The Maine Cancer Genomics Initiative: Implementing a Community Cancer Genomics Program Across an Entire Rural State

Abstract

Purpose: The Maine Cancer Genomics Initiative (MCGI) aimed to overcome patient- and provider-level barriers to using genomic tumor testing (GTT) in rural practices by providing genomic tumor boards (GTBs), clinician education, and access to comprehensive large-panel next-generation sequencing to all patients with cancer in Maine. This paper describes the successful implementation of the initiative and three key services made operative between 2016 and 2020.

Methods: A community-inclusive, hub-and-spoke approach was taken to implement the three program components: (1) a centralized GTB program; (2) a modular online education program, designed using an iterative approach with broad clinical stakeholders; and (3) GTT free of charge to clinicians and patients. Implementation timelines, participation metrics, and survey data were used to describe the rollout.

Results: The MCGI was launched over an 18-month period at all 19 oncology practices in the State. Seventy-nine physicians (66 medical oncologists, 5 gynecologic oncologists, 1 neuro-oncologist, and 7 pediatric oncologists) enrolled on the study, representing 100% of all practicing oncologists in Maine. Between July 2017 and September 2020, 1610 patients were enrolled. A total of 515 cases were discussed by 47 (73%) clinicians in 196 GTBs. Clinicians who participated in the GTBs enrolled significantly more patients on the study, stayed in Maine, and reported less time spent in clinical patient care.

Conclusion: The MCGI was able to engage geographically and culturally disparate cancer care practices in a precision oncology program using a hub-and-spoke model. By facilitating access to GTT, structured education, and GTBs, we narrowed the gap in the implementation of precision oncology in one of the most rural states in the country.

FIG 1.

Population density in Maine and per capita patient enrollment in the MCGI. (A) Population density shown per municipality. Gray areas represent uninhabited areas according to Maine state population data. (B) MCGI enrollment, per HSA defined by the Dartmouth Atlas Project. MCGI enrollment was highest on a per capita basis in rural and sparsely populated areas in the northernmost parts of the State and in coastal communities in the Mid Coast and Downeast region. Gray areas represent areas without HSA. This map also shows all 19 MCGI-participating oncology practices in the State. AR, Arthur Robinson; HSA, hospital service area, a collection of ZIP codes whose residents receive most of their hospitalizations from the hospitals in that area; MCGI, Maine Cancer Genomics Initiative; MDI, Mount Desert Island; NECS, New England Cancer Specialists; NL, Northern Light.

FIG 2.

MCGI hub-and-spoke structure. The Clinical Steering Committee, comprising 19 physician leaders representing all oncology practices in Maine, provided community feedback over all aspects of the MCGI. The MCGI Central Office acted as a research-engaged spoke site for patient enrollment and formed the hub with the CIPHR. Spoke sites consisted of 14 hospital-based practices operated by five health systems (MaineHealth, Northern Light Health, MaineGeneral Health, Central Maine Health, and Covenant Health), four oncology practices providing services for independent hospitals (Cary Medical Center, Mount Desert Island Hospital, Mid Coast Hospital, and York Hospital), and one private oncology practice serving three locations (New England Cancer Specialists). Purple circles: research-engaged hospital or private practice; blue circles: practices for which the central MCGI office provided research services. AR, Arthur Robinson; CIPHR, Center for Interdisciplinary Health Research; EMMC, Eastern Maine Medical Center; HACCC, Harold Alfond Center for Cancer Care; MCGI, Maine Cancer Genomics Initiative; MDI, Mount Desert Island; ME, Maine; MMP, Maine Medical Partners.

FIG 3.

Overview of the MCGI study protocol: To systematically evaluate the impact of the MCGI on patients and physicians, the MCGI used a study protocol focused on three outcomes: (1) feasibility and acceptability of NGS tests; (2) clinician and patient experiences with NGS tests; and (3) technical outcomes of care (treatment decision making, morbidity, quality of life, and mortality). Each clinician participated in the study both as a study subject and as a patient-enrolling subinvestigator. The protocol-designated GTT was ordered by the enrolling physician to inform clinical care of the enrolled patient. GTTs, GTBs, and the education program were the central components of the program. GTB, genomic tumor board; GTT, genomic tumor testing; MCGI, Maine Cancer Genomics Initiative; NGS, next-generation sequencing.

FIG 4.

GTB case presentation per physician. Bars represent total case result numbers per physician (each column presenting one physician, each physician assigned a random number): dark blue bars represent cases presented at GTB, and light blue bars represent cases not presented at GTB. Forty-seven of 67 (72%) physicians (enrolling at least one patient on the study) presented at least one patient at the GTB. ^aFour physicians working in rural areas with highest per capita patient enrollment in the MCGI. GTB, genomic tumor board; MCGI, Maine Cancer Genomics Initiative.

FIG 5.

Physician attitudes about the value of GTBs. Black dots represent means. Error bars represent standard error of the mean. Colored dots represent each participant's response (position jittered to avoid overplotting). The x-axis labels reflect the response labels on surveys questions: 1 = strongly disagree to 5 = strongly agree (intermediate options were not labeled). Survey responses of (A) all GTB participants and (B) physician participants only. GTB, genomic tumor board.

FIG A1.

Practice and patient enrollment, test result return, and GTB discussion data. Data broken down by calendar quarters between Q3 2017 (start of enrollment) to Q3 2020 (end of enrollment). (A) Practice enrollment starting with seven practices that were enrolled in Q3 2017 to all 19 practices by the end of Q4 2018. (B) Patient enrollment by quarter, shown as the total of each column. Dark teal: Number of case results discussed at GTB, expressed as % of patients enrolled during that calendar quarter. Light teal: Number of case results returned that were not discussed at GTBs. Gray: Number of cases that did not return a result report either to nonsubmitted specimen or specimen failure in the laboratory. GTB, genomic tumor board.

FIG A2.

Study enrollment on the basis of rurality. (A) Number of patients enrolled in the MCGI who provided their ZIP code (n = 1,313). The rurality of each participant's primary residence was determined via the USDA's Rural-Urban Commuting Area codes, which use measures of population density, urbanization, and daily commuting to map ZIP codes to urban/rural categories. Because of small numbers of participants in some categories, we collapsed the 11 primary RUCA codes (1-10, 99) into four categories: 1 = metro, 2-6 = large rural, and 7-9 and >10 = small rural, 10 = isolated rural. (B) Percent of patients enrolled by the central hub office or by the independent sites for each rurality category. MCGI, Maine Cancer Genomics Initiative; USDA, United States Department of Agriculture.

FIG A3.

Breakdown of GTB participants between 2017 and 2020. Although a number of different groups attended the GTBs (left box), the core audience for the GTB are the clinicians, that is, clinical staff and physicians (right box). GTB, genomic tumor board; JAX, Jackson Laboratory.

FIG A4.

(A) GTB attendance and (B) GTB cases in 2019 (green columns) and 2020 (gray boxes) monthly. March 2020 showed a lower number of GTB participants and GTB cases (due to start of COVID-19 pandemic) but then rebounded and exceeded 2019 numbers in the subsequent year, likely due to a switch to fully virtual GTB format. Numbers shown from March through June as a representative time period of the COVID-19 pandemic. GTB, genomic tumor board.