Geographically Clustered Mailing of Fecal Immunochemical Test (FIT) Kits

Effect of Geographically Clustered Mailing of Fecal Immunochemical Test (FIT) Kits on Colorectal Cancer Screening

This is a pragmatic, randomized, controlled trial of a geographically randomized intervention in which all eligible patients will receive a mailed fecal immunochemical test (FIT) kit (the GeoMail study). In this study, all ZIP Codes in Dallas County will be randomized to either treatment or usual care. Patients in treated neighborhoods will receive kits at the same time; patients in usual care neighborhoods will receive FIT kits randomly throughout the year. All eligible patients will receive a FIT kit each year. The trial will be conducted in a racially and socioeconomically diverse cohort of patients served by an integrated safety net delivery system. This study represents a pragmatic adaptation to the "FIT first" population health screening strategy in which all eligible patients are mailed FIT kits. All patients will also have access to whatever colorectal cancer (CRC) screening is recommended through usual visit-based care.

Colorectal cancer (CRC) screening can save lives but is under-used, especially among minority and low-income patients in urban safety-net systems. Interventions among safety-net patients have improved screening but require additional refinement. For example, the investigators demonstrated that mailing free home fecal immunochemical tests (FIT) to urban safety-net patients doubled screening uptake (59% vs. 30%, p<.001); however a full 41% who received FIT kits still did not complete screening. Drawing from economics, behavioral science, and healthcare delivery science, the investigators propose a spatially informed randomized intervention to further improve screening uptake. By targeting all patients in a single neighborhood at one time using waves of mailed invitations, the GeoMail intervention is designed to amplify social contagion among neighbors (i.e., positive communications and reinforcement among neighbors to complete screening) thereby increasing screening uptake. GeoMail will be live for 12 months. This intervention will not change the content of FIT kits that are already to be mailed under usual care conditions at Parkland, it will only change the timing of mailing. The aims of this study are to: Aim 1: Conduct and evaluate outcomes from a GeoMail intervention - a geographically clustered randomized trial comparing geographically clustered mailed screening invitations (treatment) to non-clustered invitations (usual care). Hypotheses (H): H1.1: Patient-level and H1.2: Neighborhood-level screening will be higher in the treatment neighborhoods, compared to usual-care neighborhoods. Aim 2: Evaluate whether patient-level or neighborhood-level indicators of patient density, race/ethnicity, and socioeconomic disadvantage moderate GeoMail impact on CRC screening. H2.1-3: All three types of factors will moderate the GeoMail intervention impact

Mailed outreach invitation to complete FIT include: 1) invitation letter, 2) test kit (1-sample FIT, simplified instructions on how to perform the test and return mailer with prepaid postage). Up to three "live" phone reminders from project staff 2 to 3 weeks after the invitation to encourage screening completion. Centralized processes to promote guideline-based follow-up.

The experimental arm differs from the active comparator in one way: patients living in treated ZIP codes will receive the mailed outreach at the same time.

Defined as screening completed within 6 months of when the FIT kit is mailed for each patient by any recommended screening modality (FIT, colonoscopy, sigmoidoscopy; i.e., the proportion of treated vs. usual care patients who have completed screening). Screening will be measured using electronic medical record procedure, result, and billing data (Parkland-specific, CPT, ICD9/10 codes) and will be considered complete when satisfactory (e.g., adequate sample collection) FIT or colonoscopy test results are received.

Defined for each neighborhood (defined as patient block group of residence) as the number of all screening-eligible patients who are up-to-date with screening guidelines (defined as FIT ≤18 months, sigmoidoscopy ≤5 years, or colonoscopy ≤10 years) divided by the total screening-eligible block group Parkland patient population. The analysis will determine GeoMail impact on treated block groups while differencing change in the outcome that occurs in usual care block groups using a difference-in-difference model (DID) to measure change in neighborhood level screening for each block group before and after GeoMail intervention at 2 time points: on the day prior to GeoMail going live and at 6 months after mailing the last GeoMail kit (i.e., 18 months after start of GeoMail). This outcome measures the population level impact of GeoMail across all patients in the block group, including those not mailed a FIT kit because they were already up-to-date and thus ineligible for the mailing.

Defined as any recommended screening modality. Screening will be measured using electronic medical record procedure, result, and billing data and will be considered complete when satisfactory FIT or colonoscopy test results are received.

Defined for each neighborhood as the number of all screening-eligible patients who are up-to-date with screening guidelines divided by the total screening-eligible block group Parkland patient population.

Inclusion Criteria: Males and females Age 50-64 years Seen at least one time at a Parkland primary care clinic within one year prior to randomization Participants in Parkland's medical assistance program for the uninsured (Parkland Financial Assistance), Medicaid, or private/other insurance All races and ethnicities Exclusion Criteria: Up-to-date with CRC screening, defined by: Colonoscopy in the last 10 years Sigmoidoscopy in the last 5 years FIT in the last 11 months Prior history of CRC, total colectomy, inflammatory bowel disease, or colon polyps Address or phone number not on file Incarcerated