Breast cancer screening beliefs, recommendations and practices

Primary care physicians in the united states

Authors


  • The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention or the National Institutes of Health.

  • This article is a U.S. government work and, as such, is in the public domain in the United States of America.

Abstract

BACKGROUND:

Primary care physicians (PCPs) play a key role in breast cancer screening, yet no current data exist regarding PCP practices.

METHODS:

The authors analyzed a nationally representative survey of PCPs that was fielded during September 2006 to May 2007 to investigate PCP breast cancer screening beliefs, recommendations, and practices.

RESULTS:

Most of the 1212 PCPs who participated in the survey (80%) reported that mammography for average-risk women aged ≥ 50 years was very effective in reducing cancer mortality, and 54% reported that it was very effective for women ages 40 to 49 years. Fewer respondents reported that clinical breast examination (CBE) or breast self-examination (BSE) was very effective, but the majority rated CBE and BSE as somewhat effective. The majority of PCPs routinely recommended mammography, CBE, and BSE to patients aged ≥ 40 years. In multivariate models, family/general practitioners (odds ratio [OR], 2.23; 95% confidence interval [CI], 1.57-3.17 for mammography; OR, 4.42; 95% CI, 2.60-7.52 for CBE) and internal medicine specialists (OR, 3.21; 95% CI, 2.21-4.66 for mammography; OR, 5.34; 95% CI, 3.21-8.88 for CBE) were more likely to recommend an upper age limit for screening than obstetrician/gynecologists. Physicians who reported that US Preventive Services Task Force guidelines were very influential were more likely to recommend an age at which they no longer recommend mammography and CBE.

CONCLUSIONS:

To the authors' knowledge, the current study is the first national study in over 2 decades to report the breast cancer screening practices of PCPs and provides baseline data for monitoring the impact of changes in clinical practice guidelines. The current findings suggested that virtually all PCPs routinely recommend mammography, CBE, and BSE to their patients aged ≥ 40 years, although recommendations vary by primary care specialty. Cancer 2011. Published 2011 by the American Cancer Society.

For close to 2 decades, experts have disagreed about the starting and stopping ages for breast cancer screening, the schedule on which tests should occur, and the effectiveness of different screening modalities.1-3 These debates highlight the challenge of recommending appropriate use of screening for a much feared disease while simultaneously acknowledging the limitations of both the science and the accuracy of the tests.4 Disagreements beginning in 1993 and 1997 focused primarily on the age at which women should start having mammograms and the frequency with which they should have them.5, 6 A later controversy was ignited in 2001, when publication of a meta-analysis concluded that evidence was insufficient to recommend mammography for women of any age.2 The most recent controversy arose in 2009 in response to the publication of revised recommendations from the US Preventive Services Task Force (USPSTF) questioning the value of routine mammograms for women ages 40 to 49 years who are at average risk of breast cancer.7 This led to the USPSTF advising that decisions about screening mammography for women in this age group should be made on an individualized basis. Moreover, in contrast to current guidelines of other medical organizations,8, 9 the USPSTF recommends that women should start having routine mammograms at age 50 years on a biennial instead of annual schedule. Finally, the USPSTF concluded that current evidence is insufficient to assess the benefits and risks of screening mammography in women aged ≥ 75 years. Although there is little evidence to suggest that these controversies have significantly changed consumer or provider intentions or behaviors regarding screening,1 the ongoing debates about the benefits of screening among women ages 40 to 49 years may have had an impact on screening rates among insured women.10 It is unknown whether the current controversy will prompt changes in future clinical practice.

Although much is known about the factors that influence women to get screened,11 to our knowledge, no recent national information is available regarding physicians' breast cancer screening beliefs and practices. Primary care physicians (PCPs) play a key role in performing and referring patients for cancer screening, yet research indicates that the practices of physicians and other healthcare providers are not necessarily consistent with clinical guidelines.12-15 Furthermore, it is unknown to what extent clinical guidelines motivate physician practice or which barriers or facilitators to implementation of clinical guidelines are associated with changes in practice recommendations. Our era of rapidly changing technology requires surveillance to monitor whether, when, and how new evidence influences clinical practice. Although multiple and sometimes conflicting clinical practice guidelines for breast cancer screening have existed for decades, national data to examine how PCPs implement them in practice have been quite limited.

This study is the first nationwide study since 198916 to explore the breast cancer screening beliefs, recommendations, and practices of a nationally representative sample of PCPs. Within the context of prevailing clinical guidelines in 2006 and 2007, we investigated reports from 4 types of PCPs (family medicine, general practice, internal medicine, and obstetrics/gynecology) concerning 1) which tests they recommend for breast cancer screening, at what frequency, and for which age groups; 2) their beliefs about breast cancer screening test effectiveness; and 3) physician and practice characteristics associated with recommending an age at which to no longer recommend screening. These data can serve as a baseline to monitor the impact of future changes in clinical practice guidelines.

MATERIALS AND METHODS

We analyzed data from the National Cancer Institute (NCI)-led National Survey of Primary Care Physicians' Recommendations and Practices for Breast, Cervical, Colorectal, and Lung Cancer Screening (PCP Survey), which was fielded September 2006 to May 2007. The PCP Survey was administered by contract with Westat, a research firm located in Rockville, Maryland, and was approved by their institutional review board and by the US Office of Management and Budget.

The PCP Survey used a national probability sample and had an absolute response rate of 67.5%. It used physician specialty (family practice, general practice, internal medicine, and obstetrics/gynecology) as the sampling strata and asked physicians about their demographic and practice characteristics and their beliefs, recommendations, and practices related to recommendation and use of breast cancer screening tests. We restricted our analysis in this study to the 1212 PCPs who responded to the questionnaire on breast and cervical cancer screening, all of whom reported that they routinely perform or recommend any breast cancer screening tests for their eligible patients. We combined “general practice” and family physicians, because we observed that their screening behavior here and in other studies17 was similar, and the number of general practice physicians in the PCP Survey was small (n = 50). More information about data-collection procedures and response rates for the PCP Survey is available elsewhere14 (http://www.healthservices.cancer.gov/surveys/screening_rp/ accessed December 19, 2010).

Measures

We selected independent variables that have been associated consistently with physicians' cancer screening recommendations in the literature (eg, age, medical specialty).17 We also selected measures that assess aspects of physician breast cancer screening practices (eg, the influence of screening guidelines issued by multiple organizations, use of reminder systems) that were available from the PCP Survey but that have not been examined in other national surveys.

Physician characteristics included sex, age, race/ethnicity, whether a US or international medical graduate, specialty, board certification, having a medical school affiliation, and the percentage of time spent in patient care. Practice characteristics were geographic location, size and type of practice, practice arrangement, having reminder systems for either patients or physicians, having practice guidelines for breast cancer screening, and the type of medical record system.

Physicians were asked whether screening guidelines from the USPSTF, American Cancer Society (ACS), American College of Obstetricians and Gynecologists (ACOG), American Academy of Family Physicians (AAFP), or American College of Physicians (ACP) were influential in their clinical practice. They also were asked their about beliefs regarding the effectiveness of 4 breast cancer screening tests (clinical breast examination (CBE), breast self-examination (BSE), screen-film mammography, and digital mammography) in reducing breast cancer mortality in average-risk women.

Physicians' breast cancer screening practices were assessed by asking whether they routinely recommend CBE (performed by practitioner), BSE (performed by patient), or mammography to asymptomatic, average-risk women and the frequency with which they recommend the test for women ages 40 to 49 years and for women aged ≥ 50 years. Respondents also were asked whether there is an age at which they would no longer recommend the screening modality for healthy women and, if so, to specify that age.

Analyses

Sample characteristics were assessed with descriptive statistics. Bivariate logistic regression was used to assess the association between all independent variables and the outcome measure. Bivariate associations of P < .10 were retained for inclusion in the multivariate logistic regression analysis except in cases where independent variables were highly correlated (physician specialty and the influence of practice guidelines issued by medical specialty organizations, eg, ACOG, AAFP, ACP). SUDAAN 9.0.1 statistical software (RTI International, Research Triangle Park, NC)18 was used to apply sampling weights and to account for the stratified survey design.

Past and present controversies regarding the appropriate age at which to begin recommending and stop recommending breast cancer screening led us to model physician and practice characteristics that might be associated with recommending an age at which to start and stop CBE and screening mammography. However, because an overwhelming majority (> 90%) of all physician specialties recommended CBE and mammography for women ages 40 to 49 years, there was inadequate variation to model this age to start screening. Instead, we used multivariate regression analyses to assess the associations among physician and practice characteristics and recommending an age at which to no longer recommend CBE and mammography. We did not model BSE as an outcome, because the procedure does not require physician intervention.

RESULTS

Approximately 66% of our national sample of PCPs (N = 1212) was male (66.7%), ages 40 to 59 years (63.7%), and non-Hispanic white (67.8%) (Table 1). Most were board certified (81%) and US medical graduates. Close to half (45.2%) were family medicine/general practitioners. Their medical practices were located largely in urban areas (82.9%). Most were single specialty, comprised of more than 1 physician, and physician-owned (61%). Seventy-five percent of our sample spent > 90% of their time in patient care. One-third reported having a medical school affiliation. Less than 50% of physician practices had a breast cancer screening reminder system in place for patients or physicians or had a system of electronic medical records.

Table 1. Characteristics of US Primary Care Physicians and Their Practice Settings (N=1212)a
CharacteristicUnweighted No.Weighted %
  • HMO indicates health maintenance organization.

  • a

    From the National Cancer Institute's 2007 National Survey of Primary Care Physicians' Recommendations and Practices for Breast, Cervical, Colorectal, and Lung Cancer Screening.

  • b

    Includes American Indian/Alaska Native, Native Hawaiian/Other Pacific Islander, multiple races, other race, and unknown.

  • c

    Categories are not mutually exclusive

Physician characteristics  
 Sex, men82166.7
 Age, y  
  <4024419.6
  40-4936532
  50-5938331.7
  ≥6022016.7
 Race/ethnicity  
  White, non-Hispanic83667.8
  Black, non-Hispanic655
  Hispanic745.9
  Asian18216.8
  Otherb554.5
 Specialty  
  Family medicine/general practice50245.2
  General internal medicine37636.9
  Obstetrics/gynecology33417.9
 Board certified, yes97481.5
 Time in patient care, %  
  <7512910.6
  75-8916614.1
  ≥9090674.5
 Medical school affiliation, yes41933.7
 International medical graduate, yes28425.1
 Guidelines very influentialc  
  US Preventive Services Task Force48042.4
  American Cancer Society67756
  American College of Obstetricians & Gynecologists61946.9
  American Academy of Family Physicians35931.8
  American College of Physicians2725.1
  Any guideline100882.5
Practice characteristics  
 Geographic location  
  Urban101282.9
  Large rural city/town1098.8
  Small/isolated small rural town918.3
 Practice size, no. of physicians  
  132026.2
  2-549941.2
  6-1523919.8
  ≥1614612
 Practice type  
  Single specialty87570.8
  Multispecialty30226.1
 Primary practice arrangement  
  Physician-owned practice75561
  Large medical group, HMO, health system20518.1
  University hospital/clinic, other hospital/clinic22718.8
 Has any breast cancer screening reminder system for patients51440
 Has any breast cancer screening reminder system for physicians51542.7
 Practice has guidelines for breast cancer screening59648.7
 Type of medical records system  
  Paper charts71158
  Partial electronic or in transition to full electronic32826.8
  Full electronic medical records15613.6
 Uninsured patients, %  
  0-575462.2
  6-2531826.2
  ≥26776.3
  Unknown272.5

Most PCPs identified at least 1 breast cancer screening guideline as being very influential in their practice. The ACS guidelines were cited as influential most often (56%), followed by the ACOG (47%), USPSTF (42%), AAFP (32%), and ACP (25%) guidelines.

Ninety-nine percent of all PCPs reported that, for average-risk women aged ≥ 50 years, mammography was effective in reducing cancer mortality. Eighty percent reported that it was very effective, and 19% reported that it was somewhat effective (Fig. 1). Ninety-six percent reported that mammography was at least somewhat effective for women ages 40 to 49 years (54% reported that it was very effective, and 42% reported that it was somewhat effective). Fewer physicians reported that CBE or BSE were very effective, but the majority still rated them somewhat effective. Twenty-one percent of PCPs reported uncertainty about the effectiveness of digital mammography, but 56% reported that it was very effective. A small proportion of physicians reported they were uncertain about CBE and BSE test effectiveness but still recommended the test (13% for CBE and 15% for BSE; data not shown).

Figure 1.

This chart illustrates the beliefs of primary care physicians regarding breast cancer screening test effectiveness in 2007.

Table 2 presents findings on mammography recommendations for women ages 40 to 49 years and women aged ≥ 50 years. Greater than 70% of all physicians who recommended mammography to women ages 40 to 49 years recommended it on an annual basis (69.5% of family medicine/general practitioners, 74.5% of internal medicine specialists, and 79.3% of obstetrician/gynecologists). Greater than 90% of all physicians recommended annual mammography to women aged ≥ 50 years. Family medicine/general practitioners and internal medicine specialists who recommended mammography were more likely to no longer recommend screening at a certain age (30.2% and 37.8%, respectively) than obstetrician/gynecologists (14%). Age at which to no longer recommend screening varied considerably, but < 10 percent of physicians of any specialty specified an age < 70 years.

Table 2. Primary Care Physicians' Breast Cancer Screening Recommendations: Mammographya
 Percentage of Physicians (95% CI)
Mammography RecommendationFamily Medicine/General Practice, N=502Internal Medicine, N=376Obstetrics/Gynecology, N=334
  • CI indicates confidence interval.

  • a

    Based on the National Cancer Institute's 2007 National Survey of Primary Care Physicians' Recommendations and Practices for Breast, Cervical, Colorectal, and Lung Cancer Screening.

Ages 40-49 y   
 Routinely recommend mammography   
  Yes94.6 (91.9-96.5)93.4 (89.8-95.8)98.7 (96.8-99.5)
  No4.6 (2.9-7.2)5.6 (3.6-8.8)1.1 (0.4-2.9)
 Frequency among physicians who recommend mammography, mo   
   <120.5 (0.2-1.7)0.5 (0.1-2.2)0.3 (0-2.4)
   1269.5 (65.2-73.5)74.5 (69.5-78.8)79.3 (74.7-83.2)
   From >12 to <242.3 (1.3-4.1)2.6 (1.2-5.4)4.9 (3-7.9)
   2423.8 (20.1-28)20.7 (16.5-25.7)15.2 (12-19.2)
   >243.5 (2.4-5.2)1.7 (0.7-3.8)0 (0-0)
Aged ≥50 y   
 Routinely recommend mammography   
  Yes99.5 (98.6-99.9)98.6 (95.8-99.5)99.7 (97.9-100)
  No0.5 (0.1-1.4)0.5 (0.1-2.2)0.3 (0-2.1)
 Frequency among physicians who recommend mammography, mo   
   <120.9 (0.3-2.1)0.8 (0.3-2.5)1.3 (0.5-3.3)
   1295.4 (93.4-96.7)95.3 (92.6-97.1)97.1 (94.3-98.5)
   From >12 to <241.0 (0.4-2.6)0.5 (0.1-2)0.4 (0-2.6)
   242.1 (1.2-3.8)2.1 (1.1-4.1)0.9 (0.3-2.2)
   >240.4 (0.2-1.3)0.5 (0.1-2)0 (0-0)
No longer recommend mammography screening when patients reach a certain age among physicians who recommend it for other ages   
  Yes30.2 (26.6-34)37.8 (33-42.9)14 (10.8-18)
  No67.5 (63.6-71.3)57.6 (52.3-62.7)83.4 (79-87)
Age at which physicians no longer recommend mammography screening, y   
  <709.4 (5.1-16.6)5.2 (2.6-10.1)4.7 (1.2-16.6)
  70-7414.7 (9.6-21.9)9.2 (5.5-15.2)7.2 (2.7-18.1)
  75-7921.2 (14.5-29.9)21.5 (15-29.7)23.8 (14.7-36)
  80-8431.4 (24.4-39.4)31.4 (23.3-40.8)27 (15.9-42.2)
  85-9021.3 (15.6-28.4)29.4 (22.3-37.6)28.2 (17.6-41.9)
  >900 (0-0)0 (0-0)5 (1.2-18.2)

The majority of physicians recommend annual CBE to all women aged > 40 years (Table 3). Compared with mammography, fewer physicians reported an age for no longer recommending CBE. Again, obstetrician/gynecologists (4.7%) were the least likely to specify an age for no longer recommending CBE. Practice patterns for BSE (Table 4) were quite similar to those for CBE, except that physicians recommended performing BSE with greater than annual frequency (recommendations are to perform BSE on a monthly basis).

Table 3. Primary Care Physicians' Breast Cancer Screening Recommendations: Clinical Breast Examinationa
 Percentage of Physicians (95% CI)
Clinical Breast Examination RecommendationFamily Medicine/General Practice, N=502Internal Medicine, N=376Obstetrics/Gynecology, N=334
  • CI indicates confidence interval.

  • a

    Based on the National Cancer Institute's 2007 National Survey of Primary Care Physicians' Recommendations and Practices for Breast, Cervical, Colorectal, and Lung Cancer Screening.

Ages 40-49 y   
 Routinely recommend clinical breast examination   
  Yes96.5 (94.5-97.8)90.5(86.9-93.2)99.2 (98.6-99.6)
  No3.5 (2.2-5.5)8.5(6.1-11.7)0.8 (0.4-1.4)
 Frequency among physicians who recommend clinical breast examination, mo   
  <122.2 (1.2-4.1)5.8 (3.8-8.8)2.3 (1.1-4.6)
  1294.7 (92-96.6)91.8 (88.4-94.2)96.4 (93.7-98)
  From >12 to <240.6 (0.2-1.8)0.6 (0.1-2.2)0.4 (0.1-2.7)
  242.1 (1.1-3.9)1.1 (0.4-3.1)0.9 (0.3-2.9)
  >240.4 (0.1-1.6)0 (0-0)0 (0-0)
Aged ≥50 y   
 Routinely recommend clinical breast examination   
  Yes98.1 (96.6-99)94.1 (90.9-96.2)99.5 (99.4-99.5)
  No1.9 (1.0-3.4)5.6 (3.6-8.7)0.5 (0.5-0.6)
 Frequency among physicians who recommend clinical breast examination, mo   
  <123.3 (2-5.6)7.2 (4.8-10.7)2.8 (1.5-5.3)
  1296 (93.4-97.6)91.3 (88-93.8)96.1(93.3-97.8)
  From >12 to <240 (0-0)0.3 (0-2.1)0 (0-0)
  240.4 (0.1-1.8)0 (0-0)0.3 (0-2.3)
  >240 (0-0)0.2 (0-1.6)0.4 (0.1-2.7)
No longer recommend screening with clinical breast examination when patients reach a certain age among physicians who recommend it for other ages   
 Yes18.8 (15.8-22.3)23.3 (18.7-28.5)4.7 (3-7.2)
 No80.4 (76.9-83.5)73.9 (68.7-78.6)94.3 (91.6-96.1)
Age at which physicians no longer recommend screening with clinical breast examination, y   
  <705.8 (2.6-12.3)2.7 (0.9-8.2)0 (0-0)
  70-7417.6 (10.3-28.5)9.8 (5.2-17.6)0 (0-0)
  75-7922.9 (15.8-32)18.8 (12-28.2)34 (18.1-54.4)
  80-8431.2 (22.6-41.4)36.8 (27.2-47.5)6.9 (0.8-39.7)
  85-9020.7 (13.3-30.8)31 (22.2-41.4)46.6 (25.7-68.8)
  >900.8 (0.1-5.7)0 (0-0)0 (0-0)
Table 4. Primary Care Physicians' Breast Cancer Screening Recommendations: Breast Self-Examinationa
 Percentage of Physicians (95% CI)
Breast Self-Examination RecommendationFamily Medicine/General Practice, N=502Internal Medicine, N=376Obstetrics/Gynecology, N=334
  • CI indicates confidence interval.

  • a

    Based on the National Cancer Institute's 2007 National Survey of Primary Care Physicians' Recommendations and Practices for Breast, Cervical, Colorectal, and Lung Cancer Screening.

Ages 40-49 y   
 Routinely recommend breast self-examination   
  Yes91.6 (88.5-93.9)90.4 (86.2-93.4)92.5 (89.3-94.8)
  No8.4 (6.1-11.5)8.4 (5.7-12.3)7.3 (5-10.5)
 Frequency among physicians who recommend breast self-examination, mo   
  <1298.6 (97-99.3)96 (93.6-97.6)99.6 (97.3-99.9)
  121.2 (0.5-2.8)2.6 (1.5-4.3)0 (0-0)
  From >12 to <24   
  24   
  >240 (0-0)0.3 (0-2.3)0 (0-0)
Aged ≥50 y   
 Routinely recommend breast self-examination   
  Yes92.5 (89.5-94.6)91.1 (87-93.9)94.4 (91.1-96.5)
  No7.5 (5.4-10.5)7.9 (5.4-11.4)5.6 (3.5-8.9)
 Frequency among physicians who recommend breast self-examination, mo   
  <1293.2 (90.7-95.1)93.4 (89.8-95.8)95 (92-96.9)
  126.2 (4.4-8.5)5.1 (3.1-8.1)4.4 (2.6-7.2)
  From >12 to <24   
  24   
  >240 (0-0)0.3 (0-2.2)0 (0-0)
No longer recommend screening with breast self-examination when patients reach a certain age among physicians who recommend it for other ages   
 Yes11.8 (9.1-15.2)14 (10.3-18.7)4.6 (3-7)
 No84.9 (81.1-88)80.5 (75.2-84.9)93.1 (90.1-5.2)
Age at which physicians no longer recommend screening with breast self-examination, y   
 <7011.5 (5.6-22.4)6.6 (2.3-17.6)15.4 (3.9-44.9)
 70-7419.1 (10.5-32)4.6 (1.1-16.9)6 (0.7-35.9)
 75-7923.7 (14-37.3)27.2 (15.2-43.6)22.9 (10.8-42.2)
 80-8428.6 (17.6-42.9)34.4 (20.6-51.6)12.9 (8.1-0.1)
 85-9015.7 (8.3-27.6)27.1 (16.4-41.4)29.3 (12.7-4.2)
 >901.4 (0.2-9.4)0 (0-0)0 (0-0)

Table 5 lists the PCP and practice characteristics that are associated with recommending an age at which to no longer recommend mammography and CBE. For both the mammography and CBE multivariate models, physician specialty was associated significantly with specifying an age at which to no longer recommend breast cancer screening. Family/general practice and internal medicine specialists were much more likely to recommend an upper age limit for breast cancer screening than obstetrician/gynecologists. International medical graduates were more likely than US medical graduates to not recommend mammography or CBE based on age, but this result was only marginally significant. Physicians who reported that the USPSTF guidelines were very influential were more likely to specify screening age limits. Similarly, physicians who did not consider the ACS guidelines very influential were more likely to specify a certain age for not recommending either mammography or CBE than physicians who reported that the ACS guidelines were very influential. PCPs who were men were more likely than women physicians to not recommend CBE at a certain age.

Table 5. Physician and Practice Characteristics Associated With Primary Care Physicians Who Recommend an Age at Which to No Longer Recommend Clinical Breast Examinations or Mammography (N=1212)a
 MammographyClinical Breast Examination
CharacteristicOR95% CIOR95% CI
  • OR indicates odds ratio; CI, confidence interval; NA, not available.

  • a

    From multivariate logistic regression models based on the National Cancer Institute's 2007 National Survey of Primary Care Physicians' Recommendations and Practices for Breast, Cervical, Colorectal, and Lung Cancer Screening.

  • b

    P<.05 (Wald F test).

  • c

    The NA variable was not entered into the multivariate model because there was a lack of association in bivariate analysis.

Physician characteristics    
 Specialty    
  Family medicine/general practice2.23b1.57-3.17b4.42b2.60-7.52b
  Internal medicine3.21b2.21-4.66b5.34b3.21-8.88b
  Obstetrics/gynecology1.00 1.00 
 Sex    
  Men1.360.94-1.961.91b1.21-3.01b
  Women1.00   
 Board certified    
  YesNAc 1.791.11-2.87
  No  1.00 
 International medical graduate    
  Yes1.00 1.00 
  No0.70b0.49-1.01b0.52b0.35-0.77b
 US Preventive Services Task Force very influential    
  Yes1.79b1.37-2.34b1.48b1.04-2.11b
  No1.00 1.00 
 American Cancer Society very influential    
  Yes0.52b0.40-0.67b0.63b0.44-0.90b
  No1.00   
Practice characteristics    
 Medical records system    
  Paper1.00 NAc 
  Partially electronic1.391.01-1.91  
  Fully electronic1.350.88-2.05  

DISCUSSION

Our current results suggest that there is overwhelming PCP support for breast cancer screening. Despite prior controversies surrounding test modality, target age groups, and screening intervals, virtually all physicians in our sample reported routinely recommending mammography, CBE, and BSE to their patients aged ≥ 40 years. A smaller percentage of PCPs reported recommending annual mammography to women ages 40 to 49 years than to women aged ≥ 50 years, although all prevailing clinical guidelines at the time of our survey recommended that screening mammography begin at age 40 years.8, 9, 19 Greater than 90% of all PCPs recommended initiating mammography with or without CBE at age 40 years, although only 54% reported that the test was very effective for women ages 40 to 49 years (compared with 80% who believed it was very effective for women aged ≥ 50 years). The finding that a smaller percentage of PCPs recommended annual mammography for women ages 40 to 49 years than for women aged ≥ 50 years may reflect ACOG and USPSTF recommendations when the survey was administered for a 1-year to 2-year interval for women in their 40s.

Although few physicians believed that CBE and BSE tests were very effective in reducing breast cancer mortality, the majority recommended these tests for women aged ≥ 40 years. USPSTF guidelines at the time stated that there was insufficient evidence for the use of BSE as a screening modality, and the other organizations simply stated that BSE could be considered an option. That similar practice patterns across specialties were reported suggests that physician beliefs may have relatively little influence on screening recommendations. This may result from several factors. There is some evidence that physicians practice defensive medicine. In short, when guidelines are ambiguous,20 physicians may screen more aggressively because of concerns about medical malpractice litigation.21 This is likely because missed or delayed diagnosis of breast cancer is 1 of the most litigated areas in healthcare.22 Furthermore, patients' expectations and anxiety also may influence physicians' decisions to screen even if they would not necessarily have recommended the test(s) based solely on their beliefs about test efficacy.23

New USPSTF guidelines issued in 20097 recommend against routine mammography for women ages 40 to 49 years who are at average risk for breast cancer, support biennial mammography as a single screening modality for women ages 50 to 74 years, and state that evidence is insufficient to assess the benefits and risks of screening mammography in women aged ≥ 75 years. The guidelines recommend against teaching BSE and indicate that there is insufficient evidence for recommending that providers perform CBE. Guidelines published by the specialty organizations24 and the ACS remain unchanged, making it difficult to predict how the USPSTF revisions will influence primary care practice if at all. Although physicians differentially value guidelines issued by their specialty organizations,14 our study suggests that, overall, PCPs value the ACS guidelines more than the USPSTF guidelines. However, even those who believe that the USPSTF guidelines are highly influential may be met with resistance from their patients. An informal survey recently conducted by the Annals of Internal Medicine3 suggests that clinicians are inclined to accept the new guidelines, whereas patients are more likely to stick to early and annual screening. The reality, however, is likely to be that factors in addition to guidelines will influence PCP practice, including patients' expectations,23 reimbursement, fear of malpractice litigation,22 competing demands, and time constraints.25

The only large divergence in practice across the primary care specialties that we observed was the age at which to no longer recommend breast cancer screening. At the time of our survey, ACOG did not recommend a stopping age; and USPSTF and ACS recommended continuing if women were in good health and had a life expectancy of sufficient length to benefit from early detection. We observed that obstetrician/gynecologists were significantly less likely than other specialties to consider not recommending mammography and CBE when patients reach a certain age. Our finding is new and is consistent with other research indicating that there are higher breast cancer screening rates among obstetrician/gynecologists compared with other physicians in the US.26 However, it is worth noting that only a small proportion of older women continue to use obstetrician/gynecologists for their primary care27; consequently, this finding may not be clinically meaningful. Of particular interest, however, is that all PCP specialties were more likely to not recommend mammography than CBE at older ages. This finding seems contrary to PCP beliefs about test effectiveness (with a greater proportion of PCPs reporting that mammography is very effective in reducing breast cancer mortality compared with CBE). Perhaps it is more difficult to discontinue a test that is performed face-to-face in the office on an annual basis and often is coupled with other preventive care (eg, vaccination for influenza) than it is to not recommend referrals. Another possibility is that PCPs may believe that the risk of overdiagnosis and, consequently, over treatment in older women is greater with mammography than with CBE; or they may perceive that CBE does not involve additional appointment time, financial cost, or exposure to radiation. Regardless, we know very little about how to persuade patients and physicians to discontinue established practices as new scientific evidence emerges and clinical guidelines change. More research on this is particularly important to consider, because overuse of tests carries potential harms. False-positive results, unnecessary biopsies, and overdiagnosis can impose a significant psychosocial, physical, and financial burden on patients and can increase costs to the healthcare system.28, 29

The current study had several limitations. Self-reported beliefs and behaviors may not accurately reflect what PCPs actually do in practice. Nationally, women report lower screening mammography rates30 than what might be expected based on our PCP data. It is possible that PCPs over report their recommendations to patients. However, it also is possible that some patients do not follow through with obtaining mammograms, which generally require a separate appointment with a mammography facility. In addition, only 40% of PCPs in our sample reported having a reminder system for patients. Furthermore, our PCP data refer only to the patient populations they serve. Many women in the United States do not have a usual source of healthcare, have limited access to PCPs, and also report lower rates of screening than the general population.30 Finally, our study is cross-sectional and, thus, limits our ability to make causal inferences regarding factors that drive PCP behavior.

This study represents the first national assessment of PCP breast cancer screening beliefs, recommendations, and practices since 1989. We anticipate that it will provide a baseline from which to monitor changes over time. Conflicting scientific opinion, inconsistent guidelines, and public perceptions regarding the efficacy of screening suggest that the most recent controversy surrounding revised USPSTF guidelines is unlikely to be the last.31 PCP interpretation and implementation of guidelines will continue to be critical in achieving the best possible outcomes for women of all ages. Ongoing monitoring of PCP beliefs, recommendations, and practices is needed to understand how best to incorporate the latest scientific evidence regarding breast cancer prevention and early detection into healthcare practice.

Acknowledgements

We thank Gigi Yuan of Information Management Services for expert programming support.

CONFLICT OF INTEREST DISCLOSURES

Funding support for the survey was provided by the National Cancer Institute (contract number N02-PC-51308), the Centers for Disease Control and Prevention (interagency agreement number Y3-PC-6017-01), and the Agency for Healthcare Research and Quality (interagency agreement numbers Y3-PC-5019-01 and Y3-PC-5019-02).

Ancillary