Abstract
Objective:
This study explores OB/GYN providers’ knowledge about published health and healthcare disparities in women’s reproductive health.
Method:
We collected demographic and health disparities knowledge information from OB/GYN providers who were members of ACOG District IV using an online survey (n=483). We examined differences across groups using statistical tests and regression analyses in a structural equation modeling approach.
Results:
Receiving disparities education was positively associated with higher self-reported disparities knowledge and disparities quiz performance (p<.05). African American/Black providers had higher quiz scores than their white counterparts, and providers varied in their levels of disparities knowledge across practice settings (p<.05).
Conclusions:
Differences in levels of knowledge of racial/ethnic disparities in health and healthcare outcomes among OB/GYN providers varied across race/ethnicity, practice context, and whether providers had received formal disparities education. Future research should explore these differences at a population level and develop interventions to improve health disparities education among OB/GYN providers.
INTRODUCTION
A health disparity is defined by the Centers for Disease Control and Prevention as a, “preventable difference in the burden of a disease.” 1 Although health disparities can exist due to gender, education, socioeconomic status and other factors, some of the most stark and poorly understood disparities exist across racial and ethnic groups.2 Marked racial and ethnic disparities are notable in nearly every discipline of medicine with obstetrics and gynecology (OB/GYN) being no exception. Notable racial and ethnic disparities in the incidence of infertility, preterm birth, fetal death and cervical cancer are well documented.3–7 However, the etiology of these disparities remains poorly understood. Provider perceptions and lack of awareness may contribute to and perpetuate health disparities in women’s reproductive health.
Given that by 2044 the United States population will be majority non-white, provider knowledge about health disparities is of increasing importance.8 Over half of these non-white Americans will be female sex and susceptible to disparities in reproductive outcomes. Despite the growing importance of racial and ethnic disparities in healthcare, there is little data available on OB/GYN providers’ knowledge about disparities in their field. In 2018, Moroz et al. published findings from a survey of the Society for Maternal Fetal Medicine membership.9 They noted most respondents considered themselves at least somewhat knowledgeable about disparities, over 90% were willing to learn more, and most respondents performed well on their knowledge assessment.
The current study adds to a very sparse body of research examining provider knowledge about racial and ethnic disparities in reproductive outcomes. The aim of our study was to advance the state of extant research by assessing OB/GYN providers’ knowledge about published disparities in health and healthcare outcomes. Our study adds to the literature by including both obstetric and gynecologic providers. Additionally, we constructed the questions in our knowledge assessment from an American College of Obstetricians and Gynecologists (ACOG) publication available to all of the participants prior to the survey. Our study also identifies provider sociodemographic factors that correspond with systematic differences in both self-assessed knowledge and quiz performance.
MATERIALS AND METHODS
Between June 2017 and January 2018, we invited all providers registered as members with ACOG District IV to complete an internet survey. The Duke Health Institutional Review Board approved this study (IRB# Pro00080959). We developed the content, wording, and sequencing of survey questions in consultation with survey design experts through Duke University’s Social Science Research Institute. We piloted the survey with 78 Duke OB/GYN providers and subsequently revised it to clarify construct validity based on respondents’ feedback prior to distributing the final survey to ACOG District IV members.
The final survey contained 24 questions to measure provider knowledge of racial/ethnic disparities in OB/GYN patient outcomes as well as factors related to clinical practice environment, experience and training, and background characteristics. We measured self-reported disparities knowledge using a four-point scale aimed to assess how knowledgeable respondents felt regarding racial and ethnic disparities in OB/GYN diagnoses. We also measured disparities knowledge using a quiz consisting of six true/false statements (see Appendix A). Respondents were asked whether they had received formal disparities training and the amount of time that had passed since they completed their residency training. Clinical practice environment measures included respondents’ estimation of the proportion of racial/ethnic minority patients they served, the population density of the geographic setting in which their clinic was located, and the type of clinical setting in which they practiced. Background characteristics measures included gender, race, and ethnicity. Upon survey completion, respondents received a message thanking them for their participation and provided educational resources about racial/ethnic disparities in reproductive outcomes.
We tabulated clinic practice environment, disparities knowledge self-assessment, disparities knowledge quiz assessment, and provider background characteristics. We then estimated quiz-assessed disparities knowledge as a function of observed factors using ordinary least squared regression in a structural equation modeling framework. Quiz-assessed disparities knowledge (the dependent variable of interest) was calculated by summing the number of correct responses and dividing by six (the total number of quiz questions), yielding the proportion of correctly-answered questions. Observed variables included measures of respondent race, ethnicity, gender, years since OB/GYN training, disparities education received, patient proportion minority, practice population size, practice type, and self-assessed health disparities knowledge. Race was treated as a series of binary variables indicating African American/Black, Asian, or White (reference). Ethnicity was treated as a binary variable indicating Hispanic/Latinx ethnicity or non-Hispanic/Latinx ethnicity (reference group). Gender was treated as a binary variable indicating female or male (reference group). Years since OB/GYN training was treated as a series of binary variables indicating ten or fewer years since residency (reference group), 11–30 years since residency, or more than 30 years. Disparities education received was treated as a binary variable corresponding with whether respondents reported receiving any formal training about racial and ethnic disparities. Patient proportion minority was treated as a continuous variable ranging from 1 (almost none) to 5 (nearly all). Clinic location was treated as a series of binary variables corresponding with whether providers felt their clinical practice was located in a rural environment, suburban setting, small city, or large city (reference group). Practice type was treated as a series of binary variables that corresponded with whether respondents felt their clinical practice setting was best described as solo, group, government, hospital, or academic (reference group).
We estimated four regression models using a stepwise approach. Model 1 includes provider gender, race, and ethnicity. Model 2 adds provider experience and training. Model 3 adds practice characteristics, and Model 4 adds self-rated disparities knowledge. Models were estimated in a structural equation modeling framework using the full-information maximum-likelihood (FIML) method, which uses all available data and in essence corrects for data missing because of nonresponse 10. Statistical analyses were performed using Stata 1511
RESULTS
The survey was distributed via email to 5,363 members of ACOG District IV with active email addresses. The email was opened by 1,691 members, and we received 482 survey responses. This yielded a response rate of approximately 9% of the entire population of ACOG District IV members, or 29% of those who opened the email. A majority of respondents reported that they were white (58%), were women (55%), and completed residency 11–30 years ago (51%). Nearly 40% reported working in group practice settings, while 37% reported practicing in academic medical settings. Approximately 34% of respondents reported that they had received formal training about racial/ethnic disparities. Approximately 64% reported that “a large portion” of their patient population were members of racial/ethnic minority groups, and nearly 9% reported that members of racial/ethnic minority groups made up “nearly all” of their patient populations.
A majority of respondents reported that they were either very (28.8%) or somewhat (54.2%) knowledgeable about racial and ethnic disparities in reproductive outcomes. The mean score on the six-question disparities knowledge quiz assessment was 5/6 questions correct, with a standard deviation (SD) of 0.97. Approximately 72% of respondents who completed the quiz assessment answered at least four questions correctly, while only approximately 2% answered all six questions correctly.
Among those who completed the quiz, the majority (84%) correctly indicated that the statement, “Racial/ethnic disparities in US preterm birth rates disappear after accounting for socioeconomic disparities,” was false. Only 4%, however, correctly indicated that the statement, “Black women are more likely to have had a pap smear within the past three years,” was true. Over 82% correctly indicated that the statement, “Black women are more likely to be diagnosed with cervical cancer than white women,” was true. Nearly 96% correctly indicated that the statement, “Maternal death rates are consistent across racial and ethnic groups in the United States,” was false. Approximately 47% correctly indicated false when evaluating the statement, “White women are the most likely racial/ethnic group to experience infertility,” and this statement also had the highest percentage of “unsure” responses (35%). Nearly 88% correctly indicated false when evaluating the statement, “Studies show implicit bias has very little effect on disparities in healthcare outcomes.”
Differences in individual quiz item responses are depicted in Supplementary Tables 1 and 2. We used Fisher’s exact test of independence to test for differences in quiz item responses across groups, exploring differences across race and ethnicity (Table 1) and prior disparities training (Table 2). We found no significant differences across race/ethnicity except for the statement regarding whether socio-economic disparities explain racial/ethnic disparities in U.S. preterm birth rates (the first quiz statement), for which 86% of non-Hispanic/Latinx and 55% of Hispanic/Latinx respondents selected the correct answer. Respondents who reported receiving any formal disparities training were more likely to select the correct response for quiz questions regarding preterm birth rates (p<.05), Pap smears (p=<.01), infertility (p<.01), and implicit bias (p<.05) (quiz items 1, 2, 5, and 6, respectively).
Table 1:
Descriptive Statistics
| Survey Respondents (N = 483) | |
|---|---|
| Disparities Knowledge | |
| Self-Reported Disparities Knowledge | |
| Very Little Knowledge | 5 (1.0%) |
| Minimal Knowledge | 37 (7.7%) |
| Somewhat Knowledgeable | 262 (54.2%) |
| Very Knowledgeable | 139 (28.8%) |
| Missing | 40 (8.3%) |
| Background Characteristics | |
| Gender | |
| Male | 162 (33.5%) |
| Female | 263 (54.5%) |
| Missing | 58 (12.0%) |
| Race | |
| African American or Black | 102 (21.1%) |
| Asian | 24 (5.0%) |
| White | 280 (58.0%) |
| Missing | 77 (15.9%) |
| Ethnicity | |
| Non-Hispanic/Latinx | 398 (82.4%) |
| Hispanic/Latinx | 20 (4.1%) |
| Missing | 65 (13.5%) |
| Experience and Training | |
| Years Since Residency | |
| 10 or Fewer | 134 (27.7%) |
| 11–30 | 234 (48.4%) |
| 31 or More | 98 (20.3%) |
| Missing | 17 (3.5%) |
| Disparities Education Received | |
| No | 279 (57.8%) |
| Yes | 164 (34.0%) |
| Missing | 40 (8.3%) |
| Practice Characteristics | |
| Proportion of Racial/Ethnic Minority Patients | |
| Almost None | 4 (0.8%) |
| A Few | s12 (2.5%) |
| Some | 95 (19.7%) |
| A Large Portion | 309 (64.0%) |
| Nearly All | 41 (8.5%) |
| Missing | 22 (4.6%) |
| Clinic Location | |
| Rural | 60 (12.4%) |
| Suburban | 121 (25.1%) |
| Small City | 130 (26.9%) |
| Large City | 152 (31.5%) |
| Missing | 20 (4.1%) |
| Practice Type | |
| Academic | 177 (36.6%) |
| Solo Practice | 32 (6.6%) |
| Group Practice | 187 (38.7%) |
| Government | 7 (1.4%) |
| Hospital | 32 (6.6%) |
| Missing | 48 (9.9%) |
Column percentages for individual response categories are depicted in parentheses.
Table 2:
Disparities Knowledge Measures
| Measure | Response/Level | Total |
|---|---|---|
|
Individual Quiz Items | ||
| 1: Racial/ethnic disparities in U.S. preterm birth rates disappear after accounting for socio-economic disparities. (Correct answer: False) | Correct | 362 |
| Incorrect | 38 (8.8%) | |
| Unsure | 32 (7.4%) | |
| 2: Black women are more likely to have had a Pap smear within the past 3 years than white women. (Correct answer: True) | Correct | 17 (3.9%) |
| Incorrect | 391 | |
| Unsure | 24 (5.6%) | |
| 3: Black women are more likely to be diagnosed with cervical cancer than white women. (Correct answer: True) | Correct | 356 |
| Incorrect | 36 (8.3%) | |
| Unsure | 40 (9.3%) | |
| 4: Maternal death rates are consistent across racial and ethnic groups in the United States. (Correct answer: False) | Correct | 414 |
| Incorrect | 5 (1.2%) | |
| Unsure | 13 (3.0%) | |
| 5: White women are the most likely racial/ethnic group to experience infertility. (Correct answer: False) | Correct | 202 |
| Incorrect | 80 (18.5%) | |
| Unsure | 150 | |
| 6: Studies show implicit bias (subconscious judgments or beliefs not explicitly recognized) has very little effect on disparities in healthcare outcomes. (Correct | Correct | 379 |
| Incorrect | 14 (3.2%) | |
| Unsure | 39 (9.0%) | |
| Overall Quiz Score | 0/6 Correct | 2 (0.4%) |
| 1/6 Correct | 3 (0.6%) | |
| 2/6 Correct | 22 (4.6%) | |
| 3/6 Correct | 93 (19.3%) | |
| 4/6 Correct | 164 | |
| 5/6 Correct | 140 | |
| 6/6 Correct | 8 (1.7%) | |
| Self-Reported Disparities Knowledge | Very Little knowledge | 5 (1.0%) |
| Minimal Knowledge | 37 (7.7%) | |
| Somewhat | 262 | |
| Very Knowledgeable | 139 | |
Column percentages are depicted in parentheses.
Regression results modeling overall quiz performance as a function of background characteristics, experience and training, practice characteristics, and self-reported knowledge are depicted in Table 3. Self-reported knowledge was associated with higher quiz scores, and supplementary mediation analyses indicated that receiving disparities education mediated approximately 11% of this effect.
Table 3:
Quiz Score Regression Results
| Model 1 | Model 2 | Model 3 | Model 4 | |
|---|---|---|---|---|
| Background Characteristics | ||||
| Female Gender (Ref. Male) | 0.0291 | −0.108 | −0.148 | −0.157 |
| (0.103) | (0.113) | (0.111) | (0.110) | |
| Race (Ref. White) | ||||
| Black | 0.280* | 0.257* | 0.305** | 0.232* |
| (0.109) | (0.109) | (0.110) | (0.110) | |
| Asian | 0.268 | 0.234 | 0.234 | 0.266 |
| (0.203) | (0.200) | (0.196) | (0.193) | |
| Ethnicity (Ref. Non-Hispanic) | ||||
| Hispanic/Latinx | −0.281 | −0.343 | −0.347 | −0.284 |
| (0.245) | (0.241) | (0.235) | (0.248) | |
| Experience and Training | ||||
| Years Since Residency (Ref. 10 or fewer) | ||||
| 11–30 years | −0.195 | −0.143 | −0.199 | |
| (0.102) | (0.105) | (0.104) | ||
| >30 years | −0.242 | −0.157 | −0.218 | |
| (0.158) | (0.159) | (0.156) | ||
| Disparities Education Received (Ref. No) | 0.343*** | 0.273** | 0.192* | |
| (0.0927) | (0.0911) | (0.0942) | ||
| Practice Characteristics | ||||
| Proportion of Minority Patients | 0.0803 | 0.0664 | ||
| (0.0751) | (0.0753) | |||
| Clinic Location (Ref. Large City) | ||||
| Rural | 0.184 | 0.183 | ||
| (0.159) | (0.155) | |||
| Suburban | 0.0376 | 0.0520 | ||
| (0.128) | (0.126) | |||
| Small City | −0.102 | −0.0701 | ||
| (0.117) | (0.115) | |||
| Practice Type (Ref. Academic Medicine) | ||||
| Solo | −0.527* | −0.541* | ||
| (0.237) | (0.234) | |||
| Group | −0.337** | −0.311** | ||
| (0.109) | (0.109) | |||
| Hospital | −0.0191 | 0.00707 | ||
| (0.166) | (0.151) | |||
| Government | 0.257 | 0.264 | ||
| (0.254) | (0.276) | |||
| Self-Reported Knowledge | ||||
| Self-Reported Disparities Knowledge | 0.270*** | |||
| (0.0709) | ||||
| Observations | 483 | 483 | 483 | 483 |
Main effects are depicted. Robust standard errors are shown in parentheses.
p<0.001,
p<0.01,
p<0.05
Self-identified African American or Black providers received higher quiz scores than their white counterparts. The mean number of correct quiz answers among African American or Black providers was 5.22 (SD=0.91), while the mean number of correct answers among white providers was 4.95 (SD=0.97). Regression results indicate that this difference is significant at the p<.05 level and is robust to other factors that may explain it (Model 4). Figure 1 illustrates that self-identified African American or Black providers also have higher self-rated disparities knowledge than other racial groups, and a supplementary t-test indicated that this difference is statistically significant in the bivariate at the p<.05 level. Supplementary mediation analyses indicated that approximately 26% of the difference between African American/Black and White providers in overall quiz performance was mediated by self-reported knowledge and receiving formal disparities training.
Figure 1:
Self-Reported Disparities Knowledge by Race (percentages)
Providers working in solo or group practice settings scored lower on the knowledge assessment relative to those practicing in academic settings. Supplementary mediation analyses indicated that approximately 24% of the difference between providers working in group settings and in academic settings was mediated by self-reported knowledge and receiving formal disparities training.
There were no significant differences between providers practicing in government or hospital settings, relative to those practicing in academic settings.
DISCUSSION
We conducted an online survey of OB/GYN providers who are members of ACOG District IV and found the majority of respondents felt they were “very” or “somewhat” knowledgeable about racial and ethnic disparities in reproductive outcomes. The majority of participants reported they had no formal training about racial and ethnic disparities. The mean score on the six-question knowledge assessment was 4 out of 6 questions correct. Provider race, self-reported knowledge, having previously received disparities education, and type of clinical practice were associated with differences in quiz performance.
The current study adds to a very sparse body of literature examining knowledge about racial and ethnic disparities in reproductive outcomes. Similar to our findings, a study with a similar research design of members of the Society for Maternal Fetal Medicine found that most respondents considered themselves at least somewhat knowledgeable about disparities, over 90% of respondents indicated that they were willing to learn more, and most respondents performed well on a knowledge assessment.9 Prior studies evaluating OB/GYN provider’s knowledge about disparities in women’s cardiovascular health found that African American providers were more knowledgeable about disparities in women’s cardiovascular outcomes. 12 Similarly, we found black/African American providers rated their self-assessed knowledge higher and had improved quiz performance compared to others. Studies suggest increased healthcare utilization, satisfaction, and trust among racial/ethnic minority patients when their medical provider is of a concordant race.13–15 There are likely multiple social and cultural reasons for this phenomenon. Our findings raise questions for further studies to explore, such as whether higher levels of knowledge about healthcare disparities among racial/ethnic minority providers is associated with improved outcomes among their racial/ethnic minority patients.
Despite the importance and potential implications of our findings, they must be viewed in light of several limitations. We believe that our sampling frame and the possibility of selection bias may have led to an overestimation of disparities knowledge compared to the general population of women’s health providers. Our survey was distributed to providers within ACOG District IV and represents a specific geographic region of the US. These providers may not be representative of the entire population of US OB/GYN providers. To qualify for ACOG membership, a provider must complete an application and be in good standing with ABOG and their practicing institution. Providers who elect to join ACOG also receive regular updates on clinical practice and current recommendations including publications about disparities in reproductive outcomes. Additionally, ACOG membership is geared toward obstetricians and gynecologists while many woman’s health providers are general practitioners or within other subspecialties who may be even less knowledgeable and comfortable about disparities in women’s health. Surveys of this type are often subject to response bias, such that the group that elects to participate likely represents a skewed sample that may have stronger feelings and greater knowledge about disparities in reproductive outcomes. Findings should therefore be taken as suggestive of patterns that warrant further investigation of the wider population.
IMPLICATIONS
As the US population becomes more diverse it is increasingly important that OB/GYN providers are knowledgeable about disparities among their patient populations and on the forefront of ameliorating unjust differences in outcomes based on race and ethnicity. This study advances the state of research on healthcare provider knowledge of disparities by assessing obstetric and gynecologic providers’ self-assessed knowledge of disparities, examining the extent to which their self-assessed disparities knowledge corresponded to performance on a disparities knowledge quiz assessment, and investigating systematic differences in these knowledge measures across sociodemographic characteristics of providers. Our findings provide important insights into the current state of knowledge surrounding disparities in reproductive outcomes among healthcare providers in the field.
Supplementary Material
ACKNOWLEDGEMENTS
C. W. Mueller acknowledges support from a National Institute on Aging Training Grant (T32-AG000029).
Footnotes
DECLARATIONS OF INTEREST
Declarations of Interest: none.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
REFERENCES
- 1.Center for Disease Control and Prevention. Alzheimer’s Disease and Healthy Aging. https://www.cdc.gov/aging/disparities/index.htm. Accessed September 15, 2018.
- 2.Nelson A. Unequal treatment: confronting racial and ethnic disparities in health care. J Natl Med Assoc. 2002;94(8):666–668. [PMC free article] [PubMed] [Google Scholar]
- 3.ACOG Committee Opinion No. 649: Racial and Ethnic Disparities in Obstetrics and Gynecology. Obstet Gynecol. 2015;126(6):e130–134. [DOI] [PubMed] [Google Scholar]
- 4.Grobman WA, Bailit JL, Rice MM, et al. Racial and ethnic disparities in maternal morbidity and obstetric care. Obstet Gynecol. 2015;125(6):1460–1467. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Louis JM, Menard MK, Gee RE. Racial and ethnic disparities in maternal morbidity and mortality. Obstet Gynecol. 2015;125(3):690–694. [DOI] [PubMed] [Google Scholar]
- 6.Howell EA, Zeitlin J. Quality of Care and Disparities in Obstetrics. Obstet Gynecol Clin North Am. 2017;44(1):13–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Huddleston HG, Cedars MI, Sohn SH, Giudice LC, Fujimoto VY. Racial and ethnic disparities in reproductive endocrinology and infertility. Am J Obstet Gynecol. 2010;202(5):413–419. [DOI] [PubMed] [Google Scholar]
- 8.United States Census Bureau. Projections of the Size and Composition of the U.S. Population: 2014 to 2060. 2015; https://www.census.gov/library/publications/2015/demo/p25-1143.html. Accessed September 15, 2018. [Google Scholar]
- 9.Moroz L, Riley LE, D’alton M, et al. SMFM Special Report: Putting the “M” back in MFM: Addressing education about disparities in maternal outcomes and care. American journal of obstetrics and gynecology. 2018;218(2):B2–B8. [DOI] [PubMed] [Google Scholar]
- 10.Allison PD. Missing data. Thousand Oaks, CA: Sage; 2002. [Google Scholar]
- 11.Stata Statistical Software: Release 15 [computer program]. College Station, TX: StataCorp LLC; 2017. [Google Scholar]
- 12.Jones KM, Carter MM, Schulkin J. Racial and Ethnic Disparities in Cardiovascular Disease: An Assessment of Obstetrician-Gynecologists’ Knowledge, Attitudes, and Practice Patterns. J Racial Ethn Health Disparities. 2015;2(2):256–266. [DOI] [PubMed] [Google Scholar]
- 13.LaVeist TA, Nuru-Jeter A, Jones KE. The association of doctor-patient race concordance with health services utilization. J Public Health Policy. 2003;24(3–4):312–323. [PubMed] [Google Scholar]
- 14.Boulware LE, Cooper LA, Ratner LE, LaVeist TA, Powe NR. Race and trust in the health care system. Public Health Rep. 2003;118(4):358–365. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Cooper LA, Roter DL, Johnson RL, Ford DE, Steinwachs DM, Powe NR. Patient-centered communication, ratings of care, and concordance of patient and physician race. Ann Intern Med. 2003;139(11):907–915. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.