Navigation vs Usual Care for Timely Adjuvant Therapy for Patients With Locally Advanced HNSCC

Design: The study design is a Single-Site, Parallel-Group, Randomized-Controlled Trial of Navigation Versus Usual Care for The Management of Delays and Racial Disparities Starting Postoperative Radiation Therapy in Adults with Surgically-Managed, Locally Advanced Head and Neck Squamous Cell Carcinoma

Following screening and informed consent, sociodemographic, oncologic and symptom data will be prospectively gathered about participants from validated questionnaires and the electronic medical record (EMR). Participants will then be randomized to 3-sessions of the navigation intervention (NDURE; Navigation for Disparities and Untimely Radiation thErapy) or usual care (UC) and followed until the start of postoperative radiation therapy (PORT) following surgery for head and neck squamous cell carcinoma (HNSCC). Measures of PORT delay, racial disparities in PORT delay, key cancer care delivery processes, and theoretical constructs underlying PORT will be evaluated.

Treatment Allocation: Upon enrollment, participants will be randomized 1:1 to NDURE or UC using a stratified randomization design with strata defined by race (white, African American [AA]) and location of radiation facility (Medical University of South Carolina ([MUSC], non-MUSC) because of the known association of these variables with PORT delay.

Delivery of intervention:

NDURE is a theory-based, multi-level PN intervention consisting of three in-person, clinic-based sessions of manualized PN with multiple intervention components that target system- (care coordination), interpersonal- (social support), and individual- (Health Belief Model; perceived susceptibility, severity, barriers, self-efficacy) level health behavior theoretical constructs to reduce barriers to care, increase HNSCC care delivery, and improve clinical outcomes (timely, equitable PORT). NDURE will be delivered from surgical consultation to PORT initiation (~3 months). The NDURE intervention consists of: Navigation Sessions, the Navigator Manual, the Navigator Patient Guide, structured EMR documentation, weekly conferences to facilitate care coordination, real-time patient tracking, and multidisciplinary reporting. The three in-person NDURE navigation sessions, which are expected to take 30-60 minutes each, will coincide with the presurgical consult, hospital discharge, and 1st postoperative clinic visit, time points chosen to facilitate case identification and coordination across key care transitions. Contact beyond the three prescribed in-person sessions will occur with a frequency and modality (e.g. text message, email, etc.) dictated by patient and navigator need. During the first in-person session, the navigator will 1) elicit barriers and facilitators to timely PORT from the patient, caregiver, and provider, 2) develop the personalized barrier reduction plan (BRP), review it with the patient, caregiver, and provider, and 3) implement the BRP. At the two subsequent in-person sessions, the navigator will review and update the BRP in an iterative, dynamic fashion, identifying new barriers and systematically tracking resolution of prior barriers until the start of PORT. The Navigator Manual provides a structured resource to guide intervention delivery and enhance reproducibility. The Patient Guide is 1) literacy-level appropriate, 2) personalized for each patient's care pathway and BRP, 3) updated longitudinally as the patient progresses along the cancer continuum, and 4) available to patients in print and/or electronically via the patient portal in the EMR.

UC consists of oncology visits with provider-led discussion about the referrals needed to start PORT.

Expected Effect Size and Power Calculation: Power and sample size calculations were performed using PASS version 08.0.13, "Inequality Tests for Two Independent Proportions." The primary endpoint for this pilot RCT is the rate of PORT delay, defined by NCCN Guidelines as PORT initiation > 6 weeks following surgery. Our primary objective is to compare PORT delay rates between the NDURE and UC arms. Patients (n=150) will be randomized 1:1 to NDURE or UC using a stratified randomization design with strata defined by race (white or AA) and location of radiation facility (MUSC or not MUSC). Furthermore, to facilitate evaluation of PORT delay rates in racial subgroups, the investigators will oversample AAs for a final sample size of 50 white and 25 AA patients in each arm. The investigators assume the rate of PORT delay in the usual care arm will be 45% and target an absolute reduction of 20% for a PORT delay in the navigation arm of 25%. This effect size is clinically significant and is a realistic goal given published rates of improvement in similar (non-randomized) interventions. Seventy-five patients in each arm yields 83% power to detect a 20% reduction in PORT delay (45% versus 25%) based on a two-sided Mantel-Haenszel test of two independent proportions assuming a two-sided α = 0.1. Our choice of the Mantel-Haenszel test to compare proportions is based on the trial's stratified design. Our selection of α = 0.1 and 1 - β = 0.8 is based on the desire to emphasize power over type I error at this early stage of development (single institution phase II trial) to ensure follow-up on promising interventions. The investigators therefore consider our trial to be appropriately and rigorously designed to detect a clinically meaningful reduction in PORT delay.

Statistical Methods of Analysis: Graphical displays and descriptive statistics for sociodemographic, oncologic, and baseline symptom severity characteristics will be used to examine patterns and summarize data for each arm. The percentage of patients who start PORT within 6 weeks of surgery (the primary outcome measure) and corresponding 95% confidence interval (CI) will be calculated for both arms and for white and AA subgroups within each arm separately. The rate of PORT delay will be compared between arms using a Mantel-Haenszel test of two proportions, with strata defined by race and location of radiation facility. The investigators will analyze time to PORT as a continuous variable and estimate median time to PORT for each arm and for racial subgroups within each arm using Kaplan-Meier curves with Greenwood variance estimation to construct the corresponding 95% CIs. The investigators will estimate the hazard ratio comparing the two arms using Cox proportional hazards regression controlling for the stratification variables. The investigators will compare time to PORT between intervention arms using a stratified log-rank test. For other secondary endpoints, data will be summarized using frequency and percent for categorical variables and using mean, median, standard deviation, IQR and range for continuous variables. The investigators will also construct 95% CIs to provide a measure of uncertainty in estimated proportions and means. Comparisons between trial arms of other secondary endpoints will be performed using t-tests and chi-square tests, or Wilcoxon rank sum and Fisher's exact tests as appropriate. Baseline and post-intervention values of variables measuring the theoretical constructs underlying NDURE (i.e. care coordination, self-efficacy in cancer care, interpersonal support, and knowledge) will be compared using Wilcoxon sign rank tests. Comparisons between arms of the change in scores will be conducted using Wilcoxon rank sum tests.