- © 2003 by American Society of Clinical Oncology
Public Attitudes Toward Participation in Cancer Clinical Trials
- From the Coalition of National Cancer Cooperative Groups, Philadelphia, PA; Center for Biomedical Communication, Northwestern University Medical School, Chicago, IL; Cancer Research Foundation of America, Alexandria, VA; University of Colorado Health Sciences Center, Denver, CO; National Coalition for Cancer Survivorship, Silver Spring, MD.
- Address reprint requests to Robert L. Comis, Coalition of National Cancer Cooperative Groups, Suite 1100, 1818 Market Street, Philadelphia, PA 19103; email: rcomis{at}ecogchair.org.
Abstract
Purpose: The objective of this study is to understand the attitudes of American adults toward participation in cancer clinical trials.
Methods: A national probability sample of 1,000 adults aged 18 and older living in noninstitutional settings was interviewed by telephone by Harris Interactive during March and April 2000. One participant was selected from each household selected for the study. The resulting data were weighted to reflect the full adult population of the United States as reported in Current Population Reports. An Index of Participation in a Cancer Clinical Trial was computed, using a confirmatory factor analysis and converting the factor scores into a 0-to-100 scale.
Results: Approximately 32% of American adults (64 million individuals) indicate that they would be very willing to participate in a cancer clinical trial if asked to do so. An additional 38% of adults (76 million individuals) scored in a range that indicates that they are inclined to participate in a cancer clinical trial if asked, but hold some questions or reservations about participation. Projected rates of diagnosis, eligibility, and recruitment indicate that substantially more patients are willing to participate than are actually accrued.
Conclusion: These results indicate that the primary problem with accrual is not the attitudes of patients, but rather that the loss of potential participants is the result of the unavailability of an appropriate clinical trial and the disqualification of large numbers of patients. The pool of willing patients is further reduced by the reluctance of some physicians to engage in accrual.
CLINICAL TRIALS are essential for the identification of effective therapies in modern medicine. The combination of earlier detection of virtually all forms of cancer and the application of new treatments developed through clinical trials has been responsible for the improvement in the 5-year survival rate to one-in-two from its one-in-three rate 4 decades ago.1 As the biomedical understanding of cancer increases, a growing number of new opportunities—preventive, curative, and palliative—will require increasing numbers of adults to agree to participate in a clinical trial. It is at present estimated that more than 400 new compounds, based on these advances in the understanding of biology, are now available for testing.2 The ability to recruit adult participants for future clinical trials will depend, in part, on understanding the attitudes of adults toward participation in clinical trials. This analysis uses the first national survey of American adults designed to examine the willingness of adults to participate in a cancer clinical trial.
A number of local empirical studies3–,8 have examined why patients who have been diagnosed with cancer, or are in a high-risk group, agree or refuse to enroll in a cancer clinical trial study. Reviews9–,13 of this literature concluded that a combination of altruism and a hope for better treatment motivated most of the patients who enrolled in a cancer clinical trial. The major barriers to participation in cancer clinical trials included low levels of understanding about the process of random assignment and about the meaning of the standard treatment. Some patients were concerned about the possibility of additional tests, additional discomfort, increased costs, insurance problems, and additional travel. Studies14–,18 of physician-patient communication found that patients who felt that their physicians had communicated clearly about a prospective study were more likely to enroll.
Only two studies19,20 focused on the attitude of adults who have not been diagnosed with cancer, and each looked at a single metropolitan area. Both studies found a strong belief in the benefit of clinical trials for the advancement of medical knowledge, support for the conduct of clinical trials, a substantial willingness to participate in a study, and a belief that study participants would receive the best possible care. The most recent study20 found that slightly more than half of the respondents said that they would participate in a clinical trial study if they were diagnosed with cancer. Neither of these studies can be generalized to the U.S. population because they reflect only local samples.
METHODS
The 2000 Cancer Clinical Trials Study (CCTS) is the first examination of the attitudes of a national sample of American adults toward participation in cancer clinical studies. On behalf of the 2000 CCTS, Harris Interactive21 conducted 1,000 telephone interviews with a national sample of adults aged 18 and older in February, March, April, and May of 2000. The Harris sampling procedure used a probability sample of working residential telephone numbers and attempted to secure an interview with one adult in the household who was present at the time of the call. Within each household that answered the telephone, one individual was selected by a predetermined routine that asked for the youngest male adult over age 18 who was present and then selected sequentially among the adults available to participate in the interview. Up to five attempts were made to reach each sample household who did not answer the telephone or who used an answering machine.
A total of 699 interviews were completed using the procedures outlined above. A total of 2,044 households were called to obtain the completed interviews, including 327 where the selected individual was never available for an appointment. An additional 301 minority (African-American and Hispanic-American) respondents were interviewed to provide greater racial balance to the sample. A total of 483 minority households were identified through almost 5,000 screening calls, and 301 interviews were obtained. The combined cooperation rate was 33%. The 1,000 completed interviews were weighted by age, sex, education, income, race, and ethnicity to reflect U.S. Census estimates of American adults aged 18 and older.21
The 2000 CCTS included a set of questions for respondents who had previously been diagnosed with cancer and for respondents who had previously participated in a clinical trial. A series of branches partitioned the sample into several segments. In the 2000 CCTS, 63 respondents (6.3%) reported that they had been diagnosed with cancer at some time in their life. Fifty of these respondents (80%) were unaware that a clinical trial might have been a possibility, and 13 (20%) were aware of the availability of a clinical trial. Only three (4.7%) individuals reported that they had participated in a clinical trial related to the treatment of their cancer. An additional 50 respondents (5.0%) reported that they had participated in a clinical trial unrelated to cancer, and were omitted from some of the attitude questions about cancer trials. Given these branching patterns, this analysis will focus on the 887 respondents who did not have a cancer diagnosis and who did not previously participate in any clinical trial. This 88.7% segment of the U.S. population aged 18 and over—representing 177 million American adults—is the study population in the following analysis and represents the target audience for persuasive messages about the need for adult participation in cancer clinical trials. Although none of the adults that comprise this sample have had a diagnosis of cancer and have confronted the actual decision to participate in a clinical trial, this sample represents the population of adults who are the recipients of messages about the critical role of clinical trials in the process of medical research. As discussed in later sections of this article, a positive attitude toward participation in a clinical trial does not correspond on a one-to-one basis with actual participation, but it is the best indicator that we have before actual diagnosis and choice.
RESULTS
Individuals who participated in the survey were asked about their perceived health status. Approximately 85% of participants perceived themselves to be in good to excellent health. Nearly three-quarters of participants had either an immediate family member or close friend who had been diagnosed with cancer. CCTS respondents were nearly universally in favor of medical research and the development of new treatments, and 96% of respondents agreed that new treatments should be evaluated in a relatively small number of patients before becoming available to all patients.
Approximately 60% of respondents felt that they understood what a clinical trial is “somewhat” or “very well,” but fully 40% reported that they did not understand the idea of a clinical trial. When queried about their perceptions of cancer clinical trials, there were both positive and negative impressions. For instance, 92% of respondents felt that clinical trials would benefit themselves as well as help others in the future; 61% understood that there is “very limited enrollment” onto clinical trials; 84% thought that a clinical trial might provide “better initial treatment”; and 75% felt that clinical trials were associated with “high-quality medical care.” However, 56% of those surveyed indicated it was “mostly true” that clinical trial participants were as likely to receive a placebo or sugar pill as the test treatment, and 45% indicated that patients in clinical trials are “treated like guinea pigs.”
A Measure of Attitude Toward Participation
A major purpose of the study was to assess the attitude of Americans toward participation in a cancer clinical trial. To evaluate the inclination of American adults to participate in a cancer clinical trial, four hypothetical questions were asked. The responses are shown in parentheses: If you were diagnosed with a serious disease such as cancer, how important would it be to you to participate in a clinical trial for a new treatment—very important (43%), somewhat important (38%), not very important (9%), or not at all important (6%)? If you were diagnosed with cancer, how willing would you be to participate in a clinical trial for the initial treatment you received—very willing (31%), somewhat willing (51%), not very willing (9%), or not willing at all (6%)? If you were diagnosed with cancer, how willing would you be to participate in a clinical trial if your initial treatment failed—very willing (58%), somewhat willing (29%), not very willing (6%), or not willing at all (6%)? If you did not have cancer and were asked to participate in a clinical trial to prevent cancer, how willing would you be to participate—very willing (16%), somewhat willing (49%), not very willing (18%), or not willing at all (15%)?
These results indicate that a significant portion of American adults are willing to participate in cancer clinical trials, but the precise proportion may vary according to specific circumstances. Further, some responses indicate a high degree of willingness to participate in a clinical trial, whereas other responses reflect a positive predisposition toward participation, but less than full enthusiasm.
How can these different responses be parsimoniously summarized into a single indicator of attitude toward participating in a cancer clinical trial? The most efficient and effective means available is the use of confirmatory factor analysis to verify that these four responses reflect a common underlying attitudinal dimension and to provide a metric for integrating the responses into a summary scale score.22,23 Using LISREL (Scientific Software Inc, Chicago, IL),24 a confirmatory factor analysis found that the four questions all load on a common factor at satisfactory levels (Table 1⇓). The factor loadings reflect the correlation or degree of association of each item with the underlying dimension—attitude toward participation in a cancer clinical trial—and provide a useful metric for weighting the contribution of each question to this factor, or dimension.
Confirmatory Factor Analysis* of Attitude toward Participating in a Cancer Clinical Trial
Using these loadings, a factor score was computed for each respondent in the study population and then converted into a 0-to-100 scale. An individual who responded that he or she was very willing to participate in a clinical trial in all four questions would have a score of 100, whereas an individual who indicated that he or she would be not at all willing to participate in a clinical trial in all four questions would have a score of zero. A factor score provides an efficient metric to scale the location of each individual on this scale.
By dividing this 0-to-100 scale into five equal intervals, it is possible to look at the distribution of scores in clusters that may be analytically useful. Approximately 32% of adults scored between 81 and 100 on the scale, indicating that these individuals would be very likely to participate in a clinical trial if they were diagnosed with cancer (Table 2⇓). About 38% of adults scored between 61 and 80 on the scale, reflecting individuals who would be likely to participate in a cancer clinical trial. Twenty-one percent of adults scored between 41 and 60 on the scale, indicating that these individuals are neutral toward participating in a cancer clinical trial. Five percent of adults had scores between 21 and 40, indicating that they are unlikely to participate in a cancer clinical trial, and 4% had scores between 0 and 20 on the scale, meaning that they are very unlikely to participate in a cancer clinical trial. Clearly, there is a sizable proportion of adults who are positively disposed toward participating in a cancer clinical trial if they were asked to participate.
Distribution of Adults on the Index of Participation in a Cancer Clinical Trial, 1998
The Distribution of Attitude Toward Participation
Younger adults were significantly more likely to hold positive views about participating in a cancer clinical trial than older adults, but a majority of adults aged 65 and older reported a positive attitude toward participating in a cancer clinical trial (Table 2⇑). Adults who reported that they had a high level of understanding of the concept of a clinical trial were more positively inclined toward participation than adults with lower levels of understanding of the meaning of a clinical trial. Attitude toward participation in a cancer clinical trial did not vary significantly by either sex or level of formal education. A wide range of previous studies of adult attitudes toward science and technology indicate that adults with more years of formal education often hold more positive attitudes toward science and scientific research,25–,27 and it might be expected that these general attitudes would extend to participation in a cancer clinical trial. The absence of a bivariate relationship indicates that there is a broad societal agreement on the scientific basis of medicine and that this agreement is held by adults regardless of their level of formal education.
Given this distribution of attitude toward participation in a cancer clinical trial, it is possible to estimate the number of American adults that would be positively disposed to participate in a clinical trial if they were given a diagnosis of cancer and if they were asked to enter a clinical trial. For this purpose, the six age categories used above were differentiated by men and women, producing a total of 12 strata. By applying the expected cancer incidence rate in each age classification, the number of new cancer cases expected this year or next year was estimated (Table 3⇓). The sum of these separate estimates is 1,281,203 cases—the same total number predicted by the National Cancer Institute (NCI).
Estimated Number of Likely Participants Available for Cancer Clinical Trials
The estimated number of new cancer cases in each age category was multiplied by the proportion of adults in that age classification with positive and very positive attitudes toward participation in a cancer clinical trial, producing estimates of the number of adults with new cancer diagnoses who would be very likely to become participants in a clinical trial. The results indicate that approximately 310,000 adults with new cancer diagnoses would hold very positive attitudes toward participation in a clinical trial, and that an additional 490,000 adults with new cancer diagnoses would hold generally positive attitudes toward participation, but may require more explanation and persuasion than the group with more positive attitudes.
It is important to recognize, however, that many of these potential clinical trial participants may not qualify for a protocol because of other medical conditions or because of the type or stage of cancer that they have. A 1999 study28 of 15 cancer centers in the southeastern United States found that there were available clinical trials for 936 of these patients (40%), but that there were no clinical trials available for 1,403 patients (60%). It is not known how many of these 1,403 patients might have been clinically eligible because they were not screened for any cancer clinical trial. Of the 936 patients for which one or more clinical trials were available, 573 were clinically eligible for enrollment (61%), and 166 of these patients (29%) were enrolled in a clinical trial. Approximately 40% of the 573 patients who were clinically eligible for participation refused. Seventeen percent of clinically eligible patients were excluded because their physician preferred a specific therapy, and 7% were dropped because their insurer refused to cover protocol care. Of the original population of southeast cancer patients, only 7% were eventually enrolled in a cancer clinical trial. A similar, but smaller, study29 found a similar pattern of accessibility and eligibility exclusions.
If the southeast result28 is applied to the 1.2 million new cancer cases estimated by the NCI, approximately 512,000 new cancer cases would be located in areas with an available clinical trial for their cancer, and 199,000 of these new cases would be clinically eligible. Further, it would be estimated that 58,000 of these new cases would enroll in a cancer clinical trial.
Alternately, by applying the attitude findings described Table 3⇑ to the NCI data, an estimated 74,000 new cancer patients would be clinically eligible for a clinical trial and hold a very positive attitude toward participating, and an additional 118,000 new cancer patients would be positively disposed toward participating, but would require more persuasion than the most positive group (Table 4⇓). The substantial difference between these estimates and the estimated 58,000 participants derived from a straight extrapolation of the southeastern U.S. results may be viewed as a margin of opportunity for improved communications and recruitment procedures related to cancer clinical trial accrual.
Modified Estimate of Likely Participants Available for Cancer Clinical Trials
DISCUSSION
This analysis found that a substantial proportion of American adults hold positive general attitudes toward participation in cancer clinical trials, but a limited understanding of the exact nature of a clinical trial study.9–,13 Despite the large proportion of Americans who hold positive attitudes toward clinical trial participation, the percentage of cancer patients who eventually enroll in a clinical trial is less than 10%. An extrapolation of existing incidence data from the NCI and accrual data from a southeastern U.S. study28 indicates that there are three critical points in the flow of patients from original diagnosis to enrollment onto a cancer clinical trial, and each of these decision points are critical to the final result.
First, the availability of cancer clinical trials depends on the number of trials,30 the number of physicians who join each study,29,31,32 and the frequency of presentation of the clinical trial option to patients. The number of cancer clinical trials is likely to increase because of the number of new compounds to be tested.2 The largest gains in availability of cancer clinical trials, though, could be obtained by increasing the number of physicians who join and actively participate in clinical trials relevant to their practice through the frequency with which they present the clinical trial option to patients. Physician participation in cancer clinical trials is discouraged by the increased time and effort required to find a trial, establish eligibility, explain the trial to the patient, obtain informed consent, and perform the additional work associated with the study. This additional time and effort must be appropriately reimbursed to establish the infrastructure to perform the clinical trial in a rigorous fashion.
Second, a substantial proportion of potential participants are disqualified because of existing comorbidities or other problems—such as insurance coverage or transportation.28,29,33 This is a difficult problem because a relaxation of eligibility standards to increase enrollment could complicate the findings of the trial and reduce the usefulness of the results. On the other hand, there have been arguments to reduce the eligibility requirements to a bare minimum to enhance accrual and improve the generalizability of the results.33 Most of the major cooperative oncology groups are studying ways to reduce the loss of potential participants to these factors,34 but it is particularly important to focus renewed attention on the criteria for older patients. Most of the newly diagnosed cancers occur in patients aged 65 and older, and it is important to test proposed therapies on this population to the extent possible. Presently, elderly patients are underrepresented in clinical trials.34 Clinical trials designed to address issues important to the elderly population are essential.
Finally, approximately 40% of new cancer patients who have access to a clinical trial and who qualify for enrollment decline to participate.28,29 This result is consistent with the results reported above in which 30% of the respondents in the 2000 CCTS scored 60 or lower on the Index of Attitude toward Clinical Trial Participation (Table 2⇑). A recent analysis of protocol participation at the University of Birmingham cited a virtually identical proportion. (Saleh M, personal communication, March 2000). Some portion of this differential can be attributed to the difference between an individual’s general willingness to participate in a clinical trail when asked in an interview and the need to make a choice in the context of a recent diagnosis of cancer. This linkage has been studied extensively10–12,27,35 and various interventions have been found to make modest, but statistically significant, improvements in the conversion of positive intentions into actual accruals.15
The literature points to several reasons for this decision, but it is clear that many patients hold mistaken views of the nature of clinical trials, and that many significantly overestimate the efficacy of standard therapies in making their decision. Both Cassileth19 and Trauth20 noted that there is a discrepancy between how individuals respond to general statements concerning the value and benefit of clinical trials and how they respond in making personal choices. For example, over half of the respondents in the Cassileth study19 chose “getting the best medical care” as the reason for considering clinical trials participation, yet they indicated that the treatment recommended by their physician would be superior to that provided in a clinical trial. This conflicted view of clinical trials is undoubtedly exacerbated by the stress of a new cancer diagnosis.
It is important to continue to study the decision-making process by patients and the role played by physicians in presenting and explaining the clinical trial option. The 2000 CCTS found that 59% of respondents relied on their doctor as their primary source of information about health care. Several studies18–,22 found that positive communications about participation in a clinical trial is a major factor in most patients’ decision to enroll in a cancer clinical trial. In the end, the final decision to participate in a clinical trial rests on the recommendation of the physician and how it is perceived and acted on by the patient.
Acknowledgments
We are grateful for the advice and counsel of Humphrey Taylor and Peter Risher at Harris Interactive.
Footnotes
-
This survey was supported by the Coalition of National Cancer Cooperative Groups, the Cancer Research Foundation of America, the Cancer Leadership Council, and the Oncology Nursing Society. Bristol-Meyers Squibb provided an educational grant for the study.
- Received February 19, 2002.
- Accepted September 16, 2002.
