexual history routinely were strongly associated with routine contraceptive counseling in our study population, suggesting that these are likely facilitators of contraceptive counseling for internists. We also observed other modifiable barriers to routine contraceptive counseling, including inadequate provider knowledge for contraceptive counseling and the misconception that cervical cancer screening must be up-to-date to prescribe contraception. Our finding that 18.8% of internal medicine residents routinely provide contraceptive counseling was consistent with a prior study that reported 17% of internal medicine residents routinely provide contraceptive counseling.5 Similar to our findings, Lohr et al. also found a strong association between routine sexual history taking and routinely providing contraceptive counseling (OR 6.1, 95% CI 2.38–15.49).5 Lack of time was noted as a barrier to contraceptive counseling in both our study and two other qualitative focus group studies.8,11 Akers et al. also found that providers felt that they had inadequate knowledge regarding contraceptive methods.8 Our study is the first, to our knowledge, to look at the association of contraceptive counseling by faculty vs. resident status. Our finding that only one-third of outpatient internist faculty provide routine contraceptive counseling suggests that medicine faculty are not consistently modeling contraceptive counseling behavior for residents. More than 95% of faculty and resident internal medicine respondents reported that contraceptive counseling is

712

important, but only 73.7% of residents and 84.8% of faculty report high self-efficacy, showing a gap between perceived importance and confidence to deliver counseling. In addition, 82.5% of residents and 63.6% of faculty reported inadequate contraceptive knowledge, and improved knowledge is related to improved self-efficacy for most behaviors.12 Because the frequency of taking an abbreviated sexual history was strongly associated with contraceptive counseling frequency, an important area to evaluate is simultaneously training internal medicine providers to increase their selfefficacy for performing abbreviated sexual histories and delivering contraceptive counseling. This is particularly relevant because prior studies have shown that internal medicine providers are not routinely taking a sexual history as part of prevention-focused visits13 and are less likely to take a sexual history than pediatricians or obstetricians and gynecologists.14 In taking abbreviated sexual histories, we can identify patients in need of contraceptive counseling or preconception counseling. In doing so, women may avoid unintended pregnancies, or women who are ambivalent or intend to become pregnant can make necessary plans, such as starting prenatal vitamins and stopping any teratogenic medications. Our findings should also be considered in the context of recent healthcare policy changes. The Affordable Care Act mandated that women’s preventive health visits must be offered to all health plan enrollees without individual copayment and should include the following contraception-related measures for reproductive-age women: preconception counseling, contraceptive counseling, and counseling regarding all FDAapproved contraception methods.15 Adherence to this mandate would allow internist and other primary care providers to appropriately identify patients at risk for unintended pregnancy and to discuss appropriate short-term and long-term contraceptive methods. If these Affordable Care Act mandates translate into a quality reporting system that requires primary care providers to report on how well they are meeting quality reporting measures, such as contraceptive counseling for preventive women’s health visits, it would incentivize internists and other primary care providers to provide and document preconception and contraceptive counseling. However, at the time of this publication, there were not yet any Medicare Physician Quality Reporting System (PQRS) measures with regard to preconception or contraceptive counseling. We found a strikingly low rate of preconception counseling in preventive visits with women of reproductive age: Only 7% of providers reported routinely performing preconception counseling. Women of reproductive age seen by internists may be more likely to be on teratogenic medications or have health conditions that could be worsened by pregnancy.16 The Institute of Medicine rejected listing preconception counseling as a separately reimbursable service through the Affordable Care Act, contending that it was already an integral part of the well-woman visit. Among our respondents, preconception counseling was not routinely included in preventive visits for women of reproductive age, suggesting a need for future studies to address the provider, practice, and policy-level factors related to this important omission. Increasing self-efficacy for contraceptive counseling appears to be another likely method to increase contraceptive counseling among internists. One way to improve resident self-efficacy in both abbreviated sexual history taking and contraceptive counseling is through faculty guidance and

DIRKSEN ET AL.

experience. However, only one-third of faculty reported delivering contraceptive counseling routinely in this study, so residents need other sources of guidance as well. Residents may get some exposure to contraceptive counseling through lectures and women’s health electives, but we did not find a significant association between these educational exposures and contraceptive counseling knowledge, although a women’s health elective was significantly associated with greater frequency of routine contraceptive counseling provision. In our study population, women’s health electives were required for participants of the primary care track; specific women’s health training may be one reason that residents with outpatient plans outperformed residents with nonoutpatient plans with regard to routine contraceptive counseling provision. To confirm whether mandating women’s health electives would impact routine contraceptive counseling rates would require assessing changes in contraceptive counseling rates in a randomized controlled trial across institutions or before/ after a ‘‘natural experiment’’ of implementing mandatory women’s health electives for internal medicine residents in a single institution. Other interventions to build providers’ selfefficacy to provide contraceptive counseling may include guided practice exercises, through virtual computer-based scenarios or in clinical training environments, as guided practice is known to build self-efficacy.13,17 Given the low rates of preconception counseling we observed, future interventions to improve contraceptive counseling should also address preconception counseling. Finally, not all internal medicine providers need to be experts in contraceptive management, and an important strategy may be to develop an ‘‘ask, advise, and refer’’ strategy18 so providers may refer patients to knowledgeable providers within the practice or within the women’s health community when needed.19 A strength of this study is the assessment of both faculty and resident internal medicine provider counseling practices. A weakness of this study is that our respondents were affiliated with a single institution, thus reducing the external validity of our findings. To mitigate this weakness, we surveyed providers from seven distinct internal medicine clinics affiliated with our institution, including academic clinics, safety-net clinics, a Veteran’s Affairs clinic, and a nonprofit privately funded clinic. Another possible weakness is that a social-desirability bias could have led providers to overreport their counseling frequency and self-efficacy; if this occurred, our results would overestimate actual contraceptive counseling practices. Another limitation of this study is that the results are based on provider self-report, which may be different from actual provider behavior, owing to social-desirability response bias. Thus, a potential future direction would be to assess associations of provider, patient, and system-level predictors of contraceptive counseling with objective measures of contraceptive counseling, such as chart review, or direct observation of clinical encounters via recording device or an observer. Although patient-response bias may also be an issue, patient surveys to assess provider contraceptive counseling behavior could be used to avoid provider response bias. Conclusions

Our key findings are that internists underutilize contraceptive counseling and that taking an abbreviated sexual history routinely and high self-efficacy for providing contraceptive

CONTRACEPTIVE COUNSELING BY INTERNISTS

counseling are likely predictors of contraceptive counseling. We did not find that traditional contraceptive counseling educational exposures, such as lectures and CME, were significantly associated with medicine providers’ perceived knowledge in contraceptive methods. Therefore, we need to design and test methods to improve self-efficacy for abbreviated sexual history taking and contraceptive counseling among internists to determine optimal approaches. One way this could occur is for women’s health and general internal medicine researchers to codevelop and test improved contraceptive counseling education methods. In addition to developing enhanced contraceptive counseling education, it will also likely be helpful to develop methods for medicine providers with lower contraceptive counseling self-efficacy to ask, advise, and refer patients for contraceptive counseling to women’s health colleagues, similar to methods used to improve smoking-cessation counseling in primary care settings.20 Because many internal medicine providers do not obtain abbreviated sexual histories routinely, it is important for internal medicine providers and health systems to improve standard medical history forms to include an abbreviated sexual history among other women’s health-related questions. Specifically, we propose that it would be beneficial to enhance office systems to use unobtrusive methods (e.g., paper forms, online data entry) to inquire about sexual history, contraceptive method history, and other women’s health-related history so these data are systematically collected for providers to review efficiently. Overall, we need to develop and test innovative strategies to increase contraceptive counseling by internists to ensure that all primary care providers give women the necessary tools to manage their reproductive health. Acknowledgments

Amy G. Huebschmann is supported by NIH/NCATS Colorado CTSI Grant Number KL2 TR000156. Contents are the author’s sole responsibility and do not necessarily represent official National Institutes of Health views. Contributors were Research Consulting Lab, Colorado Biostatistics Consortium, University of Colorado, and Colorado Clinical & Translational Sciences Institute. Author Disclosure Statement

No competing financial interests exist. References

1. Jones RK, Kooistra K. Abortion incidence and access to services in the United States, 2008. Perspect Sex Reprod Health 2011;43:41–50. 2. Finer LB, Henshaw SK. Disparities in rates of unintended pregnancy in the United States, 1994 and 2001. Perspect Sex Reprod Health 2006;38:90–96. 3. Lee J, Parisi SM, Akers AY, Borrerro S, Schwarz EB. The impact of contraceptive counseling in primary care on contraceptive use. J Gen Intern Med 2011;26:731–736. 4. Frost JJ. Trends in U.S. women’s use of sexual and reproductive health care services, 1995–2002. Am J Public Health 2008;98:1814–1817. 5. Lohr P, Schwarz EB, Gladstein JE, Nelson AL. Provision of contraceptive counseling by internal medicine residents. J Women’s Health (Larchmt) 2009;18:127–131.

713

6. Schreiber CA, Harwood BJ, Switzer GE, Creinin MD, Reeves MF, Ness RB. Training and attitudes about contraceptive management across primary care specialties: A survey of graduating residents. Contraception 2006;73: 618–622. 7. Cassel C, Blank L, Braunstein G, Burke W, Fryhofer SA, Pinn V. What internists need to know: Core competencies in women’s health. ABIM Subcommittee on Clinical Competence in Women’s Health. Am J Med 1997;102: 507–512. 8. Akers AY, Gold MA, Borrero S, Santucci A, Schwarz EB. Providers’ perspectives on challenges to contraceptive counseling in primary care settings. J Women’s Health (Larchmt) 2010;19:1163–1170. 9. Leeman L. Medical barriers to effective contraception. Obstet Gynecol Clin N Am 2007;34:19–29. 10. SurveyMonkey.com, LLC. Available at www.survey monkey.com (Accessed August 20, 2012). 11. Schwarz EB, Santucci A, Borrero S, Akers AY, Nikolasjski C, Gold MA. Perspectives of primary care clinicians on teratogenic risk counseling. Birth Defects Research (Part A) 2009;85:858–863. 12. Baranowski T, Perry CL, Parcel GS. How individuals, environments, and health behavior interact. In: Glanz K, Rimer BK, Marcus F, eds. Health behavior and health education: Theory, research, and practice, 3rd ed. San Francisco: JosseyBass, 2002:165–184. 13. Loeb D, Lee R, Binswanger IA, Ellison MC, Aagaard EM. Patient, resident physician, and visit factors associated with documentation of sexual history in the outpatient setting. J Gen Intern Med 2011;26:887–893. 14. Wimberly YH, Hogben M, Moore-Ruffin J, Moore SE, FryJohnson Y. Sexual history-taking among primary care physicians. J Natl Med Assoc 2006;98:1924–1929. 15. US Department of Health and Human Services. Women’s preventive services guidelines. Available at http://www.hrsa .gov/womensguidelines (Accessed November 11, 2013). 16. Schwarz EB, Maselli J, Norton M, Gonzales R. Prescription of teratogenic medications in United States ambulatory practices. Am J Med 2005;188:1240–1249. 17. American Diabetes Association. DiabetesProSM: Professional resources online—the ADA Simulation Case Program. Available at: http://professional.diabetes.org/Congress_ Display.aspx?TYP = 9&CID = 84062 (Accessed November 27, 2012). 18. Schroeder SA. What to do with a patient who smokes. JAMA 2005;294:482–487. 19. Woodhams EJ, Gilliam M. In the clinic: Contraception. Eds Cotton D, Taichman D, Williams S. Ann Intern Med 2012; 157:ITC4–14. 20. Warner DO, The American Society of Anesthesiologists Smoking Cessation Initiative Task Force. Feasibility of tobacco interventions in anesthesiology practices: A pilot study. Anesthesiology 2009;110:1223–1228.

Address correspondence to: Rachael R. Dirksen, MD Division of General Internal Medicine University of Iowa Health Care 105 9th Street Coralville, IA 52241 E-mail: [email protected]

Contraceptive counseling by general internal medicine faculty and residents.

Almost half of US pregnancies are unintended, resulting in many abortions and unwanted or mistimed births. Contraceptive counseling is an effective to...
173KB Sizes 2 Downloads 4 Views