Care Gap Closure in Value-Based Care: A Practical Workflow

Quick answer

The measure specification is the source of truth. HEDIS, Medicare Star Ratings, the Medicare Shared Savings Program, and ACO REACH are related to quality performance, but they are not interchangeable programs. Each can use different populations, exclusions, evidence rules, and reporting periods.

Automation can assemble worklists, send approved reminders, record responses, and route exceptions. It should not decide whether a service is medically appropriate, override exclusions, or close a gap without qualifying evidence. Start with one measure and one accountable owner, then track verified closures rather than calls placed.

What care gap closure actually means

A care gap is not simply a missing checkbox. It is a difference between what a defined measure expects and what the organization can demonstrate for an eligible patient. The apparent gap may represent an overdue service, but it may also result from delayed claims, incomplete clinical data, a valid exclusion, care delivered outside the network, or documentation that does not satisfy the measure.

Closure therefore requires more than outreach. A defensible program answers four questions:

1. Which exact program, contract year, measure version, and eligible population apply?
2. Does available data show that the patient truly remains open for the measure?
3. What clinically appropriate action, if any, should happen next?
4. What evidence will the reporting entity accept as proof of completion or exclusion?

Operational activity is not the same as measured performance. A scheduled visit is progress, not necessarily closure. A completed service may still look open if evidence never reaches the reporting system. A record can also appear closed internally and later fail validation if the documented evidence does not match the applicable specification.

For more context, read what care gaps are and how they form and the closed-loop referral management guide.

Do not treat HEDIS, Stars, MSSP, and ACO REACH as synonyms

Quality teams often use the phrase "care gaps" across several programs. That shorthand is convenient, but it can create reporting errors. Build separate measure maps for each contract instead of assuming one list applies everywhere.

NCQA describes HEDIS as a widely used performance measurement set. CMS separately publishes Shared Savings Program guidance and specifications and the ACO REACH model requirements. Those sources should be checked for the relevant performance year before any worklist or patient script is approved.

Label every operational record with the program, contract, measure version, reporting period, and responsible entity. If the same patient appears in more than one program, keep the logic and evidence requirements distinct even when one completed service can support multiple reporting workflows.

A six-stage care gap closure workflow

1. Translate each measure into an operational specification

Before contacting patients, convert the governing measure into a controlled playbook. Record the eligible population, denominator exclusions, numerator event, acceptable evidence, lookback period, deadline, escalation owner, and system of record. Have quality, compliance, and clinical leadership approve the interpretation.

Avoid turning a measure title into homegrown clinical guidance. A coordinator should not infer that every person on a payer list needs the same intervention. The approved playbook must show when a record needs clinical review and how staff should handle uncertainty.

2. Reconcile the apparent gap across data sources

Compare payer lists, claims, EHR data, health information exchange data where available, scanned records, and patient-reported information. Normalize identity fields and dates, then route mismatches instead of silently overwriting them.

• Confirmed open: available evidence supports that the gap remains open.
• Possible external completion: the patient reports completion elsewhere or another feed suggests it.
• Possible exclusion: documentation may support an exclusion, pending qualified review.
• Data conflict: sources disagree about identity, date, result, or service.
• Closed with acceptable evidence: the required evidence is present and validated.

The AHRQ Care Coordination Measures Atlas emphasizes that coordination spans participants, settings, and transitions. Care-gap data should be treated as a multi-source workflow, not a single static list.

3. Require clinical validation where judgment is involved

Automation may find candidates and collect facts. It should not determine medical appropriateness, interpret a clinical result, choose an alternative service, or approve a measure exclusion. Those tasks belong to qualified people operating under approved policy.

• conflicting clinical information;
• symptoms or statements that may indicate urgency;
• questions about risks, benefits, or whether a service is appropriate;
• possible exclusions that require clinical interpretation;
• refusals that prompt a clinical follow-up protocol; and
• any result that requires diagnosis or treatment decisions.

Outreach scripts should tell patients how to reach the appropriate clinical team and what to do when urgent symptoms arise, based on organizational policy. The script should never suggest that an automated conversation is medical advice.

4. Prioritize outreach with transparent rules

Prioritization should make the queue manageable without hiding eligible patients. Common operational fields include time remaining in the measurement period, scheduling lead time, prior contact history, documented communication preferences, language needs, known access barriers, and whether a referral or authorization is required.

Do not use a score that staff cannot explain. Maintain the rule version and reason each patient entered a priority band. Review outcomes by demographic and access-related groups to detect whether the workflow creates unequal opportunities to complete care.

For programs that rely on referrals, connect the gap queue to the referral management process. A reminder cannot close a gap if the patient still lacks an accepted referral, an authorized service, transportation, or an available appointment.

5. Coordinate the next actionable step

Outreach works best when it can resolve the immediate barrier. Depending on the approved workflow, that may mean confirming interest, transferring to scheduling, offering eligible appointment choices, collecting a missing nonclinical detail, or routing the record to a coordinator.

Use more than one permitted channel and make opt-out handling consistent. Each contact attempt should record time, channel, disposition, consent or preference updates, and next owner. For high-volume populations, an automated patient scheduling outreach workflow can handle routine engagement while reserving staff time for complex cases.

If the next step needs prior authorization, treat that as a separate tracked dependency. The prior authorization cycle-time guide explains why internal preparation time and payer decision time should be measured separately.

6. Capture evidence and reconcile it back to the program

Define the evidence packet before the campaign begins. It may include a claim, structured EHR field, laboratory result, procedure record, external chart document, or other contract-accepted artifact. The required evidence varies, so do not assume a scheduled appointment, outreach disposition, or generic note is sufficient.

Create a final validation gate with three outcomes: accepted closure, rejected closure with reason, or pending evidence. Keep an audit trail of source, reviewer, date, specification version, and submission destination. Reconcile accepted evidence back to every relevant program without collapsing program-specific logic.

The operating model behind reliable closure

Assign one owner to each stage, not a vague shared inbox.

The workflow also needs a daily exception queue and a weekly control review. Daily review prevents urgent or ambiguous records from waiting. Weekly review finds systemic defects such as a disconnected lab feed, an invalid payer roster, or a location with no available appointments.

Measure outcomes, not contact volume

Calls, texts, and portal messages are process counts. They help manage capacity, but they do not show whether care was completed or documented. Use a funnel that separates operational steps:

1. Apparent gaps received
2. Records reconciled
3. Confirmed open gaps
4. Patients reached
5. Next steps accepted
6. Services scheduled
7. Services completed
8. Evidence received
9. Closures validated
10. Closures accepted by the reporting process

Track time and fallout between stages. Segment results by measure, program, location, source list, outreach channel, and barrier category. Report re-opened gaps and rejected evidence separately so a superficially high internal closure count does not obscure downstream failure.

The denominator must remain visible. For example, "patients reached" should be divided by eligible outreach records, while "validated closures" should be divided by confirmed open gaps. Never present a single percentage without naming its numerator, denominator, time window, and program.

A safe 90-day rollout

Start with one measure whose specification is understood and whose next-step capacity exists.

Days 1 to 30: define and validate

• Select the program, contract year, measure, and accountable owner.
• Build the measure map and clinical escalation policy.
• Reconcile a sample manually and classify the causes of false gaps.
• Confirm that scheduling, referrals, authorization, and evidence feeds can support the workflow.
• Approve outreach content through clinical, compliance, privacy, and accessibility review.

Days 31 to 60: run a controlled pilot

• Limit the pilot to a defined population and location.
• Review every automation exception daily.
• Compare worklist classifications with human review.
• Measure the full funnel through validated evidence, not just engagement.
• Interview staff and a sample of patients about friction and clarity.

Days 61 to 90: correct and expand

• Fix recurring data and handoff defects.
• Document rule changes and revalidate them.
• Expand only when appointment capacity and exception staffing are adequate.
• Add a second measure only after the first workflow has stable evidence reconciliation.

Where automation helps and where it must stop

Automation is well suited to deterministic tasks: joining approved data sources, deduplicating worklists, applying documented nonclinical routing rules, launching approved reminders, capturing dispositions, and escalating records that exceed a time threshold. It can also help managers see where a gap is stuck.

Human review remains necessary for medical appropriateness, symptoms, urgency, diagnosis, result interpretation, contraindications, exclusions requiring judgment, and any decision that changes a care plan. The system should surface context and preserve evidence, while qualified professionals make the decision.

When comparing vendors, ask for a test using de-identified or synthetic examples that include conflicting data, external completion, language preference, opt-out, no appointment capacity, and a clinical question.

Sources

• NCQA, HEDIS Measures and Technical Resources
• CMS, Medicare Shared Savings Program Guidance and Specifications
• CMS, Medicare Shared Savings Program
• CMS Innovation Center, ACO REACH Model
• CMS, Medicare Advantage and Part D Performance Data
• AHRQ, Care Coordination Measures Atlas

Frequently asked questions