B

Occs1

S

Women’s Health Issues 15 (2005) 31–38

CP

UNDERSTANDING PHYSICIANS’ ATTITUDES TOWARDSHORMONE THERAPY

Rachel Hess, MD, MSca,b,*, Chung Chou Joyce Chang, PhDa,Joseph Conigliaro, MD, MPHa,b, and Melissa McNeil, MD, MPHa

aDivision of General Internal Medicine, University of Pittsburgh, Pittsburgh, PennsylvaniabCenter for Health Equity Research and Promotion, VA Pittsburgh Health Care System, Pittsburgh, Pennsylvania

Received 5 February 2004; received in revised form 20 April 2004; accepted 1 October 2004

Objective: We sought to understand the relationship among components of residencyeducation about hormone therapy (HT), knowledge about HT, and provider attitudes towardHT during a time of rapidly changing practice guidelines.

Methods: We surveyed residents in the University of Pittsburgh Internal Medicine residencyprograms between February to April 2002 (after the release of the Heart Estrogen/ProgestinReplacement Study and prior to the release of preliminary Women’s Health Initiative data)regarding demographics, educational (didactic and experiential) exposures to HT and meno-pause management, knowledge about HT, and attitudes toward HT.

Results: Sixty-nine of 92 (75%) eligible residents completed the survey; 38% were women. Therace and gender of responders did not differ from nonresponders. Residents had significantdidactic exposure to HT and menopause management with 80% reporting more than onedidactic exposure. Despite this, HT knowledge was low (mean knowledge score 47 � 16%)and only 26% of residents felt prepared to counsel patients about HT. We identified fourfactors related to provider attitudes toward HT: “persistence” in universally recommendingHT, confidence in “HT benefits,” concern about “HT cardiac risks,” and concern about “HTnoncardiac risks.” More appropriate attitudes were associated with attending a lecture, havinga rotation with a discussion of menopause management (i.e., Women’s Health), and acontinuity practice including more than 30% women. Pharmaceutical detailing and self-directed study were associated with less appropriate attitudes. Knowledge did not influenceattitudes. Strongly held beliefs about the benefits of HT, appropriate or inappropriate, wereassociated with increasing “persistence.”

Conclusions: In an area of rapidly changing information, such as the risks and benefits of HT,knowledge is low. Experiential learning appropriately influences attitudes, while pharma-

ceutical detailing was associated with inappropriate attitudes toward HT risks.

GtHwccur1G1

ackground

ver the past 5 years, indications for the use ofhormone therapy (HT) have dramatically

hanged. Before 1998, HT was widely accepted as botharioprotective and beneficial to bone density (David-on et al., 2000; The Writing Group for the PEPI Trial,995; The Writing Group for the PEPI Trial, 1996;

* Correspondence to: Rachel Hess, MD, MSc, 230 McKee Place,uite 600, Pittsburgh, PA 15213.

SE-mail: hessr@umpc.edu

opyright © 2004 by the Jacobs Institute of Women’s Health.ublished by Elsevier Inc.

rodstein et al., 2000; Scuteri et al., 2001). As a result,he majority of postmenopausal women were offered

T in an effort to prevent disease. Potential benefitsere assumed to be greater than risks, which included

oncerns for increases in the incidence of breast can-er, endometrial cancer in women with an intactterus, and thromboembolic complications (Collabo-ative Group on Hormonal Factors in Breast Cancer,997; The Writing Group for the PEPI Trial, 1995;rady et al., 2000; Women’s Health Initiative, 2002). In

998, the Heart Estrogen/progestin Replacement

tudy (HERS) called into question the use of HT for

1049-3867/05 $-See front matter.doi:10.1016/j.whi.2004.10.002

taiqiSWtggcaucaH

apashimbpogfsacbroafHeHtf

rHtctaawwtscdt

tatt

M

SWUtrtaanthPinhcrptrcrtem

oc

SAotraf

otwtespd

R. Hess et al. / Women’s Health Issues 15 (2005) 31–3832

he secondary prevention of cardiac disease (Hulley etl., 1998) and a reanalysis of data from one of thenitial cohort studies, the Nurses’ Health Study, alsouestioned the benefits of HT in women with preex-

sting cardiac disease (Grodstein, Manson, &tampfer, 2001). As a result of these studies and theomen’s Health Initiative (Women’s Health Initia-

ive, 2002), released after our research was conducted,uidelines for menopause management have under-one scrutiny and change. The impact of these initialhanges in the understanding of HT on physician’sttitudes about HT in menopausal management isncertain. We sought to understand the process ofhange in physician decision making as knowledgend guidelines evolved during the period betweenERS and WHI.Three previous studies have examined physicians’

ttitudes toward HT, two of which were conductedrior to the release of the 1998 HERS data (Newton etl., 2001; Nikolajevic-Sarunac et al., 1999). In the firsttudy, among almost 300 physicians in a staff-modelealth maintenance organization in the state of Wash-

ngton, Newton and colleagues found that HT wasore likely to be prescribed by female physicians and

y physicians with a positive attitude toward theerceived benefits of HT (Newton et al., 2001). In thether, Nikolajevic-Sarunac and associates surveyed aroup of physicians in Newcastle, Australia andound that these physicians were more likely to pre-cribe HT to women with higher risks of osteoporosisnd coronary artery disease and lower risks of breastancer suggesting a strong belief in the cardiac andone benefits of HT and a concern about breast cancerisk (Nikolajevic-Sarunac et al., 1999). After the releasef the HERS data, a survey of the primary authors ofrticles related to the use of HT for cardioprotectionound that one-third of the authors still believed thatT was beneficial for this purpose despite growing

vidence to the contrary (Rozenberg, Fellemans, &am, 2001). Recent evidence also shows that prescrip-

ions for HT are declining (Hersh, Stefanick, & Staf-ord, 2004).

The use of HT has continued to evolve with theelease of preliminary data from the WHI (Women’sealth Initiative, 2002). Given the changing informa-

ion about the risks and benefits of HT, as well as thehanging clinical guidelines, we sought to reexaminehe relationship among aspects of residency educationddressing HT, knowledge about the use of HT, andttitudes toward HT. We hypothesized that physiciansho had more accurate knowledge about HT useould have more appropriate attitudes: positive atti-

ude about HT benefits (alleviation of menopausalymptoms and prevention of osteoporosis) and moreoncern about HT risks (increased incidence of heartisease, breast cancer, and thromboembolic complica-

ions). Education would enhance appropriate atti- i

udes. We conducted a survey to test this hypothesisnd to explore the influence that demographics, didac-ic training, and clinical experience have on attitudesoward HT.

ethods

tudy participants and sitese surveyed the internal medicine residents at theniversity of Pittsburgh Medical Center (UPMC) be-

ween February 2002 and April 2002, prior to theelease of the Women’s Health Initiative. UPMC haswo residency programs. The university program isffiliated with an urban teaching hospital and includesVeterans Administration hospital, while the commu-ity program is based at a smaller community hospi-

al. Residents in both programs spend time at allospitals. Within each program, the University ofittsburgh offers various residency training tracks,

ncluding a traditional categorical program and threeontraditional tracks: a primary care track; a women’sealth track; and a combined, 4-year, internal medi-ine-pediatrics track. Categorical and primary careesidents can be in either the university or communityrogram, whereas all women’s health residents are in

he university program and all medicine-pediatricsesidents are in the community program. The primaryare, women’s health, and internal medicine-pediat-ics residents spend extra time in the outpatient set-ing and participate to a varying degree in specificducational programs on topics of contraception andenopause management.During the period of the survey, UPMC had a total

f 92 second-, third-, and fourth-year internal medi-ine residents.

tudy proceduresfter the Institutional Review Board of the Universityf Pittsburgh approved our study, we invited each ofhe second-, third-, and fourth-year internal medicineesidents to participate in the survey. Because of itsnonymous nature, the survey was exempt from in-ormed consent.

We approached residents in person prior to theirutpatient continuity clinics or sent them a survey viahe UPMC internal mail system. We provided them

ith written information about the study and gavehem a copy of the survey instrument to completeither during the clinic session or within 1 week ofurvey receipt. We asked them to return the com-leted survey at the site of administration or to aesignated office location. There was no monetary

ncentive for completing the survey.

STerqrirswWcdHpuwiepstl

o(rHBGtMHM2EttMaeItftt

tsewivcsao

“es

SWctlcabde

e“cnffm(rTtehr

paWftirioadsts

gsttlscuwsf

R. Hess et al. / Women’s Health Issues 15 (2005) 31–38 33

tudy instrumenthe survey consisted of four domains: (1) questionsliciting demographic data on characteristics of theesident and his or her chosen residency track; (2)uestions concerning the types of training that theesident had received about menopause management,ncluding didactic (i.e., lectures, workshops, self-di-ected study, and detailing by pharmaceutical repre-entatives) and experiential (e.g., clinic rotations inhich menopause and HT were discussed, such asomen’s Health, and number of female patients in

ontinuity setting); (3) questions evaluating the resi-ent’s knowledge about menopause management andT; and (4) questions about attitudes toward meno-ause management and HT. During the study, wesed the term “hormone replacement therapy (HRT),”hich was the more accepted phrase at that time. We

ncluded pharmaceutical detailing in the domain ofducation because other authors have found thathysicians view pharmaceutical representatives as aource, and sometimes the primary source, of educa-ion regarding medications (Jones, Greenfield, & Brad-ey, 2001; McCormick et al., 2001).

Prior to writing the knowledge assessment portionf the survey, we reviewed the current literaturethrough December 2001) on HT use, HT benefits, HTisks, contraindications to HT use, and alternatives toT use (Collaborative Group on Hormonal Factors inreast Cancer, 1997; Cauley et al., 1995; The Writingroup for the PEPI Trial, 1995; The Writing Group for

he PEPI Trial, 1996; Grodstein et al., 2000; Grodstein,anson, & Stampfer, 2001; Herrington et al., 2000;odis et al., 2001; Hulley et al., 1998; McNagny, 1999;osca et al., 2001; Schairer et al., 2000; Scuteri et al.,

001; Torgerson & Bell-Syer, 2001; Viscoli et al., 2001).ach of our questions was based on one or more of

hese topics and required a single best answer. Ques-ions followed a true-false or multiple-choice format.

ultiple-choice questions included questions aroundclinical vignette. Members of the Section on Wom-

n’s Health (a section of the Division of Generalnternal Medicine in the Department of Medicine) athe University of Pittsburgh reviewed the questionsor content and clarity. Additional faculty with exper-ise in survey design provided input regarding ques-ion structure and layout.

We received permission from members of the Seat-le Group Health Cooperative to use a modified ver-ion of an instrument that they had developed toxamine attitudes influencing the frequency withhich HT is prescribed (Newton et al., 2001). The

nstrument addressed attitudes toward HT use in aariety of hypothetical women including those with aomorbid medical condition (e.g., diabetes, hyperten-ion, coronary disease, breast cancer, or osteoporosis)nd/or a family history of heart disease, breast cancer,

r Alzheimer’s disease. The instrument also addressed s

persistence” in continuing HT in a woman whoxpressed reservations. Attitude questions werecored on 4- and 5-point Likert scales.

tatistical analysese used descriptive statistics to examine the data

oncerning demographics, residency track, and didac-ic training about menopause management. To calcu-ate knowledge scores, we divided the number oforrectly answered questions by the total number ofnswered questions for a percent correct score (possi-le range 0–100%). Knowledge scores were normallyistributed. A score of greater than 50% was consid-red “knowledgeable.”

To identify and categorize attitudinal factors influ-ncing HT use, we removed the previously identifiedpersistence” factor (Newton et al., 2001) and thenonducted principal factors analysis. Through exami-ation of the skree plot, we determined that a three-actor solution was the most parsimonious. The threeactors include: “HT benefits” (beliefs about symptom

anagement and osteoporosis); “HT cardiac risks”beliefs about cardiac protection); and “HT noncardiacisks” (beliefs about thrombosis and breast cancer).ests of reliability indicated adequate internal consis-

ency among all scales, with “persistence,” “HT ben-fits,” “HT cardiac risks,” and “HT noncardiac risks”aving Cronbach � levels of .84, .82, .71, and .71,espectively.

“Persistence” and attitudes were examined for ap-ropriateness based on information about the risksnd benefits of HT, as understood between HERS andHI. Given the changing balance of risks and bene-

its, it became clear that a universal recommendationo use HT is no longer appropriate. It is appropriate tondividualize prescribing to include concern about theisks of HT, while considering useful aspects of HTncluding management of vasomotor symptoms andsteoporosis risks. We therefore consider appropriatettitudes to include: nonpersistence (an ability to in-ividualize prescribing recommendations); highercores on the “HT benefits” scale; and lower scores onhe “HT cardiac risks” and “HT noncardiac risks”cales.

We considered “persistence” to represent a surro-ate for prescribing practices. To examine the relation-hip among demographics, residency track, didacticraining, knowledge, “persistence,” and the three atti-udinal factors, we initially conducted univariableinear regression analyses. Subsequently we con-tructed multivariable linear regression models in-luding any variable that had an � level � .25 innivariable testing. Variables in multivariable modelsere considered significant at an � level � .05. In a

imilar manner, we examined how our attitudinalactors affected the self-reported “persistence” of re-

pondents in recommending HT. The analysis was

cC

R

Ssdg.tm(csla

ilstsoebehH

po

kspiLt

wm(4riatrbora3pt

UWd

T

T

D

T

E

T

N

P

T

C

G

R

P

R

Y

P*nc

R. Hess et al. / Women’s Health Issues 15 (2005) 31–3834

onducted using STATA, version 7.0 (STATA Corp.,ollege Station, Texas).

esults

ixty-nine of the 92 residents (75%) completed theurvey. Gender and race of the respondents did notiffer significantly from those of the overall residentroup (62% vs. 58% male and 42% vs. 39% white; p �

55 and p � .71, respectively). Similar to the composi-ion of the UPMC residency program composition,

ost residents were in traditional categorical program79% vs. 21%) and more were in university thanommunity hospitals (59% vs. 41%). Because of theurvey’s anonymous nature, we were unable to calcu-ate the difference in response rate based on the site ofdministration (clinic versus mail; Table 1).

Residents had significant exposure to didactic train-ng in menopause management, with 99% attending aecture on the topic, 72% engaging in self-directedtudy, 30% reporting exposure to pharmaceutical de-ailing, and 22% attending a half-day workshop oreminar. Eighty percent of residents had more thanne didactic exposure (46% two exposures, 22% threexposures, and 12% four exposures). Increasing num-ers of exposures did not significantly impact knowl-dge (p � .55). Although 75% of residents reportaving a clinical rotation during which menopause or

able 1. Demographic characteristics

haracteristic Number Percentage

enderMale 43 62Female 26 38

aceWhite 29 42Black 3 4Hispanic 3 4Asian/Pacific Islander 28 41Other 6 9

rimary hospital locationUniversity 41 59Community 28 41

esidency track*Traditional: categorical 54 79Nontraditional

Primary care 7 10Women’s health 3 5Internal medicine and pediatrics 4 6

ear in trainingPGY-2 35 51PGY-3 31 45PGY-4 3 4

GY, Postgraduate year.Traditional residency track: categorical internal medicine training;ontraditional residency track (i.e.; women’s health, internal medi-ine-pediatrics, primary care).

T was discussed, most residents (66%) reported that

ostmenopausal women accounted for less than 30%f their outpatient population (Table 2).Despite this significant didactic exposure, overall

nowledge was poor, with an average knowledgecore of 47 � 16% SD. Only 26% of residents feltrepared to counsel women about HT (score of �2,

ndicating feeling somewhat prepared, on a 4-pointikert scale). Residents with more knowledge tended

o feel more prepared to counsel women (p � .058).Overall, despite low knowledge and low comfortith prescribing, residents were persistent in recom-ending HT to women who expressed reservations

average score of 2.9, indicating likely to persist, on a-point Likert scale). When residents were asked toate their beliefs on a 5-point Likert scale, in which 1ndicated strongly disagree and 5 indicated stronglygree with the use of HT in populations of women,hey tended to disagree with the use of HT in patientsepresented by the “HT noncardiac risks” (high risk ofreast cancer or thrombosis) questions (average scoref 2.1), to agree with the use of HT for patientsepresented by the “HT benefits” (e.g., symptom man-gement and osteoporosis) questions (average score of.6), and to feel uncertain about the use of HT foratients represented by the “HT cardiac risks” ques-

ions (average score of 3). Please see Table 3.

nivariable analysese examined 12 variables that might influence resi-

ents’ “persistence” and attitudes: the knowledge

able 2. Educational exposures: exposure present

ype of Exposure Number Percentage

idacticLecture 68 99Self-study 50 72Pharmaceutical detailing 21 30Half-day workshop addressing 15 22Menopause and HT

otal number of didactic exposures0 1 11 13 192 32 463 15 224 8 12

xperientialPostmenopausal women � 30%

of continuity practice 23 34Clinical rotation discussing

menopause and HT 52 75rackTraditional 54 79Nontraditional 14 21umber of half-days in clinical

practice1 43 622 26 38

rimary hospital siteUniversity 41 59

Community 28 41

sm(vcm(mwn

ptwrtd

eacspudmtnmsb(r“t

Hylpa

mc.

MIdst“i“Hrpwpscpic

tdsmtntw.

Tat

A

P

H

H

**a

TrrF

C

PH

H

H

Hlm*

R. Hess et al. / Women’s Health Issues 15 (2005) 31–38 35

core (knowledge about HT and menopause manage-ent); two resident-specific demographic variables

gender and race of respondent); four didactic trainingariables (lectures, self-directed study, pharmaceuti-al detailing, and workshops on HT and menopauseanagement); and five experiential training variables

traditional vs. nontraditional residency track, pri-ary hospital site, percentage of postmenopausalomen seen in the outpatient setting, rotations, and

umber of half-days in the outpatient clinic).In univariable analysis, knowledge and race of

rovider were not significantly related to “persis-ence” or any of the attitudinal factors. Male gender

as associated with less concern about the noncardiacisks (thrombosis and breast cancer) as measured byhe “HT noncardiac risks” scale (p � .002), but genderid not influence the other scales.Didactic and experiential training had mixed influ-

nces on “persistence” and attitudinal factors. Vari-bles associated with more appropriate attitudes in-luded attending a lecture (less “persistence,” p � .02),eeing a patient population with more than 30%ostmenopausal women (more disagreement with these of HT in patients represented by the “HT noncar-iac risks” scale, p � .001), and having a rotation inenopause management (more disagreement with

he use of HT in patients represented by the “HToncardiac risks” scale, p � .046). Variables with aixed profile included self-directed study (more “per-

istence,” p � .047 but more agreement with the “HTenefits” scale, P � .07) and pharmaceutical detailingmore agreement with the use of HT in patientsepresented by both the “HT benefits” scale and theHT cardiac risks” scale, p � .07 and p � .03, respec-ively).

We then examined the impact of attitudes towardT on self-reported “persistence.” In univariable anal-

sis, we found feeling increasingly positive (moreikely to use HT for beneficial effects as well as inatients with cardiac and noncardiac risks) on all

able 3. Attitudes about HT risks and benefits and persistence inecommending HT among second-, third-, and fourth-yearesident physicians at the University of Pittsburgh,ebruary–April 2002

haracteristic Mean � SD Range

ersistence in recommending HT* 2.9 � 0.50 1.80–4.00T noncardiac risks (e.g., thrombosis,breast cancer) 2.1 � 0.67 1.0–4.00T benefits (e.g., menopausalsymptoms, osteoporosis) 3.6 � 0.46 2.75–4.67T cardiac risks 3.0 � 0.78 1.25–5.00

T cardiac and noncardiac risks: 1 � most agree with risks, 5 �east agree with risks; HT benefits: 1 � least agree with benefits; 5 �

ost agree with benefits.Persistence: 1 � least persistent, 4 � most persistent.

ttitudinal scales was, as expected, associated with r

ore “persistence” (“HT benefits” p � .001, “HTardiac risks” p � .047, and “HT noncardiac risks” p �07).

ultivariable analysesn multivariable analyses, we controlled for respon-ent race and gender. The results were similar to thoseeen in univariable analyses. More appropriate atti-udes were linked with: 1) attending a lecture (lesspersistence,” p � .003); 2) having a rotation thatncludes a discussion of menopause management (lesspersistence” and more disagreement with the use ofT in patients represented by the “HT noncardiac

isks” scale, p � .04 for each); and 3) seeing a patientopulation with more than 30% postmenopausalomen (more disagreement with the use of HT inatients represented by the “HT noncardiac risks”cale, p � .002). Less appropriate attitudes were asso-iated with: 1) self-directed study (more “persistence,”� .003) and 2) pharmaceutical detailing (more will-

ngness to use HT in patients represented by the “HTardiac risks” scale, p � .03; Table 4).

Finally, we examined the relationship between atti-udes (HT benefits, HT cardiac risks, and HT noncar-iac risks) and “persistence,” our surrogate for pre-cribing practices, in a multivariable linear regressionodel. More agreement with the use of HT for pa-

ients described on both the “HT benefits” and “HToncardiac risks” scales continued to predict “persis-

ence,” while attitudes on the “HT risks cardiac” scaleere no longer significant (p � .001, p � .04, and p �

48, respectively).

able 4. Variables significantly associated with HT attitudesmong 69 second-, third-, and fourth-year resident physicians athe University of Pittsburgh, February–April 2002*

ttitude** Exposure

ExposurePresent

(Mean LikertValue)

p Value forDifferenceYes No

ersistence in recommending HTLecture 2.9 4.0 .003Clinical rotation discussing

menopause and HT 2.8 3.0 .040Self-study 3.0 2.7 .003T noncardiac risks (e.g.,

thrombosis, breast cancer)Clinical rotation discussing

menopause and HT 2.0 2.4 .040Postmenopausal women �

30% of continuity practice 1.7 2.3 .002T cardiac risksPharmaceutical detailing 3.3 2.9 .030

Controlling for gender and race.*Persistence: 1 � least persistent, 4 � most persistent; HT cardiacnd noncardiac risks: 1 � most agree with risks, 5 � least agree with

isks.

D

Sbfrrowraccisk

mwuaaoibsprocsn

2lf“fteap

ipprpasrctm2fs

iprtiat

eFdictpttvtwocppnetnpttppwtidmrtofiars

a“rvtdtfi

R. Hess et al. / Women’s Health Issues 15 (2005) 31–3836

iscussion

tudies have shown that the use of HT is influencedy physicians’ attitudes and beliefs about HT’s bene-its and risks (Newton et al., 2001). After the 1998elease of HERS (Hulley et al., 1998), when researchesults began to question the cardioprotective benefitsf HT, guidelines for menopause management under-ent serious scrutiny. Subsequently, organizations

eversed prior recommendations and counseledgainst prescribing HT for secondary prevention ofardiac disease (The North American Menopause So-iety, 2000). The timing of our research offers annteresting insight into the process of change of phy-icians’ attitudes about HT in a time of changingnowledge and recommendations.In our survey of residents at university and com-unity hospitals, we hypothesized that respondentsho had more educational exposure to HT and morep-to-date knowledge about HT would have a moreppropriate attitudes toward HT benefits (appreciatelleviation of menopausal symptoms and preventionf osteoporosis) and risks (worry about increased

ncidence of heart disease, breast cancer, and throm-oembolic complications). However, we found noignificant association between knowledge and appro-riateness of attitude. The lack of association may be aesult of generally low knowledge in our population,r it may be a reflection of physicians’ uncertaintyoncerning HT, with higher levels of knowledge re-ulting in higher levels of uncertainty. Our study doesot shed light on the cause for this lack of association.Although Newton and colleagues (Newton et al.,

001) found that male physicians were about 30% asikely as female physicians to prescribe HT, we did notind that gender played a large role in influencingpersistence” or attitudes about HT. Perhaps male andemale physicians do have different prescribing prac-ices but not different attitudes, or perhaps emergingvidence is making female physicians feel more neg-tive toward HT and will change their prescribingractices in the future.In our study reported here, only experiential train-

ng that involved direct patient contact (seeing aatient population with more than 30% postmeno-ausal women in the outpatient setting and a clinicalotation, such as Women’s Health, in which meno-ause management was discussed) had a significantssociation with appropriate attitudes. This findingupports the American Board of Internal Medicineequirement of 25% gender-specific care as a necessaryomponent of residency training to ensure compe-ence in the care of women patients (Program Require-

ents for Residency Education in Internal Medicine,002). In our previous research (Hess et al., 2004) weound an association between other experiential expo-

ures (such as being in a nontraditional internal med- p

cine residency training track, such as women’s health,rimary care, and joint internal medicine and pediat-ics) and absolute knowledge scores. It appears thathese other experiential exposures are less importantn shaping attitudes and “persistence” than is having

significant continuity experience with women pa-ients.

We found that didactic training had a greater influ-nce on attitudes. Three associations were significant.irst, physicians exposed to organized academic en-eavors, such as lectures, tended to be less persistent

n recommending HT and more likely to expressoncerns about using HT in patients with risks iden-ified by the “HT noncardiac risks” scale. Second,hysicians who engaged in self-directed study tended

o be more persistent and more apt to respond posi-ively on the “HT benefits” scale. This second obser-ation may reflect the fact that people are more likelyo independently pursue knowledge about a topic in

hich they already have a strong or positive interestr that in an area of rapidly changing knowledge, theyonsulted outdated materials. Third, physicians ex-osed to pharmaceutical detailing about HT felt moreositive about the use of HT for both its well-recog-ized benefits and its now disputed cardioprotectiveffects. Our study was conducted before the release ofhe Women’s Health Initiative results, which cast aegative light on the use of HT for primary cardiacrotection (Women’s Health Initiative, 2002) but after

he release of the HERS results, which showed thathere was no benefit to using HT for secondary cardiacrotection (Hulley et al., 1998). Industries generallylace a positive spin on the use of their products, so itould not be unusual for the pharmaceutical industry

o place a positive spin on the use of medications thatt is trying to sell. It is unclear if pharmaceuticaletailing has a different impact on residents than onore experienced physicians or if our finding in

esidents would hold true in other physician popula-ions. In any case, medical educators should take notef and utilize the types of techniques that are success-ully employed by the pharmaceutical industry tonfluence attitudes. Additionally, educators shouldcknowledge the powerful influence of industry onesident physicians and work both to minimize expo-ure and correct misconceptions.

Finally, agreeing with the use of HT in women withnd without risk factors was linked with increasedpersistence” in recommending HT despite patienteservations. Either overvaluing the benefits or under-aluing the risks may result in this effect. We believehat the attitudinal factors we identified reflect resi-ents’ current perception of the HT literature. Addi-

ionally, “persistence” may be an adequate surrogateor prescribing practices (Newton et al., 2001). Thismplies that extremely positive attitudes, if inappro-

riate, will result in inappropriate prescribing pat-

tkptm

uefoStscbtahtstuaWmoHortsp

oeai

APMHSfP

RC

C

D

G

G

G

H

H

H

H

H

J

M

M

M

N

N

P

R. Hess et al. / Women’s Health Issues 15 (2005) 31–38 37

erns and that educational efforts in residency trainingnown to influence attitudes, such as appropriateatient mix and pharmaceutical style educational

echniques, should be utilized by programs to addressisconceptions.Our study has several limitations. First, the study

ses a new survey instrument. The Section of Wom-n’s Health at UPMC reviewed all survey questionsor content validity and clarity prior to use, but somef the questions still contained prominent distracters.econd, while other studies have reported that posi-ive attitudes about HT correlate with physician pre-cribing practices, our study does not explore theseorrelations. We use “persistence,” a factor identifiedy other investigators as related to prescribing prac-ices, as a surrogate. Third, our study was conductedt a time of great uncertainty about HT, and this mayave influenced the residents’ willingness to be ex-

remely positive or extremely negative in their re-ponses. Fourth, the study was conducted in a rela-ively small sample of residents who work inniversity and community hospitals and clinics affili-ted with a single health system. The Section ofomen’s Health at UPMC is quite visible, and thisay decrease differences among groups of residents

r may improve the residents’ overall knowledge ofT and women’s health. Residents are just one subset

f physicians, and more experienced physicians mayespond differently to survey questions. To increasehe generalizability of our study, we plan to repeat theurvey in a second, larger population of practicinghysicians.We believe that the results of our study shed light

n attitudes toward HT and that they may also helplucidate the processes by which attitudes are formednd influenced in response to rapidly changing med-cal information.

cknowledgmentsresented in part at the North American Menopause Societyeeting, September 2003. During the project period, Dr.ess’s salary was supported by a Veteran’s Administration

pecial Topics Fellowship in Women’s Health. Additionalunding for supplies was provided by the University ofittsburgh’s Division of General Internal Medicine.

eferencesauley, J. A., Seeley, D. G., Ensrud, K., Ettinger, B., Black, D., &

Cummings, S. R. (1995). Estrogen replacement therapy andfractures in older women. Study of Osteoporotic Fractures Re-search Group. Annals of Internal Medicine, 122(1), 9–16.

ollaborative Group on Hormonal Factors in Breast Cancer. (1997).Breast cancer and hormone replacement therapy: collaborativereanalysis of data from 51 epidemiological studies of 52,705women with breast cancer and 108,411 women without breast

cancer. Lancet, 350(9084), 1047–1059.

avidson, M. H., Maki, K. C., Marx, P., Maki, A. C., Cyrowski, M. S.,Nanavati, N., et al. (2000). Effects of continuous estrogen andestrogen-progestin replacement regimens on cardiovascular riskmarkers in postmenopausal women. Archives of Internal Medicine,160(21), 3315–3325.

rady, D., Wenger, N. K., Herrington, D., Khan, S., Furberg, C.,Hunninghake, D., et al. (2000). Postmenopausal hormone therapyincreases risk for venous thromboembolic disease. The Heart andEstrogen/progestin Replacement Study. Annals of Internal Medi-cine, 132(9), 689–696.

rodstein, F., Manson, J. E., Colditz, G. A., Willett, W. C., Speizer,F. E., & Stampfer, M. J. (2000). A prospective, observational studyof postmenopausal hormone therapy and primary prevention ofcardiovascular disease. Annals of Internal Medicine, 133(12), 933–941.

rodstein, F., Manson, J. E., & Stampfer, M. J. (2001). Postmeno-pausal hormone use and secondary prevention of coronaryevents in the nurses’ health study. a prospective, observationalstudy. Annals of Internal Medicine, 135(1), 1–8.

errington, D. M., Reboussin, D. M., Brosnihan, K. B., Sharp, P. C.,Shumaker, S. A., Snyder, T. E., et al. (2000). Effects of estrogenreplacement on the progression of coronary-artery atherosclero-sis. New England Journal of Medicine, 343(8), 522–529.

ersh, A. L., Stefanick, M. L., & Stafford, R. S. (2004). National useof postmenopausal hormone therapy: annual trends and re-sponse to recent evidence. Journal of the American Medical Associ-ation, 291(1), 47–53.

ess, R., Chang, C. C., Conigliaro, J., Elnicki, D. M., & McNeil, M.(2004). Experiential learning influences residents knowledgeabout hormone replacement therapy. Teaching and Learning inMedicine, 16(3), 240–246.

odis, H. N., Mack, W. J., Lobo, R. A., Shoupe, D., Sevanian, A.,Mahrer, P. R., et al. (2001). Estrogen in the prevention ofatherosclerosis. A randomized, double-blind, placebo-controlledtrial. Annals of Internal Medicine, 135(11), 939–953.

ulley, S., Grady, D., Bush, T., Furberg, C., Herrington, D., Riggs,B., et al. (1998). Randomized trial of estrogen plus progestin forsecondary prevention of coronary heart disease in postmeno-pausal women. Heart and Estrogen/progestin ReplacementStudy (HERS) Research Group. Journal of the American MedicalAssociation, 280(7), 605–613.

ones, M. I., Greenfield, S. M., & Bradley, C. P. (2001). Prescribingnew drugs: qualitative study of influences on consultants andgeneral practitioners. British Medical Journal, 323(7309), 378–381.

cCormick, B. B., Tomlinson, G., Brill-Edwards, P., & Detsky, A. S.(2001). Effect of restricting contact between pharmaceutical com-pany representatives and internal medicine residents on post-training attitudes and behavior. Journal of the American MedicalAssociation, 286(16), 1994–1999.

cNagny, S. E. (1999). Prescribing hormone replacement therapyfor menopausal symptoms. Annals of Internal Medicine, 131(8),605–616.

osca, L., Collins, P., Herrington, D. M., Mendelsohn, M. E.,Pasternak, R. C., Robertson, R. M., et al. (2001). Hormonereplacement therapy and cardiovascular disease: a statement forhealthcare professionals from the American Heart Association.Circulation, 104(4), 499–503.

ewton, K. M., LaCroix, A. Z., Buist, D. S., Anderson, L. A., &Delaney, K. (2001). What factors account for hormone replace-ment therapy prescribing frequency? Maturitas, 39(1), 1–10.

ikolajevic-Sarunac, J., Henry, D. A., O’Connell, D. L., & Robertson,J. (1999). Effects of information framing on the intentions offamily physicians to prescribe long-term hormone replacementtherapy. Journal of General Internal Medicine, 14(10), 591–598.

rogram Requirements for Residency Education in Internal Medi-cine. (2002). Chicago, IL: Accreditation Council for Graduate

Medical Education.

R

S

S

T

T

T

T

W

V

A

be

ahl

as

H

R. Hess et al. / Women’s Health Issues 15 (2005) 31–3838

ozenberg, S., Fellemans, C., & Ham, H. (2001). Opinion surveytowards hormone replacement therapy in the prevention ofcoronary heart disease. Maturitas, 38(3), 273–277.

chairer, C., Lubin, J., Troisi, R., Sturgeon, S., Brinton, L., & Hoover,R. (2000). Menopausal estrogen and estrogen-progestin replace-ment therapy and breast cancer risk. Journal of the AmericanMedical Association, 283(4), 485–491.

cuteri, A., Bos, A. J., Brant, L. J., Talbot, L., Lakatta, E. G., & Fleg,J. L. (2001). Hormone replacement therapy and longitudinalchanges in blood pressure in postmenopausal women. Annals ofInternal Medicine, 135(4), 229–238.

he North American Menopause Society. (2000). A decision tree forthe use of estrogen replacement therapy or hormone replacementtherapy in postmenopausal women: consensus opinion of TheNorth American Menopause Society. Menopause, 7(2), 76–86.

he Writing Group for the PEPI Trial. (1995). Effects of estrogen orestrogen/progestin regimens on heart disease risk factors inpostmenopausal women. The Postmenopausal Estrogen/Proges-tin Interventions (PEPI) Trial. Journal of the American MedicalAssociation, 273(3), 199–208.

he Writing Group for the PEPI Trial. (1996). Effects of hormonetherapy on bone mineral density: results from the postmeno-pausal estrogen/progestin interventions (PEPI) trial. Journal of theAmerican Medical Association, 276(17), 1389–1396.

orgerson, D. J., & Bell-Syer, S. E. (2001). Hormone replacementtherapy and prevention of nonvertebral fractures: a meta-analy-sis of randomized trials. Journal of the American Medical Associa-

tion, 285(22), 2891–2897.

a

omen’s Health Initiative. (2002). Risks and benefits of estrogenplus progestin in healthy postmenopausal women: principalresults From the Women’s Health Initiative randomized con-trolled trial. Journal of the American Medical Association, 288(3),321–333.

iscoli, C. M., Brass, L. M., Kernan, W. N., Sarrel, P. M., Suissa, S.,& Horwitz, R. I. (2001). A clinical trial of estrogen-replacementtherapy after ischemic stroke. New England Journal of Medicine,345(17), 1243–1249.

uthor DescriptionsRachel Hess has an interest in the relationship

etween menopause and quality of life and in residentducation in women’s health issues.

Chung Chou Joyce Chang has provided statisticalnd methodological support for many projects andas a particular interest in modeling survival and

ongitudinal outcomes.Joseph Conigliaro has an interest in substance abuse

nd changing provider practice. He has expertise inurvey methodology.

Melissa McNeil is the Section Chief of Women’sealth in the Division of General Internal Medicine

nd has an interest in medical education at all levels.