The proposed palliative care research is responsive to PA-13-354/NOT-CA-14-016, Advancing the Science of Geriatric Palliative Care for Cancer Patients. It is significant for multiple reasons: it breaks new ground in inquiry regarding spiritual care as a part of patient-centered care for elders with serious illness before they get morbidly ill. The study uses an empirically established intervention (DT), opening chaplaincy in palliative care to rigorous health services research. The study advances palliative care research by leveraging a network of collaborators in a recently formed network for large data, and by making use of a relatively novel method, stepped-wedge design that typically reduces sample size needs and makes recruitment more feasible, especially for those who are frail and stressed by their cancer illness and treatments.

DT has its conceptual origins in Butler’s Life Review [7], which he developed as an antidote to depression in elders and understood as part of a life-cycle task. Both interventions are conceptualized as psychosocial and multidimensional for patient-centered care. But little is known about how DT works and if it can and should be used clinically.

Although DT has been established in RCTs to be beneficial to patients among multiple disease groups in multiple ways, the question of its efficacy when administered in a real-life care setting is not established. Pragmatically, within palliative care teams, it is most likely that nurses or chaplains could include DT as part of their routine work, and we, therefore, have designed a trial to assess the efficacy of DT delivered by nurses or chaplains. Considerations for either discipline include the following facts. In a prior RCT [1], a research nurse administered the DT with significant dignity impact findings. Furthermore, the nursing discipline’s focus on holistic care inclusive of spirituality makes this discipline a strong candidate. Although the ratio of nurses to patients is suitable, nurses already have a heavy flow of work that may be sidetracked by DT unless carefully scheduled. In addition, recent work indicates that a taxonomy of chaplaincy activities reasonably aligns with DT components [8,9] and recent findings [10] also indicate that clinicians from nursing and medicine look to board certified chaplains (BCCs) as the professionals with the expertise to provide spiritual care. Chaplain-to-patient ratios and their assignments are not currently suitable for routine offering of DT, but chaplains might be more interested in DT than nurses due to alignment of DT with chaplain tasks. If our findings show that DT led by nurses or chaplains has outcomes better than usual care, in the future, DT could be implemented in many places by redeploying current nursing or chaplain resources, based on their availability. Fortunately, six palliative care outpatient services are committed to test the efficacy of DT led by either nurses or chaplains and thereby provide evidence about the effects of DT compared with usual care in the context of outpatient palliative care of elderly cancer patients.

As palliative care starts earlier in the course of medical therapy and geriatric oncology care continues to seek interventions for ambulatory elders, we have asked if DT can be successfully administered in outpatient settings. The findings from a pilot study conducted in a colorectal cancer clinic indicate  feasibility in an outpatient setting [6]. This pilot study provides guidance about where DT should be done and when in the sequence of the patient’s clinical encounters. Our experience with this study also helped us plan who should transcribe the interview, how to return the legacy document, and how long an interval between transcription and delivery is suitable. None of these decisions are established for clinical practice, but the proposed study will provide evidence from 6 sites to better inform future clinical practice implementation of DT, including the impact on the flow of the team’s work and DT’s impact on illness acceptance and cancer care goals. Successful completion of the proposed study is likely to have a significant and sustained impact on gero-oncology care in the future.

Many  elders become more spiritually inclined   as a life-cycle phenomenon [11,12]. However, little research exists on how this inclination relates to their medical care. New evidence suggests that DT is primarily a spiritual intervention with implications in the other spheres of a person’s experience.

Figure 1 displays the study’s conceptual framework including the mechanisms to explore. In this individual-level model, we conceptualize physical symptoms and spiritual distress as moderators of the DT effect on spiritual outcomes, measured as dignity impact, existential tasks, and cancer prognosis awareness, and on process measures (patient satisfaction and unmet spiritual needs).

In a study of factors considered important among 340 patients, with advanced illness, most of whom were elderly, some diagnosed with cancer, being at peace with God was as important as freedom from pain [2]. This finding is consistent with the broad consensus that attention to patients’ spiritual concerns is 1 of the core dimensions of palliative care and a part of comprehensive geriatrics care [13-15]. Research with terminally ill patients makes clear that, in their view, a good death includes addressing several central spiritual issues and tasks [2,16]. A body of evidence is developing that describes moderate to high levels of spiritual concerns, unmet needs, or struggles among patients in palliative care or with advanced illness [3,17,18]. Existing evidence is limited but points to improved patient illness experience when patients’ spiritual needs are addressed, as well as greater use of hospice rather than ICU care at the end of life [19]. Furthermore, when patients feel that their spiritual needs are unmet, they have lower satisfaction with care and increased emotional distress [20,21].

Delivery of spiritual care services is a variable and the evidence-based studies on chaplaincy interventions are limited. There are few published studies of interventions designed to address the spiritual needs of palliative care patients [22] and no studies of spiritual care provided by chaplains create a gaping hole in knowledge of the effects of the interdisciplinary team. Fortunately, professional chaplaincy is currently undergoing an important transformation. Until now, chaplaincy practice has been guided primarily by tradition and expert opinion. However, professional chaplaincy associations recently embraced the importance of evidence-based care, as have practicing chaplains [23]. Presently, the major obstacle that prevents chaplains from adopting evidence-based practice in palliative care is the absence of any research about chaplains’ interventions. Although much of US health care and elsewhere is shifting from a focus on volume-based metrics (eg, number of patients seen or procedures done) to value-based metrics such as increasing patient satisfaction, chaplaincy has been very slow to make this change particularly with regard to proposing and testing appropriate outcomes. In the proposed study, we begin to fill this gap.

Few protocols- or evidence-based chaplaincy interventions have been used to meet the spiritual needs of people facing aging with chronic or serious illness or to test whether meeting those needs would improve outcomes valued by the elderly patient with cancer. Therefore, rigorous evaluation of a manualized, evidence-based intervention would constitute a landmark step in gero-oncology and palliative care research and in nursing and especially chaplaincy health services research .   DT is a well-designed, validated, and manualized intervention well-suited for clinical protocolization. For these reasons, we proposed to include both nurse-led and chaplain-led DT with each compared only with usual palliative care and hypothesize that each DT group will report higher dignity impact than their respective usual palliative care group, which would mean that either discipline could lead DT when it is implemented clinically in palliative care. Nurses, but not chaplains, have been the DT interventionists in previous feasibility studies; however, we expect both can be trained to competently lead DT in the outpatient palliative care setting with elderly cancer patients. Distinguishing effects between the DT groups is beyond the scope of our application, but our findings will provide effect size estimates for future studies, should such a comparison be warranted by our findings. Both disciplines need to be included in this study to provide an efficient test of the efficacy of DT and time of delivery in clinical care by the disciplines most likely to be available, an important issue for translation of result findings into practice.

This study is generated from a group of collaborators that came together under the auspices of the Patient-powered Spirituality-and Quality-of-life-focused Advanced-illness Research-network, a group formed in 2013. An open community of 27 organizations (health providers, researchers, patients, and other stakeholders) invited participation in research and health information giving.

Palliative care research is inherently hindered by the patients’ severity of illness. High illness burden and short life spans make participant accrual and retention challenging; the participation burden must be minimal. Using a stepped-wedge design, 1 possible design for an RCT [4], will require a  smaller number of participants and allow intervention integration into the care process while maintaining power, data quality, and desirable design features including randomization to control and intervention groups.

In summary, successful achievement of study aims will have important impact on the fields of geriatrics, comprehensive cancer care, and palliative care by showing whether DT can be optimally used in routine care settings to bring better illness experience to patients. It will further advance understanding of how DT works.

This proposed work is innovative in 6 ways. First, and most importantly, we primarily conceptualize DT as a spiritual intervention for individuals with serious illness. Other researchers have approached DT as an intervention that broadly improves the dignity of patients nearing the end of life but examined the effects of DT on a variety of outcomes and only psychosocial outcomes showed a significant effect [24,25], In contrast, our second innovation is the use of spiritual outcomes as primary and secondary indicators of the DT effect. Our focused conceptualization of DT as a spiritual intervention leads to our choice of unique outcome measures and DT’s focus at the level of the individual patient, though it could be conceptualized at the family level, which is beyond the scope of the proposed study. Other DT researchers used multiple measures of physical and psychological symptoms, as well as a broad measure of personal dignity. In contrast, our measures focus on dignity impact and existential tasks that help patients feel prepared for life completion and to have awareness of their prognosis. Furthermore, our measures of these constructs have been successfully used in previous palliative care or DT-related studies, and our recent reanalysis of RCT data shows that dignity impact differed for the DT group compared with usual care and another patient-centered intervention [5]. Although there is consensus about the central role of spiritual care in palliative care, there are few studies about spiritual care and no studies of spiritual care provided by chaplains. Thus, the third innovation of our study is that it employs a rigorous research design to test the efficacy of a manualized nurse-led or chaplain-led intervention in the palliative care context. Studies such as this are essential to clarify what spiritual care contributes to palliative care and to advance an evidence-based approach to chaplaincy care. Consistent with our rigorous test of a spiritual intervention, our fourth innovation will be to gather detailed information, in both the usual care and DT samples, about the nurses and chaplains’ assessments and activities for descriptive purposes. The development of evidence-based chaplaincy has been hampered by the lack of evidence-based assessment instruments and standardized descriptions of chaplaincy care. Our work will employ innovative approaches for nurse or chaplain assessment of the patients’ spiritual needs and for nurses’ or chaplains’ description of the care they provide. Prior studies of DT have focused on its effects on key outcomes and patients’ reports of its benefits. They have not proposed or tested hypotheses about the ways that DT improves patient dignity during serious illness like cancer. The fifth innovation is the test of 2 new hypotheses about 2 moderators of DT effects, and we will employ rigorous measures to do so. The sixth innovation is that we will employ a stepped-wedge design, which has been used in other fields, but is new to clinical trials in the palliative care context and addresses the challenges of recruitment in this clinical setting.

In summary, by focusing on DT as a spiritual intervention, using a new conceptual model and measures consistent with the model to examine its effect, examining several pathways that shape its effect, and using a stepped-wedge design, our study will contribute important novelties to palliative care research. In addition, by testing the effects of nurse-led or chaplain-led DT, we will advance an evidence-based approach to spiritual care in the gero-oncology palliative care context, which is highly responsive to the National Institute of Aging’s and the National Cancer Institute’s request for palliative care applications (PA-13-354/NOT-CA-14-016) to study elderly cancer patients.

We propose a 6-site, pre- and posttest, randomized, controlled 4-step, stepped-wedge design to compare the effects of usual outpatient palliative care and usual outpatient palliative care along with nurse-led or chaplain-led DT on patient outcomes and palliative care processes. We will assign the 6 sites to usual care during the first-step period (10 months), and randomly assign 2 sites per step to begin and continue DT during each of the next 3 steps (10 months each). Figure 2 shows the stepped-wedge study design with projected numbers of completed patients needed per site, step period, and group (usual care, DT led by either a nurse or a chaplain). Dr. Yao, a highly qualified statistician, will conduct the randomization of site from steps 2 to 4. Each step will be 10 months long, with DT training during a 1-week period between steps. For each site and step, a quota of 50% of the participants will report low or high distress on the Personal Dignity Inventory to assure that we recruit a sample with a range of problems threatening their dignity. Each patient will participate for 4 to 6 weeks. During the 10 months at each step, we expect 23 to 24 patients to participate at each site (93 to 94 total patients per site).

For this efficacy study, the settings will include Northwestern University Hospital, Rush University Medical Center, MD Anderson Cancer Center, Emory University, University of California San Francisco, and University of Florida Health. The clinical resources of each of these settings are strong. Our team members have extensive experience recruiting and retaining elderly patients with cancer for studies, including palliative care and chaplaincy research.

We will recruit the study sample from the populations receiving care at settings across the United States. For the patients who meet the eligibility criteria, we anticipate a 50% to 60% enrollment rate and a 20% to 30% attrition rate. We based these estimated numbers on accruals, withdrawals, and deaths in previous palliative care studies conducted at study settings and those in prior studies of DT. During the 40 months devoted to data collection, we anticipated a total available population of 3126 from the outpatient palliative care services of the 6 sites. From this sample, 560 patients will complete sufficient data for the planned analysis, 140 per each of 2 DT groups and 280 total control group. Therefore, we expect each of the 6 sites to complete at least 2 to 3 patients per month and a final total of 93 to 94 patients per site with data for analysis. On the basis of ethnic and racial distribution of patients served at the 6 sites, we expect the final sample to be 50% female, about 3% Asian American, 21% African American, 73% Caucasian, and 3% unknown race and 8% Hispanic ethnicity, representing substantial cultural diversity for this study of DT.

Inclusion criteria for study participation require that the patient (1) has a cancer diagnosis (receiving cancer therapy or cancer control care to be responsive to NOT-CA-14-016), (2) is receiving outpatient palliative care, (3) is aged 55 years or older (responsive to PAR-13-354), (4) is able to speak and read English, and (5) is physically able to complete the study (Palliative Performance Scale [PPS]>50 [40-45], suggesting a mean in life expectancy of greater than 53 days at the time of enrollment) [45]. Patients will be excluded if they: (1) are legally blind, (2) are cognitively unable to complete study measures (Mini Mental Status Exam [MMSE] <24), (3) have history of psychosis (medical record review), (4) have a Patient Dignity Inventory (PDI) or Religious and Spiritual Struggles Scale score that indicates their distress level falls outside the remaining quota for a given step; quota is 50% of sample, site, and step with low distress (≤ 2 problems rated >2) and 50% with high distress (≥ 3 problems rated >2 or 1 rated>3), or (5) are enrolled in another intervention study that is focused on concepts similar to the proposed study.

Retention strategies are vital to engaging participants as active research partners. Researchers in this study will always use respectful, empathic communication and schedule data collection at convenient times for participants. We will collect multiple contact information (email, cell phone, home, and extended family phone) to maximize our ability to contact them for final appointments. We also will offer a total of US $50 per patient to cover time and travel expenses to complete the study measures.

Aim 1 focuses on whether DT has an effect on patient outcomes and palliative spiritual care processes. For the 4-step, stepped-wedge design and the sample size of 23 to 24 patients per site per step (93 to 94 total), we determined from the minimum intervention effect size that the proposed study has at least 80% power to detect given a 2-sided significance level of .5. This minimum detectable effect size depends on disparity between sites as indexed by the intraclass correlation (ICC). With no site disparity (ICC=0), we can detect an effect size of 0.5 or above for either the nurse-led or chaplain-led DT intervention; with a more severe disparity (ICC=.4), we can detect an effect size of 0.6 or above. The primary outcome for this study is the 7-item Dignity Impact scale. Using Dignity Impact scale data from the preliminary study [1], the effect size of DT (relative to usual care) was 0.7 [5]. Therefore, we expect this study to have sufficient power to detect the intervention effect on the primary outcome. We do not have preliminary DT data on the secondary outcome measures; however, based on other psychosocial interventions with the measures [37-39,46], our power analysis indicates that as long as the intervention effect on these outcomes is of medium size (0.5 as defined by Cohen) or larger, they can be detected in the proposed study. Our study aims to focus on differences between the control group and each DT group, but do not focus on differences between the provider groups because, with the proposed sample, the effect size difference between the nurse-led and chaplain-led DT groups needs to be large (0.7 when ICC=0 and 0.9 when ICC=.4) to be detectable (with 80% power). As prior research shows an effect size of 0.7 between control and DT groups, it is highly unlikely that there will be a significant difference between the provider groups with the proposed sample. Aim 2 seeks to explore whether or not patient distress (physical or spiritual) moderates the effect of the intervention. The intervention effect size of patients with high distress will be compared with that of patients with low distress level. The study has power (80%) to detect an effect size difference of 1.0 if there is no site disparity (ICC=0) and 1.3 if there is severe site disparity (ICC=.4). Assuming an average effect size of 0.8 and group balance, we can detect a moderation effect if the intervention effect size for the less receptive group is 0.15 or lower.

Figure 3 shows the general flow of patients through the outpatient palliative care clinics and study procedures for this study of nurse-led or chaplain-led DT over 4 to 6 weeks. Before participant recruitment, the research assistants (RAs) will be trained in all study procedures. The investigators will introduce the study to the palliative care team members. Members of the palliative care team will refer patients to the RA on the day the patient is seen in the palliative care clinic.

The RA will recruit and consent the patient for screening. For those eligible, the RA will provide informed consent and schedule time before the next clinic visit for the patient to complete the pretest self-report measures using a tablet with smart pen. Then, the RA will inform the nurse and chaplain that the patient is available for care (usual care or DT, depending on the study step and site). The nurse or chaplain will provide the designated care on the same day as the pretest data collection. When the nurse or chaplain sessions are completed, the nurse or chaplain will inform the RA that he or she has completed all nurse or chaplain sessions. Either in-person if the patient has a clinic visit or by telephone contact, the RA will schedule the follow-up appointment and complete the posttest measures within 5 to 10 days. Upon completion of posttest measures, the RA will provide a US $50 cash per card to the patient for time and travel expenses to complete study measures.

In this trial, we will compare the efficacy of 2 interventions, both provided either by palliative care nurses or chaplains. The nurses will be the registered nurses who are trained and working with the palliative care team. Chaplains will be BCC. To become a BCC, a person must have completed a graduate-level theological education as well as 1600 hours of supervised practice in an accredited program of clinical pastoral education. BCCs must be in good standing with, and endorsed by, their faith group. In their application for Board Certification and interview, BCC candidates must demonstrate competence in 29 areas. BCC chaplains adhere to a code of ethics that includes “respect for the cultural and religious values of those they serve and refrain from imposing their own values and beliefs on those served.” BCCs must complete 50 hours of continuing education annually and a peer review every 5 years [47].

Palliative care nurses usually see patients each clinic visit to assess vital signs, function, symptoms, and to provide patient and family education. They document findings and interventions in the electronic health record (EHR). Although, usual care for palliative care chaplaincy in the outpatient setting varies by site, chaplaincy care for usual care patients in this study will follow the usual practice.

The DT intervention is detailed in Chochinov’s manualized guide; he serves as a coinvestigator. He advised the development of this application since its inception and gave us permission to use the manualized intervention as it appears here in abbreviated form. The basic questions of the DT interview appear in Textbox 1. The nurse-led or chaplain-led DT intervention involves 3 sessions, each of which follows a set process (Table 2). Nurses are familiar with use of manuals. Chaplain use of manuals is not common, but Dr. Karen Steinhauser (personal communication, February 2014) confirmed that trained chaplains used a manualized intervention guide with high fidelity. Table 3 lists exemplars of the dignity repertoire (perspectives and practices) facilitated by the interview and document preparation process. The standardize approach to the delivery of the intervention facilitates a personal process of reflection and recognition that allows the patient to make meaning of their experience.

Dr. Chochinov, who developed DT, will provide standardized training for the study nurses and chaplains at one of the study sites on 3 separate occasions; each training session will be 2 days in duration. Nurses and chaplains at all sites will be prepared to participate in the DT training during the 3 training dates that will be set early in year 1 but will first occur at the end of year 1. Only those nurses or chaplains at the randomized sites randomized to begin DT at the next step will be trained at any one date, and they will be notified before the training date. By the 4th step, all nurses and chaplains will be trained, assuring competency of all interventionists to deliver DT with fidelity. If there are new nurses or chaplains at a site randomized to the DT step, they will be allowed to attend remaining training sessions. In addition, 2 weeks before the end of each step, the statistician will divulge the 2 sites randomized to the next step. Before training, the MPI will inform the team members about the randomization. To reduce potential influence of the nurse or chaplain anticipating the training and thereby altering usual care practice, we will suspend recruitment, pretest data collection, and intervention delivery during the training periods; posttest data collection will continue.

The second RA, who is not involved with data collection, will complete the DT Adherence Form as a measure of intervention fidelity for all DT patients. Members of our team used this tool in previous studies with excellent results. The RA will read each original DT transcript and use this measure to document the nurse’s or chaplain’s adherence to the DT protocol. Scores range from 0 to 10, with 10 representing complete adherence. We will monitor the adherence scores to determine if a nurse or chaplain requires additional training by Rev. Handzo (those with scores <8).

All our measures have been used among palliative care populations, summary of timing of these measures is shown in Table 4. To collect these data, we will create a REDCap study site using the UF REDCap system, which the UF CTI supports and makes available for NIH funded studies without cost. All data collection sites will use this secure data collection site. Patients will use a tablet with a smart pen to enter the data into the REDCap system. Wilkie et al are successfully using this cost-effective approach to data collection for another palliative care study [48] focused on African American caregivers and the data manager is well versed with REDCap and current mobile devices. Wilkie et al is also successfully using mobile tablet or pen devices in a study of hospice patients in their homes and replicating prior findings of the clear feasibility of elderly patients with cancer near the end of their lives using the tablet technology, especially with use of a smart pen device [49]. We expect that patients will take 30 to 45 minutes to complete the measures at pretest and 15 to 20 minutes at posttest.

The primary study outcome is dignity impact with all study measures. The secondary outcomes are existential tasks and cancer prognosis awareness.

Our primary outcome measure is a 7-item Dignity Impact Scale [5]. We took the items from the DT Patient Feedback Questionnaire that has been used in a number of studies of DT [50]. The 7 items addressed the concept of DT and showed significant differences between those who received DT and those who did not in an RCT [1]. The items have been used in many DT studies with evidence of their validity in the target population. We modified the wording of the items to fit our pretest and posttest study design. A sample item is “The care I received during the past month has increased my sense of dignity.” The items are scored on a 5-point scale from “strongly disagree” (1) to “strongly agree” (5). The Cronbach α from a preliminary study was .85 [5]. Scores ranged from 7 to 35 in the DT group and 7 to 29 in the usual care group and showed sensitivity to DT effects at posttest with an effect size of 0.7 [5]; it was not measured at pretest, but we will do so to increase power and potential for inferences.

The existential tasks of importance in the proposed study are preparation for death and life completion. Our measures of preparation and completion are taken from the QUAL-E, a measure designed to evaluate the quality of life at the end of life and to assess the effectiveness of interventions targeted to improve the quality of life at the end of life [46]. The 4-item Preparation subscale assesses an individual’s sense of integrity and concerns about being a burden to significant others. A sample item is “I have regrets about the way I have lived my life.” The 7-item Completion subscale assesses an individual’s sense of meaning and peace, as well as any unfinished interpersonal business. A sample item is “I have been able to share important things with my family.” The items are rated on a 5-point scale from “not at all true for me” (1) to “completely true for me” (5). The items in the Preparation subscale are reverse scored. The Preparation and Completion scales have demonstrated good reliability and validity [46]. Validity of both measures was shown by correlations in expected directions with other measures of quality of life (FACT-G; Missoula-Vitas QOL) and spiritual well-being (FACIT-Sp). In a diverse sample of 248 patients near the end of life, the Cronbach alphas were .68 and .80 for Preparation and Completion, respectively. The 1-week ICCs for participants with no changes in health status were .73 for Preparation and .72 for Completion.

Two measures will provide data for peaceful awareness and treatment preference. Our approach to measuring terminal illness awareness (peaceful awareness) is taken from a study of 280 patients with advanced cancer who participated in the Coping with Cancer Study [51]. In that study, 17.5% of the sample reported being both peaceful and aware of their prognosis. Peacefully aware patients had lower rates of psychological distress and higher rates of advance care planning than those who were not peacefully aware. Peaceful awareness is also associated with modifiable aspects of medical care (eg, discussions about terminal treatment preferences).

Treatment preferences will be measured with a standardized and validated Hypothetical Advanced Care Planning Scenario that assesses scenario-based goals of care and treatment preferences. This approach, based on work by MPI LE [52], was used successfully in the study of DT in patients with advanced colorectal cancer [6]. Patients will be considered as making “life prolonging” treatment choices if they selected “I want” or “I want treatment tried. If no clear improvement, stop (only for mechanical ventilation)” for either CPR or mechanical breathing. Patients will be considered as making “non-life prolonging” treatment choices if they selected “I do not want” to both CPR and mechanical breathing. Patients selecting “I am undecided” to either CPR or mechanical ventilation and not considered “life-prolonging” in their treatment choices will be categorized as “undecided.”

We will measure 2 indicators of palliative care spiritual service processes. They are patient satisfaction with spiritual care and unmet spiritual needs.

The items in our 7-item measure of Patient Satisfaction with Nurse or Chaplain Care are taken from the work of 2 leaders in chaplaincy research. VandeCreek [53] developed the 23 item Patient Satisfaction Instrument-Chaplaincy (PSI-C) that he administered to a sample of 1440 patients who had been treated in one of 14 different US hospitals. In their study of patient satisfaction, Flannelly et al [54] used 7 items from VandeCreek’s PSI-C as well as 7 new items they developed. In their study, with 250 patients at 1 hospital, the validity of their items was supported by positive correlations with other items that assessed patients’ perception that their spiritual and emotional needs had been met. Following the work of Flannelly et al [54], we will assess 2 aspects of satisfaction; satisfaction with the process of nurse or chaplain care and satisfaction with the impact of the chaplains’ care.

As clinicians embraced an evidence-based approach to spiritual care, 1 of the challenges they faced is developing instruments that can be used for spiritual assessment [55].

Data management and preliminary data analysis procedures will be supervised by MPI DW and conducted by Dr. YY, the Co-I, and statistician, using statistical software R version 3.5.2 (R Foundation for Statistical Computing). Data will be stored in a REDCap database and will be exported to R. In the case of missing data, multiple imputations will be used to generate multiple complete datasets on which statistical inference will be performed and then aggregated. Missing at random assumption will be assessed and if necessary sensitivity analysis will be performed using pattern mixture methods. We will consider a P value less than .05 as statistically significant.

Descriptive statistics (ranges, frequencies, means, and SDs) and graphic summary (box plots, histograms, bivariate scatterplots, etc) will be first generated for both patient covariates and outcome measures. We will check patient characteristics to see if there is notable imbalance between the 3 arms, as well as whether there are significant variations between sites or over time. Descriptive statistics of patient outcomes and process data will reveal patients’ spiritual state, identify potential areas for improvement in nurse and chaplain service, and provide information on nurses’ and chaplains’ workload both in usual care and when dignity therapy is added and on family related variables (eg, legacy document disposition).

This rigorous trial of DT will constitute a landmark step in palliative care and spiritual health services research. We designed this study mindful of a number of threats to study validity as we implement the study within the workflow of outpatient palliative care in 6 sites and present plans to achieve study aims.

Anticipated numbers of potential participants are based on prior studies among similar patient populations both in in-patient and out-patient settings. However, recruitment is always difficult, especially in palliative care settings. We recruited 6 sites to ensure a generous pool of eligible patients and to use a stepped-wedge design for efficiency in the required sample size. Also, the stepped-wedge design allows flexibility in that, if recruitment proves difficult as proposed, we can add steps and sites to ensure a sufficient sample.

We include 6 sites and will recruit participants from outpatient services only. We carefully selected sites with consideration of geographic and population diversity, but our analytic model will include random effect terms to account for site differences, should they occur.

Attrition due to illness burden and death is inevitable in this population. We will monitor attrition carefully throughout the study. If the estimates we have used prove inaccurate we will adjust recruitment numbers accordingly to meet benchmarks throughout the study at each site. We will also analyze data from participants who did not complete the DT to check for differences in patient characteristics and to verify that there is no association between attrition and intervention group assignment.

Including these patients limits potential loss of patients before they can complete the intervention, for instance, during the interval between the first interview and returning the legacy document. Patients’ will be located through the doctor’s office or family; those in the hospital at posttest will finish their intervention at the hospital or by phone with electronic or express mail delivery of the legacy document and will still be included in the study. Including these patients will minimize any potential bias. Furthermore, it ensures that the intervention effect size we estimate will reflect the average effect on a realistic patient population, a portion of which will decline in health status before the intervention is complete.

We will collect preintervention data using self-report on a tablet with a smart pen. Participant operation of the device may limit data entry, but we will use Dr. Wilkie’s extensive expertise in using tablets in the homes of elderly hospice patients with cancer. The RA will be present to assist with data collection if needed. We will track incomplete data patterns and raise the topic at regular study meetings to brainstorm solutions such as an improved user interface or directions. We will collect postintervention data by phone or in-person, if the participant has a clinic visit. Obtaining correct phone information in an era when landlines are rare and patients may not be next to their cell phones may be challenging. We will ask for a backup phone number for this reason. Survey responses by phone may be particularly challenging for our sick population; our training will include protocols for taking a break and repeating questions that allow for the illness burden of our participants.

Each site has a chaplain available to the palliative care service, but others are available to cover for absences (vacation, illness, etc). Similarly, nurses are available. For each site, at least two interventionists will be trained and immediately replaced by another should there be an absence, resignation, or retirement.

We have exceptionally strong research and clinical environments to support the proposed study, and a strong, collaborative, interdisciplinary team distinguished by its rare combination of chaplaincy, nursing, and palliative care researchers. We are noted for a history of excellence in palliative care research in elderly cancer patients, including a Templeton Foundation grant to stimulate chaplaincy research in palliative care and several funded R01 level studies focused on palliative care populations. We propose a highly significant and high-quality study in which we will apply rigorous science to an area that sorely needs it: spiritual care research. Studying 560 participants in a pre- and posttest, randomized, controlled 4-step, stepped-wedge design, we will compare the effects of usual outpatient palliative care against the same plus either nurse-led or chaplain-led DT on patient outcomes (dignity impact, existential tasks, and cancer prognosis awareness). Using a multilevel analysis with site, provider (nurse and chaplain), and time (step) in the model, we will determine the efficacy and mechanism of DT when delivered by nurses or chaplains as a spiritual care therapy. Success in this landmark study will yield the first manualized intervention for chaplaincy services, its potential efficacy compared with nurse-led DT, and insights into its mechanisms of action related to spiritual care, an area of great importance to elderly cancer patients receiving palliative care. We will disseminate study findings in a variety of venues for presentation and publication to reach palliative care, oncology, gerontology, chaplaincy, nursing, and other audiences.