Mock Recruitment for the Study of Antimalarials in Incomplete Lupus Erythematosus Trial

Abstract

Objective:

Recruitment to randomized clinical trials is expensive and often falls short of goals, limiting achievement of measurable outcomes. To prepare for a trial in patients with incomplete forms of lupus, a mock recruitment protocol was carried out at four proposed study sites. The objective was to determine levels of interest in patients and to uncover potential barriers to enrollment.

Methods:

After obtaining IRB approval, study coordinators approached individuals who generally fit proposed inclusion and exclusion criteria for the trial. A standardized script was followed in a structured interview. Levels of interest were determined and any reasons for concerns were collected with an open-ended format.

Results:

73% of 45 interviewees expressed an interest in the trial, and 64% said they were likely to enroll. Concerns of those who were not interested included risk of hydroxychloroquine, desire not to receive placebo and lack of time for participation.

Conclusion:

The mock recruitment suggests that the trial will be attractive to suitable patients. The concerns raised support other data indicating that provision of information is crucial to achieving enrollment goals. Mock recruitment of potential investigators should be considered to also address referral concerns.

The condition systemic lupus erythematosus (SLE) is defined by a combination of characteristic autoantibodies along with clinical and laboratory evidence of immune-mediated pathology. While there are no diagnostic standards for SLE, criteria were developed first in 1982 by the American College of Rheumatology (ACR, revised in 1997(1)) and again in 2012 by the Systemic Lupus International Collaborating Clinics (SLICC)(2). These “classification criteria” were designed to ensure homogeneous cohorts for observational or interventional clinical research. In each case, a person previously diagnosed with an autoimmune disorder is classified as having SLE based on the accumulation of clinical and laboratory features common to expertly adjudicated cases.

The laboratory and clinical features that allow lupus classification typically evolve over time(3). Autoantibodies such as antinuclear antibody (ANA) and anti-Ro are often detectable 5 years or longer before the first clinical symptom. Retrospective studies have documented that such individuals have evidence of altered innate and adaptive immunity including elevated levels of interferon-driven cytokines and memory B cells(4). Clinically, the features of SLE can be present over several years before formal classification is possible. This has led to the term “incomplete lupus erythematosus” (ILE), which may be used to refer to individuals with clinical or serological features of lupus who are progressing to incident systemic lupus erythematosus, as well as individuals who never develop the requisite number of criteria for formal classification(5, 6).

The diagnostic concept of incomplete lupus is problematic for several reasons. While a percentage of such patients slowly develop SLE, it is not clear whether that subset can be identified and given therapy that may prevent progression of their disease. Retrospective analyses of several large cohorts of such patients reveal that two thirds or more of them are treated with antimalarials such as hydroxychloroquine (HCQ), despite lack of an FDA indication for this condition(7, 8). This drug has been thoroughly tested in SLE, where it has been shown to prevent disease flares, decrease damage, improve survival, and delay onset of disease(9). However, it has not been rigorously tested in patients with ILE. While a recent study showed that ILE patients taking HCQ had lower expression levels of type I interferon inducible genes than those not taking the drug, the long-term clinical benefit of this medication is unknown(6).

The Study of Anti-Malarials in Incomplete Lupus Erythematosus (SMILE, NCT03030118) is a multi-center randomized, placebo-controlled clinical trial of hydroxychloroquine in ILE patients. The primary outcome of this trial is the change in rate of accumulation of Systemic Lupus International Collaborating Clinics (SLICC) criteria for the classification of lupus. Approximately 200 patients are planned to be followed for 2 years while on study medication. In order to plan for this trial, it was necessary to learn whether patients with ILE would be willing to participate. While HCQ is not formally approved for this condition, it is commercially available and most physicians feel very comfortable using it in this population. Therefore, the prospect of being randomized to placebo may make it difficult to recruit sufficient numbers of participants. To study this and judge the feasibility of the clinical trial, a “mock recruitment” was done for SMILE. This has given important insights into the willingness of patients to participate and informed our recruitment strategies.

Patients and Methods

This study was performed at the University of Texas Southwestern Medical Center, the Milton S. Hershey Medical Center of Penn State University, the Oklahoma Medical Research Foundation, and the Cedars-Sinai Medical Center. It was approved by an Institutional Review Board at each site.

Study participants were identified during routine clinical care in the rheumatology and/or dermatology practices of the participating institutions. They met the provisional inclusion criteria for SMILE: 1) Age between 18 and 45 years; 2) ANA ≥ 1:160 by indirect immunofluorescence testing; 3) one or two additional SLICC criteria for the classification of lupus; 4) either sex; 5) able to provide informed consent.

Participants who met the above described inclusion and exclusion criteria were read a script (available online in Supplementary Material) that described rationale and conduct of the planned SMILE trial. The risks and potential benefits of hydroxychloroquine and placebo were explained as well as the fact that hydroxychloroquine is already available from their provider. Then, the physician or study coordinator obtained informed consent from the patient for their participation in the mock recruitment study. From the consented patients, basic demographic data and information on personal or family history of autoimmune diseases were collected. Participants were then asked the following four questions:

1. I have told you about a proposed clinical trial or study of hydroxychloroquine to treat early or mild lupus-like illness similar to the condition your doctor is treating you for. Do you have any questions about the proposed clinical trial? Can I explain it better?

   (Participants could answer “Yes” or “No”)

2. If the proposed clinical trial of hydroxychloroquine to treat early or mild lupus-like illness was enrolling patients today, how much interest do you have in learning more about the trial (for example, look at consent form, talk to family, talk to other doctors)?

   (Participants could answer 0%-no interest, 25%, 50%, 75%, or 100%-very interested)

3. If the proposed clinical trial of hydroxychloroquine to treat early or mild lupus-like illness was enrolling patients today, how likely would you be to enroll in the trial, considering all the information you have today?

   (Participants could answer 0%-definitely not enroll, 25%, 50%, 75%, or 100%-definitely would enroll)

4. If you answered 0%, 25%, or 50% to either Question 1 or 2, can you tell us what your concerns are? You may pick more than one.

   (Participants were offered the following prompts: Risk of hydroxychloroquine appear too great; Don’t want to risk getting placebo; no time to participate; already tried hydroxychloroquine in the past; want to try something else; other-free text)

Additional data, including the presence of ANA, other serology, immunological features of SLE, presence of cytopenias, serositis, musculoskeletal complaints, cutaneous complaints, and constitutional symptoms (e.g., fatigue, fever), were obtained from the clinical chart. All data were entered anonymously into a REDCap database hosted by the UT Southwestern Center for Translational Medicine.

For the purposes of analysis, participants were categorized as likely to enroll in the described clinical trial if they answered Question #3 with “75%” or “100%”. A Fisher’s exact test was performed to test the significance of differences in the proportion of such participants based on clinical and demographic characteristics.

Results

A total of 45 subjects were included. The median age was 35 (range: 16–57) years and the median duration of symptoms prompting evaluation by a rheumatologist or dermatologist was 3 (range: 1–21) years. Most participants were female (N=43; 96%) and non-Hispanic white (31; 69%). A minority were African-American (6; 13%) or Hispanic (5; 11%). Some reported that they themselves had previously been given a diagnosis of an autoimmune disease (8; 19%), while a greater number indicated that a family member had been diagnosed with an autoimmune disease (19; 42%). In some of these patients (14; 31%), autoantibodies other than an ANA were known to be present at the time of the interview. Musculoskeletal complaints (arthralgia/arthritis or myalgia) and cutaneous complaints were the most common clinical features (27; 60% and 26; 58% of participants, respectively), and constitutional complaints were also relatively common (20; 44%); these were primarily fatigue or low-grade fever. No participant had known contraindications for the use of hydroxychloroquine.

Most of the participants (42; 96%) felt that the description of the proposed clinical trial was adequate. One subject requested more information on the constituents of the placebo capsule. Overall, 33 (73%) of the interviewees expressed interest in the trial, and a majority (29; 64%) said they were likely to enroll. None of the clinical or demographic features listed in Table 1 were significantly associated with the likelihood of “interested in the trial” or “likely to enroll in the trial”.

Table 1.

Interest and likelihood of enrolling in a trial of hydroxychloroquine for incomplete lupus erythematosus

Feature	Number	Interest in Learning More About the Trial		Likelihood of Enrolling in the Trial
		“0–50%”	“75–100%”	“0–50%”	“75–100%”
Ethnicity/race1
Non-Hispanic White	31	7	24	11	20
African-American	6	1	5	2	4
Hispanic	5	2	3	3	2
Previous diagnosis of autoimmune disease	8	1	7	3	5
Family history of autoimmune disease	19	4	15	3	5
Serology other than ANA	14	2	12	3	11
Cytopenias	4	1	3	2	2
MSK symptoms	27	7	20	8	19
Cutaneous symptoms	26	6	20	9	17
Serositis	1	0	1	1	0
Constitutional symptoms	20	3	17	5	15

¹Two participants were Asian and one declined to identify their race. All comparisons of low versus high interest and low versus high likelihood of enrollment were non-significant using Fisher’s exact test.

The 16 interviewees who were not interested or were not likely to participate in the clinical trial were asked for their reasons. Some (4, 25%) were concerned that the stated risks of hydroxychloroquine appear too great and there was also concern about the risk of getting placebo (3, 19%). Half of the not interested group (8, 50%) felt that they did not have time to participate in a trial. Free text answers also included concerns about the number of medications they were on already and the logistics of participation (Table 2).

Table 2.

Free text concerns regarding potential participation in the proposed clinical trial

Participant Statement
Fear of aggravating celiac disease
Depends on job schedule, difficult to make appointments
Uncertain about starting a new medication for the patient, personally.
Just concerned about starting a new study.
Symptoms are not that significant to want to take another medication.
Give it more thought
What about comparing two doses of Plaquenil?
Wonder about side effects
On too many other medications
Want to do more research about the drug
Would like to remain off medication and would not want placebo if I needed medication

Discussion

Recruitment and enrollment in randomized clinical trials is a major challenge of clinical investigation. Clinical trials in human subjects are expensive and recruitment costs are a significant proportion of the expense. Failure to reach recruitment goals has been estimated to occur in at least 20% of trials and often results in premature study closure and decreased reliability of the data generated. As a consequence, translation of research to clinical practice is impaired, and implementation of advancements in disease treatment is slowed. Recognition of these problems has led to re-examination of approaches to recruitment, which has generally been an empiric process that is addressed only after all other elements of the study are in place. The significant and unmet need for development of innovative tools for recruitment, education and outreach has been shown by the fact that the National Institutes of Health has supported initiatives in this area (www.trialinnovationnetwork.org).

Systematic reviews of recruitment failure have shown some common themes. Most often, investigators over-estimate the number of eligible participants who meet the inclusion criteria(10). This is a possibility with the SMILE trial as well. While the presence of ANA in the general population is common, potential participants in SMILE must meet several other inclusion criteria, including age and presence of one or two objective SLICC criteria, as well as not having taken hydroxychloroquine previously. In another review of recruitment strategies, Caldwell, et al., determined that various strategies to increase the potential participants’ knowledge of the health problem under study through computer-aided instruction, videos, and seminars was the most helpful(11). Others have emphasized the greater effectiveness of recruitment to clinical trials that enroll African-Americans and Latinos when pre-consent education was provided(12). This was anticipated for the SMILE trial with the development of culturally appropriate recruitment materials and Spanish language trial information.

The proposed study in patients with incomplete lupus presents some unique challenges. One is the condition itself. Eligible patients will be those who have certain inclusion criteria but who do not satisfy the standard SLE classification criteria. The concept that this is a lupus-related trial but that it does not include patients with SLE has required outreach education with colleagues who are the major referral sources for eligible patients. Another challenge is the double-blind, placebo-controlled design. About one in five interviewees who thought they would not enroll expressed concern about placebo treatment as the reason. However, given the inherent variability in the proposed study population, placebo and active arms were deemed essential to evenly distribute the many unknown variables. The long-term trial objective of prevention of SLE, also presents challenges. A major barrier to enrollment in one previous study to prevention of cardiovascular disease in SLE patients was found to be the poor understanding of the importance of prevention, on the part of both patients and the physicians who were sources of referrals(13). That previous experience differs from the SMILE trial in that symptomatic improvement was not a likely outcome, given that the target was prevention of atherosclerosis, and the interventions proposed were unlikely to alter the patient’s perception of well-being. In contrast, individuals who receive active treatment in the SMILE trial have a likelihood of improved symptoms, which will be useful for retention as well.

It is also important to consider the baseline prevalence of persons eligible for the SMILE trial. A formal count of individuals meeting all inclusion and exclusion criteria was not possible. Preliminary estimates at the trial sites indicated that 12–25% of all referrals were for ANA positivity (not shown). At one site (UT Southwestern) an automated process monitors all patients who get ANA testing and are in the age range for SMILE. Typically, there are ~75 such people per month, with 1–2 per week meeting the other entry criteria for SMILE. This suggests that while recruitment for SMILE is numerically feasible, robust strategies such as tools that search the electronic health record and repeated outreach to referring physicians and community rheumatologists are necessary for success.

One relevant question is whether patients and providers would support enrollment in a trial of an agent that is readily available for off-label use, as is the case for hydroxychloroquine. Erkan, et al., recently reported the experience of an international multicenter trial designed to explore efficacy of hydroxychloroquine thrombosis prevention in patients with anti-phospholipid antibody syndrome(14). The trial was designed by a consortium of investigators who determined the urgent need for such a study. The design was non-blinded with hydroxychloroquine as the active agent, compared to a control group that received no treatment. The sample size estimate was for approximately 1,000 patients, but after 2 years of recruitment only 2% of this goal was reached and the study was closed without achieving interpretable results. The study encountered many problems, including cost and availability of the active drug, as well as a reluctance on the part of some of the enrollment sites to administer what was perceived as prophylactic therapy to relatively healthy individuals. Patients also expressed an unwillingness to take a medication for what they saw as a low risk of events. The enrollment goal was very high; by contrast the power calculations for the SMILE trial suggest only 192 completed patients will be required. Whether any of the issues in this previous trial might have been uncovered and mitigated by a mock recruitment exercise or by an alternative design is not known. However, it did not appear that off-label use of HCQ was a major reason for failure of recruitment.

A common theme in clinical trial strategy development is that engagement and education of both potential patients and referring providers in the health issues that are being addressed by the intervention trial are essential elements of successful recruitment. The reason for the trial appears to have greater impact than elements of the trial process itself. In the SMILE mock recruitment exercise, one-quarter of those who were not interested or not likely to participate expressed concerns about the risks of hydroxychloroquine, which are in fact low and which might be addressed with appropriate educational materials. A mock trial simulation in lupus patients at two centers reported similar findings in that patients recommended community engagement to provide information about the disease and the impact of the proposed study on patients(15). These investigators also suggested that study teams be sensitive to the needs of patients, including concerns about scheduling and time constraints, issues that were expressed by half of those who were not interested in the SMILE mock recruitment. This consideration was taken into account when planning the visit schedule and length of visits in the final protocol.

The attitudes of colleagues who would be sources of patient referrals for enrollment was not surveyed in the mock exercise and likely would be of value to include in such pre-study activities. Appropriate educational materials might be developed in response to concerns about randomization, placebo treatment and drug risk. The cited experience of the multinational study of hydroxychloroquine and thrombosis shows that even when the study question is considered by experts in the field to be of high importance, individual providers may be reluctant to participate. In the case of SMILE patients and providers will be assured that participants will be observed more closely than would be the case in routine care. If any individual progresses to an SLE classification, study exit is mandated so that appropriate therapy can be initiated.

Significance and Innovation:

• The majority of patients with incomplete lupus erythematosus were interested in a clinical trial comparing untested standard of care to placebo.

• However, only 64% said they would be likely to enroll in such a trial.

• Successful recruitment for a trial such as this, which tests a treatment available as part of standard of care, will require extra education of potential subjects and minimizing the impact of the study on participants lives.

• Study success will likely require that twice as many potential participants are screened as need to enroll for completion of the trial.