Most Organizations Find Their HCC Gaps Too Late
Picture this: it is February, the measurement year has closed, and your RAF reconciliation shows scores running 8 to 12 percent below projection. The conditions were probably there. The patients were probably seen. The documentation window is simply gone. That is the reactive model, and it is the default for a surprisingly large share of Medicare Advantage plans and provider organizations.
HCC suspecting and gap closure is the structural answer to that problem. It moves the discovery window from after the fact to before or during the patient visit, when a physician can still document, confirm, or rule out a condition within the same measurement year. The difference sounds simple. Operationally, it requires a real program with data infrastructure, clinical workflow integration, and defined accountability across multiple teams.
What a Suspecting Program Actually Does
Suspecting starts with a question: based on everything we know about this patient, what chronic conditions does she likely still have that have not been addressed or documented during this measurement year?
The data signals that feed a suspecting engine typically include prior-year diagnosis codes (both from the plan's own encounter data and from fee-for-service claims if available), active pharmacy claims (a patient filling metformin and a GLP-1 receptor agonist but without a diabetes diagnosis coded this year is a clear suspect), laboratory results that suggest uncontrolled chronic disease, and sometimes even problem lists pulled from the EHR. Some programs also incorporate predictive models that score the likelihood a given chronic condition is still clinically active for a given patient.
The output is a suspect list, typically generated at the patient level and surfaced ahead of a scheduled visit. That timing matters more than almost any other design choice. A suspect list that arrives after the appointment is a retrospective audit. One that arrives before the appointment is a clinical prompt.
What the Physician Sees
Physician-facing presentation is where many otherwise sound programs break down. A dump of thirty ICD-10 codes the system thinks might apply is not actionable. Effective suspecting programs surface a short, prioritized list of conditions, typically ranked by clinical relevance or RAF weight, with supporting evidence attached: the prior-year code, the relevant lab value, the active medication. That context lets a physician evaluate the suspect in under a minute rather than treating it as a compliance checklist someone else generated.
The format also matters. Suspect flags embedded natively in the EHR workflow generate far higher closure rates than separate portals or printed encounter prep sheets that end up in a folder nobody opens.
What Gap Closure Means Operationally
Suspecting is the identification step. Gap closure is what actually produces value.
Closing a flagged suspect means one of three things happens during or after the visit. The physician examines the patient, confirms the condition is still present and clinically relevant, and documents it to HCC-valid specificity in the current encounter note. Or the physician identifies that the condition has changed, documents the updated status, and orders appropriate follow-up. Or the physician determines the condition has resolved, which is also a valid and necessary outcome that must be documented explicitly so the flag does not carry forward indefinitely.
All three of those are closed gaps. What is not a closed gap is a suspect that sits on a list, gets scrolled past, and never receives physician attention or coder follow-up.
The Workflow Behind Closure
A defined closure workflow answers three questions: Who is responsible for following up on each flagged suspect? By what deadline? And how is the outcome recorded?
In practice, this usually means a CDI specialist or coding analyst reviews each visit after it is documented, checks whether the suspect was addressed, and routes unresolved flags through a follow-up process. That follow-up might be a query to the physician, a scheduling prompt for a gap-closure visit, or an escalation to the clinical operations team. The specific mechanism is less important than its existence. Without it, suspects accumulate and closure rates stagnate. Strong CDI program support is often what separates organizations that track closure rates from those that track only the length of their suspect list.
Why This Is Not the Same as Retrospective Chart Chasing
Retrospective chart review has a legitimate place in HCC coding quality. A coding quality audit conducted after encounters are coded can catch errors, identify training needs, and support RADV preparation. That is valuable work.
But retrospective review has a structural ceiling. If the physician did not document the condition in the encounter note, there is nothing for a retrospective coder to capture. The condition may be real and clinically active, the patient's chart may be full of supporting evidence from prior years, and it still cannot be coded for the current measurement year if it is not addressed in a current encounter. Retrospective review finds what was documented. Suspecting creates the opportunity to document what should be.
That forward-looking orientation is also why suspecting programs align naturally with quality and preventive care. Prompting a physician to address a patient's CKD stage before the visit is not just a coding intervention. It is a clinical one. Patients with complex chronic conditions benefit when their care team is actively tracking condition status, and that benefit extends to Star Ratings metrics that overlap with chronic disease management.
The Four Components That Make a Suspecting Program Work
1. Accurate, Current Data Feeding the Suspect List
Stale data produces bad suspects. If your suspecting engine is running on claims that are six months old, it will miss patients who developed new conditions recently and flag conditions in patients who are no longer attributed to your plan. Data recency and attribution accuracy are table stakes for a functional program, not enhancements.
2. A Physician-Friendly Presentation
Physicians are not opposed to HCC documentation when the workflow respects their time. They are opposed to opaque lists that feel like compliance surveillance. The best suspect presentations lead with clinical context, not administrative justification. "Patient is filling lisinopril and furosemide. CHF not yet documented this year. Prior diagnosis 2023" is a clinical prompt. A list of HCC codes with no supporting evidence is not.
3. A Defined Closure Workflow
This is the piece most organizations underinvest in. The suspect list is the easy part. What happens when a suspect is not addressed at the visit? Who follows up, and through what channel? What is the escalation path for high-RAF suspects approaching the end of the measurement year? These are operational design questions, and they need answers before launch, not after the first quarter of low closure rates.
4. Closure Rate Tracking Over Time
A program that does not measure closure rates is not a program. It is a vendor product someone bought. Tracking closure rates by provider, by condition category, and by time of year reveals where the workflow is breaking down. It also enables the kind of performance conversation that drives improvement: showing a provider group their Q3 closure rate compared to peers is more actionable than a general reminder about RAF documentation.
The Risk of Suspecting Without Closure Infrastructure
A list of flags nobody acts on is not a program. It is noise, and expensive noise at that, because it consumes data infrastructure and staff time without producing documentation outcomes.
This failure mode is more common than it should be. Organizations stand up a suspecting tool, often as part of a larger health IT implementation, but do not pair it with the operational closure workflow. Six months in, the suspect lists are generating but closure rates are under 20 percent. The RAF gap the program was supposed to close is still there. The plan-level reconciliation in February still shows scores below projection.
Sustainable risk adjustment and HCC coding performance requires treating suspecting and closure as a single integrated process, not two separate initiatives that happen to share a data source.
How This Connects to RAF Accuracy and MA Plan Revenue
RAF scores in the CMS-HCC model are payment-relevant. Conditions that are clinically present but not documented in the measurement year do not contribute to the patient's RAF score. At scale, across a panel of complex Medicare Advantage members, that gap between actual disease burden and documented disease burden translates directly into plan underpayment relative to the true cost of care.
That underpayment is not recoverable through retrospective audit alone once the measurement year closes. It represents a permanent gap for that contract year. A well-run HCC suspecting and gap closure program does not manufacture diagnoses. It ensures that conditions that are genuinely present get documented by the treating physician in the appropriate encounter during the year they are relevant. That is compliant HCC capture, and it is what RAF accuracy is supposed to reflect.
For a deeper look at how measurement year documentation connects to payment accuracy year over year, the post on HCC RAF score accuracy covers the model mechanics in detail.
Who Should Own This Program
This is not a coding department initiative. It is not a CDI initiative. It is not a clinical operations initiative. It is all three.
Coding brings the HCC model expertise and the ability to identify which conditions have documentation gaps. CDI brings the physician engagement skills and the query workflow that drives closure. Clinical operations controls the scheduling infrastructure, the EHR integration, and the provider performance conversations that move closure rates from acceptable to strong. When any one of those three is absent from program governance, the program develops predictable blind spots.
A medical director or CMO who is actively involved in the physician-facing design component will accelerate adoption faster than any training webinar. Organizations that treat this as a back-office coding function and never bring it into clinical leadership conversations consistently underperform on closure rates.
If you are evaluating whether your current suspecting program has the closure infrastructure it needs, start with the free HCC audit checklist to identify where your workflow gaps are before the next measurement year gets away from you. And if you are considering whether to build or augment that program with outside expertise, the post on when to outsource HCC risk adjustment coding walks through exactly that decision.
If your organization is ready to build a suspecting and gap closure program that actually closes gaps, contact the MedCodex risk adjustment team to discuss a workflow assessment and program design engagement tailored to your plan or provider group.