- Email: [email protected]
Communicating results of diagnostic mammography: What do patients think?
Communicating Results of Diagnostic Mammography: What Do Patients Think? 1 Kimberly S. Levin, MD, M. Patricia Braeuning, MD, Michael S. O'Malley, PhD, Etta D. Pisano, MD Eileen D. Barrett, MPH, Jo Anne L. Earp, ScD
Rationale and Objectives. The purpose of this study was to investigate women's preferences for who (radiologist or referring physician) should communicate the results of diagnostic mammography. Materials and Methods. Data from 153 women presenting to two sites for diagnostic mammography between Febru-
ary and June 1995 were collected with a 24-item, self-administered, closed-ended survey. For both normal and abnormal hypothetical results, contingency tables with Z2 tests and multiple logistic regression were used to determine the association, if any, between women's characteristics and their preferences. Results. Women undergoing diagnostic mammography preferred that their radiologists disclose their normal (90%) and abnormal (88%) mammogram results to them immediately after their examination, rather than have their referring physicians disclose results at a later time. In the case of normal findings, women whose regular physicians were specialists were less likely to want to hear first from their radiologists (odds ratio [OR] = 0.06; 95% confidence interval [CI] = 0.01, 0.77; P =.03), but women who were nervous about learning their results were more likely to want to hear first from their radiologists (OR = 4.5; 95% CI = 1.2, 17.3; P = .03). Conclusion. Radiologists may want to consider assessing women's preferences for who communicates their mammogram results, as most women in this study preferred to hear these results from their radiologists rather than waiting to hear from their referring physicians. Key Words. Mammography; communication; physician-patient relationship
Timely, accurate, and supportive communication of mammography results is an important component of quality medical care (1-5). Traditionally, the mammographer, and more generally the radiologist, has been considered a "consultant" to the referring physician. Patients are typically sent for specific radiologic examinations at the request of
Acad Radio12000; 7:1069--1076 1From the Department of Health Behavior and Health Education, School of Public Health, Rosenau Hall, CB #7400, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-7400 (K.S.L., E.D.B., J.L.E.); and the Department of Radiology (M.P.B, E.D.P.) and the UNC Lineberger Comprehensive Cancer Center (M.P.B., J.L.E., M.S.O., E.D.P.), School of Medicine, University of North Carolina, Chapel Hill. Supported in part by grants from the UNC Linebergar Comprehensive Cancer Center and the Department of Radiology, School of Medicine, University of North Carolina at Chapel Hill. Received January 18, 2000; revision requested May 15; revision received August 8; accepted August 9. Address correspondence to J.L.E. ©AUR, 2000
their primary physicians. The results of these examinations are relayed to the referring physicians, who then share that information with their patients. Studies have shown, however, that patient compliance with follow-up improves when radiologists communicate results directly and/or schedule follow-up appointments to do so (1,2,4,6,7). Informing patients of their results immediately after the examination also may alleviate anxiety (1,4,8). The Agency for Healthcare Research and Quality quality assurance guidelines recommend that radiologists notify all women in writing of their mammography test results within 10 days of the procedure. These guidelines also recommend immediate oral notification of self-referred women who have an abnormal result (1). In the past, legal opinions suggested that radiologists might be required to inform patients of their test results (9), and, in 1998, with the passage of the Mammography Standards Reauthorization Act (MQSA) by the U.S. Congress, that possible requirement became law
1069
.EVIN ET AL
(10). By requiring that results be communicated to women in writing through the mail, the MQSA minimizes potential communication gaps between mammographer, primary care giver, and patient. The MQSA still leaves room, however, for primary providers to discuss results with women before they receive written notification, so the question of who should first inform patients of their results remains. Given that radiologists are the only health care providers able to offer immediate information in these circumstances, the question of who informs the patient is inextricably tied to the timing of the disclosure. Even prior to passage of the MQSA, some radiologists modified their traditional consultant role and began communicating mammogram results directly to patients immediately after the examination; previous research, however, suggests that women's preferences for who communicates mammogram results are still not entirely clear. Although Liu et al (2) found that more than 90% of the women surveyed preferred that their radiologists disclose results, Lind et al (11) concluded that fewer than 10% of women they surveyed preferred disclosure by the radiologist. Hulka et al (12) found that when offered an informed choice between either immediate communication of results by the radiologist or delayed communication that would benefit from two radiologists' readings of the mammogram, 75% of the women surveyed preferred to wait for their results. Physician-patient communication is known to affect several aspects of health care, including patients' health outcomes (13), compliance (14), perceived severity of ambiguous results (15), readiness to di'sclose illnesses (16), recall of the physician's recommendations (15), perceived anxiety (15,17), physiological anxiety (18,19), and overall satisfaction with care (20). However, we found no previous studies that investigated the relationship between communication preferences and women's characteristics (2). Findings from past research indicate that communication style may also be important to patients when hearing the results of radiologic tests; patients have been shown to prefer that their physicians use supportive, compassionate language (21,22) and that radiologists communicate results cautiously when examinations reveal major abnormalities and refer patients instead to their primary care providers (23). In this study, we examined women's preferences for who communicates their diagnostic mammogram results, the preferred communication style, and the association between women's characteristics and their communication preferences.
1070
Academic Radiology, Vol 7, No 12, December 20(:
Setting and P a r t i c i p a n t s
As part of a larger study including women, radiologists, and primary care physicians, we surveyed women aged 20 years and older at two North Carolina radiology facilities, an outpatient department of a large public teaching hospital (n = 100) and a private clinic (n -- 99). We restricted our sample to women whom the facilities identified as presenting for diagnostic mammography, because radiologists at these sites evaluate diagnostic mammograms immediately and would, therefore, be able to communicate results when the patient is still on-site. Screening mammograms, on the other hand, are usually read at a later time. Our definition of diagnostic mammography included mammography not only in women with current breast problems or suspicious abnormalities but also in women returning for additional evaluations following abnormal results at a screening examination or for a 6-month follow-up. Between February and June 1995, one of us (K.S.L) approached consecutive female patients who had completed diagnostic mammography and were waiting for staff to determine whether any additional views were necessary. Of the 216 women approached, 199 (92%) completed a clinic intake form and a 24-item self-administered questionnaire requiring approximately 10 minutes to complete. Although both clinics were restricted to diagnostic mammography, in fact 46 of 216 women (21%) identified themselves as presenting for routine mammography and were excluded from these analyses, reducing the sample size to 153. Outcomes and Measures
The primary outcomes measured were women's preferences about who should communicate normal and abnormal results of diagnostic mammography to them. For each outcome, participants read a case scenario describing a woman who discovered a lump in her breast and whose physician sent her to undergo diagnostic mammography. In one case, the radiologist assessed the lump as not being cancer and therefore not needing further evaluation. In the second case, the radiologist thought the lump might be cancer. We then asked women what they thought the radiologist should do about communicating the results to the patient. For normal results, women could specify that the radiologist (a) tell the woman tight after mammography that the lump was not cancer or (b) say nothing and send the woman to talk to her regular physician. For abnormal results, women could specify that the radiologist (a) tell the
Table 1 Characteristics of Women in the Study Characteristic
Analysis No. of Women (n = 131)*
Sociodemographics Age > 50 y Race: white Annual family income > $30,000 Education: more than high school Medical history Prior breast cancer Prior breast surgery Family history of breast cancer Medical care use Have a regular physician Only specialist as regular physician Follow-up appointment scheduled with physician Prior mammography experience Previously underwent mammography Previous mammogram was abnormal Radiologist gave results of previous mammogram immediately Waited >2 d for results of previous mammogram Current mammography experience Undergoing mammography at a public teaching hospital Physician sent patient for today's mammography Nervous about results of today's mammography
56 85 66 70 11 43 16 96 7 40
Data from the two sites were aggregated to ensure adequate sample size for analysis; for further analyses, we included site as a variable to control for differences between the two study locations. For each outcome, we calculated the proportion of women who preferred that the radiologist communicate the results to women immediately following mammography. To investigate the association between women's communication preferences and their personal and health characteristics, we used contingency tables with X2 in bivariate analyses. Characteristics at least minimally associated with each outcome (P < .20) were then included in separate multiple logistic regression models to examine the strength of the association while controlling for other characteristics. All analyses were conducted with SPSS software (version 7.5; SPSS, Chicago, Ill).
92 51 18 46
52 90 61
*Total n ranges from 131 to 153 because not all women answered all questions.
woman right after mammography that the lump was abnormal and that further evaluation, possibly a biopsy, was needed; (b) tell the woman fight after mammography that the lump might be cancer; (c) tell the woman right after mammography that the results are abnormal but that she should talk to her physician to find out results; or (d) say nothing and send the woman to her regular physician. For analysis, we collapsed these values into two categories: radiologist communicates some type of assessment or radiologist does not disclose anything to the woman. Response choices one and two were used to define an additional variable, communication style preference. We used the study questionnaire and clinic intake forms to collect information on 17 personal and health characteristics potentially associated with women's preferences for communicating mammography results. Characteristics included sociodemographics (n = 4), medical history (n = 3), medical care use (n = 3), past mammography experiences (n = 4), and current mammography experience (n = 3). For analysis, multilevel variables were collapsed into dichotomous variables.
Women's Characteristics
Women presenting for diagnostic mammography at these two radiology sites tended to be in the middle to upper economic strata, have experienced breast problems, and be under the care of a primary care physician (Table 1). Approximately two-thirds of the women reported more than a high school education and an annual income greater than $30,000 per year. Although a minority (11%) had a prior history of breast cancer, about half had a prior history of breast surgery. Ninety-two percent reported having previously undergone mammography, with 51% saying their last mammogram was abnormal. More than half the women reported being nervous about the pending results of their test. Almost all the women had a regular physician (96%) and said that a physician had referred them for the current mammography (90%); only 40%, however, had a follow-up appointment scheduled with their physician. A minority of women (18%) said that the radiologist gave them their results immediately after the previous mammographic examination. Communication
Preferences
After reading the case scenario for normal results, 90% of women (118 of 131) said that the radiologist should immediately tell the woman her results. Similarly, 88% said that the radiologists should immediately tell the woman about her abnormal results. Most women (83%) were consistent in theft preference for radiologist disclosure of both normal and abnormal results. Only 6% of women preferred that the radiologist disclose neither normal nor abnormal results.
1071
.EVIN ET A L
Academic Radiology, Vol 7, No 12, December 2001
Table 2 Characteristics of Women Who Prefer Immediate Disclosure of Mammographic Results by the Radiologist, by Mammographic Result (n = 130)
Characteristics Sociodemographics Age _<50 y > 50 y Race Not white White Annual family income < $30,000 > $30,000 Education Up to or through high school More than high school Medical history Prior breast cancer Yes No Prior breast surgery Yes No Family history of breast cancer Yes No Medical care use Have a regular physician Yes No Only specialist as regular physician Yes No (at least one primary care physician) Follow:up appointment scheduled with physician Yes No
Normal Mammogram
Abnormal Mammogram
92 87
89 87
82 91
81 89
80* 90
89 85
87 91
93 86
81 90
94 87
86 90
90 84
86 90
86 87
90 100
88 100
67* 91
89 89
84t 93
87 88
(continued)
Notification Preferences
and Women's
Characteristics
For both normal and abnormal test results, the percentage of women preferring immediate notification by the radiologist varied somewhat by women's characteristics, although in all but two areas more than 80% of women in all subgroups said they preferred to hear from the radiologist immediately (Table 2). For notification of normal results, five of 17 characteristics were at least
1072
Table 2 (continued)
Characteristics Sociodemographics Prior mammography experience Previously underwent mammography Yes No Previous mammogram was abnormal Yes No Radiologist gave results of previous mammogram immediately Yes No Waited >2 d for results of previous mammogram Yes No Current mammography experience Radiologic facility Private clinic Public teaching hospital Who sent today Physician Other Nervous about results of today's mammography Yes No
Normal Mammogram
Abnormal Mammogram
89 100
88 92
87 89
84 91
91 87
23* 84
86 90
81' 93
85t 94
82 93*
90 81
89t 81
94* 83
89 86
* P < .05. t p < .10. * P < .01.
minimally associated with preference for radiologist disclosure, including three that were significantly associated (P < .05). Compared with those who named a primary care physician among their regular providers, women whose regular providers were specialists were significantly less likely to prefer radiologist notification (67% vs 91%), as were women with a family income less than $30,000 per year (80% vs 90%). Being nervous about the result of the m a m m o g r a m was associated with preferring radiologist notification (94% vs 83%). Having a scheduled follow-up visit with a regular physician and undergoing mammography at a private clinic were minimally associated with preference for communication by the radiologist.
~cademic Radiology, Vol 7, No 12, December 2000
COMMUNICATING RESULTS OF MAMMOGRAPH~
Table 3 Multiple Logistic Regression Analysis of Characteristics Associated with Women's Preferences for Immediate Disclosure of Normal Mammographic Results by Radiologists Characteristic Only specialist as regular physician Follow-up appointment scheduled with physician Nervous about results of today's mammography Annual family income > $30,000 Undergoing mammography at a public teaching hospital
Odds Ratio*
95% Confidence Interval
P Value
0.06
0,01, 0.77
.03
0.31
0,08, 1.2
.10
4,5 2.8
1.2, 17.3 0,72, 10.6
.03 .14
14.2
1.4, 138.6
.03
* Odds ratios below 1.00 indicate preference for hearing results from the primary care physician. Odds ratios above 1.00 indicate preference for hearing results from the radiologist.
For abnormal results, women's preferences for notification varied less often, with only four of 17 characteristics and test sites at least minimally associated with communication preferences. Having had a radiologist report the results of the previous mammography, having waited 2 days or less for the results of the last mammography, and undergoing mammography at a public teaching hospital were significantly associated (P < .05) with preference for radiologist communication. Having been referred by a physician for the current mammography tended to be associated with preference for communication by the radiologist (P <. 10). Multiple Logistic R e g r e s s i o n A n a l y s i s
For normal results, having only a specialist as a regular physician remained significantly associated with preference for disclosure by the regular physician after controlling for other related characteristics (Table 3). Being nervous about the results of the mammography and undergoing mammography at a public teaching hospital, however, were significantly associated with preference for disclosure by the radiologist. Having a follow-up appointment scheduled with a physician was minimally associated with preference for regular physician notification, while having an annual family income greater than $30,000 was minimally associated with a preference for radiologist notification. For abnormal results, after controlling for other associated characteristics, no characteristics were even minimally associated with a preference for radiologist or referring physician notification because there was not enough variation in women's preferences. Women preferred radiologist notification regardless of their personal and health characteristics.
C o m m u n i c a t i o n Style
Language style was found to be an important concern for patients when hearing their abnormal mammogram results. Most women (114 of 130, 88%) wanted the radiologist to use cautious language, such as "abnormal" or "you will need further evaluation'' when explaining the results, instead of the more direct term, "cancer." DISCUSSION
By considering ways to incorporate patients' preferences into the delivery of health care, physicians may be able to improve patient satisfaction, compliance with recommendations, and, ultimately, the quality and outcomes of care (1,6). This approach may be particularly important as the health care system moves toward a model in which patient participation is valued, or even considered necessary, for optimal care (11,24). New federal guidelines, as outlined in the MQSA of 1998, require that radiologists send patients written results of their mammograms (10), and Agency for Healthcare Research and Quality guidelines suggest that these results be sent within 10 days of the mammogram (1). These guidelines, however, were designed to increase the likelihood that patients were directly informed of their results in a timely mariner; they do not require immediate notification. Given that radiologists are in a position to disclose results immediately, patients may prefer to hear these from their radiologists. In light of the findings of this study, as well as changes in national requirements, radiologists may want to consider revising protocols that will help improve patients' timely receipt of their results.
1073
.EVIN ET A L
According to this study, however, it is important to understand that some women do not want their mammogram results communicated in this way. Even when they strongly suspected the presence of an abnormality and were given a choice of learning this right away, 10% of the patients in this study still wished to wait to hear their results from their referring physician if these results were normal, and 12% if the results were abnormal. For some patients, particularly those who had a specialist as a regular physician, relationships with their referring physicians appeared to override other considerations, such as immediate receipt of results. As radiologists begin to assume a more interactive role with their patients, they should keep this variability in mind. Instead of creating an across-the-board policy, it may be more appropriate for staff members to ask each patient individually for her preference. This study has some limitations that may have affected our results. There may be a situational bias: Had the questionnaire been administered in a referring physician's ofrice, or by mail where patients could have completed it in the privacy of their homes, more women may have favored disclosure by referring physicians. Although such a study design may have resulted in decreased bias, the likely possibility of significantly lower, and possibly systematically different, response rates associated with mailed questionnaires (25) led us to reject this method of data collection. As a result, we may have overestimated women's preferences for hearing results from their radiologists. Even so, the magnitude of the preference for radiologist disclosure was large, and having a follow-up appointment scheduled with the referring physician was only minimally associated with preference for regular physician notification. In addition, our results may not be representative of women with different ethnic, cultural, and geographic backgrounds who present at other breast cancer screening facilities. Finally, because we restricted our study to women undergoing diagnostic mammography, and a high percentage of these women (56%) were younger than 50 years, these results may have limited generalizability for older women who are undergoing routine screening mammography. Our results are consistent with those of most prior studies surveying patients' preferences for who discloses results of radiologic examinations in general, both mammographic and nonmammographic. In one study of 269 patients undergoing gastrointestinal radiologic examinations at one hospital, 253 (94%) patients preferred to know their results immediately from the radiologist (26). Another study of 391 patients undergoing diagnostic ultrasound (US) showed that, when offered the option, 348 (89%) requested a c o n -
1074
Academic Radiology, Vol 7, No 12, December 200C
sultation with a sonographer and, of tliese 348 patients, 327 (94%) preferred hearing results immediately from the sonographer rather than waiting to hear them from their referring physician (27). Our findings, however, conflict with those of Lind et al (11) who found that only 9% (49 of 558) of women preferred that their radiologists tell them their normal mammogram results and only 6% (33 of 546) their abnormal results. Lind et al surveyed women in 1988; it is possible that preferences have changed since then. Perhaps more important, their study, as with most others exploring women's preferences for reporting mammography results, surveyed women undergoing mostly screening mammography, whereas our study included women undergoing only diagnostic mammography. In a more recent study, Hulka et al (12) administered a mail-in survey and found that women chose delayed results when they were given the option of having more than one radiologist read the mammogram. One possible confounding factor, Hulka et al acknowledge, is that women who receive and fill out a survey at home may be more apt to choose delayed communication because they are free from worry or nervousness about cancer at the time they complete the survey. The opposite bias may be at play in the present study because women were surveyed immediately after undergoing mammography. Future research might correct for situational bias by using random assignment to one of multiple sites of survey administration, including a mail-in component. A future study might widen the population surveyed to include women who present not only for diagnostic, but also for screening mammography. Many researchers have argued that delayed communication of results may cause women undue anxiety (1,4,8), implying that the radiologist should be the one who discloses information. In our study, 61% of women were nervous about hearing their results, and nervousness contributed greatly to explaining women's preferences for who disclosed their normal results to them. Other researchers have found a slight patient preference for the radiologist to disclose normal (as opposed to abnormal) results (2,9,11). A policy of disclosing only normal results would not be feasible, however, because radiologists would be conveying information to patients by their silence when results were abnormal. Our study found no difference in preferences by result status. Because many mammography centers now obtain the necessary additional views and perform US examinations and even biopsies at the same visit for patients referred for diagnostic imaging, some radiologists, in effect, give pa-
dents the results whether they actually talk to the women or not. The question then becomes how this information is communicated. Radiologists' language and style may be important considerations in communicating mammography results. For those women in our study who preferred that their radiologists communicate abnormal results to them, most (88%) wanted them to do so by using cautious, instead of more direct, terms. This finding is interesting in light of findings from a study conducted by Song et al (23), who found that patients preferred receiving suggestive answers ("there is an abnormality") from the radiologist rather than a refusal to communicate any information at all. The two studies examined together suggest that women prefer language that adopts a middle ground; they neither wish to have their requests for information delayed nor wish to be told directly that they may have cancer in the event of an abnormal reading. Other studies have shown that patients are sensitive to physicians' communication styles and, not surprisingly, prefer that their physicians use supportive, compassionate language (21,22). Radiologists' training guidelines currently contain no recommendations for teaching skills in "patient interaction" (28). If radiologists are going to provide results immediately, residency programs may need to train radiologists for this activity, including helping them develop supportive and effective, as well as culturally sensitive, communication styles. Past research has shown that physicians can be taught these behaviors and interpersonal styles (3,13), and that effective communication is likely to maximize patient satisfaction and compliance with future radiologic recommendations (29). The results of this study raise several questions. Although the patients were explicitly asked about a radiologist versus a referring physician's disclosure, the question of who discloses results is confounded with the issue of immediacy of disclosure. In light of the passage of the MQSA, which requires only that the radiologist offer written (and therefore delayed) notification of results, it would be useful to tease apart whether it is "immediate" disclosure that is desired rather than disclosure by the radiologist. We were not able to disaggregate these two dimensions in our study. It also would be useful to examine what referring physicians and radiologists think about patients being immediately notified of their results by radiologists. For example, radiologists might prefer more time to prepare for discussing possible outcomes and treatments with patients and so wish to notify the patient later in writing, as is currently mandated. Patients may not benefit from immediate disclosure if radiologists are not comfortable with, or prepared for, this role. Future research could examine the effect of radiologist
communication on patients at the time'of disclosure and at subsequent visits. Our findings may also have implications for cost and scheduling in light of the increased patient contact that may be required of individual radiologists if such a program were implemented. Communicating results is likely to entail more time per patient. Extra time spent by radiologists with patients drives down reimbursement and may lead to the need to increase the cost per mammographic examination to compensate for increased time spent. In addition, unless imaging facilities are set up with rooms suitable for patient-radiologist discussions, valuable examination space may be taken up, further reducing mammography profitability. These and other issues in the changing mammography picture suggest future research on communication preferences among both patients and radiologists may be desirable. Although this research may be limited by the fact that these data were collected in 1995, before new regulations made radiologist disclosure the norm, it is important to understand the attitudes with which patients entered the present era of managed care, if only to counteract arguments based solely on "cost effectiveness." Furthermore, the new regulations do not require that radiologists communicate the results immediately or in person. Future research could compare the preferences presented here with those from patients after more extensive experience by radiologists with the new guidelines. Despite its limitations, this study has several implications for the practice of radiology and, more generally, for the expansion of communication avenues between patient and health care provider. It is reasonable to consider that most women undergoing diagnostic mammography would want to hear their results from the radiologist because they would receive their results sooner than if they heard them directly from their primary care provider or by mail. At the same time, however, radiologists might consider how they respond to those women who wish to reserve this communication for their referring physician. Assessing women's preferences as they present for mammography may be a reasonable approach; improving radiologists' patient interaction skills is likely another. Ideally, a balance can be struck between patients' needs, radiologists' skills, and the resources made available through managed care systems. ~,CKNOWLEDGMENT.
Special thanks to Claire Poyett, MD, for helping make the study possible and to the radiologic technologists for their support. Also thanks to Mary Stuart, MPH, for help with questionnaire design, to Amy Denham, MPH, MD, for
1075
editorial help, and to Anne Menkens, MA, for her revisions to the manuscript. IEFERENCE
1. Bassett LW, Henddck RE, Bassord TL, et al. Quality determinants of mammography: clinical practice guideline no. 13, 1994. Agency for Health Care Policy Research publication no. 95-0632. Rockviile, Md: Agency for Health Care Policy and Research, Public Health Service, U.S. Department of Health and Human Services, 1994. 2. Liu S, Bassett LW, Sayre J. Women's attitudes about receiving mammographic results directly from radiologists. Radiology 1994; 193:783-786. 3. Shapiro DE, Boggs SR, Melamed BG, Graham-Pole J. The effect of varied physician affect on recall, anxiety, and perceptions in women at risk for breast cancer: an analogue study. Health Psycho11992; 11:61-66. 4. Hoffman NY, Janus J, Destounis S, Logan-Young W. When the patient asks for the results of her mammogram, how should the radiologist reply? A JR Am J Roentgeno11994; 162:597-599. 5. Bassett LW, Bomyea K, Liu S, Sayre J. Communication of mammography results to women by radiologists: attitudes of referring health care providers. Radiology 1995; 195:235-238. 6. Cardenosa G, Eklund GW. Rate of compliance with recommendations for additional mammographic views and biopsies. Radiology 1991; 181 : 359-361. 7. D'Orsi C J, Debor MD. Communication issues in breast imaging. Radiol Clin North Am 1995; 33:1231-1245. 8. Levitsky DB, Frank MS, Richardson ML, Shneidman RJ. How should radiologists reply when patients ask about their diagnoses? a survey of radiologists' and clinicians' preferences. A JR Am J Roentgenol 993;161:433-436. 9. Schreiber MH, Leonard M, Rieniets CY. Disclosure of imaging findings to patients directly by radiologists: survey of patients' preferences. A JR Am J Roentgeno11995; 165:467-469. 10. Mammography Quality Standards Reauthorization Act of 1998. Pub L No. 105-248. 112 Star 1984. 11. Lind SE, Kopans D, Good M-JD. Patients' preferences for learning the results of mammographic examinations. Breast Cancer Res Treat 1992; 23:223-232. 12. Hulka CA, Slanetz PJ, Halpern EF, et al. Patients' opinion of mammography screening services: immediate results versus delayed results due to interpretation by two observers. A JR Am J Roentgenol 1997; 168:1085-1089.
13. Kaplan SH, Greenfield S, Ware JE. Assessing the effects of physician-patient interactions on the outcomes of chronic disease. Med Care 1989; 27(suppl 3):$110-$127. 14. DiMatteo MR, Hays RD, Prince LM. Relationship of physicians' nonverbal communication skill to patient satisfaction, appointment noncompliance, and physician workload. Health Psycho11986; 5:581-594. 15. Shapiro DE, Boggs SR, Melamed BG, Graham-Pole J. The effect of varied physician affect on recall, anxiety, and perceptions in women at risk for breast cancer: an analogue study. Health Psycho11992; 11:61-66. 16. Beckman HB, Frankel RM. The effect of physician behavior on the collection of data. Ann Intern Med 1984; 101:692-696. 17. Davison BJ, Degner LF. Empowerment of men newly diagnosed with prostate cancer. Cancer Nurs 1996; 20:187-196. 18. Morris KJ, Tarico VS, Smith WL, Altmaier EM, Franken EA. Critical analysis of radiologist-patient interaction. Radiology 1987; 163:565-567. 19. Smith WL, Altmaier EM, Ross RR, Johnson BD, Berberoglu LS. Patient expectations of radiology in non-interactive encounters. Radiology 1989; 172:275-276. 20. Kaplan SH, Greenfield S, Gandek B, Rogers WH, Ware JE. Characteristics of physicians with participatory decision-making styles. Ann Intern Med 1996; 124:497-504. 21. Morris KJ, Tarico VS, Smith WL, Altmaier EM, Franken EA. Critical analysis of radiologist-patient interaction. Radiology 1987; 163:565-567. 22. Fox SA, Siu AI, Stein JA. The importance of physician communication on breast cancer screening of older women. Arch Intern Med 1994; 154:2058-2088. 23. Song HH, Park SH, Shinn KS. Radiologists' responses to patients' inquiries about imaging results: a pilot study on opinions of various groups. Invest Radio11993; 28:1043-1048. 24. Emanuel LL. The consultant and the patient-physician relationship: a trilateral deliberative model. Arch Intern Med 1994; 154:1785-1790. 25. Dillman DA. Mail and Internet surveys: the tailored design method. New York, NY: Wiley, 2000. 26. Vallely SR, Manton Mills JO. Should radiologists talk to patients? BMJ 1990; 300:305-306. 27. Ragavendra N, Laifer-Narin SL, Melany ML, Grant EG. Disclosure of results of sonographic examinations to patients by sonologists. A JR Am J Roentgeno11998; 170:1423-1425. 28. Stewart DA. The radiologist as communicator: assessments and impressions. Radiology 1995; 196(2):37A-42A. 29. Inui TS, Carter WB. Problems and prospects for health services research on provider-patient communication. Med Care 1985; 23:521-538.
~.nnouncemen The Fourth International Symposium on the Costs a n d Benefits of Radiology will be held with invited participants in San Francisco, Calif, August 23-25, 2001. Individuals interested in presenting related research must submit a 200-word structured abstract (including Objectives, Methods, Results, and Conclusions) based on unpublished radiology cost-effectiveness or outcome studies by December 31, 2000, to Tyro Peters, UCSF Radiology, 3333 California St, Suite 375, San Francisco, CA 94143-0629; e-mail: [email protected].
1076
Rationale and Objectives. The purpose of this study was to investigate women's preferences for who (radiologist or referring physician) should communicate the results of diagnostic mammography. Materials and Methods. Data from 153 women presenting to two sites for diagnostic mammography between Febru-
ary and June 1995 were collected with a 24-item, self-administered, closed-ended survey. For both normal and abnormal hypothetical results, contingency tables with Z2 tests and multiple logistic regression were used to determine the association, if any, between women's characteristics and their preferences. Results. Women undergoing diagnostic mammography preferred that their radiologists disclose their normal (90%) and abnormal (88%) mammogram results to them immediately after their examination, rather than have their referring physicians disclose results at a later time. In the case of normal findings, women whose regular physicians were specialists were less likely to want to hear first from their radiologists (odds ratio [OR] = 0.06; 95% confidence interval [CI] = 0.01, 0.77; P =.03), but women who were nervous about learning their results were more likely to want to hear first from their radiologists (OR = 4.5; 95% CI = 1.2, 17.3; P = .03). Conclusion. Radiologists may want to consider assessing women's preferences for who communicates their mammogram results, as most women in this study preferred to hear these results from their radiologists rather than waiting to hear from their referring physicians. Key Words. Mammography; communication; physician-patient relationship
Timely, accurate, and supportive communication of mammography results is an important component of quality medical care (1-5). Traditionally, the mammographer, and more generally the radiologist, has been considered a "consultant" to the referring physician. Patients are typically sent for specific radiologic examinations at the request of
Acad Radio12000; 7:1069--1076 1From the Department of Health Behavior and Health Education, School of Public Health, Rosenau Hall, CB #7400, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-7400 (K.S.L., E.D.B., J.L.E.); and the Department of Radiology (M.P.B, E.D.P.) and the UNC Lineberger Comprehensive Cancer Center (M.P.B., J.L.E., M.S.O., E.D.P.), School of Medicine, University of North Carolina, Chapel Hill. Supported in part by grants from the UNC Linebergar Comprehensive Cancer Center and the Department of Radiology, School of Medicine, University of North Carolina at Chapel Hill. Received January 18, 2000; revision requested May 15; revision received August 8; accepted August 9. Address correspondence to J.L.E. ©AUR, 2000
their primary physicians. The results of these examinations are relayed to the referring physicians, who then share that information with their patients. Studies have shown, however, that patient compliance with follow-up improves when radiologists communicate results directly and/or schedule follow-up appointments to do so (1,2,4,6,7). Informing patients of their results immediately after the examination also may alleviate anxiety (1,4,8). The Agency for Healthcare Research and Quality quality assurance guidelines recommend that radiologists notify all women in writing of their mammography test results within 10 days of the procedure. These guidelines also recommend immediate oral notification of self-referred women who have an abnormal result (1). In the past, legal opinions suggested that radiologists might be required to inform patients of their test results (9), and, in 1998, with the passage of the Mammography Standards Reauthorization Act (MQSA) by the U.S. Congress, that possible requirement became law
1069
.EVIN ET AL
(10). By requiring that results be communicated to women in writing through the mail, the MQSA minimizes potential communication gaps between mammographer, primary care giver, and patient. The MQSA still leaves room, however, for primary providers to discuss results with women before they receive written notification, so the question of who should first inform patients of their results remains. Given that radiologists are the only health care providers able to offer immediate information in these circumstances, the question of who informs the patient is inextricably tied to the timing of the disclosure. Even prior to passage of the MQSA, some radiologists modified their traditional consultant role and began communicating mammogram results directly to patients immediately after the examination; previous research, however, suggests that women's preferences for who communicates mammogram results are still not entirely clear. Although Liu et al (2) found that more than 90% of the women surveyed preferred that their radiologists disclose results, Lind et al (11) concluded that fewer than 10% of women they surveyed preferred disclosure by the radiologist. Hulka et al (12) found that when offered an informed choice between either immediate communication of results by the radiologist or delayed communication that would benefit from two radiologists' readings of the mammogram, 75% of the women surveyed preferred to wait for their results. Physician-patient communication is known to affect several aspects of health care, including patients' health outcomes (13), compliance (14), perceived severity of ambiguous results (15), readiness to di'sclose illnesses (16), recall of the physician's recommendations (15), perceived anxiety (15,17), physiological anxiety (18,19), and overall satisfaction with care (20). However, we found no previous studies that investigated the relationship between communication preferences and women's characteristics (2). Findings from past research indicate that communication style may also be important to patients when hearing the results of radiologic tests; patients have been shown to prefer that their physicians use supportive, compassionate language (21,22) and that radiologists communicate results cautiously when examinations reveal major abnormalities and refer patients instead to their primary care providers (23). In this study, we examined women's preferences for who communicates their diagnostic mammogram results, the preferred communication style, and the association between women's characteristics and their communication preferences.
1070
Academic Radiology, Vol 7, No 12, December 20(:
Setting and P a r t i c i p a n t s
As part of a larger study including women, radiologists, and primary care physicians, we surveyed women aged 20 years and older at two North Carolina radiology facilities, an outpatient department of a large public teaching hospital (n = 100) and a private clinic (n -- 99). We restricted our sample to women whom the facilities identified as presenting for diagnostic mammography, because radiologists at these sites evaluate diagnostic mammograms immediately and would, therefore, be able to communicate results when the patient is still on-site. Screening mammograms, on the other hand, are usually read at a later time. Our definition of diagnostic mammography included mammography not only in women with current breast problems or suspicious abnormalities but also in women returning for additional evaluations following abnormal results at a screening examination or for a 6-month follow-up. Between February and June 1995, one of us (K.S.L) approached consecutive female patients who had completed diagnostic mammography and were waiting for staff to determine whether any additional views were necessary. Of the 216 women approached, 199 (92%) completed a clinic intake form and a 24-item self-administered questionnaire requiring approximately 10 minutes to complete. Although both clinics were restricted to diagnostic mammography, in fact 46 of 216 women (21%) identified themselves as presenting for routine mammography and were excluded from these analyses, reducing the sample size to 153. Outcomes and Measures
The primary outcomes measured were women's preferences about who should communicate normal and abnormal results of diagnostic mammography to them. For each outcome, participants read a case scenario describing a woman who discovered a lump in her breast and whose physician sent her to undergo diagnostic mammography. In one case, the radiologist assessed the lump as not being cancer and therefore not needing further evaluation. In the second case, the radiologist thought the lump might be cancer. We then asked women what they thought the radiologist should do about communicating the results to the patient. For normal results, women could specify that the radiologist (a) tell the woman tight after mammography that the lump was not cancer or (b) say nothing and send the woman to talk to her regular physician. For abnormal results, women could specify that the radiologist (a) tell the
Table 1 Characteristics of Women in the Study Characteristic
Analysis No. of Women (n = 131)*
Sociodemographics Age > 50 y Race: white Annual family income > $30,000 Education: more than high school Medical history Prior breast cancer Prior breast surgery Family history of breast cancer Medical care use Have a regular physician Only specialist as regular physician Follow-up appointment scheduled with physician Prior mammography experience Previously underwent mammography Previous mammogram was abnormal Radiologist gave results of previous mammogram immediately Waited >2 d for results of previous mammogram Current mammography experience Undergoing mammography at a public teaching hospital Physician sent patient for today's mammography Nervous about results of today's mammography
56 85 66 70 11 43 16 96 7 40
Data from the two sites were aggregated to ensure adequate sample size for analysis; for further analyses, we included site as a variable to control for differences between the two study locations. For each outcome, we calculated the proportion of women who preferred that the radiologist communicate the results to women immediately following mammography. To investigate the association between women's communication preferences and their personal and health characteristics, we used contingency tables with X2 in bivariate analyses. Characteristics at least minimally associated with each outcome (P < .20) were then included in separate multiple logistic regression models to examine the strength of the association while controlling for other characteristics. All analyses were conducted with SPSS software (version 7.5; SPSS, Chicago, Ill).
92 51 18 46
52 90 61
*Total n ranges from 131 to 153 because not all women answered all questions.
woman right after mammography that the lump was abnormal and that further evaluation, possibly a biopsy, was needed; (b) tell the woman fight after mammography that the lump might be cancer; (c) tell the woman right after mammography that the results are abnormal but that she should talk to her physician to find out results; or (d) say nothing and send the woman to her regular physician. For analysis, we collapsed these values into two categories: radiologist communicates some type of assessment or radiologist does not disclose anything to the woman. Response choices one and two were used to define an additional variable, communication style preference. We used the study questionnaire and clinic intake forms to collect information on 17 personal and health characteristics potentially associated with women's preferences for communicating mammography results. Characteristics included sociodemographics (n = 4), medical history (n = 3), medical care use (n = 3), past mammography experiences (n = 4), and current mammography experience (n = 3). For analysis, multilevel variables were collapsed into dichotomous variables.
Women's Characteristics
Women presenting for diagnostic mammography at these two radiology sites tended to be in the middle to upper economic strata, have experienced breast problems, and be under the care of a primary care physician (Table 1). Approximately two-thirds of the women reported more than a high school education and an annual income greater than $30,000 per year. Although a minority (11%) had a prior history of breast cancer, about half had a prior history of breast surgery. Ninety-two percent reported having previously undergone mammography, with 51% saying their last mammogram was abnormal. More than half the women reported being nervous about the pending results of their test. Almost all the women had a regular physician (96%) and said that a physician had referred them for the current mammography (90%); only 40%, however, had a follow-up appointment scheduled with their physician. A minority of women (18%) said that the radiologist gave them their results immediately after the previous mammographic examination. Communication
Preferences
After reading the case scenario for normal results, 90% of women (118 of 131) said that the radiologist should immediately tell the woman her results. Similarly, 88% said that the radiologists should immediately tell the woman about her abnormal results. Most women (83%) were consistent in theft preference for radiologist disclosure of both normal and abnormal results. Only 6% of women preferred that the radiologist disclose neither normal nor abnormal results.
1071
.EVIN ET A L
Academic Radiology, Vol 7, No 12, December 2001
Table 2 Characteristics of Women Who Prefer Immediate Disclosure of Mammographic Results by the Radiologist, by Mammographic Result (n = 130)
Characteristics Sociodemographics Age _<50 y > 50 y Race Not white White Annual family income < $30,000 > $30,000 Education Up to or through high school More than high school Medical history Prior breast cancer Yes No Prior breast surgery Yes No Family history of breast cancer Yes No Medical care use Have a regular physician Yes No Only specialist as regular physician Yes No (at least one primary care physician) Follow:up appointment scheduled with physician Yes No
Normal Mammogram
Abnormal Mammogram
92 87
89 87
82 91
81 89
80* 90
89 85
87 91
93 86
81 90
94 87
86 90
90 84
86 90
86 87
90 100
88 100
67* 91
89 89
84t 93
87 88
(continued)
Notification Preferences
and Women's
Characteristics
For both normal and abnormal test results, the percentage of women preferring immediate notification by the radiologist varied somewhat by women's characteristics, although in all but two areas more than 80% of women in all subgroups said they preferred to hear from the radiologist immediately (Table 2). For notification of normal results, five of 17 characteristics were at least
1072
Table 2 (continued)
Characteristics Sociodemographics Prior mammography experience Previously underwent mammography Yes No Previous mammogram was abnormal Yes No Radiologist gave results of previous mammogram immediately Yes No Waited >2 d for results of previous mammogram Yes No Current mammography experience Radiologic facility Private clinic Public teaching hospital Who sent today Physician Other Nervous about results of today's mammography Yes No
Normal Mammogram
Abnormal Mammogram
89 100
88 92
87 89
84 91
91 87
23* 84
86 90
81' 93
85t 94
82 93*
90 81
89t 81
94* 83
89 86
* P < .05. t p < .10. * P < .01.
minimally associated with preference for radiologist disclosure, including three that were significantly associated (P < .05). Compared with those who named a primary care physician among their regular providers, women whose regular providers were specialists were significantly less likely to prefer radiologist notification (67% vs 91%), as were women with a family income less than $30,000 per year (80% vs 90%). Being nervous about the result of the m a m m o g r a m was associated with preferring radiologist notification (94% vs 83%). Having a scheduled follow-up visit with a regular physician and undergoing mammography at a private clinic were minimally associated with preference for communication by the radiologist.
~cademic Radiology, Vol 7, No 12, December 2000
COMMUNICATING RESULTS OF MAMMOGRAPH~
Table 3 Multiple Logistic Regression Analysis of Characteristics Associated with Women's Preferences for Immediate Disclosure of Normal Mammographic Results by Radiologists Characteristic Only specialist as regular physician Follow-up appointment scheduled with physician Nervous about results of today's mammography Annual family income > $30,000 Undergoing mammography at a public teaching hospital
Odds Ratio*
95% Confidence Interval
P Value
0.06
0,01, 0.77
.03
0.31
0,08, 1.2
.10
4,5 2.8
1.2, 17.3 0,72, 10.6
.03 .14
14.2
1.4, 138.6
.03
* Odds ratios below 1.00 indicate preference for hearing results from the primary care physician. Odds ratios above 1.00 indicate preference for hearing results from the radiologist.
For abnormal results, women's preferences for notification varied less often, with only four of 17 characteristics and test sites at least minimally associated with communication preferences. Having had a radiologist report the results of the previous mammography, having waited 2 days or less for the results of the last mammography, and undergoing mammography at a public teaching hospital were significantly associated (P < .05) with preference for radiologist communication. Having been referred by a physician for the current mammography tended to be associated with preference for communication by the radiologist (P <. 10). Multiple Logistic R e g r e s s i o n A n a l y s i s
For normal results, having only a specialist as a regular physician remained significantly associated with preference for disclosure by the regular physician after controlling for other related characteristics (Table 3). Being nervous about the results of the mammography and undergoing mammography at a public teaching hospital, however, were significantly associated with preference for disclosure by the radiologist. Having a follow-up appointment scheduled with a physician was minimally associated with preference for regular physician notification, while having an annual family income greater than $30,000 was minimally associated with a preference for radiologist notification. For abnormal results, after controlling for other associated characteristics, no characteristics were even minimally associated with a preference for radiologist or referring physician notification because there was not enough variation in women's preferences. Women preferred radiologist notification regardless of their personal and health characteristics.
C o m m u n i c a t i o n Style
Language style was found to be an important concern for patients when hearing their abnormal mammogram results. Most women (114 of 130, 88%) wanted the radiologist to use cautious language, such as "abnormal" or "you will need further evaluation'' when explaining the results, instead of the more direct term, "cancer." DISCUSSION
By considering ways to incorporate patients' preferences into the delivery of health care, physicians may be able to improve patient satisfaction, compliance with recommendations, and, ultimately, the quality and outcomes of care (1,6). This approach may be particularly important as the health care system moves toward a model in which patient participation is valued, or even considered necessary, for optimal care (11,24). New federal guidelines, as outlined in the MQSA of 1998, require that radiologists send patients written results of their mammograms (10), and Agency for Healthcare Research and Quality guidelines suggest that these results be sent within 10 days of the mammogram (1). These guidelines, however, were designed to increase the likelihood that patients were directly informed of their results in a timely mariner; they do not require immediate notification. Given that radiologists are in a position to disclose results immediately, patients may prefer to hear these from their radiologists. In light of the findings of this study, as well as changes in national requirements, radiologists may want to consider revising protocols that will help improve patients' timely receipt of their results.
1073
.EVIN ET A L
According to this study, however, it is important to understand that some women do not want their mammogram results communicated in this way. Even when they strongly suspected the presence of an abnormality and were given a choice of learning this right away, 10% of the patients in this study still wished to wait to hear their results from their referring physician if these results were normal, and 12% if the results were abnormal. For some patients, particularly those who had a specialist as a regular physician, relationships with their referring physicians appeared to override other considerations, such as immediate receipt of results. As radiologists begin to assume a more interactive role with their patients, they should keep this variability in mind. Instead of creating an across-the-board policy, it may be more appropriate for staff members to ask each patient individually for her preference. This study has some limitations that may have affected our results. There may be a situational bias: Had the questionnaire been administered in a referring physician's ofrice, or by mail where patients could have completed it in the privacy of their homes, more women may have favored disclosure by referring physicians. Although such a study design may have resulted in decreased bias, the likely possibility of significantly lower, and possibly systematically different, response rates associated with mailed questionnaires (25) led us to reject this method of data collection. As a result, we may have overestimated women's preferences for hearing results from their radiologists. Even so, the magnitude of the preference for radiologist disclosure was large, and having a follow-up appointment scheduled with the referring physician was only minimally associated with preference for regular physician notification. In addition, our results may not be representative of women with different ethnic, cultural, and geographic backgrounds who present at other breast cancer screening facilities. Finally, because we restricted our study to women undergoing diagnostic mammography, and a high percentage of these women (56%) were younger than 50 years, these results may have limited generalizability for older women who are undergoing routine screening mammography. Our results are consistent with those of most prior studies surveying patients' preferences for who discloses results of radiologic examinations in general, both mammographic and nonmammographic. In one study of 269 patients undergoing gastrointestinal radiologic examinations at one hospital, 253 (94%) patients preferred to know their results immediately from the radiologist (26). Another study of 391 patients undergoing diagnostic ultrasound (US) showed that, when offered the option, 348 (89%) requested a c o n -
1074
Academic Radiology, Vol 7, No 12, December 200C
sultation with a sonographer and, of tliese 348 patients, 327 (94%) preferred hearing results immediately from the sonographer rather than waiting to hear them from their referring physician (27). Our findings, however, conflict with those of Lind et al (11) who found that only 9% (49 of 558) of women preferred that their radiologists tell them their normal mammogram results and only 6% (33 of 546) their abnormal results. Lind et al surveyed women in 1988; it is possible that preferences have changed since then. Perhaps more important, their study, as with most others exploring women's preferences for reporting mammography results, surveyed women undergoing mostly screening mammography, whereas our study included women undergoing only diagnostic mammography. In a more recent study, Hulka et al (12) administered a mail-in survey and found that women chose delayed results when they were given the option of having more than one radiologist read the mammogram. One possible confounding factor, Hulka et al acknowledge, is that women who receive and fill out a survey at home may be more apt to choose delayed communication because they are free from worry or nervousness about cancer at the time they complete the survey. The opposite bias may be at play in the present study because women were surveyed immediately after undergoing mammography. Future research might correct for situational bias by using random assignment to one of multiple sites of survey administration, including a mail-in component. A future study might widen the population surveyed to include women who present not only for diagnostic, but also for screening mammography. Many researchers have argued that delayed communication of results may cause women undue anxiety (1,4,8), implying that the radiologist should be the one who discloses information. In our study, 61% of women were nervous about hearing their results, and nervousness contributed greatly to explaining women's preferences for who disclosed their normal results to them. Other researchers have found a slight patient preference for the radiologist to disclose normal (as opposed to abnormal) results (2,9,11). A policy of disclosing only normal results would not be feasible, however, because radiologists would be conveying information to patients by their silence when results were abnormal. Our study found no difference in preferences by result status. Because many mammography centers now obtain the necessary additional views and perform US examinations and even biopsies at the same visit for patients referred for diagnostic imaging, some radiologists, in effect, give pa-
dents the results whether they actually talk to the women or not. The question then becomes how this information is communicated. Radiologists' language and style may be important considerations in communicating mammography results. For those women in our study who preferred that their radiologists communicate abnormal results to them, most (88%) wanted them to do so by using cautious, instead of more direct, terms. This finding is interesting in light of findings from a study conducted by Song et al (23), who found that patients preferred receiving suggestive answers ("there is an abnormality") from the radiologist rather than a refusal to communicate any information at all. The two studies examined together suggest that women prefer language that adopts a middle ground; they neither wish to have their requests for information delayed nor wish to be told directly that they may have cancer in the event of an abnormal reading. Other studies have shown that patients are sensitive to physicians' communication styles and, not surprisingly, prefer that their physicians use supportive, compassionate language (21,22). Radiologists' training guidelines currently contain no recommendations for teaching skills in "patient interaction" (28). If radiologists are going to provide results immediately, residency programs may need to train radiologists for this activity, including helping them develop supportive and effective, as well as culturally sensitive, communication styles. Past research has shown that physicians can be taught these behaviors and interpersonal styles (3,13), and that effective communication is likely to maximize patient satisfaction and compliance with future radiologic recommendations (29). The results of this study raise several questions. Although the patients were explicitly asked about a radiologist versus a referring physician's disclosure, the question of who discloses results is confounded with the issue of immediacy of disclosure. In light of the passage of the MQSA, which requires only that the radiologist offer written (and therefore delayed) notification of results, it would be useful to tease apart whether it is "immediate" disclosure that is desired rather than disclosure by the radiologist. We were not able to disaggregate these two dimensions in our study. It also would be useful to examine what referring physicians and radiologists think about patients being immediately notified of their results by radiologists. For example, radiologists might prefer more time to prepare for discussing possible outcomes and treatments with patients and so wish to notify the patient later in writing, as is currently mandated. Patients may not benefit from immediate disclosure if radiologists are not comfortable with, or prepared for, this role. Future research could examine the effect of radiologist
communication on patients at the time'of disclosure and at subsequent visits. Our findings may also have implications for cost and scheduling in light of the increased patient contact that may be required of individual radiologists if such a program were implemented. Communicating results is likely to entail more time per patient. Extra time spent by radiologists with patients drives down reimbursement and may lead to the need to increase the cost per mammographic examination to compensate for increased time spent. In addition, unless imaging facilities are set up with rooms suitable for patient-radiologist discussions, valuable examination space may be taken up, further reducing mammography profitability. These and other issues in the changing mammography picture suggest future research on communication preferences among both patients and radiologists may be desirable. Although this research may be limited by the fact that these data were collected in 1995, before new regulations made radiologist disclosure the norm, it is important to understand the attitudes with which patients entered the present era of managed care, if only to counteract arguments based solely on "cost effectiveness." Furthermore, the new regulations do not require that radiologists communicate the results immediately or in person. Future research could compare the preferences presented here with those from patients after more extensive experience by radiologists with the new guidelines. Despite its limitations, this study has several implications for the practice of radiology and, more generally, for the expansion of communication avenues between patient and health care provider. It is reasonable to consider that most women undergoing diagnostic mammography would want to hear their results from the radiologist because they would receive their results sooner than if they heard them directly from their primary care provider or by mail. At the same time, however, radiologists might consider how they respond to those women who wish to reserve this communication for their referring physician. Assessing women's preferences as they present for mammography may be a reasonable approach; improving radiologists' patient interaction skills is likely another. Ideally, a balance can be struck between patients' needs, radiologists' skills, and the resources made available through managed care systems. ~,CKNOWLEDGMENT.
Special thanks to Claire Poyett, MD, for helping make the study possible and to the radiologic technologists for their support. Also thanks to Mary Stuart, MPH, for help with questionnaire design, to Amy Denham, MPH, MD, for
1075
editorial help, and to Anne Menkens, MA, for her revisions to the manuscript. IEFERENCE
1. Bassett LW, Henddck RE, Bassord TL, et al. Quality determinants of mammography: clinical practice guideline no. 13, 1994. Agency for Health Care Policy Research publication no. 95-0632. Rockviile, Md: Agency for Health Care Policy and Research, Public Health Service, U.S. Department of Health and Human Services, 1994. 2. Liu S, Bassett LW, Sayre J. Women's attitudes about receiving mammographic results directly from radiologists. Radiology 1994; 193:783-786. 3. Shapiro DE, Boggs SR, Melamed BG, Graham-Pole J. The effect of varied physician affect on recall, anxiety, and perceptions in women at risk for breast cancer: an analogue study. Health Psycho11992; 11:61-66. 4. Hoffman NY, Janus J, Destounis S, Logan-Young W. When the patient asks for the results of her mammogram, how should the radiologist reply? A JR Am J Roentgeno11994; 162:597-599. 5. Bassett LW, Bomyea K, Liu S, Sayre J. Communication of mammography results to women by radiologists: attitudes of referring health care providers. Radiology 1995; 195:235-238. 6. Cardenosa G, Eklund GW. Rate of compliance with recommendations for additional mammographic views and biopsies. Radiology 1991; 181 : 359-361. 7. D'Orsi C J, Debor MD. Communication issues in breast imaging. Radiol Clin North Am 1995; 33:1231-1245. 8. Levitsky DB, Frank MS, Richardson ML, Shneidman RJ. How should radiologists reply when patients ask about their diagnoses? a survey of radiologists' and clinicians' preferences. A JR Am J Roentgenol 993;161:433-436. 9. Schreiber MH, Leonard M, Rieniets CY. Disclosure of imaging findings to patients directly by radiologists: survey of patients' preferences. A JR Am J Roentgeno11995; 165:467-469. 10. Mammography Quality Standards Reauthorization Act of 1998. Pub L No. 105-248. 112 Star 1984. 11. Lind SE, Kopans D, Good M-JD. Patients' preferences for learning the results of mammographic examinations. Breast Cancer Res Treat 1992; 23:223-232. 12. Hulka CA, Slanetz PJ, Halpern EF, et al. Patients' opinion of mammography screening services: immediate results versus delayed results due to interpretation by two observers. A JR Am J Roentgenol 1997; 168:1085-1089.
13. Kaplan SH, Greenfield S, Ware JE. Assessing the effects of physician-patient interactions on the outcomes of chronic disease. Med Care 1989; 27(suppl 3):$110-$127. 14. DiMatteo MR, Hays RD, Prince LM. Relationship of physicians' nonverbal communication skill to patient satisfaction, appointment noncompliance, and physician workload. Health Psycho11986; 5:581-594. 15. Shapiro DE, Boggs SR, Melamed BG, Graham-Pole J. The effect of varied physician affect on recall, anxiety, and perceptions in women at risk for breast cancer: an analogue study. Health Psycho11992; 11:61-66. 16. Beckman HB, Frankel RM. The effect of physician behavior on the collection of data. Ann Intern Med 1984; 101:692-696. 17. Davison BJ, Degner LF. Empowerment of men newly diagnosed with prostate cancer. Cancer Nurs 1996; 20:187-196. 18. Morris KJ, Tarico VS, Smith WL, Altmaier EM, Franken EA. Critical analysis of radiologist-patient interaction. Radiology 1987; 163:565-567. 19. Smith WL, Altmaier EM, Ross RR, Johnson BD, Berberoglu LS. Patient expectations of radiology in non-interactive encounters. Radiology 1989; 172:275-276. 20. Kaplan SH, Greenfield S, Gandek B, Rogers WH, Ware JE. Characteristics of physicians with participatory decision-making styles. Ann Intern Med 1996; 124:497-504. 21. Morris KJ, Tarico VS, Smith WL, Altmaier EM, Franken EA. Critical analysis of radiologist-patient interaction. Radiology 1987; 163:565-567. 22. Fox SA, Siu AI, Stein JA. The importance of physician communication on breast cancer screening of older women. Arch Intern Med 1994; 154:2058-2088. 23. Song HH, Park SH, Shinn KS. Radiologists' responses to patients' inquiries about imaging results: a pilot study on opinions of various groups. Invest Radio11993; 28:1043-1048. 24. Emanuel LL. The consultant and the patient-physician relationship: a trilateral deliberative model. Arch Intern Med 1994; 154:1785-1790. 25. Dillman DA. Mail and Internet surveys: the tailored design method. New York, NY: Wiley, 2000. 26. Vallely SR, Manton Mills JO. Should radiologists talk to patients? BMJ 1990; 300:305-306. 27. Ragavendra N, Laifer-Narin SL, Melany ML, Grant EG. Disclosure of results of sonographic examinations to patients by sonologists. A JR Am J Roentgeno11998; 170:1423-1425. 28. Stewart DA. The radiologist as communicator: assessments and impressions. Radiology 1995; 196(2):37A-42A. 29. Inui TS, Carter WB. Problems and prospects for health services research on provider-patient communication. Med Care 1985; 23:521-538.
~.nnouncemen The Fourth International Symposium on the Costs a n d Benefits of Radiology will be held with invited participants in San Francisco, Calif, August 23-25, 2001. Individuals interested in presenting related research must submit a 200-word structured abstract (including Objectives, Methods, Results, and Conclusions) based on unpublished radiology cost-effectiveness or outcome studies by December 31, 2000, to Tyro Peters, UCSF Radiology, 3333 California St, Suite 375, San Francisco, CA 94143-0629; e-mail: [email protected].
1076