Background
The cornerstone of a clinical trial is the willingness of participants to invest time and effort. However, recruitment and retention for clinical trials is challenging [
1‐
5]. In the context of a research study, recruitment is the process through which an individual is recruited as a study participant. This process involves presenting the study details to the potential participant; review of the details (by the potential participant); making a decision about participating; and providing informed consent if the participants agrees to enroll in the study. Participant retention is the continuous engagement of the participant in the research study, post enrollment [
6]. While achieving the required sample size for recruitment and retention is critical for robust statistical power, and study validity [
2,
7‐
10], achieving the targeted sample in terms of number or timespan is challenging [
2,
5,
7,
10‐
19]. It was reported that only 55% of clinical studies accrue their total targeted sample size, and 78% accrue 80% of their targeted sample size [
18]. Recruitment and retention issues can adversely affect the ability to detect intervention effects and may limit the significance of the research findings [
2,
5,
7,
20].
Patient-centered outcomes research (PCOR) is a way to achieve patient-centeredness in research. Patient-centeredness requires active participation and can lead to empowerment [
21,
22]. Focus of PCOR is on comparing the effects of treatment options on outcomes that matter most to the patients [
21,
23]. Increased participation of stakeholders in clinical trials is a key element of PCOR that can enhance concordance between research topics and patient priorities, improve patient recruitment and retention, and inform clinically relevant evidence-based policy [
21‐
24].
Community-based participatory research (CBPR) is an excellent model for engaging communities in research [
25,
26]. There is an opportunity for the PCOR model to be informed by CBPR principles for enhancing patient engagement, and the two models can complement each other [
25,
27]. Principles of CBPR can be used to inform the three phases of a PCOR study: pre, continuous and sustained phase. The CBPR principles of community as a unit of identity, building on community strengths and equal partnerships can inform the pre-engagement phase of the PCOR research. The CBPR principles of bilateral learning and integration of research and action can inform the continuous engagement phase of PCOR research; and finally, the CBPR principles of system development, and involvement of all partners in the dissemination of findings can inform the sustained engagement phase of PCOR research.
Research indicates that recruitment strategies that evolve and adapt to the needs of the study and use local resources appropriately are successful [
28]. In addition, successful strategies are those that enhance a participant’s awareness of the health issue [
2,
29‐
32]. At the same time, operationalization of these insights into research protocols is challenging. Furthermore, while important, these studies do not describe a patient-centered approach for recruitment and retention. Patient-centered recruitment and retention implies continuous engagement of stakeholders to identify the resources available, develop appropriate strategies and address the challenges to recruitment and retention. The objective of our paper is twofold. First, we present a conceptual model of patient-centered recruitment and retention. Next, in the context of this conceptual model, we describe the recruitment and retention strategies, and related challenges in our PCOR-based comparative effectiveness randomized controlled trial (RCT) in localized prostate cancer patients.
Discussion
Involvement of patient and other stakeholders’ in research is an essential component of PCOR. Additionally, the PCORI encourages stakeholders’ involvement over the entire research spectrum. Thus, for a patient-centered outcomes study, it is necessary that researchers have a collaborative relationship with both participants and stakeholders. Our work offers new insights into patient-centered recruitment and retention strategies in the context of a large, patient-centered RCT study in the following ways: (1) a novel patient-centered recruitment and retention model to operationalize concepts of trust, attitude, communication and expectations; (2) active engagement of all stakeholders including patients to enhance recruitment and retention; (3) sensitivity to the needs and challenges; and (4) appropriate use of available resources at study sites.
Our conceptual model of patient-centered recruitment and retention consists of patient-, physician-, hospital-, and community-level factors that have implications for recruitment and retention. We observed that strategies related to our concepts of trust (e.g., physician involvement, ensuring protection of information), communication (e.g., brochures and pamphlets in physicians’ offices, continued contact during regular clinic visits, resending assessments), attitude (e.g., emphasizing the altruistic value of research, positive attitude of providers and research staff), and expectations (e.g., full disclosure of study requirements and time commitment, update letters) facilitated recruitment and retention. Screening patient records, identifying eligible patients, preparing recruitment material and ensuring that the relevant clinicians were informed about the study, were useful strategies in our study, and have been identified by prior research [
9,
28]. The in-person recruitment among eligible patients was more effective than telephone recruitment (87% vs. 41%). However, despite having trained research staff, time and resource constraints prompted use of diverse techniques to augment the recruitment and retention efforts. Contacting potential participants prior to their scheduled appointment helped streamline the process at one site. Participants arrived 1 h early for their appointment, and this offered sufficient time to discuss questions and concerns.
Collaboration between patients, clinicians, and researchers helped build trust. Most of our patients were in a sensitive period, having recently learned about their cancer diagnosis. The treating physician was fully knowledgeable about the study and the patients’ circumstances. The physician introduced the study and the research staff to the participant. Because of this process, the participant was able to appreciate the physicians’ involvement with the study. Additionally, the research staff discussed the measures that were in place for protection of health information and identity. The physician maintained ongoing involvement with the study and along with the research staff, remained accessible to the participant. Together, these measures generated trust and helped the recruitment and retention efforts. Studies report that up to 76% of patients expected their physician to alert them about appropriate clinical trials [
31], and that physician referral was one of the most useful recruitment strategy [
30]. One RCT of localized prostate cancer patients analyzed the comparative effectiveness and cost-effectiveness of nurses and surgeons in recruiting patients. The recruitment rates between urology consultants and nurses were not significantly different, and although nurses spent longer time with patients on average, they were more cost-effective recruiters [
46]. Due to the uncertainty that is inherent in clinical research, time pressures and lack of information about study details, physicians may be reluctant to discuss clinical research with patients [
38]. However, interaction between physician and patient is an important aspect of study recruitment [
28], whereas medical distrust is often cited as a barrier to recruitment [
11,
18,
20,
28].
Attitude was another of our core concept that aided recruitment and retention. By learning the flow of each physician’s office and working closely with the staff, we were able to enhance in-person recruitment. Altruism is one of the reasons for participating in a clinical trial [
38]. Across all study sites, we observed that patients agreed to enroll after they learnt that study has the potential to help future prostate cancer patients. Many of the enrolled patients were approached fairly soon after learning about diagnosis and were usually accompanied by family members. While the family members were naturally protective of the patient initially, when presented with the opportunity to help future patients, the family members played a pivotal role in encouraging the patient to participate.
We encountered some challenges in the course of recruitment and retention activities. For example, identifying and recruiting eligible active surveillance patients was time consuming as it mostly involved telephone recruitment. Usefulness of telephone recruitment was site specific. While at site 3 it was helpful in reaching and recruiting patients, at site 1, it was challenging to contact the patients from Penn Data Store list. Often the patient requested a call back at a specific time, which sometimes was inconvenient. One study attempted recruitment using targeted telephone and mail lists, and found that calling was time consuming and inefficient, similar to our findings [
12].
Research indicates that geographical location has an effect on recruitment for clinical trials, and that large cities were associated with poorer recruitment [
9,
28,
47]. The likely reasons are a more diverse population, traditionally more difficult to engage subgroups, greater population mobility, and presence of multiple research entities resulting in participant exhaustion [
9,
47]. Although recruitment for our study took place at urban, academic hospitals, we did not face these obstacles. While we attempted to recruit participants from all socio-economic groups, most of our participants had at least a college degree, and an annual household income of more than US$75,000.
Maintaining engagement at desired levels needs monitoring and tailored solutions for study-specific challenges [
28]. In general population, the response rates for self-completion surveys was as little as 11% [
47,
48]. It is likely that respondents have better health statuses, compared to nonrespondents and this can affect the generalizability of the results [
47,
49,
50]. Our retention strategies were also rooted in our core concept of “expectations”, e.g., providing study updates, and offering monetary incentive. Research has identified participant incentive as a helpful strategy to enhance retention [
51]. In a study of lung cancer patients, acknowledging and appreciating the efforts of participants motivated continued participation [
51].
Limitations
We note some limitations to our study. All three sites in our study are urban academic institutions and our sample consisted of men with localized prostate cancer. Most of our participants were college graduates, and had an annual household income of more than US$75,000. Thus, the generalizability of the findings may be limited. Additionally, of the four levels of factors (patient, provider, hospital, and community) depicted in our conceptual model, we focused mostly on the patient- and provider-level factors. There may be core concepts in addition to the ones that we have identified, and these need to be addressed in future studies. While we observed some differences in the yield of different strategies, it is not fully possible to separate the effects of different recruitment and retention approaches because several strategies were in place concurrently and no single strategy was adopted in isolation. Each research study faces specific conditions, and some recruitment and retention strategies we have presented may be more relevant than other strategies. Finally, in this paper our focus describing our novel patient-centered recruitment and retention activities. While we present the demographic differences between intervention and usual care groups to demonstrate that randomization worked, the outcomes data will be presented in our future work.