Post-Study Aspirin Intake and Factors Motivating Participation in a Colorectal Cancer Chemoprevention Trial¹

Factors influencing participation in cancer chemoprevention trials are poorly understood. Motivational factors that influence an individual’s decision to participate in a prevention trial, and whether or not study participation impacts post-study health-related behaviors have not been adequately studied. These issues are critical to the recruitment and retention of prevention study subjects. Furthermore, it is important to determine whether study participation can impact future health-related behaviors, such as continued intake of the study drug, if available.

Some published reports have examined the perceptions of individuals participating in chemoprevention studies (1, 2, 3, 4, 5, 6). In a basal cell carcinoma prevention trial using isotretinoin, subjects reported that perceived benefits of participation included careful medical follow-up and being part of a research effort (1). Participants were also queried as to whether they would be willing to take the study drug post-trial if it were shown to be effective, and most (78%) answered “definitely” or “probably.” In another study, Hudman et al. (2) explored perceived benefits and barriers to participation in a colon cancer chemoprevention trial using calcium. The most important potential study benefit identified was the perception of potential colon cancer prevention. Over 90% of the participants indicated they would continue taking calcium if the drug was shown to be effective. In each of these studies, the question of continuation of the study drug was placed in the context of “if the drug were found to be effective.” Neither of these studies, however, explored whether or not subjects would choose to continue taking the study drug regardless of benefit awareness.

In a breast cancer chemoprevention study evaluating tamoxifen, a questionnaire was given to women attending information sessions to assess predictability of enrollment (3). Several factors were reported to be barriers to study enrollment including concern about side effects of the study drug, out-of-pocket expenses, the possibility of receiving a placebo, and not being allowed to take estrogen replacement therapy. Enrollment was correctly predicted for 72% of respondents to the questionnaire, with not being able to take estrogen replacement therapy being the strongest predictor.

We recently completed a randomized, double-blinded, placebo-controlled trial to determine the optimal dose of aspirin (81 mg, 325 mg, or 650 mg taken once daily for 4 weeks) for chemoprevention using mucosal prostaglandin E₂ levels as a biomarker in subjects with prior colorectal adenomas (elsewhere in this issue). A total of 60 subjects were enrolled in the study, of which 55 were evaluable for biomarker analysis. This study provided the opportunity to examine behavioral factors related to motivation for study participation and the impact of participation on post-study aspirin intake.

Specifically, we sought to identify whether or not these same participants chose to take aspirin post-study and if so, why this decision was made before disclosure of study results. Additionally, we examined whether pre-study motivating factors or health-related behaviors were correlated with post-study aspirin use.

The study protocol was reviewed and approved by the Institutional Review Board of the University of Texas M. D. Anderson Cancer Center. Eligible subjects for this exploratory cross-sectional study consisted of 55 individuals with prior colorectal adenomatous polyps who were participants in a study to determine the optimal dose of aspirin for chemoprevention. The subject population was recruited from a health management organization (HMO) in Houston, Texas.

This randomized and double-blinded trial evaluated the effects of three doses of aspirin (81mg, 325mg, 650mg) versus placebo given daily for 4 weeks upon rectal mucosal prostaglandin E₂ levels. Patients were regarded as evaluable if they completed both baseline and end-of-study (4-week) endoscopic exams with rectal mucosal biopsies at both time points. A behavioral component was included in this study with the objective of determining motivators for study participation and the impact of participation on post-study aspirin intake. To this end, questionnaires were mailed to all of the 55 subjects in a single mailing after completion of the parent trial. Self- addressed stamped envelopes were included with each questionnaire. When possible, individuals not returning forms within 4 weeks of mail-out were contacted by telephone and asked to complete and return the questionnaire. None of the participants were aware of their treatment arm nor any study results at the time of questionnaire completion.

The 16-item questionnaire was created using Pender’s Health Promoting Model (HPM) as its theoretical framework (7, 8, 9) and incorporated information derived from previous studies of participation in chemoprevention trials (1, 2, 3, 4, 5, 6), thereby establishing the instrument’s construct validity (10). Pender’s Health Promoting Model is used to predict the likelihood of engaging in health-promoting behavior based on two phases: decision-making and action. Elements from both the cognitive-perceptual and modifying factors were incorporated into the questionnaire design to facilitate investigation of the research objectives. Questions were presented in a Likert or multiple-choice format. Specific questions included those associated with motivating factors for participation, and instructions were as follows: please circle the number that best describes how important each item was to you when you enrolled in the aspirin study; number “1” being not important at all and number “10” being very important.“ Potential motivational factors for study participation included in the questionnaire are shown in Table 2. Other queried items included: history of health-related behaviors, personal perception of health status, personal control over health, and post-study use of aspirin.

The survey was reviewed and pilot-tested by a panel of experts and several patients participating in other chemoprevention trials to ensure content validity (11). Demographic information had been previously collected as part of the parent trial and was available on all participants. These data included gender, age, race, education, income, type of employment, medical history, personal and family history of cancer, smoking and drinking history, and diet history. Opportunity was consistently provided for participants to add comments or additional variables not otherwise identified with each question. Questionnaire responses were entered into a database according to a subject identifier code and systematically reviewed for accuracy (data were entered into two databases, the data listings printed and compared, and any discrepancies corrected). Demographic information available from the parent study was then merged with the questionnaire data.

Analyses were computed using the Statistical Program for Social Sciences (SPSS, 1999). Descriptive, nonparametric statistics were used to describe the demographic data and responses to the questionnaire. Motivators for participation in the trial and for the use of aspirin post-trial were calculated using measures of central tendency, percentages, frequencies, and variance. Use of aspirin post-trial was described in terms of frequency and was compared with nonusers (the pre-study amount of aspirin used was zero for all of the trial participants). Queries were posed in a bipolar Likert-type format and scored using a scale of 1–10 with the positively worded adjective associated with higher scores. A semantic differential technique was also used and scores for each question were then summed to yield a total score. The highest possible score for a particular query was dependent on the number of individuals answering the question. For example, if all 43 subjects responded to a query, the MDS³ would be 430. Analysis of scores using measures of central tendency or semantic differentials yielded similar results. Those choosing to take aspirin post-study were compared with those who did not, on the basis of demographics, pre-study motivators, and health-related behaviors using the Mann-Whitney U test. The level of significance was set at 0.05.

A total of 43 (78%) of 55 poststudy questionnaires were returned constituting a response rate of 78%. The demographic characteristics of the 43 individuals who completed and returned the questionnaire are shown in Table 1. Data are presented by group, i.e., those who chose to take aspirin post-study and those who did not, and overall. The sample included 23 males and 20 females with a mean age of 58.8 years (range, 40–76 years). Thirty-four (79.1%) were married and ∼one-half (55.8%) held at least a bachelor’s degree. Employment status included full-time workers (51.2%), retirees (27.9%), and homemakers (16.3%). Most classified their occupations as professional (74.4%), and many reported a family income in excess of $50,000.00 annually (58.2%).

Participants were queried on perceived health status, use of cancer-screening procedures, and current health-related behaviors. Using the scale of 1–10, with 10 being “very healthy,” participants rated themselves as healthy (mean, 8.43 ± 1.73; MDS, 150), and their perceived control over personal health status was high (mean, 9.14 ± 0.98; MDS, 220).

Thirty-nine (91%) of 43 respondents reported compliance with recommended surveillance colonoscopic examinations. Ninety percent of women respondents underwent annual mammography and 91% of men received annual or biannual prostate examinations. Fifty-six percent of respondents reported eating low-fat foods and 49% eating high-fiber foods daily or every other day. Twenty-three (41.8%) reported modifying their diets in the past 10 years in an attempt to lower cholesterol or to eat healthier. Fifty-three percent exercised daily or every other day. Vitamin/mineral and herbal supplements were taken daily or every other day by 56 and 30% of respondents, respectively. Smoking history revealed that 14.5% were current smokers, 38.2% were former smokers, and 44% had never smoked.

Of note, 86% of participants reported never having participated in clinical research before the aspirin trial. We determined potential motivators that contributed to the decision to participate in the aspirin chemoprevention trial and identified several factors that were strongly associated with this decision. Using a scale of 1–10, with 10 being most important, participants applied scores to bipolar statements pertaining to potential factors influencing study participation. Of the five factors receiving the highest scores (mean values, >8), two were altruistic, i.e., helping doctors learn to prevent colon cancer and helping future generations at risk; two were personal and included the desire to lower personal risk of developing colon cancer and becoming more informed about cancer prevention; and another being the fact that the trial was conducted at a major cancer center (Table 2). The most important factor influencing study participation, as indicated by the highest mean score and MDS, was reported for the response “helping doctors learn to prevent colon cancer.” The lowest mean score (3.4) received for all of the items tested was financial compensation for participation. When each of these queries was examined using the semantic differential, analysis of the results were concordant with mean values (Table 2).

Participants were also queried as to whether their decision to participate was influenced by study duration, personal cost, or evaluation of an investigational, prescription or over-the-counter drug. To defray cost associated with travel to the study site, we supplied parking vouchers and subjects were paid $200.00 on study completion. Responses were again determined using a scale of 1–10 with 10 indicating a “very different decision,” and MDSs were calculated. Respondents indicated that a longer trial duration and a higher personal cost would negatively impact the likelihood of participation, as would evaluation of an investigational or prescription drug, versus an over-the-counter medication (data not shown). Most (74.4%) of the respondents expressed interest in taking part in similar trials in the future.

We determined the frequency of post-study aspirin use as a measure of the impact of study participation on a presumed health-promoting behavior. Participants were queried about regular aspirin use post-study in the absence of efficacy information. Nineteen (44%) of 43 respondents reported taking daily aspirin after study completion (Table 3). The 325-mg daily dose was most commonly chosen by former participants. Most (74%) respondents who chose to take aspirin began taking the drug immediately after trial participation. At a mean follow-up of 17.3 months (range, 3 to >24 months), 18 of 19 former participants continued to take aspirin regularly. Potential motivating factors for post-study aspirin use were evaluated (Table 4). Using a scale of 1–10 with 10 being most influential, highest scores were given to “desire to lower personal risk of developing colon cancer” (mean, 9.83 ± 3.93; MDS, 170) and “participation in the aspirin trial” (mean, 6.50 ± 4.27; MDS, 110). Of note, 42% of these respondents reported that their health care providers were unaware of their decisions to take aspirin post-study. Of the 24 respondents who elected not to take aspirin post-study, explanations most often cited were uncertainty of benefit (mean, 7.0 ± 3.99; MDS, 100) and lack of knowledge of the appropriate dose for chemoprevention (mean, 7.21 ± 3.74; MDS, 100; maximum MDS, 240).

We also explored pre-study factors influencing the decision to take aspirin post-study. A nonparametric test was used to assess differences in responses to queries between those who took aspirin post-study (n = 19) and those who did not (n = 24). None of the demographic variables were significantly correlated with post-study aspirin use. Factors associated with the decision to participate in the trial were then examined in relation to post-study aspirin use (Table 5). Of these variables, only a desire to help future generations at risk for cancer was associated with post-study aspirin use (P = 0.014), and a trend was observed for those desiring to help doctors learn to prevent colon cancer (P = 0.072). Among health-related behaviors, only regular use of pre-study vitamin or mineral supplements was significantly (P = 0.048) associated with the decision to take aspirin post-study. Compliance and utilization of cancer-screening procedures were similar between those who took aspirin post-study and those who did not. Assessment of personal health and perceived control over personal health did not differ between these groups.

We conducted an exploratory cross-sectional behavioral study as a component of a clinical trial that was designed to determine the optimal dose of aspirin for chemoprevention against colorectal cancer. We sought to determine motivating factors for study participation and impact of participation on post-study health-related behaviors. We queried participants using a posttrial questionnaire and found that altruism was rated as the most important motivator, and concerns were both immediate, i.e., helping doctors to learn to prevent colon cancer, and delayed, i.e., helping future generations at risk for the disease. After altruism, personal factors received the next highest ranking and included the desire to lower personal risk of colorectal cancer and to become more informed about cancer prevention. These findings expand on published data in the chemoprevention literature, in which subjects have been queried concerning motivators for participation. A desire to support a research effort and the possibility of personal disease prevention have been reported to be important factors influencing subject enrollment in clinical trials (1). The perceived benefit(s) from participation is regarded as a key determinant of recruitment success (4).

To date, no published reports have specifically addressed the impact of prevention trial participation on subsequent health-related behaviors. More specifically, such studies have not explored whether study subjects choose to take the study drug, if available, after trial completion. To address this critical issue, we determined the frequency of post-study aspirin use as a measure of the impact of study participation on a presumed health-promoting behavior. We found that 44% of the respondents reported taking aspirin on a regular basis post-study, and all but one participant continued doing so at a mean follow-up of 17.3 months. Participants did so in the absence of efficacy data from the prevention trial. The explanation for continuing aspirin intake that was rated highest by respondents was to lower their personal risk of colon cancer. To our knowledge, this is the first time that participation in a prevention trial has been reported to lead to a behavior change.

Interestingly, subjects who chose to take aspirin post-study were statistically more likely to consume vitamins or mineral supplements on a regular basis, but other health-related behaviors such as physical activity and diet were not associated with post-study aspirin use. Our important finding of continued use of a study drug after study completion requires replication in other chemoprevention trials using a larger sample size. The study findings apply to patients who have a history of colorectal adenomas and are at increased risk of colorectal cancer. A limitation of this study is its small sample size and limited generalizability. However, this population is similar to participants in other chemoprevention studies as evidenced by demographic variables, high rates of compliance with colonoscopic surveillance/cancer-screening recommendations, and motivation toward preserving and improving their personal health. Furthermore, we are only interested in generalizing the study findings to patients at increased risk of cancer who are participants in chemoprevention studies. The 78% rate of questionnaire return is this study was higher than the expected rate for mail surveys (11, 12), wherein ≥65% return rates are considered acceptable. Post-study aspirin use would still be at 34.5%, even if all of the participants not returning questionnaires chose not to take aspirin post-study. We emphasize that this study is exploratory in nature, but that our study findings are provocative and warrant further examination in prospective studies.

Studies are needed to determine whether prevention trial participation may have long-term benefits and/or risks by impacting health-related behaviors. As in this trial, the decision to take aspirin post-study in the absence of efficacy data are not without risk. Patients taking aspirin post-study did so generally without a recommendation from a physician or without the awareness or supervision from their primary health care providers. The potential for adverse effects and the risk benefit ratio must be carefully weighed, in consultation with medical personnel, when a drug is continued outside the confines of a clinical trial for a nonapproved indication. This issue must be carefully considered by prevention researchers and as an important and potential byproduct of trial participation. Patient education and post-study follow-up are, therefore, needed to guide safe and informed decision-making for trial participants interested in continuing to take a study drug, if available to them, after prevention trial enrollment.

We thank Dr. Judy Headley from the University of Texas-Houston Health Science Center for support and assistance with this study.