Your organization has already invested in retrieving and reviewing medical records. The retrospective coding file captures one part of the value those records contain. But what happens to the clinical information that could help your prospective team decide which conditions need evaluation next?
Consider a prior-year specialist note documenting a chronic condition absent from the plan’s available comparison data. A retrospective review may identify a supported diagnosis addition for that service year. Preserving the evidence, date of service, and provider context gives the prospective team a useful starting point: has the condition been evaluated this year, or does it still need clinical follow-up?
When a retrospective review ends with coding results alone, that context may never reach the team responsible for prospective suspecting. They may retrieve the same chart, review the same documentation, or pursue a less specific lead from administrative data.
AI-enabled chart review creates an opportunity to preserve more value from work already performed. Alongside retrospective coding outputs—or through a separate clinical chart scan—findings can be organized for subsequent review and evaluation.
For risk adjustment leaders, the opportunity is practical: more informed prioritization, less repeated work, and evidence that helps teams understand why a finding deserves attention.
Enriching administrative signals with clinical context
Claims and pharmacy data remain valuable inputs for prospective suspecting. They reveal reported diagnoses, specialist visits, and medication patterns. Medical records add context: what the provider documented, whether the condition was evaluated or treated, and whether later notes clarified or contradicted it.
Retrospective clinical intelligence brings those details together across dates of service. Rather than receiving only a potential condition, the prospective team can receive the supporting text, source page, encounter date, provider, and reason for follow-up.
The output should distinguish findings that require different actions:
- Diagnoses supported for retrospective coding: Eligible for reporting in the reviewed service year, subject to applicable requirements.
- Previously documented chronic conditions: Candidates for current-year evaluation, when they have not already been addressed.
- Potential conditions requiring clarification: Clinical questions that need provider evaluation before a diagnosis can be established.
- Findings to suppress or deprioritize: Duplicate, resolved, ruled-out, or contradicted findings that could otherwise generate unnecessary outreach.
These distinctions help teams prioritize meaningful questions and avoid treating every historical reference as an active condition.
Reducing repeated retrieval and review
Separate retrospective and prospective programs may use different vendors, systems, and deliverables. Without a shared record of reviewed findings, each team can end up reconstructing the same patient history.
Preserving clinical evidence allows available charts to serve more than one purpose. Subsequent teams can build on prior work, check findings against current-year information, and focus new retrieval or review on what remains unresolved.
Updated records may still be needed. The benefit is knowing what to look for and why.
For providers, the value can be fewer duplicate requests and more specific questions. For program teams, it can mean less time searching through charts and reconciling disconnected findings.
Making the intelligence actionable
An additional suspect file creates value only when someone can use it.
Returning to the specialist-note example, the prospective team first checks available current-year records and encounter data. If the condition has already been evaluated, the finding can be reconciled. If it remains unresolved, the historical evidence can support an appropriately framed question in the provider’s existing workflow.
Each finding should make the next step clear. That requires the condition, evidence, encounter context, finding status, and reason for follow-up—not simply an HCC label.
The workflow also needs an owner and a way to record the outcome. Was the condition evaluated and confirmed? Ruled out? Clarified? Does it remain unresolved? Recording those outcomes helps prevent the same question from resurfacing without new information.
Across charts, recurring documentation issues can also guide provider education. Specific examples give engagement teams a stronger basis for education than broad reminders to document chronic conditions.
Preserving evidence and clinical judgment
Traceable evidence makes findings easier to review, but historical documentation does not establish that a condition remains active in a later year.
Retrospective coding findings must meet applicable documentation, coding, provider, encounter, and service-period requirements. Prospective findings must preserve the distinction between an established historical diagnosis and a question requiring current clinical evaluation.
The applicable CMS-HCC model determines how eligible diagnoses map to HCCs. It does not determine whether the patient has the condition or whether the documentation supports reporting it.
A useful output therefore preserves uncertainty as carefully as it preserves supporting evidence. Providers should be able to confirm, rule out, or clarify findings, and contradictory or resolved documentation should remain visible.
Measuring value beyond the number of suspects
A larger suspect list does not demonstrate a better program. The meaningful question is whether the intelligence helps teams identify relevant findings, act efficiently, and reach a resolution.
Risk adjustment leaders should evaluate:
- Finding quality: How often are findings clinically relevant, and are the evidence, dates of service, and provider attribution accurate?
- Workflow usefulness: Can teams interpret and act on the findings without substantial additional chart review?
- Follow-through: What proportion of findings are evaluated, confirmed, ruled out, or left unresolved?
- Efficiency: How much duplicate retrieval, review time, and unnecessary outreach is avoided?
A ruled-out finding is a useful outcome. It resolves uncertainty and can prevent repeated pursuit. Measuring only confirmed diagnoses or RAF yield misses that value.
Getting more from each chart review
Retrospective chart review can leave behind more than a coding file. It can preserve an organized clinical history that helps prospective teams ask better questions and focus their next steps.
For organizations evaluating vendors, three questions make the opportunity concrete:
- What findings and evidence will we receive?
- How will our teams use them?
- How will we measure their value?
The investment in chart retrieval and review has already been made. Preserving useful clinical intelligence can extend its value into the next encounter and the next program cycle.
To see how retrospective suspecting intelligence can close the gaps in your prospective program, reach out to schedule a demo with one of Charta’s clinical chart intelligence experts.
At the Becker's 5th Annual Fall Payer Issues Roundtable, taking place November 2–3 in Chicago, payer executives and healthcare leaders will come together to discuss value-based care, regulatory changes, cost management strategies and innovations shaping the future of payer-provider collaboration. Apply for complimentary registration now.
