Morgan, Maria A. PhD; Lawrence, Hal III MD; Schulkin, Jay PhD
The appropriate scope of medical care provided by obstetrician–gynecologists is frequently debated, particularly in regard to health care reform. The issue encompasses what services obstetrician–gynecologists should provide outside of reproductive health care, and what services they are willing and qualified to provide. Obstetrician–gynecologists provide the majority of non-illness visits for women of childbearing age, but these visits typically address reproductive health, whereas non-illness visits to general practitioners are more likely to address nonreproductive topics.1 A sample of recent obstetrics and gynecology residency graduates was ambivalent regarding whether primary care should remain part of residency education, and one-quarter indicated their primary care training was inadequate.2 Regarding patients' needs, women who saw a gynecologist were more likely to receive a range of preventive care services than were those who saw a generalist,3 and receiving routine nonobstetric care from both a generalist and obstetrician–gynecologist was associated with more comprehensive care than seeing a generalist alone.4
The current study examined how obstetrician–gynecologists define well-woman care within obstetrics and gynecology, what they include in periodic well-woman care visits, and how they manage patients with acute and chronic conditions within and outside of reproductive health. We also looked at how factors such as sex, age, and how they define well-woman care are associated with opinions and practices.
MATERIALS AND METHODS
A questionnaire on practices and opinions related to scope of practice was developed by the research department at the American Congress of Obstetricians and Gynecologists (ACOG). Questions were developed in consultation with practicing obstetrician–gynecologists. The questionnaire was pilot-tested on a sample of practicing obstetrician–gynecologists and was adjusted accordingly before final distribution. The study was approved by the ACOG Institutional Review Board.
The survey was sent to 1,000 ACOG fellows. Of these participants, 600 were members of the Collaborative Ambulatory Research Network (CARN). The Network members are ACOG fellows and junior fellows in practice who have volunteered to participate in survey studies on a regular basis without compensation; they are typically recruited through advertising or random selection from ACOG's membership rolls. The Collaborative Ambulatory Research Network was established to improve the response rate on ACOG Research Department survey studies while maintaining a participant pool representative of practicing ACOG members. The remaining 400 participants consisted of a computer-generated random sample of ACOG fellows and junior fellows in practice who had not received a survey from ACOG during the previous 2 years (group non-CARN). The survey was administered by a mixed-mode method; both e-mail and paper mail were used to distribute the survey. Of the total sample of 1,000 physicians, all of those for whom we had a valid e-mail address (n=875) were sent an e-mail in mid May 2009. The e-mail included information about the study, a link to the survey, and a password unique to each participant that they would need to log-on to the electronic survey. Four reminder e-mails were sent to nonresponders. Paper mailings, which included a cover letter, a questionnaire, and a stamped return envelope, were sent July 16, 2009, to any of the original 875 participants who had not yet responded and to those for whom we did not have a valid e-mail address on record (n=125). Those who did not respond to the paper mailing were sent one reminder (including cover letter, questionnaire, and envelope) via paper mailing on August 11, 2009. Electronic responses received after July 2009 and paper surveys received after mid October 2009 were not included in analyses. The 1,000 participants had a mean age of 50 (range 31–83 years).
Survey responses were anonymous and provided demographic details of physicians and their patient population. Questions assessed respondents' opinions about what constitutes well-woman care, what they include in a well-woman examination, how they manage chronic and acute conditions that patients present with, and how they perceive nonpregnant patients to be utilizing their services. We avoided using the term “primary care provider” to minimize bias attributable to reference to the structure of the health care and insurance systems. Because we were interested primarily in the practices that comprise routine well-woman care and the range of practices that providers of such care engage in, the data of respondents who do not provide routine gynecologic care are not included in the final analyses.
Question formats included multiple choice, yes/no, check all that apply, Likert-like scale, and complete the blank. For three questions, participants were presented with a statement and asked to indicate their level of agreement on a 5-point scale, with 1 being strongly disagree and 5 being strongly agree, with 3 indicating a neutral point. When participants were asked how they were most likely to manage nonpregnant patients presenting with 1 of 17 medical conditions, they were provided with the response options of: counsel, prescribe medication, refer to a primary care physician, refer to a specialist, watch and wait, or other; they were asked to select only one treatment per condition. For one set of analyses, we categorized treatment methods into two groups: personally treating (combining counsel, prescribe, and watch and wait responses) or referring (combining refer to primary care physician and refer to specialist responses).
The data were analyzed using a personal computer- based software package (SPSS 16.0; SPSS). Descriptive statistics were computed for the measures used in the analyses and reported as mean±standard error of the mean. Student t test was used to compare group means of continuous variables. Analysis of variance was used to compare group means of continuous variables by sex, with age as a covariate. Differences on categorical measures were assessed using χ2. Group differences on ordinal measures were assessed using the Mann-Whitney U test. Age, when used as a categorical variable, was divided into two groups: those aged less than or equal to the mean respondent age of 51 years and those aged 52 years or older. We used binomial logistic regression with how respondents define well-woman care as a dichotomous dependent variable; age (continuous) was included as a covariate to check for the possibility that the relationship between the independent and dependent variables varied as a function of age. Analyses were tested for significance using an alpha of 0.05, except when indicated.
A total of 579 questionnaires was returned. Questionnaires from 15 respondents were judged invalid (physician retired, returned blank questionnaire), resulting in a valid response rate of 57% (564 of 985). There were responding physicians from all ACOG Districts except District X (Armed Forces), including from the District of Columbia and from every state of the United States except Montana. Respondents' mean age (51±0.43) closely matched that of the population to whom the survey was sent (50±0.34). Men and women did not differ significantly in response rates (women=59%, 291/497, men=56%, 273 of 488, P=.408), and 52% of those returning a survey were female. Of the 564 valid responses, 327 (58%) were in an electronic format and 237 (42%) were in paper format. The final analyses presented in the study are based on the responses of the 513 participants (91% of respondents) who provide routine gynecologic care. Respondents who provide only obstetric care (n=24) and those who provide neither routine obstetric nor routine gynecologic care (n=27) were not included in further analyses. Characteristics of respondents included in the study are presented in Table 1.
The Collaborative Ambulatory Research Network had a higher response rate than non-CARN (65% compared with 45% of all participants) and were older (average of 2 years), but they did not differ by sex (P=.18), practice type (P>.60), practice location (P>.56), proportion of patient visits devoted to different types of care (P>.46), patient race (P>.52), patient age proportions (P>.25), patient insurance types (P>.21), or the proportion of nonpregnant patients they believe have them as their primary care physician (P>.60). Before conducting final analyses, we compared CARN and non-CARN members on all nondemographic questions to determine if their responses differed. They differed on four questions: CARN members were less likely than non-CARN members to say they would refer nonpregnant patients with irritable bowel syndrome to a primary care physician as the most likely course of management (25% compared with 42%; P<.01); CARN members were more likely to agree (4 and 5 on a 5-point scale) that well-woman care is a high priority in their workload (80% compared with 70%; Mann-Whitney U = 24,248; P<.01); and CARN members were more likely to include the thyroid/neck (95% compared with 88%; P<.01) and lipids (83% compared with 71%; P<.01) in a periodic well-woman examination. Because of the limited number of differences and the clinical insignificance of those differences, data are collapsed across membership groups throughout the results.
Doctors were asked several questions about how they view the practice of obstetrics and gynecology and how their patients utilize their services. Asked how they define well-woman care within the context of obstetric and gynecologic practices, 61% of respondents said it was “medical care related to the overall health of the asymptomatic woman/primary care,” and 39% said it was “medical care related to reproductive health of the asymptomatic woman.” One-third of respondents (33%) agreed (4 and 5 on a 5-point scale) and 44% disagreed (1 and 2 on a 5-point scale) that obstetrician–gynecologists should limit their medical care to issues directly related to women's reproductive health. More than one-third (35%) of respondents said they would prefer to spend less time devoted to patient issues not directly related to reproductive health or obstetrics, whereas 11% said more time, and 54% said the same as now (note: because of a technical glitch, this question was asked only of the 219 participants who responded to the questionnaire in a paper format). Three-fourths of respondents agreed (77%) that well-woman care is a high priority in their workload, and the majority (82%) agreed that women they see typically stay under their care throughout their reproductive years. Twenty-one percent of respondents did not know what proportion of their nonpregnant patients had them as their primary physician; of those providing a response, almost half (47%) estimated that more than 30% of their nonpregnant patients see them as their primary physician. Women were more likely than men to agree that well-woman care is a high priority in their workload (81% compared with 71%; P<.01), independent of age. Younger respondents (51 years or younger) were more likely to agree that obstetrician–gynecologists should limit care to reproductive health (40% compared with 25%; P<.001), independent of sex.
Participants were presented with a list of procedures and asked to indicate which ones they include in an age-appropriate well-woman examination (Table 2). Although almost all include breast (100%), pelvic (99%), and abdominal (97%) examinations, few examine the oral cavity (19%). Most (89%) indicated that they screen for sexually transmitted diseases (STDs), although far fewer (69%) screen for human immunodeficiency virus (HIV). Participants were then asked how many times per month, on average, nonpregnant patients present with 1 of 17 medical conditions, some of which are clearly within the domain of reproductive health (eg, menopausal, STDs), some are common general health issues (eg, obesity, high blood pressure), and some are acute medical conditions (eg, earache, skin lesions; Table 3, column 2). Respondents reported encountering obesity and menopausal issues most frequently and asthma and earaches least frequently. Analysis of variance of condition by sex, with age as a covariate, showed that participants differed by sex on a number of conditions. Female doctors reported seeing more cases of high cholesterol, obesity, menopausal issues, STDs (all P<.005), as well as more cases of migraine, skin lesions, interest in hormone therapy (all P<.05), and musculoskeletal issues (P=.05). When age was a factor, it tended to be the older female physicians who reported a greater number of cases (menopausal issues, interest in hormone therapy, obesity).
Doctors were asked how they were most likely to manage a patient with these conditions (Table 3, last 6 columns). For those conditions most clearly within the domain of reproductive health and, other than obesity, accounting for the most frequently reported visits, most respondents were likely to prescribe medications (urinary tract infection 95%, STDs 91%, interest in hormone therapy 73%, menopausal 69%); for many respondents, prescribing medications was also a first line of treatment for migraines (45%), earaches (42%), and generalized anxiety disorder (39%). For other conditions, they were most likely to refer the patient to a primary care physician (high blood pressure 73%, high cholesterol 65%, diabetes 61%, asthma 59%, cardiovascular 57%, earache 47% [also prescribe], musculoskeletal 42%) or to a specialist (skin lesions 58%, major depressive disorder 50%, irritable bowel syndrome 34%). Sixty-two percent would most likely counsel an obese patient. When we categorized treatment methods into two groups (personally treating or referring), we found that female respondents, when they differed from men (obesity, high blood pressure, high cholesterol, skin lesions, earache [P<.01], generalized anxiety disorder, irritable bowel syndrome [P<.05]), were more likely to refer patients than were men, except for obesity, which women were more likely to personally treat (80% compared with 64%; P<.001). Respondents aged 51 or younger were more likely, when they differed from older respondents (high cholesterol, generalized anxiety disorder, skin lesions, major depressive disorder [P<.01], high blood pressure, earache [P<.05]), to refer than were older respondents, except for major depressive disorder, which younger respondents were more likely to treat (43% compared with 26%; P<.001). The sex effect of increased referral by women was largely independent of age, although the difference was more pronounced in the younger group for management of earaches, skin lesions, and high cholesterol.
When doctors do refer nonpregnant patients to another doctor, do they request a report if one is not automatically sent to them? More than one-quarter (28%) of respondents said yes, they always do, 45% said it depended on the type or severity of the disorder, and 27% said not typically. One-fifth (22%) of doctors always contact the patient to discuss test results from a medical referral, and 17% typically do not. The remainder selected various conditions under which they contact the patient, such as a positive test result (20%), the disorder type (33%) or severity (13%), or the patient's level of concern (12%). Doctors aged 52 or older were more likely to say they always contact the patient (29% compared with 16%; P<.001).
We were interested in whether respondents' practices are associated with how they define well-woman care. Do those who define it as “medical care related to the overall health of the asymptomatic woman/primary care” (overall health group 61%) differ from those who define it as “medical care related to reproductive health of the asymptomatic woman” (reproductive health group 39%) in their management of conditions outside of the strictly reproductive health domain (eg, earaches, mood disorders, high blood pressure)? The overall health group estimated a greater percent of their patient appointments as periodic well-woman care appointments than did the reproductive health group (mean of 31%±0.95% compared with 27%±1.16%; t=2.73; P<.01). The overall health group was less likely to agree that obstetrician–gynecologists should limit their medical care to issues directly related to women's reproductive health (15% compared with 62%; Mann-Whitney U=10,369; P<.001), was more likely to agree that well-woman care is a high priority in their workload (82% compared with 69%; Mann-Whitney U=22,348; P<.001), and was somewhat more likely to agree that women they see typically stay under their care throughout their reproductive years (84% compared with 77%; Mann-Whitney U=25,769; P<.01). More than two-fifths (44%) of reproductive health respondents (who were presented with the question) said they would prefer to spend less time devoted to patient issues not directly related to reproductive health or obstetrics, compared with 30% of overall health respondents (P<.05). Further, 55% of the overall health group, compared with 32% of the reproductive health group, indicated that at least 30% of their nonpregnant patients have them as their primary physician (t=4.84; P<.001).
Those who define well-woman care as overall health were significantly more likely to say they include a number of items in an age-appropriate periodic well-woman examination, such as a skin survey (overall health 81%, reproductive health 62%), a mouth examination (overall health 22%, reproductive health 13%), and a number of laboratory tests and vaccinations (Table 2). Regarding management of various conditions with which patients present, doctors in the reproductive health group were more likely, when they differed (obesity, high blood pressure, high cholesterol, generalized anxiety disorder, musculoskeletal, migraines, irritable bowel syndrome, asthma, earache [P<.001], skin lesions, diabetes, major depressive disorder [P<.02]), to refer patients to a primary care physician or specialist than were doctors in the overall health group (Table 3). Finally, when respondents receive patients' test results or a report from a referral, the reproductive health group was somewhat more likely to indicate that they do not typically contact the patient (23% compared with 13%; P<.005).
To describe overall health and reproductive health respondents, we used χ2 tests to differentiate them on a number of demographic characteristics. There was an association between how they define well-woman care and age, with younger respondents (mean age 51 years or younger) more likely than older respondents to define it as reproductive health (44% compared with 34%; P=.017). There was a marginal association with sex, with women more likely than men to define it as reproductive health (43% compared with 35%; P=.055). There was not an association with whether they practice obstetrics and gynecology or gynecology only (P=.743), nor with their type of practice (P=.311), nor with the location of their practice (P=.478). A binary logistic regression analysis using how they define well-woman care as the dichotomous dependent variable and age (continuous) and sex as covariates found no effect of sex (P=.504) and a modest effect of age (P<.05). There was a 3-year mean age difference between the overall health (age 52±.56 years) and reproductive health (age 49±.75 years) groups (t=3.12; P<.005).
The objective of this study was to estimate how obstetrician–gynecologists define well-woman care within the field of obstetrics and gynecology and to examine their opinions and practices regarding the scope of well-woman care. We found that the majority of respondents consider well-woman care to include the overall health of the asymptomatic woman/primary care (61%) rather than care limited to reproductive health (39%). In practice, obstetrician–gynecologists are, predictably, most consistently focusing on reproductive health issues; breast examinations and Pap tests are nearly universally performed during age-appropriate periodic well-woman examinations. Less consistently performed are assessments of nonreproductive organs: less than three-fourths of respondents examine heart and lungs; two-thirds offer flu vaccinations; and one-third measure liver function. This is consistent with findings that although internists and family physicians were more likely to order tests for nonreproductive health issues during well-woman care, obstetrician–gynecologists were more likely to focus on reproductive health.1
Far from the majority of nonpregnant patients were estimated to utilize their obstetrician–gynecologist as their primary physician, although four of five respondents believe their patients typically stay under their care throughout their reproductive years. A 2008 study found that more than 80% of office visits to internists and family physicians were considered visits to the patient's primary care provider, whereas 24% of visits to obstetrician–gynecologists were.5 In a previous study of obstetrician–gynecologists, the majority of respondents reported that most or all of their pregnant patients initially made contact with them once they were pregnant.6 These findings suggest that many women do not establish a routine doctor–patient relationship with their obstetrician–gynecologist before obstetric care.
How doctors define well-woman care distinguishes two groups of doctors who differ regarding the scope of care involved in obstetrics and gynecology. Overall health respondents reported screening for more nonreproductive issues during a well-woman examination. Most respondents manage reproductive health conditions, and nearly all respondents refer patients to another doctor for cardiovascular conditions. For all other conditions, from an acute earache to mental health, overall health respondents were more likely to personally treat the conditions than were reproductive health respondents and were less likely to refer. When age and sex were factors, male and older respondents tended to have an increased likelihood of personally treating nonreproductive conditions. More than one-quarter of respondents said they do not typically request a report from a referral if one is not automatically sent, suggesting a lapse in follow-up and reduced opportunity for coordination of care.
Our study has limitations. The response rate was 57%, indicating a potential nonresponse bias. However, this rate is similar to that of recent ACOG studies,7,8 it is at the high end of the 35% to 60% rate typical of our studies, and responses were derived from diverse geographical locations and practice types. A subset of participants was CARN members, who may differ from the general ACOG membership. However, CARN members are representative of the ACOG membership, and only 1 of 17 CARN studies examined showed more than minor differences between the groups, a difference ultimately attributed to an age effect. When we asked participants their one most likely course of management, it is highly possible that respondents initially counsel and report the course of management they subsequently take. We attempted to minimize the potential for response bias by stressing the anonymity of responses.
The American Congress of Obstetricians and Gynecologists recommends that periodic assessments include age- and condition-based screening, evaluation, and counseling on a range of issues not directly related to reproductive health.9 We did not ask physicians if they clarify for patients whether they are performing a comprehensive well-woman examination or one limited to reproductive health, and if the latter, if they advise patients to establish a relationship with an internist for routine and acute care. The opinions and practices presented in this article suggest that a large subset of obstetrician–gynecologists is receptive to acting as a comprehensive source of overall health care for women. Younger respondents were more likely to express opinions defining well-woman care as limited to reproductive health. This is consistent with findings that recent obstetrics and gynecology residency graduates were ambivalent as to whether primary care should remain a part of residency education.2 Further research is needed to determine whether this indicates a trend away from providing primary care services by newer obstetrician–gynecologists, particularly given the perceived shortage of doctors providing primary care10 and the need to better-define obstetrics and gynecology in regard to current health care reform. It should be made abundantly clear to patients whether their obstetrician–gynecologist views well-woman care as encompassing overall health or as restricted to reproductive health so patients can plan their wellness care accordingly and know where to turn when an acute condition arises.
© 2010 by The American College of Obstetricians and Gynecologists.