The Head and Neck Survivorship Tool (HN-STAR) Trial (WF-1805CD): A protocol for a cluster-randomized, hybrid effectiveness-implementation, pragmatic trial to improve the follow-up care of head and neck cancer survivors.

Abstract

Survivors of head and neck cancer (HNC) can have multiple health concerns. To facilitate their care, we developed and pilot-tested a clinical informatics intervention, HN-STAR. HN-STAR elicits concerns online from HNC survivors prior to a routine oncology clinic visit. HN-STAR then presents tailored evidence-based clinical recommendations as a clinical decision support tool to be used during the visit where the oncology clinician and survivor select symptom management strategies and other actions. This generates a survivorship care plan (SCP). Online elicitation of health concerns occurs 3, 6, and 9 months after the clinic visit, generating an updated SCP each time. HN-STAR encompasses important methods of improving survivorship care (e.g., needs assessment, tailored interventions, dissemination of guidelines) and will be evaluated in a pragmatic trial to maximize external validity. This hybrid type 1 implementation-effectiveness trial tests HN-STAR effectiveness while studying barriers and facilitators to implementation in community oncology practices within the National Cancer Institute Community Oncology Research Program. Effectiveness will be measured as differences in key survivorship outcomes between HNC participants who do and do not use HN-STAR over one year after the clinic visit. The primary endpoint is HNC-specific quality of life; other outcomes include patient-centered measures and receipt of guideline-concordant care. Implementation outcomes will be assessed of survivors, providers, and clinic stakeholders. The hybrid design will provide insight into a dose-response relationship between the extent of implementation fidelity and effectiveness outcomes, as well as how to incorporate HN-STAR into standard practice outside the research setting.

1. Introduction

Delivering guideline-concordant survivorship care for the heterogeneous, growing population of of head and neck cancer (HNC) survivors is a complex challenge. There are over 380,000 HNC survivors in the United States, with more than 66% of patients surviving beyond 5 years.^1–3 Multi-modality treatment for HNC can result in serious late effects, commonly including dental issues, hearing loss, dry mouth, neck and shoulder dysfunction, and lymphedema in the neck and face.^4–16 Other debilitating late effects include destruction of the jaw, inability to speak, difficulty swallowing, and difficulty opening the mouth.^4–8 Furthermore, HNC survivors commonly experience psychological distress.^17–20

The epidemiology of HNC is changing, with nearly 90% of oropharynx cancers now caused by human papilloma virus (HPV); previously, tobacco and alcohol use were the major causes of HNC.²¹ Patients with HPV-related HNC, on average, are younger, are less likely to be heavy tobacco and alcohol users, have fewer comorbidities, and live longer than their HPV-negative counterparts, thereby forming a large subgroup of long-term HNC survivors who must cope with late effects of therapy.

Post-treatment follow-up in HNC has historically been limited to early identification of recurrent and second HNCs.²² However, in 2016, the American Cancer Society (ACS), with endorsement by the American Society of Clinical Oncology (ASCO), published the first comprehensive guidelines for the multidisciplinary care of HNC survivors.^23,24 In addition to recommending oncologic surveillance and symptom management, the ACS/ASCO guidelines recommend primary care to manage comorbidities, modify risk factors (e.g. tobacco cessation), complete recommended screening for other cancers, assess and manage depressed mood, and deliver preventive care. Other healthcare providers(e.g. dentists, speech therapists, physical therapists, nutritionists, and mental health specialists) are important in the post-treatment care of HNC survivors.^23–25 With the support of healthcare providers, survivors commonly need to engage in self-management, including alcohol reduction, tobacco cessation, nutrition, physical activity, and rehabilitation for symptoms.^8,26–41

These ACS/ASCO guidelines – comprising over 30 individual recommendations – are complex and potentially difficult to enact during routine clinical follow-up visits. We developed a clinical informatics intervention to facilitate guideline implementation in the oncology setting, improve care coordination, and support self-management. The Head and Neck Survivorship Tool: Assessments and Recommendations (HN-STAR) uses patient-reported health concerns to guide routine oncology visits and create a survivorship care plan (SCP), a document that summarizes treatment and a plan for ongoing care.⁴² Prior pilot work demonstrated usability of HN-STAR.⁴³

The current protocol aims to test the effectiveness of HN-STAR while studying barriers and facilitators of implementation in community oncology practices.The trial will evaluate differences in key survivorship outcomes between HNC participants who use HN-STAR and those who receive usual care.

2. Methods

2.1. Study overview and aims.

The objective of this trial is to evaluate the effects of HN-STAR on patient-centered outcomes among HNC survivors receiving post-treatment care at community oncology practices within the National Cancer Institute Community Oncology Research Program (NCORP), a National Cancer Institute-funded network of community-based oncology practices.

Our central hypothesis is that the use of HN-STAR will improve patient-centered outcomes (including HNC-specific quality of life) and receipt of ACS/ASCO guideline-concordant care compared to usual care.

This pragmatic cluster randomized controlled trial evaluates the effects of HN-STAR on post-treatment HNC survivors (N=350), using a site-randomized trial design (≥15 intervention practices and ≥15 usual care practices). We will use a hybrid effectiveness-implementation type 1 design to simultaneously 1) test effectiveness on patient-centered outcomes and receipt of ACS/ASCO guideline-concordant care and 2) evaluate the implementation of HN-STAR in community oncology practices.⁴⁴

2.1.1. Aim 1.

Evaluate the impact of HN-STAR, compared to usual care, on change in patient-centered outcomes from baseline to one-year follow-up. Our primary endpoint is HNC-specific quality of life, as measured by the Trial Outcome Index from the Functional Assessment of Cancer Therapy – Head and Neck (FACT H&N); other outcomes include symptom burden, patient activation (knowledge, confidence, and skills to manage one’s own health), and perceived quality of cancer care.^15,45–55

2.1.2. Aim 2.

Evaluate the impact of HN-STAR, compared to usual care, on receipt of ACS/ASCO guideline-concordant care in the year after enrollment, including a visit to a primary care provider in the year following the designated clinic visit.

2.1.3. Aim 3.

Conduct a robust, mixed methods evaluation of the implementation process, including assessments of survivor, health care provider (oncology and primary care), and organizational barriers and facilitators that may influence integration of HN-STAR in various NCORP settings.

2.2. Setting.

Practices will include outpatient medical, surgical, or radiationoncology practices affiliated with the Wake Forest NCORP Research Base. Oncology practices have been enrolled in prior cancer care delivery research studies.^56–60 We created HN-STAR to be useful in diverse settings, particularly for clinics that have the greatest need for provider education and survivor support. To that end, we are conducting a pragmatic trial, minimizing restrictions on survivor eligibility, settings of care, and provider expertise.

2.3. Practice and designated clinician eligibility

Eligible oncology practices within each site must have treated ≥12 incident HNC patients in the last 12 months and identify at least one “designated clinician” who can implement HN-STAR or usual care for enrolled survivors. The designated clinician must be a provider (medical doctor, doctor of osteopathy, nurse practitioner, or physician assistant) who routinely provides follow-up care for cancer patients or survivors.

2.4. Practice enrollment and randomization

Eligible practices may enroll at any time after the study opens. Randomization of enrolled practices will take place at the practice level; most practices will include a single clinic, but some may include multiple locations that share providers and patients. WF NCORP RB will randomize the eligible practices with a 1:1 allocation. Randomization will be stratified by practice size (12 to 39 vs. ≥40 incident HNC cases in the prior year).

2.4. Survivor eligibility

Eligible survivors are individuals ≥18 years old and diagnosed with head and neck squamous cell carcinoma, specifically including oral cavity, larynx, any part of pharynx and unknown squamous cell carcinoma primary. They must have completed chemotherapy and/or radiation therapy (with or without surgery) for head and neck squamous cell carcinoma ≤ 6 months prior to enrollment. Eligible participants must have been deemed free of disease at last assessment. They must be cognitively and physically able to complete study surveys and plan to attend a follow-up clinic visit with the designated clinician who has agreed to participate in this study. The “designated clinic visit” may take place in person or via telehealth, including via telephone call or video. Survivors must be willing to complete study assessments after the designated clinic visit either 1) remotely (via smartphone, tablet, or computer) or 2) at the clinic on a clinic tablet or computer. Because the HN-STAR intervention is currently only available in English, survivors are ineligible if they do not speak or read English. Survivors are ineligible if they have ever been diagnosed with a prior primary HNC or any other cancer other than non-melanoma skin disease. Survivors enrolled on another interventional symptom management study protocol are ineligible.

2.5. Survivor screening, consent, and enrollment

NCORP staff located at each practice (“local staff”) will obtain information on follow-up care appointments for survivors of HNCs and will review medical records to identify potentially eligible participants. Local staff will contact potentially eligible survivors, screen survivors, and ascertain interest in participating. Screening can also take place in the clinic on the day of the appointment. During the consent process, survivors will be asked if they would like to participate in additional interviews.

Consent and enrollment is a two-step process. The first consent may be conducted verbally over the phone to enable the survivor to complete the baseline survey online prior to their designated clinic visit. The second consent requires a signature and takes place directly prior to the designated clinic visit.

Practices will continue to enroll survivors until they reach a maximum of 30 enrolled survivors or until study accrual goals are met.

To enhance inclusion of survivors from racial and ethnic minorities, we will monitor minority recruitment rates at our monthly WF NCORP RB committee meetings and provide feedback to the NCORP participating sites via bi-monthly study teleconference calls. Sites with strong minority recruitment will be asked to share their experiences with other sites during these calls.

2.6. Key stakeholder enrollment

Key stakeholders include members of the practice’s clinical or administrative team identified by designated clinicians as involved in the oversight of the delivery of HNC survivorship care or who would make decisions about implementing HNC survivorship tools such as HN-STAR. This includes clinic administrators, nurse navigators, key clinical team members involved in the care of HNC patients or survivors, program directors, information technology specialists, cancer care delivery leads, and other staff (e.g., service line or nursing leaders). They must be employed for at least one month at the practice and be able to speak and read English. Local staff will identify up to 3 stakeholders within each practice and will obtain informed consent.

2.7. Primary care provider eligibility and enrollment

Survivors are not required to have a primary care provider to participate in this study, but for a subset of survivors who identify a primary care provider, survivors may provide consent to have the study team contact their primary care provider for an interview about their provision of survivorship care. We will stratify by intervention arm and contact primary care providers within each arm, with 30 primary care providers sampled from the HN-STAR arm and 10 sampled from the usual care arm. Each selected primary care provider will be contacted by the Memorial Sloan Kettering (MSK) study team a year after the designated clinic visit to obtain consent remotely. Primary care providers will be sent the consent form via email and review and verbally consent over the phone with the MSK study team. Primary care providers will complete a brief phone interview (5–10 minutes) with MSK, which will elicit feedback on the provision of care for the survivor participant and their perceptions of the survivorship care plan, if they received one.

2.8. Interventions

2.8.1. Usual care.

In order to provide a minimal standard for usual care, designated clinicians in both arms will be invited to view an hour-long publicly available educational video of HNC survivorship, including an overview of the ACS/ASCO guidelines (CME offered).⁶¹ The standard of care for routine post-treatment visits will be followed for all survivors receiving follow-up care at sites assigned to the usual care arm.

2.8.2. HN-STAR

HN-STAR collects patient-reported outcomes from survivors and use them (along with clinical data) to present tailored, evidence-based guidance to clinicians providing routine follow-up care to survivors of HNC. HN-STAR generates one SCP based on decisions made during a single clinic visit (the “designated clinic visit”) and three updated SCPs based on quarterly patient-reported self-assessments. See Figure 1 for the timing of each component of HN-STAR in relation to the designated clinic visit.

Figure 1.

Timing of Study Activities

Platform.

HN-STAR was built by the Wake Forest Center for Healthcare Innovation to use on a tablet or computer. The Innovations platform enables survivors and designated clinicians to take surveys online and report outcomes in a user-friendly display. Survivors who do not have reliable internet access at home may come to the clinic to complete surveys on a tablet provided by the study.

HN-STAR components:

1. Survivor Self-Assessments.

Survivor Self-Assessments at baseline, 3, 6, and 9 months elicit the presence and burden of symptoms from survivors using relevant items from the NCI’s Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE™).^54,62–64 When no PRO-CTCAE™ item was available for a symptom, such as for hearing loss or jaw pain, we developed items in the style of the PRO-CTCAE™. A recent study found that cancer patients’ concerns were commonly not addressed in the pre-specified battery of PRO-CTCAE™ items, and write-in symptoms captured valuable information.⁶⁵ Because symptoms for HNC are known from the literature, developing items in the wording style of PRO-CTCAE™ contributes to comprehensive symptom elicitation. Items for alcohol misuse, tobacco use, physical activity, and distress are from validated screening instruments.^66–74 We ascertain distress using an adapted version of the Distress Thermometer, a clinically valid distress scale.^68–71 The Distress Thermometer will identify anxiety, depression, and sources of distress using a problem list. We adapted the problem list to exclude physical symptoms not relevant to HNC or those assessed elsewhere, in order to avoid duplication of assessments. (Figure 2)

Figure 2.

Survivor Self-Assessment (Excerpt)

Items from the Survivor Baseline Self-Assessment will inform the designated clinic visit. The 3-, 6-, and 9-month QuarterlySurvivor Self-Assessments are intended to monitor survivors’ concerns outside the clinic setting..The QuarterlySurvivor Self-Assessments will inform updated SCPs, as described in the SCP section below.

2. Treatment Summary Checklist.

HN-STAR presents a structured checklist for the designated clinician to enter diagnostic and treatment details prior to the clinic visit. (Figure 3)

Figure 3.

Treatment Summary Checklist (Excerpt)

3. Clinical Decision Support Tool.

Using data from the Baseline Survivor Self-Assessment and the Treatment Summary Checklist, HN-STAR generates a clinician-facing Clinical Decision Support Tool, which presents tailored recommendations for the designated clinic visit on a computer or tablet. Recommendations come from the ACS/ASCO HNC survivorship guidelines and address the topics symptom management, wellness (tobacco and alcohol use, nutrition, and physical activity), and prevention of lafe effects. ^23,24 Strategies may include exercises and other self-management approaches, medications, testing, and recommendations to visit other providers. All survivors receive an explicit recommendation to visit a primary care provider to receive general preventive care and comorbidity management. The designated clinician uses the Clinical Decision Support Tool as a point-of-service discussion guide during the designated clinic visit. The designated clinician and survivor choose actions from among the recommendations, which the designated clinician enters into HN-STAR. (Figure 4)

Figure 4.

Clinical Decision Support Tool (Excerpt)

4. SCPs.

HN-STAR automatically generates four SCPs based on ASCO guidelines for developing SCPs.⁷⁵ Each SCP is intended as a plain-language stand-alone document to improve follow-up care coordination. The Visit-Based SCP is generated when the designated clinician completes the Clinical Decision Support Tool, and it presents each action selected during the designated clinic visit (for cancer surveillance, symptom management, wellness, and preventive care). It contains a cancer history, contact information for the oncology team, and an explicit recommendation to visit a primary care provider. The designated clinician or clinical staff gives the Visit-Based SCP to the survivor and the primary care provider (electronically or on paper). Local staff will add the Visit-Based SCP to the medical record. (Figure 5)

Figure 5.

Visit-Based SCP (Excerpts)

Immediately after completing each Survivor Quarterly Self-Assessment (at 3, 6, and 9 months after the designated clinic visit), the survivor will be provided with an electronic Updated SCP, which tracks the trajectory of each symptom the survivor reported. It includes recommendations to visit providers as suggested by the most recent guidelines, but the emphasis will be on self-management strategies. The Updated SCP contains the same cancer history, oncology team contact information, and recommendation for primary care as in the Visit-Based SCP. Each Updated SCP will be provided electronically to the designated clinician at the practice, who reviews the SCP prior to adding it to the medical record. Each Updated SCP will be available in the medical record to inform any subsequent care provided as part of the standard of care. Survivors are instructed to contact their oncology team if symptoms are getting worse or if there are new symptoms.

2.8.3. HN-STAR training.

Designated clinicians in the HN-STAR arm will be provided with training to use HN-STAR. Designated clinicians will receive written, step-by-step instructions in an electronic slide deck with brief (less than two-minute) video vignettes included. Vignettes demonstrate how to log in and complete the Treatment Summary Checklist before the designated clinic visit and how to use the Clinical Decision Support Tool during the designated clinic visit. The study team will be available to answer questions about using HN-STAR.

2.9. Assessments and outcome measures.

Outcome measures will be collected via surveys and interviews of survivors, designated clinicians, stakeholders, and primary care providers, in addition to data abstracted from oncology medical records, primary care medical records, and tracking logs. (Table 1) Survivors will complete three surveys: at baseline (within two weeks prior to the designated clinic visit, directly after the designated clinic visit, and a year after the designated clinic visit). In addition, 40 survivors who opt to be considered to participate in additional interviews (30 from the intervention arm and 10 from the usual care arm) may be selected to participate in one interview after the designated clinic visit and another interview after the follow-up survey. Each designated clinician will complete an initial survey and interview before enrolling the first survivor from the practice into the study and a final survey and interview after the tenth survivor from the practice is enrolled in the study. In addition, for each survivor enrolled, the designated clinician will complete a brief survey directly after the designated clinic visit. Each stakeholder will complete an initial survey prior to enrolling the first survivor from the practice into the study and a final survey and interview after the tenth survivor from the practice is enrolled in the study. Finally, 40 primary care providers identified by survivors who agree that their primary care providers can be contacted may be selected to participate in an interview one year after the survivor’s designated clinic visit (30 from the intervention arm and 10 from the usual care arm). Interviews with survivors, designated clinicians, and stakeholders will follow the completion of a survey in an explanatory sequential design, providing the opportunity to gather in-depth perspectives on relevant topics identified in survey responses.^76,77 Additional outcomes will be ascertained from oncology medical records, primary care medical records, and site tracking logs.

Table 1.

Summary of Study Assessments

Assessment	Timing	Constructs (Aim)
Survivors
Survivor Baseline Health Assessment	Before designated clinic visit	HNC-specific quality of life; mental health; symptoms; patient activation; perceived quality of cancer care (1)
Survivor Post-Visit Survey	After designated clinic visit	Acceptability, appropriateness, and feasibility (3)
Survivor Post-Visit Interview	After designated clinic visit	Acceptability, appropriateness, and feasibility (3)
Survivor Follow-up Health Assessment	1 year after designated clinic visit	HNC-specific quality of life; mental health; symptoms; perceived quality of cancer care; patient activation (1)
		Symptom management, preventive care, wellness (2)
		Acceptability, appropriateness, and feasibility (3)
Survivor Follow-up Interview	1 year designated clinic visit	Acceptability, appropriateness, and feasibility (3)
Designated Clinicians
Designated Clinician Initial Survey	Before practice’s first survivor enrolled	Acceptability, adoption, appropriateness, feasibility, sustainability (3)
Designated Clinician Initial Interview	Before practice’s first survivor enrolled	Adoption, sustainability (3)
Designated Clinician Post-Visit Survey	Immediately after each clinic visit	Symptom management, preventive care, wellness (2)
Designated Clinician Final Survey	After practice’s 10th survivor enrolled	Acceptability, adoption, appropriateness, feasibility, sustainability (3)
Designated Clinician Final Interview	After practice’s 10th survivor enrolled	Acceptability, adoption, appropriateness, feasibility, sustainability (3)
Stakeholders
Stakeholder Initial Survey	Before practice’s 1st patient enrolled	Acceptability, adoption, appropriateness, feasibility, sustainability (3)
Stakeholder Final Survey	After practice’s 10th survivor enrolled	Acceptability, adoption, appropriateness, feasibility, sustainability (3)
Stakeholder Final Interview	After practice’s 10th survivor enrolled	Acceptability, adoption, appropriateness, feasibility, sustainability (3)
Primary Care Providers
Primary Care Provider Interview	1 year after clinic visit	Acceptability, adoption, appropriateness, feasibility (3)
Other data sources
Medical Record Abstraction	1 year after clinic visit	Primary care, oncology (2, 3)
Study tracking	Throughout study	Adoption, costs, fidelity, reach/penetration (3)
Automated HN-STAR tracking (HN-STAR arm)	Throughout study	Adoption, costs, fidelity (3)

Surveys will be electronically self-administered via REDCap and the Innovations platform (HN-STAR). Because the trial is measuring the implementation of HN-STAR, and the Survivor Self-Assessment is administered to survivors in the HN-STAR arm as part of the baseline survey, the baseline survey must be self-administered. For all other surveys, survivors can request assistance from the local NCORP site staff, who can conduct the surveys by phone. Interviews will be conducted by the Memorial Sloan Kettering Cancer Center (MSK) research assistant (RA) via WebEx or phone using semi-structured interview guides. All survivors, providers, and stakeholder participants in the trial will be compensated for completion of surveys and interviews.

2.9.1. Aim 1: Patient-centered outcomes.

For Aim 1, we will compare changes in patient-centered outcomes from baseline to one year among survivors using data from surveys. We hypothesize a clinically significant greater mean improvement in these outcomes over time in the HN-STAR arm compared to the usual care arm. Individual constructs and survey items are described in Table 2.

Table 2.

Aim 1 Endpoints and Instruments

Aim 1 Endpoint	Instrument
Head and neck cancer-specific QOL	Functional Assessment of Cancer Therapy-Head and Neck Version (FACT H&N TOI)45,46
Neck swelling	European Organization for Research and Treatment of Cancer (EORTC) HN43 Lymphedema scale15,49–51
Neck or shoulder stiffness	European Organization for Research and Treatment of Cancer (EORTC) HN43 Problems with shoulder scale15,49–51
Difficulty opening mouth (trismus)	EORTC HN43 Trismus scale15,49–51
Dry mouth (xerostomia)	EORTC HN43 Dry mouth scale15,49–51
Altered taste	EORTC HN43 Sense problems scale15,49–51
Bad taste in mouth	Item created in style of Patient Reported Outcomes version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE™)54,63,78,79
Mouth pain	EORTC HN43 Pain in mouth scale15,49–51
Difficulty chewing	EORTC HN43 Chewing scale15,49–51
Jaw swelling	Item created in style of PRO-CTCAE™
Jaw pain	Item created in style of PRO-CTCAE™
Difficulty swallowing (dysphagia)	EORTC HN43 Swallowing scale15,49–51
Cough	EORTC HN43 Coughing scale15,49–51
Heartburn	PRO-CTCAE™ Heartburn scale54,63,78,79
Speech	EORTC HN43 Speech scale15,49–51
Hearing loss	Item created in style of Patient Reported Outcomes version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE™)
Ringing in ears (tinnitus)	PRO-CTCAE™ Ringing in ears scale54,63,78,79
Dizziness	PRO-CTCAE™ Dizziness scale54,63,78,79
Numbness or tingling (peripheral neuropathy)	PRO-CTCAE™ Numbness and tingling scale54,63,78,79
Fatigue	EORTC HN43 Fatigue scale15,49–51
Insomnia	EORTC HN43 Insomnia scale15,49–51
Distress	EORTC HN43 Emotional well-being scale15,49–51
Body image	EORTC HN43 Body image scale15,49–51
Tobacco use	C-TUQ67
Alcohol use	Audit-C Questionnaire72,73
Sexual health	EORTC HN43 Sexuality scale15,49–51
Physical Activity	The Rapid Assessment of Physical Activity (RAPA)74
Pain	Brief Pain Inventory Short Form (BPI)52,53
Patient activation	Patient Activation Measure Short Form (PAM)47
Perceived quality of care	Consumer Assessment of Healthcare Providers and Systems (CAHPS®) Cancer Care Quality Measure55

HNC-specific quality of life.

The primary endpoint is HNC-specific quality of life. We anticipate that the primary benefit of HN-STAR will be improvement in HNC-specific quality of life at 1-year follow-up as measured by the Trial Outcome Index (TOI) from the Functional Assessment of Cancer Therapy Head and Neck (FACT H&N).^45,46 The TOI is a composite score of the Physical, Functional, and Head and Neck scales of the FACT.

Symptom burden.

We will evaluate physical symptom burden using the European Organisation for Research and Treatment of Cancer (EORTC) Module for Head and Neck Cancer (QLQ-H&N43). The EORTC QLQ-H&N43 captures symptoms relevant to HNC patients across tumor sites and treatment modalities.^15,49–51 Unlike the FACT H&N, the EORTC QLQ-H&N43 generates multiple symptom scales rather than a composite score. We will use the Brief Pain Inventory (BPI) to capture pain at baseline and follow-up.^80–82 Physical symptoms not included in the EORTC QLQ-H&N43 were assessed using items from the PRO-CTCAE™ amd the items created in the style of the PRO-CTCAE™. To measure mental health-related quality of life, we will use the EORTC QLQ-H&N43 Emotional well-being scale.

Patient activation.

Patient activation describes the extent to which patients are active and engaged in their health – a critical construct for HNC survivors who require self-management. We will use the Patient Activation Measure (PAM), which measures whether the patient believes an active role is important, the patient’s confidence and knowledge to take action, taking action, and staying the course under stress.⁴⁷

Perceived quality of care.

The Consumer Assessment of Healthcare Providers and Systems (CAHPS®) survey assesses patient experiences with their health care, and the CAHPS® Cancer Care Survey is used to evaluate the patient-centeredness of cancer care.⁵⁵ Relevant subscales of the CAHPS® Cancer Care Survey include composite measures of how well the cancer care team communicates with the patient, cancer care team’s use of information to coordinate patient care, cancer care team supports patients in managing the effects of their cancer and treatment, and patient’s rating of overall cancer care.

2.9.2. Aim 2: Receipt of ACS/ASCO guideline-concordant care.

For the analysis of receipt of guideline-concordant care in the year after the designated clinic visit, we hypothesize that survivors in the HN-STAR arm will have greater adherence to symptom management and wellness recommendations than survivors in the usual care arm, including a visit to primary care, which is recommended for all survivors. These data come from surveys and medical records. (Table 1)

Primary care.

We will ascertain whether each survivor had ≥1 visit to a primary care provider in the year between baseline and follow-up and, if so, what issues were addressed (e.g., management of comorbidities, risk factor modification, preventive care, management of late effects). Receipt of primary care will be established from primary care provider medical records abstracted a year following the designated clinic visit; when unavailable, we will supplement with self reported visits to a primary care provider in the survivor follow-up survey. Exploratory metrics of primary care will come from the primary care medical record, including whether visits addressed 1) management of comorbidities, 2) risk factor modification, 3) preventive care, or 4) management of late effects.

Cancer surveillance.

We will ascertain receipt of guideline-concordant cancer surveillance from the oncology medical records, as measured by visits with oncology providers that incorporate history and physical exam, including direct visualization of the oral cavity, pharynx, or larynx, as appropriate per guidelines. We will also measure survivors’ receipt of SCPs.

Symptom management.

For each symptom reported at baseline that meets a threshold of concern, a measure of minimal adherence to ACS/ASCO-recommended management will be whether the survivor spoke to a provider about that concern one year after the designated clinic visit, assessed in the survivor follow-up survey. Receipt of guideline-concordant care in the year after the designated clinic visit will also come from oncology and primary care medical records pertaining to the year following the designated clinic visit. Self-report and medical record self-management assessments will be reconciled and will form a composite symptom management score that is the proportion of symptoms reported at baseline that were brought to the attention of a provider in the following year.

A similar method will be used to evaluate receipt of guideline-concordant care during the designated clinic visit, which will be assessed using self-reported data from the post-visit surveys completed by survivors and the designated clinician, supplemented with oncology and primary care medical records.

Preventive care.

In the survivor follow-up survey, we will assess self-reported receipt of a minimal set of salient, individualized wellness approaches, including tobacco cessation, alcohol reduction, visit to a dentist, and discussions with a provider about nutrition, physical activity, and screening for other cancers, as appropriate. These items will form a composite preventive care score that is the proportion of relevant prevention approaches that were addressed.

2.9.3. Aim 3: Implementation of HN-STAR

Outcomes related to the implementation of HN-STAR will be used to:

1. Gauge how survivors and providers implement HN-STAR to assess a postulated dose-response relationship between implementation fidelity and effectiveness.

2. Gather quantitative and qualitative data to enhance future dissemination of HNSTAR, if proven effective.

The theory of organizational readiness for change posits that there is a shared psychological state within an organization that reflects a commitment to implementing practice changes, and the confidence to do so.⁸³ For HN-STAR, this theory posits that the extent to which those involved in decision making about HNC survivorship care (stakeholders and designated clinicians) feel they have the support and resources to implement HN-STAR in clinical practice, the more successful the implementation of HN-STAR. We will follow Proctor and colleagues’ recommended taxonomy for measurement of implementation outcomes.¹⁹ Specifically, we will measure acceptability (satisfaction with content and delivery), adoption (program uptake), appropriateness (relevance), cost, treatment fidelity, penetration (reach), and sustainability. These outcomes will describe the implementation process across all stakeholders (ie., survivors, oncology and primary care providers, and non-clinical stakeholders). We also specifically hypothesize that organizational readiness for change will be associated with adoption, fidelity, and sustainability of the HN-STAR intervention. Implementation outcomes will be assessed using mixed methods, with qualitative and quantitative data collected from survivors, stakeholders, designated clinicians, primary care providers, study tracking documents, and data captured in HN-STAR. (Table 1) Interviews will be used to enrich the understanding of implementation outcomes, asking respondents both broadly and specifically about their impressions of each component of HN-STAR or HNC survivorship care. Survey responses will be used to probe for further detail.

Organizational readiness for change.

We will assess clinicians’ and stakeholders’ readiness for changing practice prior to survivor enrollment using the Organizational Readiness for Implementing Change instrument during initial surveys of designated clinicians and stakeholders.⁸⁴ To better understand organizational characteristics that support or pose challenges to the adoption and sustainability of HN-STAR, we will evaluate the clinical services relevant to HNC survivors available at each practice using a structured checklist of services (e.g., speech language pathologists, mental health professionals) as part of the Designated Clinician Initial Interview. We will gain a more thorough understanding of services typically offered and provided to HNC survivors in the initial year after treatment completion, including any changes due to the COVID-19 pandemic, during the interview.

Acceptability, appropriateness, and feasibility.

The acceptability, appropriateness, and feasibility of crucial aspects of the educational video, HN-STAR, and usual care will evaluated through surveys of survivors, designated clinicians, and stakeholders, using the Acceptability of Intervention Measure (AIM), Intervention Appropriateness Measure (IAM), and Feasibility of Intervention Measure (FIM).⁸⁵ Acceptability of HN-STAR will also be measured using the User Acceptance of Information Technology Survey in the designated clinician and stakeholder surveys. ⁸⁶ In addition, semi-structured interviews with survivors and providers will supplement our understanding of acceptability, appropriateness, and feasibility. For example, each survivor in the HN-STAR arm who completes interviews will be asked about the ease of reporting their own concerns, whether all relevant topics were addressed in clinic, their thoughts about their clinician using a computer in clinic, and whether the clinic visit was too long, among other topics. Interviews of designated clinicians in the HN-STAR arm will include impressions of the ease and usefulness of using HN-STAR in clinic, among other topics. We will use total length of time required for SCP creation and designated clinic visit (see Cost) as a measurement of feasibility.

Adoption.

The MSK research associate will track uptake and utilization of the study intervention at each practice in both arms (i.e. participation in monthly calls, survivor enrollment, use of the Clinical Decision Support Tool and dissemination of the SCP in the HN-STAR arm, and completion of the designated clinician and stakeholder surveys and interviews).

Costs.

Each participating NCORP practice will maintain a study-tracking document to record costs of adopting each intervention at the practice and survivor levels. Practice-level costs include time devoted to training and general supplies (e.g., tablets for use in the HN-STAR arm). Survivor-level costs include length of designated clinic visit and time for clinician to complete the study intervention (Treatment Summary Checklist or completion of an SCP in the usual care arm). Staff time costs will be estimated based on national average wages by job type.

Fidelity.

In the HN-STAR arm, extent of completion of individual components of HN-STAR (Treatment Summary Checklist and Clinical Decision Support Tool) will be measured. In the usual care arm, the MSK RA will use a checklist to record whether the designated clinician conducted designated clinic visits and provided an SCP. If an SCP was provided, the components of the SCP will be recorded. In both arms, participation in training (including watching the educational video) will be documented, and the local NCORP site staff will use the study tracking document to record whether the SCP was delivered to the survivor and primary care provider (if the survivor has one).⁸⁷ Survivors’ completion of the Survivor Baseline Self-Assessment and Survivor Quarterly Self-Assessments will be recorded automatically in the Innovations platform.

Penetration.

Local NCORP site staff will document penetration – the numbers of eligible survivors approached and enrolled at each practice. Reasons for ineligibility, refusal, and characteristics of refusers and dropouts will be documented.

Sustainability.

Sustainabilitywill be assessed in the Designated Clinician and Stakeholder Final Interviews, which will explore challenges encountered in integrating HN-STAR in routine care and resources needed to maintain integration. The interviews will also explore changes to clinical practices during the study period, including ongoing changes due to the Covid-19 pandemic.

2.10. Design and Analysis.

This is a clustered randomized controlled trial study to understand the effectiveness of HN-STAR in improving patient-centered outcomes and receipt of guideline-concordant care. The study will also address barriers and facilitators of implementation of HN-STAR and usual care.

2.10.1. Statistical Power.

Our study aims to evaluate a service delivery intervention, and as such, we have designed a cluster randomized trial to avoid contamination.⁸⁸ We have powered our cluster randomized study based on the primary objective developed in Aim 1.

The primary analysis on which our power calculation is based will be a comparison of the TOI, which will be evaluated at baseline and at follow-up one year later. For our sample size calculation, we assume a two-sided type I error of 5%, 5 units as the minimal clinically significant difference, 12 as the overall standard deviation, and an average cluster size of 10. This corresponds to an effect size of 0.42, which can be interpreted as a medium effect size and is in line with the scope of our study.⁸⁹ Because we are using a cluster randomization, we also require an estimate of the intracluster correlation (ICC), which we assume is 0.05, which is in line with median empirical ICC estimates calculated in a range of primary care cluster randomized trials.⁹⁰ Using the approach and R function proposed by Donner & Klar, for each arm, we require a minimum of 15 practices with an average of 10 survivors per practice to achieve 80% power at a 5% type I error.^89,91,92 This totals 30 randomized practices, with a sample size of 300 evaluable survivors. Practice sizes vary in the sampling frame, and to ensure that we have wide representation of practices, we will require a minimum of 10 survivors per practice and will cap accrual at 30 survivors for any given practice. We will enroll 350 survivors at baseline to account for ~15% attrition from baseline to follow-up due to recurrence, second cancer, death, or loss to follow-up.

2.10.2. Analysis Aim 1: Patient-centered outcomes.

We will use hierarchical models to evaluate our primary outcome, the FACT H&N TOI, at one year, with practice-level covariates (randomization arm, stratification variable) and survivor-level covariates (e.g., baseline TOI, time from diagnosis to eligibility, and other demographic and disease characteristics).⁹³ Secondary outcomes for Aim 1 (symptom burden, mental health-related quality of life, patient activation, and perceived quality of cancer care) will be evaluated using a similar methodology. Missing data will be imputed according to conventions published alongside scoring systems for each scale.

2.10.3. Analysis Aim 2: Receipt of guideline-concordant care.

The primary goal of Aim 2 is to evaluate whether HN-STAR increases adherence to guideline-concordant care compared to controls. We will use hierarchical models and graphical and descriptive statistics to evaluate the outcomes for Aim 2. The models will evaluate each outcome, with practice-level covariates (randomization arm, stratification variable) as well as survivor-level covariates such as quality-of-life measurements at baseline, time from diagnosis to eligibility, and other demographic and disease characteristics.⁹³ Receipt of primary care and secondary outcomes (receipt of guideline-concordant cancer surveillance, composite symptom management score, and composite preventive care score) will be evaluated using a similar methodology. To understand patterns of care, we will use descriptive statistics to compare the frequency of each visit (primary care, oncology, or other provider) over the year between the study visit and follow-up and between arms.

2.10.4. Analysis Aim 3: Implementation of HN-STAR and usual care.

Quantitative Descriptive Analyses.

We will use descriptive statistics to summarize implementation outcomes. Additionally, we conduct the following analyses:

Adoption:

We will conduct paired t-tests to assess pre/post changes in organizational readiness to adopt survivorship care.

Penetration:

We will compare the proportion of eligible survivors who participate at each practice using chi-square tests. We will compare survivor sociodemographic characteristics between enrollees and refusers using univariate analyses (chi-square or Fisher’s exact tests and Wilcoxon rank sum tests) with enrollment.

Fidelity:

We will create a dichotomous composite indicator of SCP delivery that describes whether the SCP was delivered to the survivor (and, if the survivor has one, to the primary care provider). Another indicator will describe whether the Survivor Baseline Self-Assessments and Survivor Quarterly Self-Assessments were completed fully (Y/N). Similar to Aim1, we will use a hierarchical model to examine the association of fidelity on TOI adjusted for baseline TOI and other survivor-and provider-level covariates.

Cost:

We will evaluate the relationship between cost and TOI in a hierarchal model as described in Aim 1. This model will include survivor-and practice-level costs as defined above. Specifically, we will calculate the average practice and per-survivor cost of the study for each arm. We will compare average practice costs between the study arms and evaluate the association of practice characteristics.¹⁹²

Interview Analyses.

Interviews will provide a detailed and thorough description of the implememtation process. Analyses will occur coincident to data collection so that we can ascertain when thematic saturation is achieved for each study stratum. Interviews will be audio-recorded, transcribed, and analyzed using Atlasti v11 qualitative software using a structured thematic content analysis approach. Coders will develop the thematic framework and a comprehensive coding scheme. To ensure coding reliability, inter-coder discrepancies will be resolved through discussion and comparison. Coding will continue until a high level of reliability (Kappa= >0.80) is established between coders. After coding is complete, coders will review statements assigned to each major code to identify themes. Survivor post-visit and follow-up interviews will be analyzed for each study arm.

Interviews will be analyzed according to Appropriateness, Adoption, Feasibility, and Sustainability domains. Qualitative and descriptive quantitative findings will be presented together to illustrate HN-STAR implementation. Consistent with an explanatory sequential design, after analyzing the quantitative data we will modify the qualitative interview guide, adding probing questions to gain insight into interesting trends, associations, or unexpected outliers found in the quantitative data. Once all interviews are coded, the themes will be summarized and merged with trends from the survey data in a joint display, a standard approach in mixed-methods integration.⁹⁴ We will compare the two data sources side by side to identify key findings across the dataset. Each row will list the implementation constructs evaluated in the survey and interview (e.g., feasibility, acceptability, sustainability) and each column will contain corresponding findings from each data source (e.g., themes, summary statistics). After constructing the table, the analysis team will meet to achieve consensus on major themes across each construct, supported by the qualitative and quantitative data sources.

3.0. Discussion

The HN-STAR intervention combines multiple approaches that are designed to improve the quality of care provided to HNC survivors, including the use of patient-reported outcomes, the dissemination of guidelines, and the provision of tailored and updated survivorship care plans. This implementation-effectiveness hybrid trial is equipped to measure both HN-STAR’s effectiveness at improving health and healthcare, and the implementation of HN-STAR into community oncology practices.^44,95

Interventions for HNC survivors have typically addressed single symptom or symptom clusters, rather than the wide range of HNC survivorship needs identified in the guidelines.⁹⁶ Only one prior study examined a digital health tool to support self-management across multiple symptoms; this trial enrolled, but was not specific to HNC and did not embed the tool within a clinic visit or include a survivorship care plan; this trial found limited improvement in some HNC symptoms over six months.⁹⁷ The HN-STAR trial fills an important gap in the evidence regarding the effectiveness of digital tools for symptom management for HNC survivors.

The ACS recently described a framework for equitable improvement of outcomes for cancer survivors and their caregivers, in which there are three priority areas for improvement: routine needs assessment, personalization of care, and facilitating dissemination and implementation of survivorship interventions. Each priority area is are directly addressed by HN-STAR.⁹⁸

The ACS asserts that cancer survivors should have routine assessment of their needs, with repeated assessments of patient-reported outcomes suggested as a way of systematically monitoring symptoms and healthcare needs. HN-STAR uses patient-reported outcomes to inform clinical practice in the setting of a routine cancer survivorship visit. Patient-reported outcome assessment prior to a routine visit is supplemented by routine assessments outside the clinic visits using quarterly assessments and updated survivorship care plans. This feature of HN-STAR enables longitudinal tracking of symptoms and well-being, facilitating the identification of emerging or worsening issues in a timely manner. Integrating patient-reported outcomes into active cancer treatment enables patients to report their own concerns and have them acted upon directly, and has been shown to improve health-related quality of life, reduce premature stopping of treatment, and extend survival.^62,98–103 By recommending the extension of patient-reported outcomes from active cancer treatment to cancer survivorship care, the ACS acknowledges the importance of symptom management for cancer survivors, many of whom face long life expectancy and significant symptom burden.^98,104 The new ACS/ASCO clinical practice guidelines for survivors of HNC highlight supportive care. However, the long list of possible late effects addressed in the guidelines makes it difficult to identify and manage all issues. Assessing patient-reported outcomes is critical to this identification and prioritization of supportive care needs.

HN-STAR also addresses the second priority area in the ACS framework: the provision of personalized care tailored to the needs of survivors and caregivers. HN-STAR uses patient-reported outcomes to tailor care in two ways – to clinicians at the point of care and to survivors and primary care provider in survivorship care plans. In oncology settings, studies using patient-reported outcomes typically generate symptom reports for clinicians prior to visits.^105–107 HN-STAR extends the concept of a symptom report by identifying areas of unmet need and presenting corresponding evidence-based recommendations for care for each identified symptom during the designated clinic visit. After the designated clinic visit, patient-reported outcomes are used to personalize care using tailored survivorship care plans. In typical oncology practice, survivorship care plans vary in their content and implementation, but they typically are not tailored to specific toxicities and late effects experienced by survivors.¹⁰⁸ In contrast, HN-STAR’s tailored survivorship care plans only present recommendations for ongoing care for identified issues.

The third priority area of the American Cancer Society is to facilitate the dissemination and implementation of survivorship interventions, including guidelines for the care of cancer survivors. HN-STAR disseminates and measures the implementation of clinical practice guidelines. The ACS/ASCO clinical practice guidelines for HNC survivorship care address dozens of topics and may be particularly challenging to implement. At the same time, the guidelines also reflect the complexity and diversity of these survivors, and only a subset of the recommendations are applicable to each survivor. The Clinical Decision Support Tool brings only relevent recommendations from recent guidelines directly into the clinic visit, with the goal of facilitating implementation of evidence-based care.

Importantly, the effectiveness of survivorship care plans in practice has not yet been demonstrated, necessitating the careful study of both the effectiveness and implementation of HN-STAR.^109–114. Our hybrid type I study design not only assesses the effectiveness of HN-STAR in improving patient-centered outcomes and care delivery, but also collects data on the mechanisms, barriers, and facilitators relevant to the implementation of HN-STAR in real-world clinical practices.⁴⁴ Well-documented shortcomings of existing trials of survivorship care plans include the lack of attention to care coordination outcomes,^115,116 absence of data on the process of implementing survivorship care plans,⁹⁵ and the focus on early-stage, common cancers.^{115,117–119} In practice, adoption of survivorship across cancers has been poor, due to the resources required to create and review survivorship care plans.^{108,109,120–125} These barriers may be salient for complex cancer survivors, who have complicated clinical histories and whose survivorship care plans need to address the management of numerous late effects, treatment of comorbidities, modification of multiple risk factors, and surveillance for recurrence. Only one trial of survivorship care plans (the ENHANCES trial) focused on HNC survivors; survivors in the intervention arm were provided with an individualized written survivorship care plan within the context of a nurse-led session that targeted self-efficacy. ^126,127 The survivorship care plan was then shared with the primary care provider.

While the protocol described a comprehensive approach to management of multiple health issues, the content of survivorship care plan and the specifics of its implementation were not described; the trial found no difference in head and neck cancer-specific quality of life or self-efficacy.

The trial of HN-STAR aims to resolve these research gaps by focusing on survivors of a complex cancer and by measuring key components of both care coordination (including primary care involvement) and implementation as key trial endpoints. Furthermore, the trial aims to reduce practical barriers to survivorship care plan dissemination by automatically generating survivorship care plans. The hybrid design will provide insight into whether there is a dose-response relationship between the extent of implementation fidelity and study outcomes, as well as informing how to incorporate HN-STAR into standard practice outside of the research setting.

Ultimately, HN-STAR is a patient-centered clinical informatics tool, developed with input from HNC survivors, that uses patient-reported concerns to inform the content of a survivorship visit and survivorship care plans. Oncologists often under-report the burden of toxicities for HNC patients, and using patient-reported symptoms to guide the clinic visit prioritizes the survivor’s needs.¹²⁸ The survivorship care plan generated during the designated clinic visit reflects decisions made actively with the survivor, focusing on current, salient issues. Beyond the clinical encounter, HN-STAR continues to engage the survivor using periodic, online self-assessments. With this additional contact, HN-STAR enables the survivor to track symptoms over time and, by providing an updated SCP, creates reminders and encouragement for ongoing self-care. In this way, not only does HN-STAR aim to facilitate the use of evidence-based care by providers, it also focuses on engaging HNC survivors in their own care. Our trial of this two-pronged approach will inform whether we can improve survivorship care and health outcomes among these complex survivors.

Abbreviations

HNC

head and neck cancer

PCP

primary care provider

SCP

survivorship care plan