Fax: (310) 794-6097
Health care disparities in older patients with breast carcinoma
Informational support from physicians
Article first published online: 3 MAR 2003
Copyright © 2003 American Cancer Society
Volume 97, Issue 6, pages 1517–1527, 15 March 2003
How to Cite
Maly, R. C., Leake, B. and Silliman, R. A. (2003), Health care disparities in older patients with breast carcinoma. Cancer, 97: 1517–1527. doi: 10.1002/cncr.11211
- Issue published online: 3 MAR 2003
- Article first published online: 3 MAR 2003
- Manuscript Accepted: 25 OCT 2002
- Manuscript Revised: 23 OCT 2002
- Manuscript Received: 1 AUG 2002
- California Breast Cancer Research Program of the University of California. Grant Number: 4PB-061
- Robert Wood Johnson Foundation Generalist Physician Faculty Scholars Program. Grant Number: 036833
- health care disparities;
- patient-physician interaction;
- breast carcinoma;
- older patients;
- ethnic minorities;
- patient-centered care;
- information giving
Little is known about how disparities in the treatment of patients with breast carcinoma based on patient age and ethnicity are effected or mitigated at the patient-physician interaction level. The objectives of this study were to document physician provision of informational support to patients at the time of a new diagnosis of breast carcinoma and to assess differences according to patient age and ethnic group in terms of the information received and desired.
Participants were 222 patients with breast carcinoma in Los Angeles County, California, age ≥ 55 years who were interviewed within 6 months of their diagnosis of breast carcinoma and/or within 1 month posttreatment. Respondents were asked about receipt and helpfulness of 10 tangible informational support items (e.g., whether booklets, videotapes, medical records, etc. were provided by physicians) and 15 interactive informational support items (e.g., whether physicians discussed breast cancer topics, such as risk of recurrence or treatment options). An index of the tangible informational support items and a scale of the interactive informational support items received were created for summary analyses. Patients' medical records were abstracted for breast carcinoma stage and treatment type; surgeons also were surveyed about sociodemographic and practice characteristics.
In multiple linear regression analyses, older age (β coefficient [β] ± standard error [SE], − 0.08 ± 0.02; P = 0.001) and Latina ethnicity (β ± SE, − 1.21 ± 0.40; P = 0.003) had a negative association with physician provision of interactive informational support, controlling for patient and physician sociodemographic characteristics, practice characteristics, breast carcinoma stage, comorbidity, number of physicians seen, visit length, social support, and patient self-efficacy in interacting with physicians (adjusted correlation coefficient [R2] for the model, 0.33; P < 0.00001). Both older patients and ethnic minority patients, as well as their respective comparison groups, rated most breast cancer information as at least as helpful. Both groups preferred interpersonal sources of information to written sources, although they received interpersonal sources less frequently.
Older patients and Latina patients with breast carcinoma received less interactive informational support from their physicians compared with younger patients, differences that persisted after controlling for a wide range of sociodemographic, psychosocial, and physician factors. Improving the quality of communication at the patient-physician interaction level may be an important avenue to reducing age and ethnic group treatment disparities among patients with breast carcinoma. Cancer 2003;97:1517–27. © 2003 American Cancer Society.
The population of the united states is aging rapidly and is becoming more ethnically diverse.1, 2 Cancer is expected to become the leading cause of morbidity and death as the population ages.3 Health and health care disparities have been well documented in both older patients4–6 and ethnic minority patients7 with malignant disease. Women comprise the majority of the older population and, among older women, breast carcinoma is the most common malignancy.8, 9 Because of its prevalence in the older population, breast carcinoma may serve as a valuable model for examining sources of health care disparities in vulnerable populations and for identifying ways to reduce them.
Older patients with breast carcinoma are more likely to present with distant disease and are less likely to receive appropriate diagnostic evaluation and therapeutic intervention compared with younger patients,4, 5, 10–13 despite evidence that older women who do not receive definitive treatment for breast carcinoma have a greater likelihood of dying compared with patients who receive definitive treatment.14 Ethnic differences also have been noted in the care of patients who have breast carcinoma, with African-American (AA) women less likely compared with white women to receive appropriate and/or aggressive treatment for breast carcinoma15, 16 and less likely to undergo breast-conserving surgery (BCS) and radiation therapy after BCS.17, 18 Both older patients and ethnic minority patients with breast carcinoma are less likely compared with their respective counterparts to receive breast reconstruction after undergoing mastectomy.19 The underlying causes of these discrepancies are not clear.20, 21
Central to the process of health care delivery is the patient-physician interaction. Ultimately, it is at this level that disparities in health care, whether they originate socially, organizationally, or individually, are either effected or mitigated. An essential component of the patient-physician interaction is the physician providing information, which has been associated with many desirable outcomes of medical care, including reduced postoperative pain and hospital stays,22–24 improved functional status,25–27 and better physiologic outcomes.28 In patients with breast carcinoma, the treatment decision-making process is complicated by the need to consider a panoply of treatment options at different junctures. At each decision point, providing effective information is essential to involving the patient and to appropriate decision-making.
Disparities in the treatment of older patients and ethnic minority patients with breast carcinoma may be overcome by effective physician communication and the provision of information tailored to the needs and preferences of these groups. However, a comprehensive assessment of breast cancer information that is provided by physicians, whether tangible (e.g., booklets, copies of medical records, or videotapes), or interactive (e.g., discussion of topics like tumor aggressiveness or radiation treatment), and that is most beneficial to older patients and ethnic minority patients is lacking. A detailed understanding of group differences in the receipt of, and preferences for, the provision of specific types of breast cancer information could provide the basis of potential interventions at the patient-physician level to mitigate health care disparities.
The study reported here is among the first to comprehensively survey newly diagnosed older patients and ethnic minority patients with breast carcinoma in detail about physician-provided breast cancer information and their informational preferences. The research questions addressed here are as follows: 1) Does the information provided and/or communicated by physicians to newly diagnosed patients with breast carcinoma differ according to patient age or ethnicity, and do age or ethnic group differences exist when potential confounding factors are controlled? 2) What types of information did these groups find most helpful after a new diagnosis of breast carcinoma? 3) Are there group differences in perceptions of helpfulness of informational support by provider specialty? Assessing differences in the information received and informational preferences will assist in understanding and reducing health care disparities in vulnerable populations of patients with breast carcinoma.
MATERIALS AND METHODS
This study was a cross-sectional survey of all eligible patients with breast carcinoma age ≥ 55 years who were newly diagnosed with breast carcinoma and who were referred from three sources in Los Angeles County, California. Patients' surgeons also were surveyed about their sociodemographic and practice characteristics, and patients' medical records were abstracted for breast carcinoma stage and treatment type.
Study Setting and Recruitment
Participants in this study were 222 English-speaking and Spanish-speaking women age ≥ 55 years who had been diagnosed with any stage of breast carcinoma within the previous 6 months and/or were 1 month posttreatment. Patients with a history of previous breast carcinoma, patients who actively were receiving treated for another malignancy, or patients who were impaired cognitively were excluded. Informed consent was obtained from all patients who participated in the study. Three sources of potential participants were used in this study. Initially, 17 community-based, private surgical and medical oncology practices and cancer centers throughout Los Angeles County agreed to identify and send recruitment letters to all eligible patients in their practices. Because of difficulty in identifying adequate numbers of ethnic minority patients in these community practices, two additional sources of older patients with breast carcinoma were added. The first supplemental source was the Rapid Ascertainment Program of the Los Angeles County Cancer Surveillance Program (CSP) (the county tumor registry), which became available to the study for a 2-month period for African-American (AA) patients and Latina (L) patients with breast carcinoma. The second supplemental source was the California Breast Cancer Treatment Fund (BCTF), a nonprofit organization that funds the treatment of low-income, uninsured patients.
In total, 579 recruitment letters were sent to all potentially eligible patients by the study sites. We were unable to contact 22 of these patients by telephone, and it was found that another 197 patients were ineligible. One hundred twelve patients refused to participate, leaving 248 patients who responded. Of these 248 respondents, 26 patients participated in a preliminary qualitative interview study that was used to help with interview design, and 222 patients completed the interviews for this study. The overall response rate among patients who were not determined ineligible by telephone (including patients whose eligibility could not be confirmed either because they refused by mail or because they could not be contacted otherwise) was 64.9%. The response rates by source were 63.9% for the private practices (138 interviews), 64.3% for the BCTF (36 interviews), and 66.1% for the CSP (48 interviews). We were unable to compare responders with nonresponders due to Institutional Review Board restrictions, which prevented any data collection on patients who did not consent specifically to participate in the study.
We anticipated that age and ethnicity may be significant barriers to patient recruitment into the study, especially given the personal nature of the interview questions with regard to a stigmatized disease. Therefore, to facilitate participation, we offered patients the option of face-to-face interviews versus telephone interviews, methods of administration that have demonstrated equivalent results.29 Forty-four percent of participants chose to be interviewed over the telephone, and 56% of participants chose to be interviewed face-to-face. Interviews were conducted on average 7.2 months after the participant was diagnosed with breast carcinoma. Participants were asked to sign releases for all of their breast carcinoma-related medical records. These medical records were abstracted for breast carcinoma treatment received, treatment dates, and breast carcinoma stage. Finally, all 155 surgeons who treated the participants were surveyed using mailed or faxed, self-administered questionnaires. One hundred seventeen surgeons completed these questionnaires, for a response rate of 76%; we obtained essential sociodemographic data on the remaining 38 surgeons from public electronic sources, specifically, the websites of the California Medical Board (http://www.medbd.ca.gov) and the American Medical Association (http://www.ama-assn.org).
Receipt of information and perceived helpfulness
Based on a previous qualitative interview study of 33 newly diagnosed, older patients with breast carcinoma,30 we identified two primary modalities by which patients receive information about breast carcinoma from physicians. In the first mode, patients were given or referred to tangible sources of information, such as booklets or videotapes. Within the tangible mode, we identified another 10 possible sources of information that either were used or were perceived as desirable by older patients with breast carcinoma in the qualitative study. We included concrete behaviors, such as a physician follow-up telephone call or office staff contact (i.e., only the behavior and not what was communicated), within the grouping of tangible information sources. The tangible information sources assessed in this study were 1) books, booklets, and pamphlets; 2) videotapes about breast cancer; 3) medical records, including breast carcinoma pathology reports and/or dictated consultations from cancer care providers; 4) in-person office staff contact; 5) follow-up phone calls from physicians or their staff; 6) a visit tape recording; 7) linkage with a breast cancer survivor; 8) information about home health assistance; 9) information about cancer hotlines and/or agencies; and 10) information about cancer support groups. Items asked participants whether they were given each type of information by their physicians or staff. For analytic purposes, we created a tangible information support index by summing the total number of tangible information sources provided to each patient by their physicians as a simple count with a range from 0 to 10. The items represented diverse information sources, and no internally consistent scale was hypothesized; therefore, a factor analysis was not performed.31
The second mode by which patients receive information was through direct patient-provider interaction, i.e., physician discussion of specific breast cancer topics (such as tumor aggressiveness). Within the interactive information-giving mode, 15 breast cancer topics that participants in our previous qualitative study30 reported were discussed by cancer care providers and/or that were perceived as important were identified. Patients in the current study were asked which providers, if any, talked to them about 1) the aggressiveness of their tumor, 2) chance of their breast carcinoma reoccurring, 3) how far their disease had or had not spread, 4) the reason for lymph node removal, 5) all treatment possibilities, 6) the specific option of lumpectomy versus mastectomy, 7) wound care, 8) why someone may need chemotherapy, 9) why someone may need radiation therapy, 10) what to expect during breast carcinoma treatment, 11) the patient's likely appearance after treatment, 12) possible causes of breast carcinoma, 13) side effects of lymph node removal, 14) why someone may need to take tamoxifen, and 15) how quickly the breast tumor needed to be treated. Binary indicator variables were formed for each of the 15 interactive information items. These variables had a value of 1 if any physician had discussed the topic with a given patient and 0 if no physician had discussed the topic. For analytic purposes, it was hypothesized that these items formed one dimension of related breast cancer topics. To test this hypothesis, binary factor analysis was performed on these 15 indicator variables using the Mplus software package.32 This analysis indicated that the 15 interactive information items were unidimensional. A summative interactive informational support scale, with a range of 0 to 15, was formed from these 15 items. The reliability of internal consistency (Cronbach α33) for this scale was 0.71.
Regardless of whether or not they were provided with a particular type of information, all participants were asked to rate how helpful each informational support item was or would have been (if it was not actually received) on a scale of 0–10, with 10 indicating extremely helpful and 0 indicating not helpful at all. In addition, participants were asked to rate each physician they saw (e.g., surgeon, medical oncologist, radiation oncologist, and primary care physician) on the overall helpfulness of the kind or amount of information provided by that physician, also on the 0–10 helpfulness scale.
Patient and physician variables
In addition to age and ethnicity, patient variables that were measured in this study included patient education, income (as measured by an item asking whether the patient's financial resources were adequate to cover their financial needs), Health maintenance organization (HMO) insurance, breast carcinoma stage (abstracted from the patients' medical records), and comorbidity (as measured by an adaptation of the Charlson comorbidity index for patient self-report by Katz et al.34). The total number of physicians seen and the visit length at the time of diagnosis (i.e., ≥ 15 minutes vs. < 15 minutes) were assessed by patient report. In addition, patients' perceived self-efficacy in patient-physician interactions, which is likely to facilitate successful elicitation of information from physicians, was measured by the Perceived Efficacy in Patient-Physician Interactions Questionnaire (PEPPI).35 This instrument consists of five statements about the patient's confidence in her ability to obtain needed medical information and attention to her chief medical concern from physicians: Each item was rated by the patient on a 0–10 point scale, ranging from 10 (extremely confident) to 0 (not confident at all). The internal consistency of the PEPPI scale in this sample was 0.94. The patients' social support also may be a factor in effectively soliciting informational support from physicians. Social support was assessed using the Lubben Social Network Scale36 and had an α of 0.79 in this sample. Finally, participants' surgeons were surveyed by mail and by fax about their age, gender, ethnicity, specialty, and type of practice (cancer center, private office, etc.).
Age was dichotomized for unadjusted analyses at ≥ 75 years versus < 75 years due to evidence that disparities in the treatment of patients with breast carcinoma occur among the oldest women.37 Because of skewed distributions, patient ratings of the helpfulness of each of the individual information items were dichotomized at 10 (extremely helpful) versus < 10. Associations of age and ethnicity with receipt of specific types of information and the helpfulness of that information were assessed with chi-square tests. T tests and analyses of variance followed by Duncan multiple-range tests were used to examine age and ethnic differences in the tangible information index and the interactive information scale.
Multiple linear regression analyses were performed to examine the effects of age and ethnicity on the overall tangible and the interactive informational support measures, controlling for the patient and physician predictors described above. Age was used as a continuous variable in the regression models to increase its predictive power, because age was a chief focus of the study.
Finally, we assessed independent age and ethnic group differences in patient ratings of the overall helpfulness of the informational support provided by specific types of physicians using multiple logistic regression analysis. Patient report of the overall helpfulness of information they received from each of the providers by specialty (i.e., surgeon and, where applicable, medical oncologist, radiation oncologist, and primary care physician) constituted the dependent variable in four separate regression models that were dichotomized into scores of 10 (extremely helpful) versus scores < 10. Socioeconomic and health variables that potentially may have confounded the relation between age or ethnicity and perceived helpfulness of informational support were included as independent variables, i.e., education, income (as measured by financial adequacy), comorbidity, and disease stage.
All unadjusted statistical analyses, except for the binary factor analysis described above, were performed using the SAS statistical package (version 6.12; SAS Institute Inc., Cary, NC). Regression modeling was done using Stata (version 7; Statsoft, Tulsa, OK) to obtain estimates of associations corrected for clustering of patients within surgeons. The level for statistical significance was 0.05.
Sociodemographic and health characteristics of the entire sample of 222 participants are shown in Table 1. The average age was 67 years, with 100 participants age 55–64 years, 85 participants age 65–74 years, and 37 participants age ≥ 75 years. Although the sample was predominantly white, there was substantial representation (36%) of AA and L participants. Of the 52 L participants, 34 women (65.4%) reported speaking Spanish better than English. Approximately 31% of the sample reported that their financial resources were not adequate to cover their expenses. Over three-quarters (78%) of the sample had health insurance; a little more that one-third (35%) had HMO insurance.
|Characteristic||No. of patients||%|
|Mean ± SD||66.7 ± 7.9||—|
|Education (≥ college graduate)||74||33.3|
|Marital/partner status (married/partnered)||115||51.8|
|Comorbidity (any major)||86||38.7|
|Breast carcinoma stage (≥ Stage III)||20||9.1|
Tangible Informational Support
Table 2 displays the physician-provided tangible breast cancer information sources, as well as the frequency with which they were given to patients, by patient age and ethnicity. Information sources are listed in rank order of the frequency with which they were provided to the overall sample. The three sources most commonly provided to older women (i.e., age ≥ 75 years) were 1) books, booklets, and pamphlets (81%); 2) support group information (54%); and 3) cancer hotline and agency information (30%) (Table 2). In contrast, the three information sources older women most commonly rated as extremely helpful (i.e., 10 on a scale of 0–10) were 1) physician or staff follow-up phone call (53%); 2) books, booklets, and pamphlets (33%); and 3) contact with an office staff member (32%).
|Sources of breast carcinoma information provideda||Age (%)||Ethnicity (%)|
|< 75 yrs||≥ 75 yrs||White||Black||Latina|
|1. Books, booklets, pamphlets||83.8||81.1||85.8||81.5||76.9|
|2. Support groups||56.8||54.1||58.9||40.7||57.7|
|3. Cancer hotline/agency||47.6||29.7b||36.2||59.3||59.6b|
|4. Medical records/pathology reports||42.7||35.1||39.0||44.4||48.1|
|5. Physician/staff phone call||41.6||36.1||46.4||33.3||30.8|
|6. Office staff member||34.1||35.1||29.1||40.7||46.2|
|8. Visit tape recording||14.6||2.7b||17.0||3.7||3.9b|
|9. Breast carcinoma survivor||13.5||5.4||9.2||22.2||15.4|
|10. Home health agency||7.0||18.9b||7.8||18.5||7.7|
Unadjusted analyses on each of the 10 informational sources demonstrated some age group differences in tangible information provided by physicians. Women age ≥ 75 years received cancer hotline and agency information and a visit tape recording less frequently compared with younger women (Table 2). In contrast, women age ≥ 75 years were more likely to receive home health agency information compared with their younger counterparts. Overall, physicians provided older women and younger women with roughly similar amounts of tangible informational support, as measured by the summary index of tangible information sources (see Table 3).
|Measure||Age (mean ± SD)||Ethnicity (mean ± SD)|
|< 75 yrs||≥ 75 yrs||White||Black||Latina|
|Tangible Information Support Index (range, 0–10)||3.6 ± 1.9||3.1 ± 1.5||3.4 ± 1.8||3.7 ± 2.1||3.6 ± 1.8|
|Interactive Information Support Scale (range, 0–16)||10.5 ± 2.7||9.2 ± 2.6a||11.0 ± 2.7||9.7 ± 2.3||9.0 ± 2.5b|
Women age ≥ 75 years rated all but one information source extremely helpful (i.e., 10 on a scale of 0–10) as frequently or more frequently than younger women (data not shown). The one exception was cancer hotline or agency information, which older women rated extremely helpful 8% of the time, compared with 22% for younger patients (P = 0.047). Older women were especially more likely to rate home health agency information extremely helpful compared with younger patients (30% vs. 14%; P = 0.02).
For AA and L patients with breast carcinoma, the three types of tangible informational support most frequently provided were 1) books, booklets, and pamphlets; 2) cancer hotline/agency information; and 3) support group information (see Table 2). The top three types in terms of helpfulness ratings were 1) physician/staff phone call (AA patients, 63%; L patients, 60%), 2) office staff member (AA patients, 48%; L patients, 52%), and 3) referral to another patient with breast carcinoma (AA patients, 52%; L patients, 48%).
Ethnic group differences in physician provision of tangible informational support were mixed (see Table 2). Physicians provided significantly more cancer hotline and agency information to AA and L patients with breast carcinoma compared with white patients and provided more breast cancer videotapes to AA patients compared with white and L patients. However, ethnic minority patients with breast carcinoma received visit tape recordings less frequently. There were no important ethnic group differences in overall tangible support provided (Table 3).
More consistent ethnic group differences were found for helpfulness of tangible informational support. Both ethnic minority groups rated physician/staff phone calls, referrals to an office staff member for more information, breast cancer videotapes, and home health agency information extremely helpful more often compared with white patients with breast carcinoma (all P < 0.05; data not shown). However, minority patients were less likely compared with white patients to rate receiving a tape recording of their visit and receiving medical records extremely helpful.
Interactive Informational Support
The three topics discussed most frequently by physicians with older women were 1) wound care (98%), 2) tumor aggressiveness (84%), and 3) breast carcinoma spread (81%) (Table 4). In terms of being extremely helpful, the older women rated 1) wound care (83%), 2) breast carcinoma spread (68%), and 3) reason for lymph node dissection (64%) as their top three priorities.
|Topic discusseda||Age (%)||Ethnicity (%)|
|< 75 yrs||≥ 75 yrs||Caucasian||Afr-Amer||Latina|
|All treatment choices||88.1||78.8||85.8||89.0||86.5|
|Breast carcinoma spead||84.3||81.1||85.8||81.5||82.7|
|What to expect during treatment||82.6||81.1||88.7||77.8||68.6b|
|Radiation treatment rationale||82.2||73.0||91.5||59.3||61.5b|
|Lymph node dissection rationale||79.2||78.4||84.9||81.5||63.5b|
|Lumpectomy vs. mastectomy||70.3||64.7||68.2||88.9||63.5c|
|Side effects–lymph node removal||63.1||46.0c||55.5||70.4||68.0|
|Risk of recurrence||56.0||37.8c||67.4||37.1||23.1b|
|Breast carcinoma cause||23.8||13.5||29.8||3.7||11.5b|
Some notable age group differences were demonstrated in interactive informational support provided by physicians. Specifically, physicians discussed four breast cancer topics less frequently with older women compared with younger women: the likelihood of recurrence, side effects of lymph node removal, chemotherapy rationale, and posttreatment appearance (Table 4). Table 3 shows that there also was an age group difference in the scale measuring overall interactive informational support provided. Nonetheless, with the exception of chemotherapy, women age ≥ 75 years were at least as likely as younger women to rate discussions of all topics extremely helpful. Specifically, older women rated discussions of chemotherapy rationale extremely helpful far less often compared with younger women (30% vs. 50%, respectively; P = 0.03). In contrast, older women rated discussions of wound care extremely helpful far more often compared with younger women (83% vs. 63%, respectively; P = 0.02).
With regard to ethnic minority group differences in interactive informational support, physicians discussed a number of breast cancer topics less often with AA and L patients with breast carcinoma compared with white patients with breast carcinoma. These topics included the risk of recurrence; posttreatment appearance; what to expect during treatment; and radiation, chemotherapy, and tamoxifen rationales (Table 4). Less than two-thirds of L patients said their physicians discussed reasons for lymph node dissection and using tamoxifen, compared with over three-quarters of white and AA women whose physicians discussed those issues. Table 3 shows that, compared with white women who had breast carcinoma, women in ethnic minorities who had breast carcinoma received less overall interactive informational support from their cancer care physicians.
There were no ethnic group differences in helpfulness ratings for 13 of 15 breast cancer topics that may have been discussed by cancer care physicians (data not shown). Both AA patients and L patients rated discussions of radiation treatment rationale extremely helpful less often compared with white patients (P = 0.001). L patients rated discussions of the risk of disease recurrence as less helpful compared with white or AA patients (P = 0.001).
The three breast cancer topics discussed that were most frequently by physicians with AA and L patients with breast carcinoma were 1) wound care, 2) tumor aggressiveness, and 3) all treatment choices (see Table 4). The three topics they rated extremely helpful most frequently were 1) wound care (AA patients, 67%; L patients, 64%), 2) all treatment choices (AA patients, 63%; L patients, 69%), and 3) reason for lymph node dissection (AA patients, 52%; L patients, 61%).
Multiple Linear Regression Analyses: Provision of Informational Support
Predictors of physician provision of informational support, both tangible and interactive, using multiple linear regression models containing a wide range of patient and provider characteristics are shown in Table 5. The only significant predictors of tangible informational support were having HMO insurance, a surgeon of Latino ethnicity, and social support. Important predictors of interactive informational support included age (older age was predictive of less support), patient ethnicity (L patients received less support compared with white patients), number of physicians seen, and social support.
|Tangible informationa||Interactive informationb|
|β||SE||P value||β||SE||P value|
|No. of physicians||0.08||0.10||0.43||0.47||0.15||0.00|
|Surgeon: CA center||0.47||0.43||0.28||−0.02||0.44||0.96|
|Time spent at diagnosis (≥ 15 minutes vs. less)||−0.25||0.25||0.33||1.07||0.40||0.01|
Multiple Logistic Regression Analyses: Helpfulness of Informational Support by Provider Specialty
In terms of overall helpfulness of information giving by specific health care provider specialties, older women and L women were particularly likely to rate their surgeon extremely helpful with regard to the amount or kind of information they gave, controlling for education, income, comorbidity, and disease stage (older women vs. younger women: adjusted odds ratio, 2.4; 95% confidence interval [95%CI], 1.1–5.0; P = 0.02; L women vs. white women: adjusted odds ratio, 2.4; 95%CI, 1.1–5.4; P = 0.03). There were no age or ethnic group differences in perceived overall helpfulness of information provided by medical or radiation oncologists in similar logistic regression models. However, older women had 5.5 times greater adjusted odds (95%CI, 1.4–21.3; P = 0.02) compared with younger women of rating information given by their primary care provider extremely helpful. There were no ethnic group differences in helpfulness ratings of information received from primary care providers.
Health care and outcome disparities that have been documented for older patients and for ethnic minority groups have highly complex roots.7, 21 Gaining a clearer understanding of the sources of these differences will require examining health care delivery to these groups at every level, i.e., from the macro level of social and health care systems to the micro level of patient-physician interactions. An essential component of the patient-physician interaction is physician information-giving. Because the treatment of patients with breast carcinoma often is complicated by many modalities and choices, information may play an especially critical role in making appropriate treatment decisions for them. This study is among the first to examine comprehensively informational support provided by physicians to newly diagnosed older patients and ethnic minority patients with breast carcinoma.
The findings of this study indicate that older patients with breast carcinoma receive less interactive informational support from their physicians compared with younger patients. This support disparity does not appear to be due to age group differences in preferences, because older patients rated most informational sources and topics at least as helpful as younger patients rated them, whether or not they were provided with the informational sources. Nor was this difference mitigated by differences in financial resources, education, or comorbidity or in psychosocial factors, such as patient assertiveness in communicating with physicians or the extent of a patient's social support. Furthermore, the age disparity persisted regardless of the number of providers seen or the length of time spent with the patient at the time of diagnosis. A possible reason for the discrepancy in physician information-giving by patient age may be negative, stereotypical attitudes about older people known as ageism.38
Similarly, physicians provided less informational support to ethnic minority women, specifically AA and L women with breast carcinoma. This disparity was particularly notable in the discussion of specific breast cancer topics, although minority patients considered discussions of most breast cancer topics helpful, as did white patients. However, the results of the regression analysis for interactive information-giving suggest that, for AA patients, socioeconomic and psychosocial factors underlie the observed differences in physician provision of informational support. For L patients, ethnic group identity appears to be an independent factor in physician provision of interactive informational support. Whether this finding is due to bias on the part of providers or whether it is due to language barriers (nearly two-thirds of the L patients in this study reported speaking Spanish better than English) cannot be discerned from this study.
Overcoming informational support disparities, and perhaps treatment disparities, will require not only providing more information but also crafting that information to meet specific group needs and preferences. In this study, discrepancies also were present not only in the quantity of information given but also between group informational preferences and what information physicians provided most frequently. In older patients with breast carcinoma, behavioral/interpersonal informational support, such as physician follow-up phone calls or office staff contact, was most preferred, a finding consistent with previous research in older patients with breast carcinoma.39 However, concrete informational support, such as booklets and cancer hotline and support group information, was provided most frequently by physicians. Older patients also found information on home health agency referrals more helpful and found information on cancer agency/hotline referrals less helpful compared with younger patients. These differences in preference likely reflect age-related home support needs after treatment (due to physical frailty and/or less support at home) and stronger preferences for privacy (speaking with strangers on a hotline may be much less preferable than speaking with personal physicians and their staff).
In terms of interactive informational support, i.e., discussions of specific breast cancer topics, the preferences of older patients with breast carcinoma were essentially no different compared with the preferences of younger patients, although older patients appeared to desire more wound care instructions. This finding is consistent with the work of others indicating a greater preference among older patients with breast carcinoma for self-care information compared with younger patients.40
Preferences for interpersonal health professional contacts were marked in particular among ethnic minority patients, with AA and L patients rating follow-up telephone contact by physicians and/or their office staff as more helpful compared with white patients. In contrast, written information, such as medical records or pathology reports, was rated less helpful by patients in both ethnic minority groups compared with white patients. This finding of an ethnic minority preference for information-giving through interpersonal contact, rather than written information, is supported by an Institute on Medicine report detailing cultural preferences for interpersonal contact over written contact for recruitment into cancer clinical trials.7
In terms of interactive informational support preferences, AA and L patients with breast carcinoma rated a large majority of breast cancer topics at least as helpful or more helpful compared with white patients. However, AA and L patients rated discussions of radiation treatment rationale less helpful and L patients rated discussions of breast carcinoma recurrence risk as less helpful compared with white patients. The reasons for these differences are unclear. With regard to radiation treatment, higher rates of mastectomy in the ethnic minority groups in this study (63% of AA patients and 44% of L patients vs. 21% of white patients) may have made the issue of radiation less germane. A greater degree of fatalism described among L patients41 may account for their relative lack of interest in discussing the risk of breast carcinoma recurrence.
It is interesting to note that both older patients and ethnic minority patients with breast carcinoma rated their surgeons, and older patients rated their primary care physicians, as considerably more helpful in their overall information giving compared with younger patients and white patients, although they received less information from them. This discrepancy between actual information provided and patient ratings of provider information giving is of concern for a number of reasons. First, inflated patient perceptions of the informational support provided may perpetuate disparities by masking patient demand or need for information. Second, other researchers have cited patients' illusion of competence in treatment decision-making in documenting patients' exaggerated perceptions of the information provided relative to what actually was discussed.42 The illusion that a patient is more informed than she really is particularly may undermine appropriate treatment decision-making in a patient with breast carcinoma, given its potentially complicated nature with many options, thereby perpetuating treatment disparities. Nonetheless, the primary care physician appears to play a particularly important role in providing information to older patients and specifically should be incorporated into their care by physicians as much as possible.
Limitations of this study include the fact that, although we were able to sample a fairly diverse population ethnically, it was not a probability sample of patients with breast carcinoma. Therefore, the external generalizability of our findings to the larger population of all older patients with breast carcinoma is limited. However, the fact that age-related and ethnic group-related differences in the provision of information were found in a relatively well-educated sample suggests that such differences may be even more prominent in the general population of older patients who are newly diagnosed with breast carcinoma.
Another limitation is that patient-physician interactions in this study were measured by retrospective patient report, making recall bias a potential confounder of our results. However, it has been noted that people who have undergone a sudden and life-threatening health crisis manifest very clear recall of the details surrounding the event: This recall has been referred to as flashbulb memories.43 Others have reported minimal recall bias for medical visits made in the previous year.44 Moreover, a recent study of patients with prostate carcinoma demonstrated that 6-month recall of prediagnostic functioning was identical, item-by-item, to that reported at baseline in a large majority of men.45 Finally, we maintain, along with others, that patient perceptions of the communication, not reality, are the basis of action, including breast carcinoma treatment decision-making and treatment received.42
Because older patients with breast carcinoma often are less assertive in communicating with physicians,46 and because it has been shown that they actively seek less information compared with younger women in making treatment decisions,47 physicians may assume that they do not need the level of communication and information compared with younger patients who have breast carcinoma. This study indicates that the informational needs of older patients with breast carcinoma are quite similar to the needs of younger patients, although older patients are given less informational support. Overcoming these informational deficits in the form of exposure to more providers has been related to the receipt of definitive therapy for breast carcinoma therapy in older women.48 Our results suggest that it may not be necessary to refer older patients with breast carcinoma to multiple providers to bridge the information gap and thereby reduce treatment disparities. The same amount of information that is provided to younger patients with breast carcinoma by any given individual physician may be all that is required, which also was suggested in a study that indicated a correlation between the number of treatment options discussed and the receipt of BCS.49 In addition, delivering age pertinent and/or ethnic group pertinent information on a personal basis, rather than relying on inanimate or other less private interpersonal sources of information, may further pave the way to optimizing the provision of information and reducing breast carcinoma treatment disparities in these vulnerable patient populations.
The authors thank Drs. M. Douglas Anglin and David B. Reuben for their helpful critique of the article and Dr. Patricia A. Ganz for her invaluable advice on the design of the study.
- 1Institute for the Future. Health and health care 2010. Princeton: Robert Wood Johnson Foundation, 2000: 17.
- 2U.S. Bureau of the Census. Current population report, series P25-1130: population projections of the United States by sex, race, and Hispanic origin, 1995–2050. Washington, DC: U.S. Department of Health and Human Services, 1996.
- 7Institute of Medicine. The unequal burden of cancer: an assessment of NIH research and programs for ethnic minorities and the medically underserved. Washington, DC: National Academy Press, 1999.
- 27A randomized controlled trial of facilitating information giving to patients with chronic medical conditions: effects on outcomes of care. J Fam Pract. 1999; 5: 356–363., , .
- 31Clinimetrics. New Haven: Yale University Press, 1987..
- 32Mplus: the comprehensive modeling program for applied researchers. Los Angeles: Muthén and Muthén, 1999., .
- 36Assessing social networks among elderly populations. Fam Commun Health. 1988; 11: 42–52..
- 43Flashbulb memories. In: NeisserU, editor. Memory observed. San Francisco: Freeman and Company, 1982: 23–40., .
- 44AndersonR, KasperJ, FrankelM, et al., editors. Total survey error: applications to improve health surveys. San Francisco: Jossey Bass, 1979: 131.
- 46Problems in communication between physicians and older patients. J Geriatr Psychiatr. 1996; 29: 13–32., , .