Strategies and Opportunities to Stop Colon Cancer in Priority Populations
Aim 1 Conduct preliminary analyses of EHR data at two pilot OCHIN clinics.
- We will utilize the OCHIN patient database to identify patients (ages 50 - 74) who are
eligible for CRC screening. Automated codes will be used to manage auto interventions
so that eligible patients get initial CRC screening on time, are prompted to repeat
screening as recommended, and obtain follow-up care or surveillance over time
- Define inclusion and exclusion criteria definition of inclusion and exclusion criteria
to identify sets of patients with average CRC risk at two pilot OCHIN clinics
o These criteria will likely consider previous screening history (e.g., receipt of
recent screening), comorbid conditions that make patients poor candidates for screening
(e.g., end-stage renal disease), and other factors. For the pilot test, language will
also be a criterion, as we plan to pilot-test the program only in English and Spanish.
- Define data sources and methods - definition of data sources and methods to extract EHR
data on socioeconomic and demographic variables, receipt of CRC screening tests, test
results, receipt of follow-up care, and diagnoses
- In Phase I, we will perform sensitivity analyses, cross tabulations, and natural
language processing-assisted chart audits (of 50 charts from each pilot clinic) to
determine whether claims and referral data can be used to track colonoscopies.
- We will provide support and training to enter these data within each safety net
organization, so that both the patient registry and Epic will have complete data
- Define intervention moderators - identification of phenotypes (e.g., race/ethnicity,
age, sex) and potential moderators of intervention effectiveness (e.g., native
language, income, co-morbidities, continuity of care, insurance status) and creating
codes to extract EHR data relevant to these moderators o We will define methods and
data sources for our proposed intervention moderators: Hispanic ethnicity, primary
language, poverty status (% above FPL), and insurance status. This will be the basis
for moderator analysis planned for Phase II (Section F.3).
Aim 2: Use codes and methods developed in Aim 1 to test the feasibility, reach,
effectiveness, and cost of an EHR-based CRC screening intervention using a subset of 100
patients at each of two pilot clinics. We will use codes defined in Aim 1 to create a
Solutions-based, Epic-integrated system to track CRC screening, mail FIT kits, and track
patient test results and receipt of follow-up care. The system will be pilot-tested in two
- Create EHR-integrated CRC screening registry. We will use the Solutions panel
management support tool to create the database registry for automated interventions
(Auto). Solutions will be customized to include tools needed for tracking and
completion of CRC screening at the two VGMHC clinics. The registry will list patients
who meet the eligibility criteria (e.g., no personal history of CRC, no recent
screening). The registry will list patient CRC screening history as well as key
demographic variables for carrying out the intervention (e.g., name, address, preferred
language). For the pilot, we will select only patients whose preferred language is
English or Spanish; at VGMHC only 2% of patients speak a language other than English or
Spanish. We will select a cohort of 100 patients from each of two clinics (50 Hispanic
and 50 non-Hispanic white) while ensuring relatively equal distribution of gender and
age (<65 vs. 65+).
o Reports: The registry will generate two real-time reports: (1) CRC screening outcomes
(date of mailing, date of screening events, and results); and (2) CRC follow-up
diagnostic care and treatment. Reports will be used to monitor and maintain quality
assurances during intervention implementation. Registry data will also be available in
Epic (since they are directly generated from lab and other data entered into Epic).
- Assessment of program reach. To refine our power calculations and protocols, we want to
know how many patients we reached through mailed FITs (Auto) and how many through other
clinic activities (Auto-Plus). We will consider (1) the proportion of patients who meet
the study's eligibility criteria (e.g., no personal history of CRC); (2) the proportion
who agree to the programs (i.e., do not opt out); and (3) the proportion who are
reached by the programs. To assess the percentage of patients unreachable by mail, the
clinic coordinator at each pilot site will mail an introductory letter to all eligible
patients announcing our program.
o Reports: We will report the percentage of postcards and FIT kits returned by the post
office (recognizing that the return rate may differ for postcards and FIT kits).
- Assessment of program effectiveness. Using codes defined in Aim 1, we will assess
program effectiveness, calculated as the proportion of eligible patients who completed
a FIT within 3 months of the pilot intervention.
- Reports: We will also assess the percent of patients who screen positive by FIT.
These estimates will be used to recalculate power, assess the number of needed
colonoscopies, and select the number of clinics for our pragmatic trial.
- Reports will be generated weekly listing the number of letters, FIT kits, and
reminders mailed; number of FIT and other CRC tests completed; positive FITs and
completion of colonoscopy afterwards; and advanced adenomas and CRCs diagnosed.
These reports will be reviewed regularly by the research team and the Planning and
Implementation Advisory Groups.
- Cost assessment for pilot intervention and economic evaluation plan for Phase II. We
will conduct a Phase I cost assessment for the pilot intervention and plan for our
Phase II economic evaluation.
Aim 3: Use results from the pilot intervention to prepare for a large-scale,
cluster-randomized pragmatic trial across 18 OCHIN clinics (see Phase II).
- Define inclusion and exclusion criteria for clinic participation in Phase II. We will
work with the Planning Advisory Group to develop clinic-level inclusion criteria for
participation in the pragmatic trial. These criteria will consider statistical
requirements for powering our analysis (i.e., minimum number of clinics and minimum
number of age-eligible patients per clinic), generalizability (e.g., how overall
demographic characteristics of the selected clinics match the national FQHC
population), and potential for large-scale spread. Other criteria may include the
available capacity (via clinic or community resources) to provide colonoscopy services
among patients who screen positive for FIT.
- Select additional FQHC clinics. The research team and Planning Advisory Group will
apply the defined criteria to OCHIN clinics and document reasons for exclusions. Among
FQHCs meeting the criteria (an estimated 30 in 5 or 6 organizations), Drs. Coronado and
DeVoe will conduct one-on-one semi-structured interviews (in-person or via phone) to
assess capacity and readiness to adopt the program and participate in the trial.
Capacity questions will be specific to CRC screening: e.g., "Does your network have a
centralized site for processing FOBTs/FITs?" and "What resources are available for
providing colonoscopies to patients who screen positive?" We will draw on the published
literature, specifically on Weiner and others,85 to develop these questions. After
applying the inclusion and exclusion criteria and confirming funding for Phase II, we
will report the number of FQHCs that agreed to participate in the program and the
number that refused (Section F.4.b).
Allocation: Randomized, Endpoint Classification: Efficacy Study, Intervention Model: Parallel Assignment, Masking: Open Label, Primary Purpose: Screening
Increased Screening Rates
Investigators will determine if the auto and auto-plus interventions increase the rates of screening compared to usual care at FQHC clinics.
up to 4 years (study period)
Gloria Coronado, PhD
The Center for Health Research, Kaiser Permanente Northwest
United States: Institutional Review Board
|Virginia Garcia Memorial Health Center||Hillsboro, Oregon 97124|