The Prevention Adoption Initiative starts with one colorectal cancer screening pilot in a Michigan health center. Every later stage waits on that pilot’s evaluation. Nothing here has a date, because none of it has a credible schedule yet.
One health center, three transparent ranking rules, the site’s own staff. Three months of feasibility at no fee, then a funded pilot to month twelve.
By month twelve the routine is written down and handed to the site’s staff. The written part is what can travel to another clinic. None of it exists yet; each item is something the first pilot is designed to produce.
The gate passed, and the site agrees what can be shared.
A second site willing to run it from the written toolkit, which is Stage 3.
The same routine at another health center.
One site cannot show that the result travels.
The gate passed, plus a host and an evaluator for the second site.
A direct EHR connection in place of the nightly coded file.
Manual exports and a one-day lag.
A business associate agreement, a HIPAA-eligible setup, and the site’s IT sign-off.
A model suggests who is least likely to finish, next to the rules.
Some patients at risk of not finishing do not stand out on days overdue alone.
The rules evaluated, enough outcome data, a fairness audit, staff still deciding, and a model that beats the rules.
Two-way text reminders in the patient’s language that help book the test.
Staff time spent on routine reminders instead of the hard cases.
The site’s consent process, a BAA, and a tested hand-off to a person. Nothing automated before then.
Regular checks that outreach is not reaching easier patients first.
Automation can widen the gap between language or payer groups.
A review board, agreed measures, and enough patients in each group to read.
The same routine could run on other recommended steps that people start and do not finish, and in other health systems. Five are sketched below. Each would need its own local design and validation.
Same four steps, same division of work: your staff act, HSREP keeps the list, an independent evaluator judges. Each line below is a recommended step that people start and do not finish, with the published size of the gap. Pick one.
A repeat reading and a plan within four weeks of a high reading (CMS measure 22, screening for high blood pressure and follow-up).
30.5%
of patients whose readings met the hypertension threshold had no recorded diagnosis during the study, and another 14.6% were diagnosed late. One health system, 311,743 patients, 2010 to 2021.
Lu, Brush et al., JAMA Network Open 2025.
Everyone with a reading of 140/90 or above and no recheck on the books, from the nightly file.
Days since the high reading, how high it was, prior no-shows.
A nurse blood-pressure visit, a loaned home cuff with a call-back, or a pharmacist visit.
A diagnosis made or ruled out, and the days from the first high reading to that decision.
Invented patientRosa M. · two readings at 140/90 or above · 61 days, no recheck booked → nurse BP visit booked, cuff lent, call-back in a week.
A retinal exam within the year for every adult with diabetes (HEDIS measure EED).
54.1%
of Medicare beneficiaries with diabetes had an eye exam in 2017. Healthy People 2030 puts the national baseline at 64.8% and the target at 70.3%.
Lundeen et al., MMWR 2019; Healthy People 2030 objective D-04.
Everyone with diabetes and no eye exam in the last twelve months.
Months since the last exam, a known retinopathy history, prior no-shows.
A retinal photo at the clinic if it has a camera, or a booked eye appointment with a ride.
The exam result in the chart, and the days it took.
Invented patientDmitri K. · last eye exam 23 months ago · one prior no-show → retinal photo added to the visit already booked next week.
Both kidney tests, a blood test (eGFR) and a urine test (uACR), within the year (HEDIS measure KED).
Under 40%
of adults with diabetes received both tests in 2017. The urine test is the one that gets missed. Seven million adults, insurance claims data.
Mayo Clinic Proceedings: Innovations, Quality & Outcomes 2023.
Everyone with diabetes missing either test in the last twelve months.
Months since the last urine test, whether a blood draw is already scheduled, prior no-shows.
Add the urine test to a lab order that already exists, or a lab-only visit with no appointment.
Both results in the chart.
Invented patientAmina S. · blood test done, no urine test in 14 months → urine test added to the lab order already scheduled for Thursday.
Diagnostic imaging or a biopsy within 60 days of an abnormal screening mammogram (BI-RADS 0, 4 or 5; measure PQM 4700e).
66.4%
of 2.9 million abnormal (BI-RADS 0) screening mammograms from 2008 to 2021 had documented diagnostic follow-up. A third did not.
2024 study cited in the Brigham and Women’s measure submission, PQM 4700e.
Every abnormal screening result with no diagnostic imaging or biopsy recorded.
Days since the result, the BI-RADS category (4 and 5 first), prior no-shows.
A booked diagnostic appointment, a navigator call, a ride.
The diagnostic result in the chart (benign, probably benign, or biopsy done), and the days.
Invented patientGrace O. · BI-RADS 0 · 48 days, nothing booked → diagnostic ultrasound booked, ride arranged, result back in the chart on day 57.
A colposcopy after an abnormal result that calls for one. Guidelines expect it within months; the study below allowed a year.
75.3%
received a colposcopy within twelve months of the abnormal result. One in four did not, even with a year. 28,706 patients, 2010 to 2018.
Kamineni, Feldman et al., Preventive Medicine 2022 (PROSPR).
Every abnormal result that calls for colposcopy, with none recorded.
Days since the result, the grade of the result (high-grade first), prior no-shows.
A call in the patient’s language, a booked colposcopy, help with childcare or transport.
Colposcopy done, and the days from result to procedure.
Invented patientLan T. · high-grade result · 95 days, no colposcopy → called in Vietnamese, booked, reminder two days before.
Nothing here is proposed as a pilot. Colorectal screening is the only pilot being asked for. The patients are invented and the gap figures come from the sources named, not from any HSREP work. Each line would need its own local design, evidence review and evaluation before anyone ran it.
More health centers after the Michigan pilot, through health-center networks and professional bodies.
After the pilot is evaluated.
Community clinics under the national cervical and breast screening program, and private diagnostic networks.
A future direction, in development.
For a private network, a natural first partner, higher completion compounds commercially: more screenings lead to more diagnostics, more appropriate downstream care, stronger patient retention, and better-positioned corporate-wellness and payer contracts. The initiative raises the completion rate the business is already working to grow, advancing patient health and commercial performance together.
For the DGHS Non-Communicable Disease Control Programme (NCDC), the national cervical & breast screening leadership, and the Community Clinic Health Support Trust, the completion loop embeds into the existing network of roughly 14,000 community clinics as of 2019 and the national electronic registry, closing the uptake and follow-up gaps the program’s own evaluation identifies, piloted in a few upazilas and aligned to the national strategy.
BRAC and Marie Stopes bring community reach; icddr,b, BSMMU, and NICRH bring evaluation and clinical authority. Together, these partners can turn a pilot into publishable, scalable evidence, and open pathways to both government and donor funding.
Experience in one setting does not show that the routine will work in another. Where the work could begin abroad is where I already have standing and screening-service experience; that is a reason to start there, not evidence that it will work there.
All of this depends on the first pilot.
If your organization could host it, that is the conversation to have first.
An HSREP initiative in development. Everything on this page is future direction and illustrative; no patient data has been collected.
