HCC Gap Closure Software for ACOs: How to Evaluate It

By Jay Chowdappa, MD, co-founder and CEO, Zynix AI · Updated November 3, 2026

Key takeaways

HCC gap closure software for ACOs should find suspect and recapture gaps from claims and clinical data, put dated evidence in front of the treating clinician before a visit, get patients with open gaps scheduled, and count a gap closed only when a supported diagnosis reaches a claim. Clinicians decide every diagnosis; the software only surfaces evidence.

In the Medicare Shared Savings Program, prospective CMS-HCC risk scores adjust an ACO’s historical benchmark for changes in case mix, subject to a cap on positive adjustments.
Suspect gaps and recapture gaps are different problems and need different rules, owners and evidence.
Outpatient coding guidelines bar coding a condition documented only as “suspected” or “probable”, so a suspect list is a prompt for evaluation, never a diagnosis.
Test every vendor on your own data with the ten-point scorecard below, and judge closure on claims, not on clicks.

What is HCC gap closure software?

HCC gap closure software is a tool that finds patients whose chronic conditions are likely present but not documented in the current year, and helps the care team get those patients evaluated so that any condition the clinician confirms is documented accurately. It is a workflow tool, not a coding engine.

In practice it does four jobs: it builds a list of open gaps, it prepares the clinician with evidence before a visit, it gets patients with no visit booked onto the schedule, and it tracks each gap until it is confirmed closed. Most products do the first job well. The difference between products shows up in the other three, which is where this guide spends its time. If you want the background on why the work has to run all year, start with moving HCC risk adjustment from a Q4 scramble to year-round work.

Why HCC gaps matter to an ACO

For a Shared Savings Program ACO, HCC documentation affects the benchmark that savings are measured against. CMS uses prospective HCC risk scores to adjust the historical benchmark for changes in severity and case mix between the third benchmark year and the performance year, subject to a cap on positive adjustments.

Those risk scores come from risk adjustment-eligible diagnoses on final action fee-for-service claims (inpatient, outpatient and physician). For agreement periods that began on January 1, 2024 or later, CMS scores both the benchmark years and the performance year with the CMS-HCC model that applies to the performance year’s calendar year. The positive adjustment is capped at the ACO’s aggregate demographic risk score growth plus 3 percentage points. A chronic condition that is present but never documented makes the population look healthier than it is; a condition coded without support creates audit and compliance exposure. The goal is accuracy in both directions.

Many ACOs also hold Medicare Advantage contracts through their practices. There, CMS uses the 2024 CMS-HCC model (often called V28) for 2027, and it has finalized excluding diagnoses from audio-only encounters and from unlinked chart review records. Our explainer on V24 vs V28 covers the model changes in detail.

Suspect gaps vs recapture gaps

A recapture gap is a condition documented in a prior year that has not yet been documented this year. A suspect gap is a condition the data suggests but no clinician has documented, such as a lab pattern consistent with chronic kidney disease or a medication usually prescribed for heart failure.

The two need different handling. Recapture gaps are mostly a scheduling and preparation problem: the patient needs a visit where the clinician evaluates the condition again. Suspect gaps need more clinical judgment, because the evidence may point to a condition that is not present, already resolved or better explained another way. Good software keeps them on separate worklists and shows the evidence behind each one.

The coding rules draw a firm line here. The ICD-10-CM outpatient guidelines say not to code diagnoses documented as “probable”, “suspected” or “rule out”, and to code all documented conditions that coexist at the encounter and require or affect patient care, treatment or management. A suspect flag is therefore a reason to evaluate the patient, nothing more. Only the clinician’s own assessment, written in the note, turns it into a diagnosis.

The ten-point evaluation scorecard

Judge HCC gap closure software on what happens after the list is built. Use this scorecard in every demo, and ask each vendor to show each row on a sample of your own patients rather than on a prepared demo dataset.

What to test	Ask the vendor to show you	What good looks like
Data inputs	Which feeds the gap list is built from: CCLF or other claims files, EHR problem lists and notes, ADT feeds, labs and pharmacy, and how often each refreshes.	Claims and clinical data in one patient record, with the source of every piece of evidence named.
Suspect vs recapture logic	How a suspect gap (evidence but no diagnosis) is told apart from a recapture gap (documented last year, not yet this year).	Two labelled gap types, each with its own rules and its own worklist.
Evidence for clinician review	The exact screen a clinician sees before the visit for one patient.	Dated evidence (lab value, medication, specialist note, prior claim) next to each gap; no pre-filled diagnosis.
MEAT documentation support	How the tool helps the clinician show the condition was monitored, evaluated, assessed or treated at the visit.	Prompts that point to what is missing in the note, never text that states a diagnosis for the clinician.
Getting the visit scheduled	What happens to a patient with an open gap and no visit booked.	An owner, an outreach step and a booking, tracked to completion; not just a list exported to a spreadsheet.
Audit trail	The history of one gap from creation to closure.	Who saw it, when, what evidence was shown, and what the clinician decided, including “not present”.
RADV readiness	How a submitted diagnosis links back to the encounter note that supports it.	Every closed gap traceable to a face-to-face encounter and a signed note a reviewer could read.
Model support	Which CMS-HCC model each gap is scored under, and how the tool handles a model change.	The 2024 CMS-HCC model (V28) for Medicare Advantage and the model CMS applies to your MSSP performance year, with mappings updated each payment year.
Confirmed closure	When the tool marks a gap closed.	Only when the diagnosis appears on a final action claim or accepted encounter record, or the clinician documents that the condition is not present.
Reporting	Open, addressed and closed gaps by practice and clinician, month by month.	Numbers you can reconcile to your own claims, not a vendor-calculated revenue estimate.

MEAT (monitor, evaluate, assess or address, treat) is a documentation convention widely used by coders and auditors to check that a note supports a diagnosis. It is not a term defined in CMS regulation, but it is a practical test of whether a note would stand up when a reviewer asks if the condition was actually addressed at the visit.

How to run the evaluation on your own data

Run a short, structured test instead of comparing slide decks. A two-to-four-week bake-off on your own feeds tells you more than any reference call.

Send the same data to every vendor. A recent CCLF extract, an EHR export for a few practices and your ADT feed, under a BAA.
Pick a sample of patients your team already knows. Ask your clinical documentation lead to judge whether each flagged gap is plausible and whether the evidence shown is enough to act on.
Watch one gap end to end. From creation, through the pre-visit summary, outreach and the visit, to the claim that confirms closure.
Ask who does the work. For each step, name the owner: vendor staff, your care coordinators or the software’s outreach. A list with no owner does not close gaps.
Check the exit. Ask how you get your gap history, evidence and audit trail back if you leave.

Common mistakes when buying HCC gap closure software

Most failed rollouts trace back to buying the list instead of the follow-through. These are the mistakes we see most often.

Buying on the size of the suspect list. A longer list is not a better one. Ask how many flagged gaps your own clinicians would accept as worth evaluating.
Letting software pre-fill diagnoses. Suggested codes dropped into a note invite unsupported diagnoses and audit findings. Evidence belongs next to the gap, not in the assessment.
Counting a signed note as closure. A gap is closed when a supported diagnosis reaches a final action claim or accepted encounter record, or when the clinician documents the condition is not present.
Leaving unseen patients off the plan. Patients with open gaps and no visit booked need outreach and a slot on the schedule, or their conditions drop out of next year’s score.
Ignoring the model year. Scoring gaps under the wrong CMS-HCC model sends outreach to conditions that no longer affect the risk score.
Treating chart review as a safety net. For Medicare Advantage, diagnoses from unlinked chart review records are excluded for 2027. Documentation at the visit carries the weight.

Worked example (fictional)

This scenario is invented to show the workflow. It describes steps and owners only, and makes no claim about results.

A physician-led MSSP ACO with practices on three different EHRs is reviewing gap closure tools in the autumn, ahead of the next performance year. Its quality director runs the bake-off above with two vendors.

January: the software builds the year’s list from the latest CCLF files, EHR problem lists, labs and pharmacy data. Recapture gaps and suspect gaps land on separate worklists, each scored under the CMS-HCC model for the performance year.
Before each scheduled visit: the clinician receives a one-page summary per patient: each open gap, the dated evidence behind it and the documentation elements the last note was missing. No diagnosis is pre-filled.
For patients with no visit booked: a care coordinator owns the list, and patients are invited for an Annual Wellness Visit or a chronic care visit in the first half of the year, with reminders until they are booked.
At the visit: the clinician evaluates each gap and documents what is present, what is not and what needs more workup. A suspect gap for heart failure, for example, may end as “not present” with the reason noted.
After the visit: coders review the note before the claim goes out. The gap is marked closed only when the diagnosis appears on the final action claim, and the audit trail keeps the evidence, the note and the decision together.
Every month: the ACO reviews open, addressed and closed gaps by practice and clinician, and moves outreach staff toward the practices with the most unseen patients.

Where Zynix AI fits

We built Zynix AI around the follow-through half of this scorecard. Its HCC gap closure workflow for ACOs ranks open gaps, prepares clinicians before scheduled visits and invites patients with no visit booked. Population Intelligence brings claims, EHR, ADT, lab and pharmacy data into one patient record, and the Embedded Care Management Platform gives each gap a task, an owner and escalation by rule. Clinicians decide what to document and coders decide what to submit; the software never chooses codes. Health plans can see the member-facing version in risk adjustment outreach for MA plans. Hold us to the same scorecard you use for everyone else.

This guide summarizes Medicare rules as CMS published them as of November 3, 2026. It is not legal or coding advice. Check current CMS guidance, the ICD-10-CM Official Guidelines and each payer’s policy before you change documentation or coding workflows.

Frequently asked questions

What is HCC gap closure software for ACOs?

It is software that finds patients whose chronic conditions are likely present but not yet documented this year, prepares the treating clinician with the evidence before a visit, gets patients with open gaps scheduled and tracks each gap until a supported diagnosis reaches a claim. It supports accurate risk adjustment for Shared Savings Program benchmarks and Medicare Advantage contracts. It does not diagnose or code on its own.

What is the difference between a suspect gap and a recapture gap?

A recapture gap is a condition documented in a prior year that has not been documented yet this year, so the patient needs a visit where the clinician evaluates it again. A suspect gap is a condition the data suggests, from labs, medications or outside notes, that no clinician has documented. Suspect gaps need more clinical judgment, because the condition may not be present.

Does HCC documentation matter for MSSP ACOs that are not in Medicare Advantage?

Yes. CMS uses prospective CMS-HCC risk scores, built from diagnoses on final action fee-for-service claims, to adjust a Shared Savings Program ACO’s historical benchmark for changes in case mix between the third benchmark year and the performance year. Positive adjustments are capped. Conditions that are present but never documented make the assigned population look healthier than it is.

Can HCC gap closure software code diagnoses automatically?

It should not. The ICD-10-CM outpatient guidelines bar coding conditions documented only as suspected, probable or rule out, so a flag from software is a reason to evaluate the patient, not a diagnosis. The treating clinician decides what is present and documents it, and coders review what is submitted. Pre-filled diagnoses create risk in a CMS risk adjustment data validation audit.

How do you know an HCC gap is actually closed?

Count a gap closed when the diagnosis appears on a final action claim or accepted encounter record from an eligible face-to-face visit and the note supports it, or when the clinician documents that the condition is not present. A signed note alone is not closure. Ask every vendor to show one gap traced from the evidence to the claim.

Source notes

CMS: Medicare Shared Savings Program, Shared Savings and Losses, Assignment and Quality Performance Standard Methodology Specifications, Version 14 (April 2026), sections 3.2.1 and 3.2.4. https://www.cms.gov/files/document/medicare-shared-savings-program-shared-savings-losses-assignment-methodology-specifications-version.pdf-0

CMS (April 6, 2026): 2027 Medicare Advantage and Part D Rate Announcement (fact sheet). https://www.cms.gov/newsroom/fact-sheets/2027-medicare-advantage-part-d-rate-announcement

CMS: 2027 Model Software/ICD-10 Mappings. https://www.cms.gov/medicare/payment/medicare-advantage-rates-statistics/risk-adjustment/2027-model-software-icd-10-mappings

CMS: Medicare Advantage Risk Adjustment Data Validation Program. https://www.cms.gov/data-research/monitoring-programs/medicare-risk-adjustment-data-validation-program

CMS (May 21, 2025): CMS Rolls Out Aggressive Strategy to Enhance and Accelerate Medicare Advantage Audits. https://www.cms.gov/newsroom/press-releases/cms-rolls-out-aggressive-strategy-enhance-and-accelerate-medicare-advantage-audits

CMS and NCHS: ICD-10-CM Official Guidelines for Coding and Reporting, FY 2027 (effective October 1, 2026), Section IV.H and IV.J. https://www.cms.gov/files/document/fy-2027-icd-10-cm-coding-guidelines.pdf

About the author

Jay Chowdappa, MD is co-founder and CEO of Zynix AI. He is a physician, and ACOs led by him generated $300M+ in shared savings before he started Zynix AI to help care teams follow through on what their data already shows.

Evaluating HCC gap closure software?

See how Zynix AI ranks open gaps, prepares clinicians before visits and brings in patients with no visit booked, with clinicians deciding what to document.

Book a 30-min walkthrough →

HCC Gap Closure Software for ACOs: What It Should Do and How to Evaluate It