OR WAIT 15 SECS
Analysis suggests age, condition, and treatment satisfaction have most significant effect on participation.
Recruitment of subjects for clinical trials remains an ongoing and increasing challenge, with nearly 90% of industry-sponsored clinical trials experiencing delayed enrollment.1,2 Since these delays often last six months or more, patient recruitment is one of the most important determinants of the clinical program timeline.3 In addition to being an important timeline driver, patient recruitment also represents a significant portion of program costs. A large amount of resources are spent starting-up and monitoring sites, approximately 50% of which will recruit one or no patients in their studies.4
In order to expedite study timelines and decrease costs, researchers are actively seeking opportunities to speed enrollment, and supplemental patient recruitment is one such option. For the purposes of this article, supplemental recruitment is defined as the practice of pre-qualifying community-based patients and referring them to clinical trial sites for study screening.
The portion of patients available for supplemental recruitment is large, with studies suggesting that fewer than 5% of all eligible adult patients are enrolled in therapeutic studies and this proportion is even lower among the elderly, women, and racial and ethnic minorities.2 To better understand the reasons why these patients do not participate in clinical trials, ECRI conducted a meta-analysis of 14 published studies and concluded that three factors were important: fear of placebo or randomization; too far to travel; and the desire to have the physician choose the treatment rather than accept a process that uses randomization to assign patients to a treatment group.5
Primary research conducted by MediGuard.org and published in the 2009 Industry Standard Research report titled "Success in Patient and Investigator Recruitment," however, showed different results. The MediGuard survey of clinical trial naïve patients found that the primary reason that patients do not participate in clinical trials is simply that they are unaware of opportunities to do so (69% of respondents).6 As a result, physician investigator sites and other researchers are actively seeking opportunities to make community-based patients aware of clinical trials. With ever-increasing numbers of patients seeking health information on the Internet, digital outreach programs through search engines, social media, and online patient communities are becoming more commonplace. Although the number of clinical trials using supplemental patient recruitment has grown substantially, there is little information available in the literature on patient level of interest in studies and patient characteristics associated with study interest. This article seeks to contribute to researchers' understanding of patient interest in clinical trials by analyzing the results of patient feasibility and recruitment studies conducted across a broad range of medical conditions.
Design overview. All feasibility and recruitment screening studies conducted with patients between January 2009 and July 2011 in the United States or United Kingdom by MediGuard.org were searched. Specifically, patient questionnaires that contained a question regarding patient consent to contact if a clinical trial were available at a nearby site were identified. The response to this question on consent to referral (yes, no) was extracted into a patient-level, de-identified study database, along with any other patient characteristics available from the screening survey including: treatment history (disease, number of medications, side effects, and treatment satisfaction), patient demographics (age, gender, race, country, and previous clinical trial experience), and other socioeconomic variables (employment and insurance status).
Setting and participants. All surveys were conducted online with members of MediGuard.org, a free online service that monitors the safety of prescription medicines, over-the-counter medicines, and healthcare supplements for over 2.5 million patients in the United States, United Kingdom, France, Germany, Spain, and Australia. These online surveys were sponsored by the owner of MediGuard.org Quintiles Inc., a pharmaceutical services company offering clinical trial support solutions.
Patients enrolling in the service consent to be contacted about research opportunities as part of the registration process. When a specific research opportunity arises, an e-mail is sent to potentially eligible members who can click on a link in the e-mail to read about study details, screen for eligibility, and provide further consent to have their contact details sent to an investigator site.
Statistical analysis. Univariate logistic regression was used to assess the association of patient interest in referral with each study variable. All analyses were conducted using SAS version 9.2 for Windows (SAS Institute Inc, Cary, NC). Subsequent to the univariate analysis, a multivariate, step-wise logistic regression was conducted to identify the characteristics that were predictors of a patient's interest in referral to a nearby clinical trial site.
Sample characteristics. Information was extracted from 206 feasibility and recruitment screening studies involving 8,599 patient responses. Characteristics for the sample are displayed in Table 1. Of the 8,599 patients responding, 68% were aged under 65, and 68% were female. Only 14% of respondents did not have health insurance, although a majority (55%) was unemployed (i.e., unemployed, working in the home, retired, or disabled).
Table 1. The demographics of the survey participants.
Patient interest in referral. Across all conditions, the average rate of patient interest in being referred for a clinical trial was 72% (see Table 2). Patient interest ranged from a low of 56% among cancer patients to a high of 83% for those with chronic infectious diseases (e.g., HIV and hepatitis).
Table 2. Patient interest ranged from a low of 56% among cancer patients to a high of 83% for those with chronic infectious diseases.
Factors correlated with patient interest. Table 3 displays the results of the univariate regression where characteristics are grouped based on relationship with interest in referral. Disease condition and previous clinical trial experience both had a significantly positive effect on patient interest in study referral. Two additional variables were also significantly related to interest: presence of bothersome side effects (the more bothersome the side effects, the greater the interest) and satisfaction with current treatment (the lower the treatment satisfaction, the greater the study interest). Patient age, gender, race, number of medications, employment status, and insurance status were not significantly related to a patient's interest in study referral.
Table 3. Disease condition, side effects, treatment satisfaction, and previous clinical trial experience had a significantly positive effect on patient interest in study referral.
The results of the step-wise regression, however, show that the three variables that best explain patient interest in studies include: age, condition, and treatment satisfaction. Specifically, we find that patients age 65 and over are less interested in participating in a clinical trial than those who are under age 40 (odds ratio (OR)=0.549, p=0.0300). Additionally, patients who are dissatisfied with their medications are significantly more interested in studies (OR=2.105, p=0.0015). Finally, with respect to disease, in thinking of cardiovascular disease/diabetes as a reference condition, we find that patients with cancer (OR=0.056, p=0.0097) and gastrointestinal (OR=0.414, p=0.0420) conditions are significantly less interested in studies as compared to the reference of cardiovascular disease/diabetes.
The current analysis is believed to cover the most extensive sample on patient feedback ever examined because of both the large sample size and broad range of conditions included. In the analysis, it was found that individual factors such as condition, side effects, treatment satisfaction, and previous clinical trial experience, are all significantly correlated to patient interest in study referral. However, age, treatment satisfaction, and condition emerge as the best predictors of study interest in the multivariate regression model.
The finding that age is negatively correlated with interest in clinical trial participation is consistent with studies previously conducted in hypertension (Halpern)7 and kidney disease (Israni).8 Because both the Halpern and Israni studies used physician interview as the primary method of data collection, this also suggests that the finding of age as correlated to patient interest in referral is not limited to the online approach used to collect data for this study.
Similarly, factors found as having an insignificant relationship in the current study such as gender, race, and employment status were also found to be insignificant in the Israni analysis. The one inconsistency identified in comparing the findings from this study with previous research is that the Halpern study found in their multivariate analysis that previous participation in a research study was significantly correlated with willingness to participate. This factor, while significant in the univariate analysis, was not significantly correlated to interest in referral in step-wise, multivariate regression.
Although the bothersome nature of side effects appeared as significant in the univariate model, it is not surprising that the factor disappeared from the multivariate analysis due to correlation of this variable with treatment satisfaction. Both the side effects and treatment satisfaction questions are based largely on a validated patient-reported outcomes instrument called the Treatment Satisfaction Questionnaire for Medications, where correlation between the side effects and global satisfaction domains is known to exist.9
Due to the lack of existing literature examining patient interest in clinical trial referral across conditions, it is difficult to compare the findings with previous studies. One possible explanation for the variance in interest between CV/metabolic disease and cancer or gastrointestinal conditions is the patient's relationship with their managing physician. This relationship manifests itself in two different ways. In the case of cancer, for example, feedback from qualitative discussions with patients and responses to other quantitative survey questions suggests that cancer patients have such a close relationship with their physician that the patient's lack of interest in referral may be driven by their desire to remain under the care of their personal oncologist until the time that their managing physician recommends a clinical trial.
Unlike cancer patients, data on hand from MediGuard.org patient surveys suggests that gastrointestinal patients do not place a high level of importance on their physician's recommendation for (benefit of participation) or against (drawback of participation) a potential clinical trial. In fact, many of our patients screened for heartburn/GERD studies are self-treating with OTC medicines and are not in close contact with their primary care physician about their current symptoms or treatments. As a result, we believe that the low interest in clinical trial referral amongst these gastrointestinal patients stems from an overall low level of engagement with their underlying condition.
While approximately 250 responses in the analysis came from the United Kingdom, there was not sufficient sample size to confidently draw conclusions on potential differences based on country of origin. United States versus United Kingdom will be a focus of future analyses following the addition of more UK samples.
The data should be viewed within the context of some limitations. First, interest in referral was studied based on responses to screening questionnaires, not through actual clinical trial enrollment. While it is possible that the factors correlated with patient interest in referral may differ when actual participation in clinical trials is measured, the MediGuard.org experience in supplemental recruitment suggests that other factors such as travel distance to the site, responsiveness of the site in contacting the patient for screening, protocol inclusion/exclusion criteria applied at the site, etc., also emerge as important.
It is also important to remember that these study results are generated from the opinion of community based patients being offered the opportunity for referral to a different physician for a clinical trial. Thus, the findings could differ from studies examining feedback from patients screened within a physician's own practice.
Another limitation of this analysis is that it is difficult to control for all patient, disease, and protocol-specific factors that could influence a patient's decision to participate in a study. Patient characteristics such as race or education level, disease factors such as unmet medical need, and protocol-driven elements such as the study time commitment (e.g., study duration, number of visits, and frequency of visits), burdensome study procedures (e.g., lumbar punctures), or medication washout/switching are frequently mentioned by patients in MediGuard.org surveys as drawbacks of participating in a clinical trial. The current study was unable to control for these factors.
Finally, it is possible that the unique characteristics of the MediGuard.org patient population account for its high interest in clinical studies. As of September 2011, MediGuard.org had over 2.5 million members and is much larger than other online health-oriented social networks.2 While this may limit generalization of findings to non-digital patients, recent statistics from the Pew Foundation suggest that nearly 60% of all US adults are actively seeking health information online (75% have Internet access and 80% of those search for information on their condition or medicines).10 As the population ages and Internet-enabling technology continues to evolve (e.g., smartphones and iPads), remaining concerns around the generalizability of digital patients should continue to diminish.
The data from this analysis suggest that age and treatment satisfaction have the most significant effect on a patient's decision to participate in a clinical trial in any disease area. Other factors such as experiencing bothersome side effects, treatment satisfaction, and previous clinical trial experience are also correlated to patient interest in referral. These findings have important implications for several elements of the clinical trial recruitment plan, including advertising concept design, messaging, and channels for patient outreach for supplemental recruitment. Specifically, the findings suggest that physician investigators and other researchers that tailor their recruitment strategies to appeal to patients who are younger and have a greater dissatisfaction with their current treatment (due to bothersome side effects or unmet treatment need) should see a better return on investment for their patient outreach activities.
Looking ahead, as more robust data on supplemental recruitment for clinical trials becomes available within and outside of the United States, it will be interesting to include additional variables likely correlated to patient interest as well as compare metrics across countries to see if these same factors—age, treatment satisfaction, and disease area—remain most important.
Elisa Cascade*, Vice President, MediGuard.org, Quintiles, 1801 Rockville Pike, Suite 300, Rockville, MD, e-mail: email@example.com. Andrew Burgess is Statistical Scientist, Late Phase Biostatistics and Mark Nixon is Associate Director, Late Phase Biostatistics both at Quintiles, Reading, UK.
1. B. R. Luce, et al., "Rethinking Randomized Clinical Trials for Comparative Effectiveness Research: The Need for Transformational Change." Annals of Internal Medicine, 151 (3) 206-209 (2009).
2. Mount Sinai School of Medicine, "IOM and Mount Sinai Collaborate to Examine Issue of Declining Participation in Clinical Trials," http://bit.ly/p2Fp67.
3. Elisa Cascade, et al., "Direct-to-Patient Enrollment Strategies," Applied Clinical Trials, 19 (9) 44-50 (2010).
4. CISCRP, "Clinical Trial Facts & Figures," http://www.ciscrp.org/professional/facts_pat.html
5. ECRI, "Should I Enter a Clinical Trial? A Patient Reference Guide for Adults with a Serious or Life-Threatening Illness," http://www.pfizer.com/files/research/AAHP_CT_Guide.pdf.
6. News Medical, "Challenges in Patient Recruitment for Clinical Studies," (2009) http://bit.ly/3kspou.
7. S. D. Halpern, et al., "Hypertensive Patient's Willingness to Participate in Placebo-Controlled Trials: Implications for Recruitment Efficiency," American Heart Journal, 146 (6) 985-992 (2003).
8. A. K. Israni, et al., "Willingness of Patients to Participate in a Randomized Controlled Trial of Daily Dialysis," Kidney International, 65 (3) 990-998 (2004).
9. M. Atkinson, et al., "Validation of a General Measure of Treatment Satisfaction, the Treatment Satisfaction Questionnaire for Medication (TSQM), Using a National Panel Study of Chronic Disease," Health and Quality of Life Outcomes, 26 (2) 12 (2004).
10. S. Fox, "The Social Life of Health Information, 2011," Pew Research Center, http://bit.ly/m8Q7D3.
Related Content:Peer-Reviewed Articles