Recruitment,Retention and Intervention Delivery Experiences With Hospice Family Caregivers

Design

We conducted a pilot study to test an educational intervention for hospice caregivers, NOVELA, using a single-group pretest-posttest design. ⁷ Results of NOVELA intervention are published elsewhere. For the purpose of this study we tracked data related to recruitment, retention and intervention delivery and performed post-hoc comparison with SMM framework to describe lessons learned.

Participants

Study participants were HFCGs of patients who were receiving care from one of two hospice sites in the mid-Atlantic United States during November 2020 to June 2021. Inclusion criteria included: (1) HFCG older than 18 years, and (2) access to an internet device. Exclusion criteria included: (1) patients with a Palliative Performance Score of less than 20%, and (2) HFCGs unwilling to be recorded. The full trial protocol was approved by the Johns Hopkins University Institutional Review Board and is registered at clinicaltrials.gov (NCT 04533594).

Intervention

This intervention tested the use of a four-chapter telenovela video series, “To Care,” as part of a theory-based educational intervention, NOVELA. Each chapter addresses one challenge, including, decisions regarding hospice care, pain management, healthcare decision-making, and finally the dying process. Four online educational intervention sessions were arranged over 4 weeks. HFCGs met with a trained interventionist via synchronous web-conferencing. Participants used a personal web-enabled device (computer, smartphone or tablet). During the sessions, the interventionist introduced the purpose and topic of the video, facilitated video viewing, and then answered questions and reinforced the main messages for each topic. Each virtual session was video recorded, with participant permission, to enable monitoring of intervention fidelity and feasibility.

Recruitment Process

Hospice agency leadership determined social workers should present the program to potential participants during home visits. The study Principal Investigator (PI) held virtual meetings to train social workers in study protocol and demonstrated the intervention by presenting one video. A brochure with this information and a suggested hospice staff script were provided. Using recruitment strategies proven effective in our prior research,^1,8,9 all caregivers were screened by hospice staff within 48 hours unless they had opted out or their loved ones were actively dying (PPS < 20%). Hospice staff emailed the research team with referral information, including, patient name, caregiver name and phone number. Then a phone session was scheduled by the research staff to go over the study in detail and obtain verbal consent to participate in research activities. During this call, an informed consent form was read aloud to the participant, with contact numbers for additional information as well as contact information for the University Institutional Review Board (IRB). If eligible and willing to participate, participants verbally consented and were scheduled for their first intervention session. The consent form was attached to a follow-up email with the link to complete the baseline measures.

Fidelity Protocol

The fidelity plan was based on the NIH Behavior Change consortium. ¹⁰ We addressed fidelity by creating a theory-based intervention (design), and developing an intervention manual for the purpose of interventionist training (training). In addition, we provided personalized feedback based on the audio and video recordings of discussion content from this pilot (delivery), used a checklist for study session about intervention engagement (receipt), and participants discussed with interventionist questions regarding video topics and intention-to-change behaviors (enactment). During weekly study team meetings, the interventionist and study team discussed study sessions, reviewed content and opportunities to improve intervention fidelity. To further evaluate receipt and enactment, two research team members used a fidelity checklist as a tool. We then took a sample of intervention video recordings to calculate protocol compliance using the checklist.

Data Collection and Sources

Data on the number of referrals, recruitment (consented) and retention (received intervention), number of sessions attended, video viewing (total time viewed), time spent in discussing each video and content of discussion were obtained from video recording reviews. To evaluate intervention fidelity (receipt and enactment), two investigators (GM and DC) reviewed a random subset of intervention videos using a fidelity checklist (See Appendix) to calculate interrater reliability (IRR).

Data Analysis

Descriptive statistics for referrals, enrollment, sessions attended, and delivering the intervention were computed with SPSS^® v24 ¹¹ using frequencies and percentages for categorical variables, and means and standard deviations (SD) for continuous variables. Interrater reliability analysis using Cohen’s Kappa (κ) statistics^12,13 and percent agreement ¹⁴ was performed to determine consistency among raters (GM, DCO) for intervention fidelity.

Lessons Learned

Study procedures were designed to minimize burden to participants and hospice staff. Careful selection of inclusion and exclusion criteria helped us to avoid enrolling HFCGs of patients close to death (Participant). Even with careful planning, 8 patients died before the fourth video. The most important retention strategy was the virtual nature of NOVELA and the interventionist’s scheduling flexibility to work around HFCGs’ other demands (Place). Intervention duration of less than 15 mins/session and weekly frequency were acceptable to HFCGs (Price). This is in accordance with other studies^17,18 suggesting that investigators must carefully weigh maintaining low burden against the desire for more participation time. It is important to lay the groundwork of connection from the first point of contact for an ongoing reciprocal relationship, using person-centered retention strategies. ¹⁷ Development and maintenance of a procedural manual helped to ensure that all members of the research team adhered to the approved protocol and that the interventionist’s delivery of NOVELA was consistent with the fidelity checklist (Product).

Preparation allowed us to establish relationships with hospice leadership and negotiate the recruitment process (universal screening vs referral-based). The support of the hospice medical directors helped us gain support from hospice leadership and teams (Partners). Training and involving hospice staff members in the study preparation process helped to minimize staff gatekeeping documented in other hospice studies. ¹⁹ Given the competing issues that emerged during study (e.g. leadership staff turnover and COVID-19 restrictions), adaptations to planned strategies included, engagement with hospice staff expanded beyond originally identified study champions, gift campaign to encourage 3 referrals per staff provided and recurrent participation in interdisciplinary group meetings established.

Both sites chose to use a referral process where their agency staff determined who was referred to the study, rather than an automatic referral process where the research team was allowed to screen all admissions. (Promotion). However, the referral process still allows a gatekeeping role and places the study referral burden on overworked hospice clinicians. The relationship with the staff was critical to a successful referral process. Our team was in constant contact with hospice staff and provided study insights as often as possible to help busy hospice clinicians remember to refer. Common limitations associated with provider referral studies include poor prognostic accuracy, gatekeeping, individual variation in interpreting eligibility, lack of time and incentive and poor generalizability. ¹⁷ In our study recruitment was hindered due to lateness of study referral, short hospice LOS and HFCGs’ time restraints. When a patient was too close to death the HFCGs’ ability to watch all 4 videos was impacted by their caregiving burden. As a result, only half of participants fully completed NOVELA. While universal screening on admission was not approved by either of our hospice sites, our results justify further hospice site engagement to consider alternative recruitment procedures. We have learned there may be value in obtaining the necessary privacy waivers and implementing a referral process that places the screening burden on the research staff at the time of hospice admission.

Limitations and Implications

This study is limited because the sample size was small and full analysis of the effects of time and intervention dosage was not possible. However, this pilot study provides important opportunities for modifying recruitment and intervention delivery in a follow up clinical trial. During study start-up ensuring hospice staff engagement was crucial for the initiation and maintenance of recruitment. The established relationships with hospice partners are ongoing and will be re-engaged into exploring ways to decrease hospice staff burden by considering a form of universal screening on admission. Further, a research team will be housed within each hospice to foster a sense of teamwork and promote sustainability of these efforts. Research referral earlier in the trajectory of hospice care could potentially enable delivery of at least 2 sessions per participant, as we think that one session is insufficient. Therefore, although virtual intervention delivery over 4 weeks was feasible, a shorter timeframe will increase the potential that more videos are seen. As we move forward, further exploration will allow us to determine the best way to tailor the videos to effectively meet the unique needs of each HFCG.