Factors That Determine Satisfaction With Surgical Treatment of Low-Income Women With Breast Cancer

Objective  To analyze the relationship between patient satisfaction with surgical treatment and 4 consultation skills and processes of the surgeons (time spent, listens carefully, explains concepts in a way the patient can understand, and shows respect for what the patient has to say), controlling for a range of patient, surgeon, and treatment characteristics.

Design  Cross-sectional survey.

Setting  The Breast and Cervical Cancer Treatment Program for the state of California.

Patients  A statewide sample of 789 low-income women who received treatment for breast cancer from February 1, 2003, through September 31, 2005.

Main Outcome Measure  Satisfaction with surgical treatment.

Results  Three of every 4 women reported being extremely satisfied with the treatment they received from their surgeon. African American women and those with arm swelling were less likely to be satisfied, whereas those reporting that the surgeon always spent enough time and explained concepts in a way they could understand were more likely to report greater satisfaction.

Conclusion  Our findings highlight the importance of 2 relatively simple behaviors that surgeons can easily implement to increase patient satisfaction, which can be of potential benefit in the litigious world of today.

Surgeons play a key role in the treatment of patients with breast cancer. In the clinical environment of today, the traditional indicators of clinical care quality (such as 5-year survival rate) are increasingly being complemented with patient-centered outcomes, such as patient satisfaction. There is a rich body of literature that focuses on determinants of patient satisfaction with their physicians, their treatment, and the decision-making process.^1-5 In terms of surgical treatment of breast cancer, the literature focuses on patient satisfaction with body image, decision regret, treatment decision making, and concordance with treatment wishes.^1,5-8 A few studies^1,5 have ascertained the relationship of overall patient satisfaction with the treatment provided by consultation process and skills of the surgeon. In this limited subset, most studies control for patient characteristics (eg, age, education level, income, disease stage) and those of the surgeon, but few control for confounding caused by factors such as treatment delay, patient self-efficacy, surgical concordance, and surgical outcomes. In addition, none of the studies examines this relationship in a population of low-income women.

In a statewide sample of 789 low-income women in California who received treatment for breast cancer under the Breast and Cervical Cancer Treatment Program (BCCTP) of California, we analyzed the relationship between satisfaction with surgical treatment and 4 consultation skills and processes of the surgeon (time spent, listens carefully, explains concepts in a way the patient can understand, and shows respect for what the patient has to say), controlling for a range of patient, surgeon, and treatment characteristics.

The details of the study design and sample have been described in a previously published article⁹ and are reviewed briefly herein. We conducted a cross-sectional survey of low-income women 18 years and older living in California who were newly diagnosed as having breast cancer. The study was approved by the UCLA Human Subjects Protection Committee and the California State Department of Health Services Human Subjects Protection Committee. A consecutive sample of all women treated through the California BCCTP from February 1, 2003, through September 31, 2005, was recruited for the study. The BCCTP was legislated by the federal government as part of the Breast and Cervical Cancer Prevention and Treatment Act of 2000 to fund the treatment of breast and cervical cancer for uninsured and underinsured, low-income women (at or below 200% of the federal poverty level). The BCCTP in California is supported via federal funding through Medicaid and separate state funding.

A total of 921 women were recruited into the study. Women who did not speak English or Spanish, had a previous history of breast cancer, or were receiving treatment for another type of cancer were excluded from the study. Our overall response rate was 61.0%. Compared with survey responders, nonresponders tended to be older (52 vs 50 years, P < .001), were more likely to be Asian/Pacific Islanders (11.6% vs 7.4%), and less likely to be Latina (37.6% vs 53.4%, P < .001). Further details about the design and flow of the study can be found in a previous article.⁹ Our analysis for this article is based on 789 women who completed telephone interviews approximately 18 months after diagnosis and for whom medical records from a systematic medical record abstraction were available.

Variables were constructed by means of both patient self-report and medical record abstraction data. Our dependent variable was based on the response to the question, “Altogether, how satisfied are you with your treatment by the surgeon?” A similar format has been used in the literature to assess patient satisfaction.^6,10-13 Response categories were based on a 5-point Likert scale: extremely satisfied, somewhat satisfied, neither satisfied nor dissatisfied, somewhat dissatisfied, and extremely dissatisfied. Because of the skewed nature of the responses (74.0% reported being extremely satisfied), we dichotomized the variable (extremely satisfied vs not extremely satisfied).^8,13

The key independent variables that captured the consultation skills and process were based on the 4 questions from the Consumer Assessment of Healthcare Providers and Systems survey,¹⁴ which evaluates the surgeon-patient interaction. These questions were as follows: (1) “How often did [the surgeon you saw the most] spend enough time with you?” (2) “How often did he/she listen carefully to you?” (3) “How often did he/she explain things in a way you could understand?” and (4) “How often did he/she show respect for what you had to say?” Responses to these questions were on a 4-point Likert scale: never, sometimes, usually, and often. Because of the sparse responses in the “never” category, we combined the “never” and “sometimes” responses into 1 group.

Control variables included characteristics of the patient, surgeon, and treatment. Patient characteristics included age (≤50 or >50 years), education level of less then grade 12 (no or yes), ethnicity (white, African American, Latina, or other), annual household income (<$10 000, $10 000-$19 999, $20 000-$29 999, or ≥$30 000), and stage of disease (0, 1, 2, 3, or 4). Self-reported general health status was assessed with the question, “In general, would you say your health is . . . ?” and measured on a 5-point scale from 1 (poor) to 5 (excellent). Depression was evaluated by means of the Center for Epidemiologic Studies–Depression (CES-D) scale.^15-17 Patient self-efficacy was measured by means of the validated Perceived Efficacy in Patient-Physician Interactions (PEPPI) questionnaire.¹⁸ The PEPPI questionnaire measures the perceived ability of patients to obtain needed medical information and attention to their chief medial concerns from physicians. The PEPPI sum scale ranges from 0 to 50, and the Cronbach α in our sample was 0.96.

Treatment characteristics included type of surgery (lumpectomy, mastectomy, or other/unsure), treatment delay (coded “yes” if the woman reported a period of ≥60 days between the time she first became aware that something was wrong and the first invasive diagnostic procedure [ie, biopsy or surgery]), surgical concordance (“yes” if the woman received the type of surgery she preferred and “no” if she did not), and presence in the previous 4 weeks of arm swelling and decreased arm movement. We also controlled for sex of the surgeon and time elapsed since surgery to adjust for confounding by time.

Descriptive statistical analyses were used to understand the characteristics of the patients and surgeons. Unadjusted bivariate relationships between the dependent and independent variables were examined. Logistic regression was used to obtain the odds ratios between the dependent variable and the regressors of interest, controlling for other covariates. Regression diagnostics were conducted by means of the Hosmer-Lemeshow goodness-of-fit tests. Second-order interactions were tested for, but none were found to be significant in the final models. P < .05 was considered statistically significant. All data analysis was performed by means of Stata/SE statistical software, version 10.0 (Stata Corp, College Station, Texas).

Our sample of 789 women was predominantly Latina (53.9%), with whites accounting for approximately a third (31.9%) of patients (Table 1). Half the women (50.7%) were older than 50 years, and most (58.5%) had an education level of more than grade 12. More than half (60.1%) had an annual household income of less than $20 000. Almost one-third of the women (31.2%) had early-stage disease (0 or 1), and more than one-third had stage 2 disease (36.4%). Approximately half the women reported that they had undergone lumpectomy (51.6%), and more than half (51.3%) reported at least a 60-day delay in undergoing biopsy or surgery from the time they noted something was wrong.

Most women reported being treated by a male surgeon (76.4%). Most believed that the surgeon always spent enough time with them (51.9%), always listened carefully to what they had to say (63.9%), always explained concepts in a way they could understand (62.5%), and always showed respect for what they had to say (62.6%). Less than three-fifths of women (58.2%) received the type of surgical procedure they preferred. Almost three-quarters of women (73.5%) did not report having any arm swelling during the preceding 4 weeks, and three-fifths (60.1%) did not report any decreased arm movement during that time. The mean (SD) score on the self-reported general health scale was 3 (1), and the mean (SD) score on the CES-D scale was 9.2 (7.5).

In the unadjusted bivariate analyses, among patient characteristics, statistically significant associations were noted between satisfaction and ethnicity, patient self-efficacy (as measured by the PEPPI scale), self-reported general health, and CES-D score (Table 2). African American women and those categorized as “other” were less likely to report being extremely satisfied. Women who scored higher on the self-efficacy scale were more likely to report being extremely satisfied, whereas an increase in the general health and CES-D scores were associated with lower likelihood of being extremely satisfied. Among surgeon characteristics, all 4 surgeon-patient communication measures (spending enough time, listening carefully, explaining concepts, and showing respect) were strongly associated with satisfaction. None of the treatment characteristic variables was associated with satisfaction.

After controlling for other variables in the logistic regression model, these findings were attenuated, as depicted in the last column of Table 2. African American women had significantly lesser odds of being extremely satisfied with treatment from the surgeon compared with white women. The PEPPI score remained significant, with effect size nearly identical to that in the unadjusted analyses. Among the 4 measures of surgeon-patient communication, only spending enough time and explaining concepts remained significant, with an attenuation of the effect size compared with the unadjusted analyses. Women who had arm swelling were less likely to report being satisfied with treatment from the surgeon compared with those who had not had arm swelling in the previous 4 weeks.

In our statewide sample of low-income women in California, 3 of every 4 women reported being extremely satisfied with the treatment they received from their surgeon. This high percentage of satisfaction is similar to that reported in other studies.^1,2,5,8,19 The BCCTP is a safety net for women who “fall through the cracks” and are not eligible for other insurance coverage options. Thus, low-income women in this sample are relatively enabled in terms of their access to care by virtue of their enrollment in the BCCTP. Thus, it is highly probable that this high rate of satisfaction may reflect a sense of gratitude among women for the receipt of care for a life-threatening condition.

Among the 4 measures of patient-physician communication, 2 were statistically significantly associated with satisfaction: time spent and explanation of concepts. The breast cancer diagnosis is often experienced as being devastating, and patients are likely to seek succor from their treating physicians. Spending enough consultation time with the patient is 1 way in which surgeons can underscore the gravity of the situation; in addition, it allows them adequate time to address the concerns and questions of the patient. Evidence from a regional breast cancer center suggests that surgeon consultations average approximately 20 minutes,²⁰ and a recent report¹ suggests that breast cancer patients whose consultation was longer than 15 minutes were more likely to characterize the surgeon-patient interaction as extremely helpful. To explain things in a manner that the patient can understand may be helpful to allay and address patient concerns and questions and to decrease patient anxiety about surgery that is often invasive and sometimes mutilating.²¹

Compared with white women, African American women were significantly less likely to report being extremely satisfied with treatment from their surgeon; this result is similar to that of Lantz et al.⁸ The finding could be owing in part to the well-documented disparities in health care that African Americans have historically experienced.^22-24 Even in our sample, a significant difference was found in the self-reported treatment delay; African American women were significantly more likely to report at least a 60-day treatment delay (data not shown). The self-efficacy of a woman was related to higher satisfaction; women who were more self-efficacious could presumably obtain the information they needed from the surgeon during the consultation, thus increasing their satisfaction with treatment from the surgeon. Self-efficacy has previously been shown to be a predictor of satisfaction.^1,20,25

Women who had had arm swelling during the preceding 4 weeks were significantly less likely to report being satisfied compared with women who had not had arm swelling. Mandelblatt et al⁶ found a similar relationship in a sample of older women (>67 years) with stage 1 or 2 disease. The mean number of days since surgery in our study was 473 (median, 499 days), so presence of arm swelling at the time of the interview would imply a persistent postsurgical sequela, which would certainly affect the satisfaction level of a patient. However, the presence of arm swelling was not statistically significant in the bivariate (unadjusted) analysis. This is because arm swelling was confounded by ethnicity; in the presence of arm swelling, African American women were more likely to report lower satisfaction compared with all other groups (data not shown). When the effect of ethnicity was controlled for in the regression model, the true effect of the arm swelling became apparent.

From a policy perspective, our findings highlight the importance of 2 relatively simple behaviors that surgeons can easily implement to increase patient satisfaction. Although our study does not provide any indication of how long an ideal consultation should be, it indicates that more is always better; to the extent the busy schedule of a surgeon allows, the benefits of spending more time with a patient are significant in terms of patient satisfaction. In addition, the surgeon should explain concepts, be they surgical options, postoperative care, or relative risks of treatment options, in a way that the patient understands. This usually entails eschewing medical jargon and not overestimating the capacity of the patient to absorb, interpret, and retain information; the corollary here is that simpler is better.^26-28 Such behaviors from surgeons have the potential to not only improve patient satisfaction but also to improve compliance and thereby outcomes, which will decrease potential future litigation.²⁹ For example, the American College of Surgeons' Closed Claims Study²⁹ found that better communication could have prevented complications in 22% of patients; in addition, for those patients among whom the expected quality of surgical care was satisfactory, communication and practice-pattern violations were the most common deficiencies in provision of care.

Some caveats should be borne in mind as one interprets the results of our study. First, our data are cross-sectional, which precludes any interpretation of causality. Second, our sample is a low-income population in California that was relatively enabled because of enrollment in a specific program; thus, our results may not be applicable to other states or populations. Third, a relatively small number of women in our study underwent breast reconstruction, so we were not able to assess the effect of breast reconstruction with regard to satisfaction. Reconstruction has been linked to satisfaction in the literature.³ Fourth, our data were obtained 18 months after the enrollment of the women in the BCCTP. Recall bias could, therefore, be an issue, although some reports^30,31 suggest that patients can accurately report treatment received in the past. Controlling for time since surgery did not change our regression results.

Despite these caveats, our study has many strengths. It is among the few, to our knowledge, that has analyzed a statewide sample of low-income women. Our sample size is relatively large, and our data set allows us to control for important confounders that have not been controlled for in the past, such as treatment delay, patient self-efficacy, surgical concordance, and surgical outcomes, thus lending credence to our findings.

In conclusion, analysis of our sample of 789 low-income women who received care funded through the California BCCTP suggests that most are extremely satisfied with the treatment provided by their surgeon. Spending enough time and explaining things in a way that the patient can understand are 2 clinical skills or processes that significantly affect this level of satisfaction. In addition, being African American, reporting the presence of arm swelling in the past 4 weeks, and having a low self-efficacy score are associated with lower satisfaction. Future research would be helpful in identification of the duration of an ideal consultation.

Correspondence: Amardeep Thind, MD, PhD, Center for Studies in Family Medicine, Schulich School of Medicine, The University of Western Ontario, 245-100 Collip Cir, London, Ontario N6G 4X8 Canada (athind2@uwo.ca).

Accepted for Publication: November 14, 2008.

Author Contributions:Study concept and design: Thind, Diamant, and Maly. Acquisition of data: Diamant, Liu, and Maly. Analysis and interpretation of data: Thind, Diamant, and Maly. Drafting of the manuscript: Thind. Critical revision of the manuscript for important intellectual content: Thind, Diamant, Liu, and Maly. Statistical analysis: Thind and Liu. Administrative, technical, and material support: Diamant and Maly.

Financial Disclosure: None reported.

Funding/Support: The study was funded by grant TURSG-02-081 from the American Cancer Society and grant 7PB-0080 from the California Breast Cancer Research Program. Dr Thind is funded by a Canada Research Chair in Health Services Research.