|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2001 American Society for Clinical Oncology Influences on Oncologists’ Adoption of New Agents in Adjuvant Chemotherapy of Breast CancerFrom the Wellmark Foundation, Des Moines, and University of Iowa, Iowa City, IA. Address reprint requests to William R. Doucette, PhD, University of Iowa, S518 PHAR, Iowa City, IA 52242; email: william-doucette@ uiowa.edu.
PURPOSE: Little is known about how oncologists’ adopt new treatments for breast cancer. This study investigated influences on oncologists’ adoption of paclitaxel as adjuvant chemotherapy for early-stage breast cancer, 9 months after presentation of phase III data suggesting improved disease-free and overall survival when paclitaxel was added to doxorubicin and cyclophosphamide for such patients. METHODS: Self-reported data were collected with a mail survey of a random sample of 1,200 oncologists practicing in the United States. Using Rogers’ model, we measured four types of influences on adoption of innovation: (1) communication channels, (2) innovation characteristics, (3) a practitioner’s social system, and (4) physician characteristics. Multiple regression analysis assessed the associations between oncologist adoption of paclitaxel for early-stage breast cancer patients and variables representing the modeled influences on adoption.
RESULTS: On average, respondents (n = 181) reported having adopted paclitaxel for 37% of their early-stage breast cancer patients. The overall model was significant, with seven variables associated (P CONCLUSION: As new modalities become available to treat cancer, it is vital to understand what factors influence oncologists and patients when choosing to use them. Those parties interested in fostering the adoption of new breast cancer treatments should address features of communication channels (eg, use of symposia), characteristics of new treatments (eg, perceived advantage in efficacy), physicians’ social systems (eg, patient requests), and characteristics of potential adopters (eg, previous experience with the treatment).
ADJUVANT CHEMOTHERAPY after primary surgical resection for early-stage breast cancer improves relapse-free and overall survival in select populations of affected women. The optimal combination and schedule of chemotherapy drugs remains the focus of investigation. In the United States, although there is no consensus regarding the optimal schedule, combinations of cyclophosphamide, methotrexate, fluorouracil, and doxorubicin represented the backbone of virtually all commonly used regimens until the recent introduction of taxanes.1 Attention is turning from the relative value of the anthracyclines to the optimal role of the taxanes in combination with the traditional agents. Given the ubiquitous nature of breast cancer, improvements in treatment are of strong interest to practitioners and policy makers. As a result, significant resources are devoted to the development and rapid dissemination of new knowledge regarding this disease. Despite the activities of the National Cancer Institute, the American Cancer Society, and others, we know little about the factors that lead to physicians’ acceptance of new information or how that information influences their adoption of new cancer treatments. The focus of this study was the adoption by oncologists of paclitaxel as a component of adjuvant polychemotherapy for early-stage breast cancer after reports of paclitaxel efficacy in this setting. An abstract presented at the Thirty-Fourth Annual Meeting of the American Society of Clinical Oncology (ASCO) in May 1998 suggested a statistically significant disease-free and overall survival advantage associated with paclitaxel as an adjuvant treatment for select women with node-positive breast cancer.2 This randomized study stemmed from promising activity demonstrated for taxanes in metastatic disease and on promising phase II data in high-risk early-stage patients.3 The ASCO presentation was potentially significant as the first randomized study in years to demonstrate a survival advantage in early-stage breast cancer using a fundamentally new class of drugs, yet it was not subject to full peer review. The availability of this new evidence regarding the efficacy of paclitaxel for treating patients with early-stage breast cancer presented an opportunity to evaluate the factors influencing adoption of this new treatment strategy by practicing oncologists. Previous studies have shown that there are differing rates of adoption for new medical treatments for cancer. Several studies have specifically addressed the differences in adoption of newer breast-sparing surgical treatment in women with breast cancer.4-9 There also have been a few studies addressing the issue of variation of adjuvant chemotherapy treatments.10-12 These studies, however, have not specifically addressed which factors affect physicians’ decisions to adopt a chemotherapy treatment. This study was designed to identify factors associated with oncologists’ adoption of paclitaxel for the adjuvant treatment of women with node-positive early-stage breast cancer during a time period closely following the report made at the 1998 ASCO Annual Meeting. The model for this study was adapted from Rogers’ diffusion of innovation model.13 This model has been used to guide the diffusion of guidelines for cancer pain control,14 as well as to compare the adoption of cancer control techniques with the adoption of cancer treatments.15 Rogers’ model posits that physician adoption behavior is influenced by four types of factors: (1) communication channels, (2) innovation characteristics, (3) the social system, and (4) physicians’ characteristics. A communication channel is a means by which information about a treatment gets from one person to another.13 In this case, we were interested in which possible communication channels an oncologist uses regularly to stay informed about the evidence for breast cancer treatments. Possible communication channels could include scientific articles, symposia, the National Cancer Institute, other oncologists, continuing education programs, and advertisements. Innovation characteristics for a chemotherapy agent involve relative advantages in efficacy, patient survival, patient tolerance, and cost-effectiveness. Oncologists form their perceptions of the relative advantages of a drug through evaluation of the published clinical evidence, from their own experience with the drug, and through interactions with other practitioners. A drug that shows a relative advantage in efficacy is likely to be adopted more quickly than an agent that does not offer such an advantage. Similarly, a relative advantage in patient survival should be an important influence on the adoption of a chemotherapeutic agent. An advantage in patient tolerance can be important but may not be sufficient to overcome the absence of an advantage in efficacy or patient survival. Although cost-effectiveness has been important for the adoption of drugs in some categories, an advantage in cost-effectiveness may not be vital for adopting a chemotherapy agent. The life-and-death nature of cancer may have a dampening effect on the influence of cost-effectiveness. A social system refers to individuals and organizations that work together toward a common goal.13 For cancer treatment, a social system could involve providers and patients, as well as broader factors that could influence them, such as the degree of rurality in a setting. In this study, the social system was an oncologist’s practice setting, including breast cancer patient volume, active physician involvement in promoting new cancer treatments, the number of oncologists in the practice group, patient treatment preference variables, and urban versus nonurban practice location. Physicians’ characteristics are the fourth category of modeled influences on physician adoption behavior. These characteristics could include demographics, education and training, practice specialty, and practice experience. For this study, oncologist characteristics were the number of years in practice, sex, and adoption of paclitaxel for stage IV breast cancer patients. The main objective of the study was to determine whether variables representing Rogers’ four characteristics influenced the frequency with which oncologists incorporated the use of paclitaxel for adjuvant chemotherapy of node-positive early-stage breast cancer patients in the 9 months after the initial report of the first large randomized study of paclitaxel in this population.
This study was a national mail survey of 1,200 oncologists in the United States, randomly selected from all 50 states. The sample frame was the ASCO mailing list, which contains the names of all oncologists in the United States who are members of ASCO. The main data collection occurred in February 1999, approximately 9 months after the sentinel report regarding paclitaxel was made at the 1998 ASCO Annual Meeting and published in the Program and Proceedings, which was distributed to all ASCO members. The mail survey contained 23 items: one for the dependent variable (adoption of paclitaxel for early-stage breast cancer patients) and 22 variables that made up the four influences on adoption. Seven variables measured the communication channels, four assessed innovation characteristics, eight measured social system, and three items asked about physician characteristics. The items were pretested with a small panel of oncologists to improve reliability. As shown in Table 1, the measures used a variety of scales and quantitative data.
The empirical model of oncologists’ adoption of paclitaxel was tested using multiple linear regression analysis with oncologists’ adoption of paclitaxel for node-positive early-stage breast cancer as the dependent variable. Adoption was measured by the percentage of node-positive early-stage breast cancer patients reported as being treated with paclitaxel. The 22 independent variables in the regression were the measures of communication channels, innovation characteristics, the social system, and physicians’ characteristics listed in Table 1. Before the regression analysis, descriptive statistics were calculated for each measure. For the variables measured with scales, the numerical responses were used to calculate means and SDs. SPSS for Windows (Version 8.0, SPSS, Inc, Chicago, IL) was used on all statistical analyses.
Of the 1,200 surveys that were mailed, four were returned as undeliverable. Two hundred forty-seven responses were received, for a response rate of 21%. Of these 247 responses, 12 were unusable because of missing data. Of the 235 usable responses, 181 respondents (77%) reported that they had determined treatment for a patient with breast cancer in the past 6 months. These 181 responses were used in the analyses reported here. Descriptive analyses ( Table 2) show that the mean percentage of physician adoption of paclitaxel for stage II/III node-positive breast cancer was 37.2% (SD, 35.6%). That is, on average, respondents reported using paclitaxel for 37.2% of their early-stage breast cancer patients. In contrast, physician adoption of paclitaxel for stage IV breast cancer patients was 62.4% (SD, 31.0%). Mean length of time in practice of the responding oncologists was 16.6 years, and almost 79% were male. The average number of breast cancer patients cared for per year was 115.2 (SD, 125.8), and the average number of breast cancer patients participating in clinical trials was 8.64 (SD, 14.0).
The multiple regression model for oncologist adoption of paclitaxel for node-positive early-stage breast cancer patients was significant, with an F-test value of 4.26 and a significance level of .001 ( Table 3). This means that the set of 22 independent variables significantly explains the variation in oncologists’ adoption of paclitaxel for early-stage breast cancer patients. The R-square value 0.456 estimates that the model explains approximately 45.6% of the variation of the adoption measure. In addition, the t values test the association between adoption and each of the 22 independent variables. Coefficients (ie, standardized beta) were significant (P < .05) for seven independent variables: use of print advertisements as therapy information source (ß = -0.213), use of symposia on breast cancer as therapy information source (ß = 0.259), number of times interacted with oncologists outside practice setting (ß = 0.165), perceived advantage of paclitaxel in efficacy (ß = 0.217), percentage of patients requesting a treatment about which they obtained outside information (ß = 0.163), working with a pharmaceutical company to bring treatment regimens into a practice community (ß = -0.235), and physician adoption of paclitaxel for stage IV breast cancer patients (ß = 0.176). In addition, two independent variables were significant at P < .10: perceived advantage of paclitaxel in patient survival (ß = 0.218) and the number of oncologists on staff in main practice setting (ß = 0.157).
Our results show that many oncologists use paclitaxel for some of their node-positive early-stage breast cancer patients. This is a remarkable result, because at the time of the survey, the data released on using paclitaxel for early-stage breast cancer patients were in a non–peer-reviewed report of an ongoing study involving a distinct subset of patients. However, oncologists likely were obtaining firsthand experience with paclitaxel by using it to treat late-stage breast cancer patients. Knowledge gained from such practice experience seems to have supported oncologist adoption of paclitaxel for early-stage breast cancer patients. Three variables in oncologists’ communication channels were significantly associated with their adoption of paclitaxel for treating node-positive early-stage breast cancer patients: use of symposia on breast cancer as therapy information source, number of interactions with oncologists outside of practice, and use of print advertisements as therapy information sources. The first two factors had a positive influence on oncologist adoption of paclitaxel, whereas the use of print advertisements had a negative association. Symposia tend to provide therapy updates, often presenting the most recent clinical information, even if not yet peer reviewed. Thus oncologists who regularly use such a communication medium may be more informed about the latest information about cancer treatments, even preliminary findings. Similarly, interactions with oncologists outside of their practice setting are an informal communication channel that can raise the likelihood of learning about advances in cancer treatment based on practice experiences or ongoing trials. Oncologists who used print advertisements as a therapy information source tended to have a lower rate of adoption of paclitaxel for early-stage breast cancer patients. Such advertisements are highly regulated by the United States Food and Drug Administration and would contain only approved information, which may not reflect the latest clinical evidence (ie, findings presented at a meeting). Oncologists’ perceptions about any treatment will weigh into their chemotherapy choices. These results show that a perceived advantage in efficacy was a positive influence on their adoption of paclitaxel for early-stage breast cancer patients. Also, we found that oncologists’ beliefs about paclitaxel’s influence on patient survival approached significance (P = .054). In contrast, any perceived advantage in cost-effectiveness had almost no association with physician adoption of paclitaxel. Given the life and death nature of cancer, it is not surprising that cost-effectiveness has less influence on physician adoption than does efficacy or survival advantages. The results also show that patients’ involvement in their treatment influenced oncologists’ adoption of paclitaxel. That is, the percentage of patients who had requested a treatment for which they had sought outside information was positively associated with the adoption of paclitaxel for node-positive early-stage breast cancer patients. This is an important sign of the times, as it signals that patients are being increasingly proactive in their treatment. In addition, materials available to educate them, especially those available on the Internet, are becoming more readily accessed and used. An oncologist’s use of paclitaxel with stage IV breast cancer patients was a significant positive influence on adoption of paclitaxel for early-stage breast cancer treatment. Knowledge gained through experience is highly credible. Thus oncologists’ observations of the clinical effectiveness of paclitaxel in late-stage breast cancer probably gave them some confidence of its utility in treating early-stage cancer. There was no apparent influence by physician demographics on adoption behavior. Similarly, the rurality of a practice was not a significant influence on oncologists’ adoption of paclitaxel, suggesting a reassuring efficiency of information dissemination across the country. A variable of oncologists’ social system, working with a pharmaceutical company to bring treatment regimens into a practice community, showed a significant negative association with adoption of paclitaxel for early-stage breast cancer patients. Perhaps oncologists who are involved in such activities become advocates for the therapy in which they are involved and have low likelihood of adopting a competing treatment (eg, paclitaxel). Alternatively, such oncologists could be more experienced scientists and view preliminary findings more skeptically than other oncologists. The role of change agents in oncology (ie, advocates for a particular treatment) should be further investigated. Several limitations should be considered when interpreting the findings of this study. One is that self-reported data were collected at one point in time only. We were unable to examine physician adoption over time. Longitudinal study of oncologist adoption of treatments would provide more details on this important process. Also, a combination of self-reported and objective data (eg, patient chart review, claims data) would allow assessment of the validity of the self-reported data. A second limitation is the relatively low response rate. This raises concern that nonrespondents differed from those oncologists who did respond. However, our respondents did not greatly differ from the ASCO membership in terms of sex (80.3% male v 75.0% male) and oncology specialty (75.2% medical or medical/hematology v 63.6% medical or medical/hematology). Also, the obtained response rate did not differ much from those typically reported for surveys of physicians. The model explained almost one half of the variation in oncologists’ adoption of paclitaxel for early-stage breast cancer patients, which suggests that improvements can be made in the model. One example would be to account for the financial incentives that are present in a physician’s practice setting (ie, social system). The presence of capitation could influence physicians to use one treatment over another, affecting adoption of a particular treatment. Future research should address the financial incentives facing oncologists when adopting treatments for their patients. Understanding the influences on oncologists’ adoption behavior may make it possible to increase the efficiency of adopting a new treatment regimen or medical innovation. Data for other new treatment modalities incorporated into adjuvant therapy (eg, monoclonal antibodies, high-dose chemotherapy) will soon be available, and it will be useful to understand which factors influence practicing physicians and patients in choosing to use new modalities. Those parties interested in fostering the adoption of new breast cancer treatments should address features of communication channels (eg, use of symposia), characteristics of new treatments (eg, perceived advantage in efficacy), physicians’ social systems (eg, number of oncologists in practice; patient requests), and characteristics of potential adopters (eg, previous experience with the treatment).
1. Early Breast Cancer Trialists’ Collaborative Group: Polychemotherapy for early breast cancer: An overview of the randomised trials [see comments]. Lancet 352: 930-942, 1998[Medline] 2. Henderson IC, Berry D, Demetri G, et al: Improved disease-free and overall survival from the addition of sequential paclitaxel but not from the escalation of doxorubicin dose level in the adjuvant chemotherapy of patients with node-positive primary breast cancer. Proc Am Soc Clin Oncol 17: 101a, 1998 (abstr 390A)
3.
Hudis C, Seidman A, Baselga J, et al: Sequential dose-dense doxorubicin, paclitaxel, and cyclophosphamide for resectable high- risk breast cancer: Feasibility and efficacy. J Clin Oncol 17: 93-100, 1999 4. Albain KS, Green SR, Lichter AS, et al: Influence of patient characteristics, socioeconomic factors, geography, and systemic risk on the use of breast-sparing treatment in women enrolled in adjuvant breast cancer studies: An analysis of two intergroup trials. J Clin Oncol 14: 3009-3017, 1996[Abstract]
5.
Grilli R, Scorpiglione N, Nicolucci A, et al: Variation in use of breast surgery and characteristics of hospitals’ surgical staff. Int J Qual Health Care 6: 233-238, 1994 6. McKinlay JB, Burns RB, Feldman HA, et al: Physician variability and uncertainty in the management of breast cancer: Results from a factorial experiment. Med Care 36: 385-396, 1998[Medline] 7. Nattinger AB, Gottlieb MS, Hoffman RG, et al: Minimal increase in use of breast-conserving surgery from 1986 to 1990. Med Care 34: 479-489, 1996[Medline] 8. Committee S: Mastectomy or lumpectomy? the choice of operation for clinical stages I and II breast cancer. Can Med Assoc J 159: 15-21, 1998 9. Young WW, Marks SM, Kohler SA, et al: Dissemination of clinical results: Mastectomy versus lumpectomy and radiation therapy. Med Care 34: 1003-1017, 1996[Medline] 10. Ford L, Kaluzny AD, Sondik E: Diffusion and adoption of state-of-the-art therapy. Semin Oncol 17: 485-494, 1990[Medline] 11. Howe HL, Lehnherr M, Katterhagen JG: Education of rural physicians about breast cancer through an oncology outreach program. Public Health Rep 109: 804-808, 1994[Medline] 12. Warnecke RB, Johnson TP, Kaluzny AD, et al: The community clinical oncology program: Its effect on clinical practice [comment]. Jt Comm J Qual Improv 21: 336-339, 1995[Medline] 13. Rogers EM: Diffusion of Innovations (ed 4). New York, NY, Free Press, 1995 14. Hagen N, Young J, MacDonald N: Diffusion of standards of care for cancer pain. Can Med Assoc J 152: 1205-1209, 1995[Abstract] 15. McKinney MM, Barnsley JM, Kaluzny AD: Organizing for cancer control: The diffusion of dynamic innovation in a community cancer network. Int J Technol Assess Health Care 8: 268-288, 1992[Medline] Submitted January 18, 2000; accepted October 16, 2000.
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2001 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
.05) with adoption of paclitaxel. Significant influences on adoption included use of symposia as a therapy information source, physician experience with paclitaxel to treat late-stage breast cancer, and perceived advantage in efficacy of paclitaxel. 
