Underdiagnosed Condition Care Coordination Analysis Case Study

Client question

A client wanted to understand how Medicare beneficiaries with an underdiagnosed condition were identified, what other conditions appeared alongside the diagnosis, which provider specialties were involved, and whether patients appeared to move through coordinated care pathways or fragmented multi-specialty care.

Data foundation

FastHSR used 100% Traditional Medicare claims and beneficiary enrollment data across a recent two-year period. Claims sources included carrier, outpatient, inpatient, and other Medicare files. Beneficiary-level demographics and enrollment characteristics were drawn from the Medicare beneficiary summary file.

• Cohort definition: beneficiaries were included if they had qualifying diagnosis evidence for the target condition.
• Claims settings: the cohort was built across inpatient and non-inpatient claim types.
• Index event: each beneficiary received an index date based on the first qualifying diagnosis claim.
• Follow-up window: claims after the index date were used to study provider involvement, treatment patterns, and care journeys.

Cohort construction

The cohort algorithm required stronger evidence than a single incidental diagnosis in routine outpatient billing. FastHSR used a claims-based approach that distinguished inpatient evidence from repeated non-inpatient evidence, then assigned each beneficiary an index date for downstream analyses.

• Identify claims with the target diagnosis code family in any diagnosis position.
• Require one qualifying inpatient claim or repeated qualifying non-inpatient claims on separate service dates.
• Collapse claims to the beneficiary level to create a patient cohort file.
• Flag demographic and enrollment subgroups for stratified analysis.

Comorbidity mapping

FastHSR mapped co-occurring diagnoses to understand the clinical context surrounding the condition. The analysis retained diagnosis codes from claims associated with the target condition, collapsed them to the beneficiary level, and summarized co-occurrence patterns overall and by subgroup.

• Frequency tables of co-occurring diagnoses in the overall cohort.
• Stratified comorbidity tables by age and sex groups.
• Condition co-occurrence edge lists showing which diagnoses appeared together on the same claims.
• Network maps limited to the strongest co-occurrence relationships so the patterns remained interpretable.

Provider network analysis

The condition was not managed by one specialty. FastHSR therefore mapped the provider landscape to identify which specialties were involved, how often patients moved between specialties, and whether any specialty appeared to function as a coordinating hub.

• Provider specialty frequencies across all qualifying claims.
• Evaluation and management visits analyzed separately from procedure-heavy claim volume.
• Unique patient counts by specialty to distinguish reach from claim intensity.
• Specialty-to-specialty handoff pairs built from each beneficiary's chronological visit sequence.
• Provider network maps showing patient sharing and referral flow across specialties.

Fragmentation and continuity metrics

To quantify coordination, FastHSR measured how many clinicians and specialties each patient encountered after the index date, then calculated continuity of care ratios. These measures helped separate concentrated care from fragmented multi-provider evaluation patterns.

• Distinct provider NPIs per beneficiary.
• Distinct provider specialties per beneficiary.
• Metrics calculated for all qualifying services and for evaluation and management visits only.
• Continuity ratio defined as the share of visits accounted for by the beneficiary's most-seen provider.

Patient journey archetypes

FastHSR classified each beneficiary into a care journey archetype based on timing and sequencing of specialist contact, conservative management, invasive procedures, and provider fragmentation. Archetypes were tested under multiple claim-filtering scenarios to compare narrow diagnosis-specific views with broader clinically related views.

• Time from index diagnosis to first relevant specialist contact.
• Time from index diagnosis to conservative treatment.
• Time from index diagnosis to invasive procedure.
• Time from first conservative treatment to first invasive procedure.
• Journey categories including early procedure use, delayed procedure use, stable conservative care, fragmented care, and mixed pathways.
• Comparison of emergency department utilization and Medicare payments across journey categories.

Findings

The analysis showed that the condition was embedded in a broader pattern of clinical complexity and multi-specialty care. The provider network did not show a simple, dominant coordinating pathway, and continuity measures suggested that many patients moved across multiple clinicians and specialties during evaluation and management.

Journey analyses helped distinguish patients with more stable, conservative patterns from patients with fragmented or procedure-oriented pathways. The results were descriptive and intended to guide further clinical validation, pathway design, and care coordination strategy.

Deliverables

• Patient cohort definition and index-date file.
• Comorbidity frequency tables by subgroup.
• Co-occurrence edge lists and network maps.
• Provider specialty distribution tables.
• Specialty handoff matrices and provider network maps.
• Fragmentation and continuity metrics at the beneficiary level.
• Patient journey archetype file with utilization and cost summaries.
• Methodology notes describing code logic, claim filters, and limitations.

Use cases

• Evidence generation for an underdiagnosed condition.
• Care pathway design and coordination strategy.
• Provider network and referral pattern analysis.
• Patient journey segmentation.
• Market education for conditions with fragmented diagnosis and treatment pathways.

Frequently asked questions

Why use claims data for an underdiagnosed condition?

Claims data can show how diagnoses appear in real-world billing, which providers participate in care, and whether patients follow coherent pathways or move across disconnected settings.

Why build provider networks?

Provider network analysis shows whether care is organized around a clear coordinating specialty or spread across many specialties with repeated handoffs.

Why create patient journey archetypes?

Archetypes simplify complex longitudinal claims histories into interpretable patterns that can guide care model design, clinical validation, and future intervention planning.

For underdiagnosed condition analysis, care coordination analytics, provider network mapping, or patient journey analysis using Medicare claims, please email us.

See more case studies