Problem list and diagnosis cleanup in eClinicalWorks for FQHCs: one traceable chart at a time

    Open any long-time patient's chart. Hypertension is on the list three times. Two of them are codes that were retired years ago. The pneumonia from 2021 is still active. The diabetes has no status at all.

    The provider scrolls past all of it. Every visit.

    Nobody owns the problem list. So nobody cleans it. This is the fix for FQHCs and other 330 centers on eClinicalWorks, and it is one of the MediBOT automations running in production today.

    The list nobody owns

    Problem lists rot for boring reasons:

    • Multiple providers add the same condition with different codes.
    • Nobody wants to delete another provider's entry.
    • Acute problems never get resolved.
    • Legacy and retired codes survived the last conversion.
    • Free text entries with no code at all.
    • A migration from the old EHR dumped everything in as active.

    Each one is small. Across a panel, it is a chart nobody trusts.

    What a dirty list costs a 330 center

    • Quality measures read diagnoses. Diabetes and hypertension measures start from who has the condition. Wrong list, wrong denominator. That flows straight into UDS reporting in eClinicalWorks.
    • Clinical Rules fire on bad data. Alerts show up for conditions the patient no longer has. Providers learn to ignore alerts. Then they ignore the real ones.
    • Claims and risk adjustment tell a stale story. Medicare Advantage is a big share of the island market. The diagnoses on the chart are the story the payer reads.
    • Provider time. Every visit starts with scrolling and second-guessing.

    Clean lists do not fix health outcomes by themselves. They stop the chart from hiding what matters.

    Why a one-time cleanup fails

    The usual play: a project, a spreadsheet, a few providers on overtime, a weekend. The list is clean for a month. Then the same habits refill it.

    Or the play is a report. It lists every bad entry. Nobody has time to act on it.

    If the root cause is build and templates, start with eClinicalWorks optimization. If the build is fine and the backlog is the problem, you need something that does the clicks, every night, forever.

    How it works inside eClinicalWorks

    MediBOT is digital staff with a named user account in eCW. Its own credentials. Its own role. Its own queue.

    Step one is human. Your clinical leadership approves the mapping: which codes get updated to which, which statuses change, which entries are left alone. MediBOT does not decide what a patient has. It applies the rules your clinicians signed off on.

    Then the nightly job:

    1. Read the upcoming appointments.
    2. Open each appointment.
    3. Locate the patient by account number.
    4. Apply the approved code and status mapping.
    5. Save the record.
    6. Log the change as one traceable queue item.

    If an entry does not fit the mapping, it does not guess. It goes to an exception queue for a person to review.

    Because it is appointment-driven, the patients coming in next are the patients cleaned first. The provider opens a current list. Over time the whole panel gets there without a weekend project.

    What staff stop doing

    • Manually reconciling duplicates before a visit.
    • Converting retired codes one chart at a time.
    • Resolving stale acute problems nobody closed.
    • Running a cleanup project every year that undoes itself.

    Providers still own the diagnosis. MediBOT owns the clerical pass. Pair it with lab results automation so the chart is current on both sides: what the patient has and what the last result said.

    Proof

    Problem list and diagnosis cleanup is live in production in Puerto Rico. It runs nightly inside eClinicalWorks.

    Across 3 Puerto Rico community health centers, from December 2024 to July 2026, MediBOT processed 16,990 automated transactions, redirected 1,088 staff hours, and touched 13,059 patient interactions, for an estimated $20,525 in labor savings. That is the combined total across the automations running in those centers. Not one center. Not one workflow.

    Those automations include problem list cleanup, document handling, and eligibility verification. We do not publish a problem-list-only number until it is measured on its own.

    Who this is for

    • FQHCs, CHCs, and other 330 centers on eClinicalWorks.
    • CMOs and QI leads who stopped trusting what the problem list says.
    • Centers that came through an EHR migration with a messy import.
    • Leadership that wants cleaner charts without pulling providers off the schedule.

    If nobody on your side owns EHR data quality, that is a virtual CIO conversation too.

    FAQ

    Is problem list cleanup live or a pilot?

    Live. Problem list and diagnosis cleanup runs nightly in production inside eClinicalWorks in Puerto Rico.

    Does MediBOT decide which diagnoses to change?

    No. Your clinical leadership approves the code and status mapping first. MediBOT applies that mapping chart by chart. Anything that does not fit the mapping goes to a person for review.

    Who can see what MediBOT changed?

    Every change is made by a named MediBOT user in eCW and logged as its own queue item, so auditors and providers can see who changed what and when. Document the new account in your HIPAA security risk analysis like any other user.

    Why not just run a one-time cleanup project?

    Because the list refills. A one-time project cleans the backlog and stops. A nightly, appointment-driven job keeps the patients coming in next clean, every night, without pulling providers off the schedule.

    Does a cleaner problem list help UDS?

    UDS clinical measures start from who has a condition, so a list with duplicates, retired codes, and stale actives can distort denominators. Cleanup does not guarantee a measure result. It removes one source of bad data.

    How do we start?

    Book a discovery call. Export a sample of duplicate or conflicting problems from a few charts. We start with the mapping, then turn on the nightly job.

    Book a discovery call

    Pick ten charts at random. Count the duplicates, the retired codes, and the stale actives. That is your baseline.

    Book a discovery call or ask for an automation assessment. We agree on the mapping with your clinicians, then MediBOT works the list every night. See it at medibot.health. Meet our leadership team.

    Humans first. Tech second.

    Millennial Networks Metro Office Park Lot. 6, Suite 204 Guaynabo, PR 00968 Florida office: 4850 Tamiami Trl N, Suite 301, Naples, FL 34103 787-945-2260 corp@mnetpr.com