NCT07141901

Life Expectancy-informed Colorectal Cancer Screening

Improving Colorectal Cancer Screening Decisions Through Consideration of Life Expectancy

• Status: Not Yet Recruiting
• Phase: N/A
• Study type: Interventional
• Enrollment: 15,000 (estimated)
• Sponsor: Kathryn Martinez · Academic / Other
• Sex: All
• Age: 76 Years
• Healthy volunteers: Accepted

Summary

The purpose of this study is to improve life-expectancy concordant colorectal cancer screening for adults over 75 years through design and testing of a life expectancy notification to clinicians delivered via the electronic health record. This study has three components: 1. The investigators will develop a life expectancy prediction model using patient data from the Cleveland Clinic electronic health record (EHR) and test it against two existing life expectancy prediction models to determine which should be used in clinical care. This is an observational cohort study. 2. Concurrently with the development and testing of the life expectancy prediction model, the investigators will conduct interviews with clinicians to generate knowledge regarding the optimal way to integrate life expectancy information into decision making about colorectal cancer screening in patients over 75 years. This is a qualitative study. 3. The investigators will then conduct a cluster randomized trial of a clinical decision support-delivered life expectancy notification on life expectancy-congruent colorectal cancer screening orders by primary care clinicians.

Detailed description

The purpose of this study is to improve life-expectancy concordant colorectal cancer screening for adults over 75 years through design and testing of a life expectancy notification to clinicians delivered via the electronic health record. The investigators will pull patient data from the Cleveland Clinic Electronic Health Record (EHR) and use this to build and test a prediction model. The investigators will compare this model against two other existing life expectancy prediction models using these data. Once the investigators have identified the model with the best predictive ability, the investigators will test the model in the Metro Health System study population. Concurrently with the development and testing of the life expectancy prediction model, the study team will develop the interview guide, which will be semi-structured; the investigators will use pre-defined questions based on study interest but will also allow for new topics to emerge. The investigators will invite clinicians for interviews based on their prior two-year colonoscopy ordering for patients \>75 years, identifying clinicians in the top quarter and bottom quarter of colonoscopy orders compared to their peers. Akin to a "positive deviance" sampling approach, this method will capture clinician perspectives at both ends of the distribution and will provide richer information about how varied clinicians consider life expectancy in colonoscopy decision making. As the investigators want to collect diverse clinician perspectives, unique clinicians will be recruited for each phase of interviews. Interviews will last between 45 minutes and an hour take place via Microsoft Teams. Interviews will be audio recorded and transcribed verbatim. For each phase, following coding of 10 interviews, investigators will share findings with the study team and develop a preliminary coding dictionary. Subsequent interviews will be coded using this dictionary, adding additional codes as they emerge. Interviews will continue until thematic saturation is reached. Based on prior similar studies the investigators anticipate this will require 20 clinician interviews per phase. The investigators will work with the Cleveland Clinic EHR team to implement the life expectancy algorithm and the Best practice alerts (BPA) into EHR. The BPAs will be piloted over a one-month period by 10 clinicians. During the pilot, the investigators will monitor use to ensure the BPAs are working correctly, in keeping with the 5 Rights framework. (e.g. displaying for the right patients in the right location). The investigators will conduct clinician interviews to elicit feedback on the notification, including making changes to the suggested language to support screening discontinuation for patients with \<10 years life expectancy. The investigators will then modify the placement and/or the wording of the notifications and conduct another pilot. Based on experience, 1-2 pilots will be sufficient, but the investigators can conduct a third round, if needed. The investigators will randomize the clinics 3:2, with 30 sites in the intervention arm and 20 sites in the control arm. The investigators will then turn on the Clinical Decision Support at the intervention sites, where it will fire automatically for any eligible patient. By pulling data from the EHR, the investigators will assess colonoscopy orders, and other outcomes between 1 and 365 days following the index visit.

Conditions

• Colorectal Cancer Control and Prevention

Interventions

Type	Name	Description
BEHAVIORAL	Decision Tool	For patients with life expectancy \<10 years: Active (i.e. interruptive) BPA that will fire when the clinician orders colorectal cancer screening. The BPA will advise that screening is not recommended due to \<10 year life expectancy. An active alert is required to prevent the clinician from ordering a potentially inappropriate test. The BPA will contain a box with suggested language with which to express this to the patient. For patients with life expectancy ≥10 years: Passive alert on the Storyboard indicating that patient could benefit from screening as their predicted life expectancy is ≥10 years. A passive alert does not interfere with workflow and is available whenever it is convenient for the clinician to address it.

Timeline

• Start date: 2026-08-01
• Primary completion: 2029-01-01
• Completion: 2029-03-01
• First posted: 2025-08-26
• Last updated: 2025-09-24

Locations

1 site across 1 country: United States