Feasibility and Acceptability of a Nurse-Led Telehealth Intervention (BOLSTER) to Support Patients with Peritoneal Carcinomatosis and their Caregivers: A Pilot Randomized Clinical Trial

Abstract

Objective:

Patients with advanced gynecologic (GYN) and gastrointestinal (GI) cancers frequently develop peritoneal carcinomatosis (PC), which limits prognosis and diminishes health-related quality of life (HRQoL). Palliative procedures may improve PC symptoms, yet patients and caregivers report feeling unprepared to manage ostomies, catheters, and other complex needs. Our objectives were to (1) assess the feasibility of an efficacy trial of a nurse-led telehealth intervention (BOLSTER) for patients with PC and their caregivers; and (2) assess BOLSTER’s acceptability, potential to improve patients’ HRQoL and self-efficacy, and potential impact on advance care planning (ACP).

Methods:

Pilot feasibility RCT. Recently hospitalized adults with advanced GYN and GI cancers, PC, and a new complex care need and their caregivers were randomized 1:1 to BOLSTER or enhanced discharge planning (EDP). We defined feasibility as a ≥50% approach-to-consent ratio and acceptability as ≥70% satisfaction with BOLSTER. We assessed patients’ HRQoL and self-efficacy at baseline and six weeks, then compared the proportion experiencing meaningful improvements by arm. ACP documentation was identified using natural language processing.

Results:

We consented 77% of approached patients. In the BOLSTER arm, 91.0% of patients and 100.0% of caregivers were satisfied. Compared to EDP, more patients receiving BOLSTER experienced improvements in HRQoL (68.4% vs. 40.0%) and self-efficacy for managing symptoms (78.9% vs. 35.0%) and treatment (52.9% vs. 42.9%). The BOLSTER arm had more ACP documentation.

Conclusions:

BOLSTER is a feasible and acceptable intervention with the potential to improve patients’ HRQoL and promote ACP. An efficacy trial comparing BOLSTER to usual care is underway.

Trial Registration: ClinicalTrials.gov: NCT03367247; PI: Wright.

BACKGROUND

Peritoneal carcinomatosis (PC) is a frequent complication of advanced gynecologic and gastrointestinal cancers.¹ Patients with PC experience bowel, lymphatic, or ureteral obstructions that lead to recurrent ascites, nausea, vomiting, pain, and obstipation.² Palliative procedures such as ostomy formation and indwelling catheter placement may relieve symptoms,^3,4 yet the management of these devices is complex. PC is a poor prognostic indicator,^5,6 and patients with PC and their caregivers have described struggling to adapt to new complex care needs in the setting of disease progression and a dearth of anticipatory guidance.⁷

We developed the Building Out Lifelines for Safety, Trust, Empowerment and Renewal(BOLSTER) intervention to support patients with PC and a new complex care need and their family caregivers.⁸ BOLSTER is a nurse-led telehealth intervention that aims to improve health-related quality of life (HRQoL) in patients with PC by fostering self-efficacy and providing longitudinal nursing support and tailored educational tools. Consistent with Phase I of the ORBIT Model for Behavioral Intervention Development,⁹ we began by adapting an existing home care intervention for patients with cancer.¹⁰ Next, we refined BOLSTER in an open pilot study.⁸ In this manuscript, we describe our research activities and findings during Phase II of the ORBIT Model, which assesses the potential for a fixed treatment package to produce a clinically meaningful improvement in one or more health outcomes and the feasibility of a Phase III efficacy trial.⁹

The primary objectives of this study were to assess the feasibility of conducting a Phase III efficacy trial of BOLSTER and determine the extent to which BOLSTER is acceptable to patients and their family caregivers. The secondary objective was to compare the proportion of patients experiencing clinically meaningful improvements in HRQoL and self-efficacy from baseline to follow-up across study arms. Our exploratory objective was to inform secondary outcome selection for a Phase III trial by comparing changes in patients’ and caregivers’ anxiety and depression symptoms; patients’ experiences with health care services, healthcare utilization, and advance care planning (ACP); and family caregivers’ health status, caregiving burden, and preparedness from baseline to follow-up across study arms.

METHODS

Study Design

We conducted a pilot two-arm, parallel-group, randomized clinical trial with a 1:1 allocation ratio (ClinicalTrials.gov: NCT03367247; PI: Wright). The Consolidated Standards of Reporting Trials (CONSORT) extension to randomized pilot and feasibility trials informed the preparation of this manuscript.¹¹

Participants and Setting

Eligible participants included English-speaking adults with PC and a new complex care need (e.g., an ostomy or ileostomy, venting gastric tube, PleurX catheter, or percutaneous nephrostomy tubes) and their caregivers. We initially enrolled patients with gynecologic cancers. One year into the study, we expanded eligibility criteria to include patients with gastrointestinal cancers at the request of clinicians and in anticipation of a larger trial. Patients were eligible for recruitment between hospitalization and their second post-discharge outpatient visit. Patients were ineligible if they planned to be discharged to rehabilitation or hospice.

Participants were identified at Dana-Farber Cancer Institute and Brigham and Women’s Hospital using an electronic health record (EHR) dashboard. With oncologists’ permission, we approached patients and their caregivers in the hospital, in the outpatient clinic, or via telephone. Patients were asked to identify a family caregiver or friend involved in their care at the time of enrollment; however, caregivers were not required to enroll. All participants provided written informed consent. Study procedures were approved by the DFCI IRB (#17–475).

Interventions

Participants randomized to BOLSTER received six visits over four weeks with a bachelor’s-prepared nurse, augmented by a smartphone-based remote symptom monitoring application that used items from the Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE).¹² Nurse visits primarily took place over HIPAA-compliant Zoom, but patients had the option of meeting the nurse in the clinic or by phone. Caregivers were encouraged to attend each visit. Each visit followed a routine that included agenda-setting, assessing symptoms, providing tailored symptom education and skills training, and goal-setting. Specific visit topics are detailed in Table 1. Following each visit, the nurse sent patients and caregivers tailored educational materials through the smartphone application, e-mailed an update to the patient’s oncologist, and documented the visit in the EHR. Between visits, patients received daily prompts to report their symptoms using the smartphone application. The application advised patients with severe symptoms, defined as a grade 3 symptom on the PRO-CTCAE, to contact their oncologist.

Table 1.

BOLSTER visit schema

Week	Visit	Topic(s)
1	1	Introduction Discharge education and coordination
	2	Goal setting Symptom management Skills training
2	3	Goal setting Review symptom management and skills training, if needed Coping, stress, and relaxation
	4	Goal setting Review symptom management and skills training, if needed Advance care planning Optional: Fatigue, nutrition, relaxation
3	5	Goal setting Debrief on advance care planning Optional: Fatigue, nutrition, relaxation
4	6	Coordination and closure

Participants randomized to the Enhanced Discharge Planning arm received standard of care discharge planning, a single visit with the nurse, and written educational materials related to the reason they were hospitalized. The single visit focused on discharge education and care coordination. Following the visit, the nurse e-mailed an update to the patient’s oncologist and documented the visit in the EHR.

Data Collection

Participants completed assessments prior to randomization and six weeks after enrollment. Participants self-reported their age, gender, race, ethnicity, marital status, educational attainment, and employment status. Caregivers self-reported their relationship to the patient, caregiving role, and time spent caregiving. Twelve weeks after enrollment, we extracted information related to patients’ diagnoses, treatments, complex care needs, healthcare utilization, and ACP from the EHR.

Primary Outcomes

Feasibility and acceptability benchmarks were defined a priori to judge whether to proceed with a future definitive trial. To ensure we would have adequate resources to conduct a full-scale efficacy trial, we defined feasibility as a ≥50% consent-to-approach ratio. To establish buy-in among this highly burdened patient population, we defined acceptability as ≥70% of participants in the intervention arm agreeing they would recommend BOLSTER to other patients and were satisfied with BOLSTER. We elicited feedback on BOLSTER and Enhanced Discharge Planning in optional debriefing interviews, the details of which have been reported.⁷ Debriefing interviews were audio recorded, de-identified, and professionally transcribed.

Secondary Outcomes

We assessed health-related quality of life (HRQoL) using the Functional Assessment of Cancer Therapy-General (FACT-G).¹³ The FACT-G comprises 27 items across four subscales: physical well-being, emotional well-being, social and family well-being, and functional well-being. Higher scores represent better HRQoL. Consistent with a meta-analysis of FACT-G effect sizes,¹⁴ we defined a clinically meaningful improvement as an increase of at least 9 points for the total scale, 4.1 points for physical well-being, 1.9 points for emotional well-being, 0.8 points for social and family well-being, and 3.8 points for functional well-being.

The PROMIS Self-Efficacy short forms assess self-efficacy for managing medications and treatments and for managing symptoms.¹⁵ Scores on each four-item measure range from 0 to 100; higher scores indicate greater self-efficacy. Prior literature suggests a clinically meaningful difference across PROMIS measures ranges from 2–6 t-score points;¹⁶ we defined a clinically meaningful improvement as an increase of at least 4 t-score points.

Exploratory Outcomes

Patients and caregivers completed the Hospital Anxiety and Depression Scale (HADS), which comprises two seven-item subscales, one for anxiety symptoms and one for depressive symptoms.¹⁷ Scores range from 0–21 for each subscale; higher scores indicate higher symptom burden. We defined a borderline case as a score of 8–10 and a case as a score >10 on either subscale.¹⁷

Patients’ care experiences were assessed with the CAHPS Cancer Care Survey, which assesses how well the cancer care team communicates with patients, uses information to coordinate patient care, supports patients in managing the effects of their cancer and treatment, is available to provide information when needed, and involves family members and friends in care discussions.¹⁸ Higher scores indicate a better care experience.

To assess patients’ healthcare utilization, we tallied the number of invasive procedures experienced by each patient during the 12-week enrollment period. We also calculated the number of days spent at home by subtracting the number of days spent in an emergency department or hospital from the 84 days of study participation.

To assess ACP, we used rule-based natural language processing (NLP) to identify documentation of goals of care conversations, hospice discussions, subspecialty palliative care involvement, and decisions to limit life-sustaining treatment in outpatient and inpatient clinical notes written during the enrollment period. NLP allows investigators to analyze large amounts of unstructured text more quickly than manual chart review with equal accuracy.¹⁹ We used the text annotation software Clinical Regex²⁰ to review clinical notes for keywords associated with each of the four ACP outcomes. Keyword libraries were previously developed and validated by a member of our study team (CL) to identify ACP documentation in our health system’s EHR.²¹ Two investigators (RAP, KS) used a pre-established codebook to verify whether each instance of a keyword represented the ACP outcome of interest.

Caregivers completed the Short Form 12-Item Survey (SF-12), a measure of physical and mental health status.²² Mental and physical summary scores range from 0 to 100; higher scores indicate better functioning.

We assessed caregiver burden using the Caregiver Reaction Assessment (CRA), a 24-item measure of the impact of caregiving on a person’s self-esteem, finances, family support, and schedule.²³ Totals for each of these four subscales range from 1–5. Higher scores indicate a more positive impact on self-esteem and a more negative impact on finances, family support, and schedule.

Finally, the Caregiver Preparedness Scale measures how confident the caregiver is that he or she can meet the patient’s needs.²⁴ Consistent with prior research,²⁵ we administered the first four items of the scale. Response options range from 0 (“not at all confident”) to 4 (“extremely confident”); higher scores represent greater confidence.

Sample Size

Consistent with published guidance for pilot studies of interventions with a small anticipated effect size,²⁶ we aimed to recruit at least 25 patients per arm for a total sample size of 50 patients. There were no planned interim analyses or stopping guidelines.

Randomization

Patients were randomized to BOLSTER or Enhanced Discharge Planning (EDP) using a computer-generated random allocation sequence with a block size of four. A statistician who was not involved in the daily operations of the study generated group assignments and placed them in sequentially numbered and sealed opaque envelopes. Trained research coordinators enrolled participants, collected baseline data, and opened the sealed envelope after data collection.

Analysis

We summarized enrollment data, participant characteristics, acceptability ratings, healthcare utilization, and ACP using descriptive statistics and frequency distributions. Four members of the study team (RAP, JAW, AT, AAW) analyzed debriefing interview transcripts using conventional content analysis;²⁷ we resolved coding discrepancies through discussion.⁷ We calculated the proportions of patients experiencing clinically meaningful improvements in HRQoL and self-efficacy in each arm using the above-referenced cutoff scores. Finally, we calculated the mean change in anxiety and depression scores, patients’ care experiences, and caregivers’ health status, caregiving burden, and preparedness across arms. Given our small sample size and the exploratory nature of these outcomes, we did not conduct formal hypothesis testing. Data were analyzed using IBM SPSS 28.

RESULTS

Recruitment took place from December 2020 - May 2022, and follow-up data were collected from March 2021 - August 2022. The baseline characteristics of our sample have been previously described⁷ and are presented by study arm in Table 2. Briefly, patients were predominantly women with gynecologic cancers, while family caregivers were predominantly male spouses or partners of patients. Most patients and caregivers were college-educated, white, and non-Hispanic.

Table 2.

Participant characteristics by study arm

	Patients		Caregivers
	EDP	BOLSTER	EDP	BOLSTER
Demographic characteristics	(n = 32)	(n = 29)	(n = 22)	(n = 17)
Age, years, mean (SD)	61.9 (7.5)	60.1 (10.2)	60.4 (10.2)	59.3 (9.5)
Gender, n (%)
Female	31 (96.9)	25 (86.2)	7 (31.8)	7 (41.2)
Male	1 (3.1)	4 (13.8)	15 (68.2)	9 (52.9)
Not reported	0 (0.0)	0 (0.0)	0 (0.0)	1 (5.9)
Race, n (%)
White	26 (81.3)	24 (82.8)	20 (90.9)	14 (82.4)
Black or African American	3 (9.4)	0 (0.0)	2 (9.1)	0 (0.0)
Asian	0 (0.0)	2 (6.9)	0 (0.0)	1 (5.9)
Other	3 (9.4)	2 (6.9)	0 (0.0)	1 (5.9)
Not reported	0 (0.0)	1 (3.4)	0 (0.0)	1 (5.9)
Ethnicity, n (%)
Hispanic or Latinx	2 (6.3)	1 (3.4)	1 (4.5)	0 (0.0)
Not Hispanic or Latinx	27 (84.4)	27 (93.1)	21 (95.5)	16 (94.1)
Not reported	3 (9.4)	1 (3.4)	0 (0.0)	1 (5.9)
Education, n (%)
Advanced degree	14 (43.8)	14 (48.3)	7 (31.8)	11 (64.7)
College graduate	11 (34.4)	11 (37.9)	8 (36.4)	3 (17.6)
Some college or technical school	2 (6.3)	2 (6.8)	6 (27.3)	1 (5.9)
High school or less	4 (12.5)	1 (3.4)	1 (4.5)	2 (11.8)
Not reported	1 (3.1)	1 (3.4)	0 (0.0)	0 (0.0)
Relationship status, n (%)
Married or partnered	27 (84.4)	23 (79.3)	20 (90.9)	16 (94.1)
Divorced	2 (6.3)	2 (6.9)	2 (9.1)	0 (0.0)
Single, never married, or widowed	2 (6.3)	3 (10.3)	0 (0.0)	0 (0.0)
Not reported	1 (3.1)	1 (3.4)	0 (0.0)	1 (5.9)
Employment status, n (%)
Full or part-time work	8 (25.0)	9 (31.0)	13 (59.1)	10 (58.8)
Retired	13 (40.6)	11 (37.9)	7 (31.8)	5 (29.4)
Disabled	5 (15.6)	4 (13.8)	0 (0.0)	0 (0.0)
Unemployed	3 (9.4)	1 (3.6)	2 (9.1)	1 (5.9)
Other	3 (9.4)	3 (10.3)	0 (0.0)	1 (5.9)
Not reported	0 (0.0)	1 (3.4)	0 (0.0)	0 (0.0)
Relationship to patient, n (%)
Spouse or romantic partner			15 (68.2)	13 (76.5)
Child			3 (13.6)	2 (11.8)
Parent			2 (9.1)	0 (0.0)
Friend			1 (4.5)	1 (5.9)
Sibling or sibling-in-law			1 (4.5)	1 (5.9)
Clinical characteristics
Cancer diagnosis, n (%)
Ovarian / fallopian tube / peritoneal	22 (68.9)	20 (68.9)
Endometrial	4 (12.5)	3 (10.3)
Cervical	2 (6.3)	1 (3.4)
Colorectal	3 (9.3)	0 (0.0)
Esophageal	1 (3.1)	2 (6.9)
Pancreatic	0 (0.0)	3 (10.3)
Complex care need, n (%)
Colostomy or ileostomy	8 (25.0)	11 (37.9)
Total parenteral nutrition	11 (34.4)	9 (31.0)
Percutaneous nephrostomy	5 (15.6)	4 (13.8)
Venting gastric tube	3 (9.4)	3 (10.3)
Jejunostomy feeding tube	1 (3.1)	2 (6.9)
PleurX catheter	3 (9.4)	2 (6.9)
Indwelling drain	2 (6.3)	1 (3.4)
Urostomy	1 (3.1)	1 (3.4)
Vacuum-assisted closure of wound	1 (3.1)	0 (0.0)
More than one need (included in count above)	3 (9.4)	4 (13.8)
Recurrence status, n (%)
No recurrence	13 (40.6)	11 (37.9)
Recurrence status	14 (43.8)	14 (48.3)
Progression during primary treatment	5 (156)	4 (138)

Abbreviations: EDP = Enhanced discharge planning

Feasibility of BOLSTER

We approached 95 patients, 73 (76.8%) of whom consented. Of these, eight (11%) withdrew or were excluded prior to randomization (three enrolled in hospice, three did not complete the baseline survey, and two withdrew). Sixty-five patients were randomized (32 to BOLSTER and 33 to EDP). Of these, 41/65 (63.1%) completed six-week outcome measures. The most common reasons for attrition after randomization were hospice / death (9/24, 37.5%) and illness (8/24, 33.3%). Three patients in the BOLSTER arm who were too ill to complete six-week outcome measures responded to the acceptability surveys. A CONSORT flow diagram is provided in Figure 1.

Figure 1.

CONSORT 2010 flow diagram

Acceptability of BOLSTER

Twenty-one of 22 (95.5%) patients and 10/10 (100%) caregivers in the BOLSTER arm agreed they would recommend BOLSTER to other patients. Twenty of 22 (90.9%) patients and 10/10 (100%) caregivers in the BOLSTER arm were satisfied with BOLSTER. In debriefing interviews, participants in the BOLSTER arm found the nurse visits and materials helpful. Patients and caregivers appreciated the anticipatory guidance and personalized support they received from the nurse and spoke to the need for caregiver resources. Participants valued the flexible timing, content, and modality of the visits. Some patients recommended limiting the BOLSTER visits to 30 minutes. In contrast, participants in the EDP arm rarely recalled participating in the single visit with the nurse and had difficulty differentiating it from the other visits that followed discharge.

HRQoL and Self-Efficacy

As depicted in Table 3, the proportion of patients experiencing clinically meaningful improvements in overall HRQoL, physical well-being, functional well-being, self-efficacy for medications and treatment, and self-efficacy for managing symptoms was numerically higher in the BOLSTER arm than in the EDP arm. Changes in social and family well-being and emotional well-being were similar across arms.

Table 3.

Proportions of patients experiencing clinically meaningful improvements in outcomes

	Total (n = 41)		EDP (n = 22)		BOLSTER (n = 19)
	n	(%)	n	(%)	n	(%)
Clinically meaningful improvement in FACT-G scores
Overall health-related quality of life	21/39a	(53.8)	8/20	(40.0)	13/19	(68.4)
Physical well-being	25/39	(64.1)	10/21	(47.6)	15/18	(83.3)
Social and family well-being	11/41	(26.8)	6/22	(27.3)	5/19	(26.3)
Emotional well-being	13/41	(31.7)	7/22	(31.8)	6/19	(31.6)
Functional well-being	18/40	(45.0)	8/21	(38.1)	10/19	(52.6)
Clinically meaningful improvement in PROMIS scores
Self-efficacy for medications and treatment	18/38	(47.4)	9/21	(42.9)	9/17	(52.9)
Self-efficacy for symptoms	22/39	(56.4)	7/20	(35.0)	15/19	(78.9)

Abbreviations: EDP = Enhanced discharge planning

^aDenominators are lower than total sample size when missing data precluded scale or subscale calculation.

Exploratory Outcomes

Exploratory outcomes are detailed in Table 4. Patients in both arms experienced small reductions in anxiety and depressive symptoms from baseline to follow-up. Patients in the BOLSTER arm reported greater mean improvements in care coordination, symptom management, clinician availability, and communication with the health care team. Patients in the EDP arm reported an increase in caregiver involvement over time; in the BOLSTER arm, caregiver involvement did not change.

Table 4.

Exploratory outcomes by study arm

	Total		EDP		BOLSTER
Mean change scores	m	(SD)	m	(SD)	m	(SD)
Hospital Anxiety and Depression Scale a
Change in patients’ depression	−0.90	(4.24)	0.62	(3.28)	−2.58	(4.62)
Change in patients’ anxiety	−1.64	(3.29)	−0.95	(3.12)	−2.37	(3.39)
Change in caregivers’ depression	0.65	(2.31)	0.85	(1.99)	0.40	(2.76)
Change in caregivers’ anxiety	0.09	(3.19)	−0.38	(3.25)	0.70	(3.16)
CAHPS Patient Experience Measures b
Change in care coordination	1.46	(18.57)	−5.56	(18.52)	8.48	(16.18)
Change in symptom management	10.53	(24.68)	9.04	(27.93)	12.03	(21.61)
Change in clinician availability	5.70	(19.48)	4.39	(18.29)	7.02	(21.02)
Change in caregiver involvement	4.05	(34.11)	7.89	(38.24)	0.00	(29.70)
Change in communication	1.97	(12.91)	0.44	(11.27)	3.51	(14.52)
Short Form-12 b
Change in caregivers’ physical health	−1.87	(4.87)	−2.03	(5.00)	−1.67	(5.03)
Change in caregivers’ mental health	−1.49	(10.22)	−2.68	(12.91)	−0.01	(5.91)
Caregiver Reaction Assessment
Change in self-esteemb	0.06	(0.39)	−0.05	(0.30)	0.20	(0.45)
Change in lack of family supporta	0.02	(0.55)	0.15	(0.54)	−0.18	(0.52)
Change in impact on financesa	0.00	(0.55)	0.17	(0.54)	−0.22	(0.50)
Change in impact on schedulea	0.27	(0.57)	0.38	(0.69)	0.12	(0.34)
Change in impact on healtha	0.20	(0.65)	0.29	(0.77)	0.08	(0.47)
Caregiver Preparedness Scale b
Change in caregiver preparedness	0.34	(0.70)	0.35	(0.63)	0.33	(0.82)
	n	(%)	n	(%)	n	(%)
Healthcare utilization
Procedures during enrollment	0.9	(1.0)	1.1	(1.0)	0.7	(1.0)
Days at home	67.1	(19.0)	65.9	(20.5)	68.4	(17.4)
Advance care planning
Goals of care conversation	33	(54.1)	13	(40.6)	20	(69.0)
Hospice discussion	17	(27.9)	7	(21.9)	10	(34.5)
Palliative care involvement	29	(47.5)	14	(43.8)	15	(51.7)
Limitation of life-sustaining treatment	13	(21.3)	7	(21.9)	6	(20.7)

^aLower scores are more desirable

^bHigher scores are more desirable

Abbreviations: EDP = Enhanced discharge planning, EHR = electronic health record

The proportions of patients with documented goals of care conversations, hospice discussions, and palliative care consultations were numerically higher in the BOLSTER arm than the EDP arm. A higher proportion of patients in the EDP arm had documented limitations on life-sustaining treatment. However, patients in the BOLSTER arm underwent fewer invasive procedures and spent approximately three additional days at home.

There were few differences between caregivers’ anxiety and depression symptoms from baseline to follow-up, or by study arm. Caregivers in both arms experienced a decline in physical and mental health from baseline to follow-up; average declines were smaller in the BOLSTER arm than in the EDP arm. The average impact of caregiving on caregivers’ self-esteem, lack of family support, and finances improved in the BOLSTER arm and worsened in the EDP arm. The average impact of caregiving on caregivers’ schedules and health worsened in both arms; declines were smaller in the BOLSTER arm than in the EDP arm. Caregiver preparedness improved to a roughly equal extent in both groups.

DISCUSSION

In this study, we found that BOLSTER was feasible and acceptable for patients with PC-related complex care needs and their family caregivers. Although this pilot study was underpowered and not designed to test the significance of between-arm differences in study outcomes, a higher proportion of patients in the BOLSTER arm experienced clinically meaningful improvements in overall HRQoL, physical well-being, functional well-being, and self-efficacy compared with a control arm. Based on this preliminary evidence, we plan to assess the impact of BOLSTER on HRQoL and test the role of self-efficacy as a mediator of this effect in a full-scale efficacy trial.

Our findings suggest BOLSTER may positively impact patients’ perceptions of care coordination, symptom management, and clinician availability and communication. In our formative research, patients with PC and their family caregivers described needing support to navigate the healthcare system and safely manage patients’ needs at home.^7,8 Accordingly, a major focus of the BOLSTER intervention was to support care coordination and symptom management through longitudinal contact with the BOLSTER nurse. Prior studies have shown navigation can improve HRQoL and patient satisfaction during cancer treatment and survivorship.^28,29 Our study suggests it may also be possible to improve HRQOL in patients with serious illness, complex care needs, and a high competing risk of death.

We saw a promising signal that BOLSTER may enhance illness understanding and facilitate ACP. At baseline, fewer than 40% of the patients in our sample acknowledged the terminal nature of their illness, and fewer than 12% had discussed the type of care they would want to receive if they were dying with their clinicians.⁷ At follow-up, the proportion of patients with documented goals of care conversations, hospice discussions, and palliative care involvement was higher in the BOLSTER arm than in the control arm. Nevertheless, our findings need to be interpreted with caution as illness understanding can be difficult to change and prior interventions have had limited impact.³⁰ Similarly, our exploratory analyses of anxiety and depressive symptoms in patients and caregivers produced mixed results. Interventions that aim to reduce symptoms of anxiety and depression may need to explicitly incorporate evidence-based approaches such as cognitive behavioral therapy and acceptance and commitment therapy to achieve a consistent effect.³¹

In the BOLSTER arm, patients’ perceptions of caregivers’ involvement in their care did not change from baseline to follow-up. Indeed, we found it difficult to engage family caregivers in the BOLSTER visits. Caregivers may have viewed the patients’ visits with the BOLSTER nurse as a respite or an opportunity to tend to other pressing tasks. Alternatively, caregivers may not have felt that BOLSTER addressed their unique needs. Prior interventions that have enrolled patients and caregivers together have failed to improve caregiver burden and mood, suggesting it may be difficult to address both patient and caregiver needs in a single intervention.³² Prior to a full-scale efficacy trial, we plan to expand the caregiver-facing elements of BOLSTER to ensure the intervention adds value for caregivers. Given that participants in the EDP arm often did not recall their single visit with the nurse, we will forgo the resource-intensive EDP visits in the full-scale efficacy trial and compare BOLSTER to usual care.

In this pilot feasibility study, we did not assess the impact of BOLSTER on clinician workload, healthcare costs, or facility resources. Follow-up interviews with clinicians whose patients participate in the full scale-efficacy trial will identify barriers to and facilitators of implementing BOLSTER in the clinical setting. Consistent with the ORBIT model for behavioral intervention development,⁹ if the results of the efficacy trial suggest BOLSTER is efficacious, we will assess implementation outcomes in an effectiveness trial.

LIMITATIONS

Our findings should be interpreted with caution, given several limitations. First, the study design and small sample size limit our ability to make causal inferences. Second, our sample was predominantly white, non-Hispanic, and college-educated, limiting the generalizability of our findings. In preparation for a Phase III trial, we are culturally adapting and translating BOLSTER into Spanish and plan to recruit from sites that serve more diverse patients and caregivers. Finally, while we achieved our target approach-to-consent ratio, we lost approximately one-third of our sample to attrition, consistent with prior trials.³³ Our experiences underscore the importance of conducting pilot feasibility trials to plan efficacy studies. The sample size calculations for our full-scale efficacy trial of BOLSTER have incorporated our estimated attrition rate. Notably, we cannot assess the extent to which missing data from patients who were more highly burdened by symptoms or nearer to the end of life may have biased our results. Our experience is common across behavioral trials in patients with life-limiting illness, in which attrition due to death or illness is the leading cause of missing data.³⁴

CONCLUSION

BOLSTER was feasible and acceptable and showed potential signs of efficacy. We will assess the effect of BOLSTER on HRQoL in patients with PC and their caregivers in a Phase III trial.

Supplementary Material

2

Supplement 1. Mean baseline and follow-up scores by study arm

HIGHLIGHTS.

• An efficacy trial of a nurse-led telehealth intervention (BOLSTER) for patients with peritoneal carcinomatosis is feasible.

• Patients and caregivers were satisfied with BOLSTER and said they would recommend it to others.

• Patients in the BOLSTER arm experienced clinically meaningful improvements in quality of life and self-efficacy.

• Patients in the BOLSTER arm had more advance care planning documentation than those in the control arm.