Your organization’s risk-adjusted factor scores increased 6.2% year-over-year. Leadership is celebrating. Finance is projecting higher revenue. You’re getting kudos for strong risk adjustment performance.

Then CMS sends an investigation notice. Your RAF growth pattern triggered algorithmic flags. Now you’re explaining to federal regulators why your population got sicker so fast.

Not all RAF growth is created equal. Some growth patterns are defensible. Others are red flags that invite scrutiny. Here’s how to distinguish between the two.

The Demographic-Disconnected Growth Problem

Your member population demographics stayed stable. Average age changed by 0.3 years. Gender mix shifted 1%. Geographic distribution is essentially identical. Chronic disease prevalence in the broader community remained flat.

Yet your RAF scores increased 6.2%.

CMS asks: If your population didn’t age significantly, didn’t shift geographically, and isn’t experiencing regional health trends, why did average risk increase?

The defensible answer requires documenting specific clinical reasons: new high-acuity members enrolled, existing members experienced disease progression, better identification of previously undocumented conditions through improved screening.

The indefensible answer: “We implemented a new retrospective coding vendor” or “Our documentation improvement program matured.” Those answers tell CMS your population didn’t get sicker, you just started coding more aggressively.

The Sudden Onset Chronic Disease Spike

Your diabetic nephropathy coding rate was 18% of diabetic members last year. This year it’s 34%. Your COPD coding increased from 22% to 41% of members with respiratory conditions. Your vascular disease coding doubled.

These are chronic conditions that develop gradually. Population-level prevalence doesn’t double in one year without specific clinical drivers.

CMS asks: What clinical interventions, screening programs, or population health initiatives drove these specific prevalence increases?

The defensible answer requires documentation: “We implemented diabetic retinopathy screening for all diabetic members, which led to identification of previously undiagnosed complications” or “We launched a COPD case-finding program using spirometry for high-risk members.”

The indefensible answer: “Our coders were trained to look for complications more carefully” or “We started using AI to identify documentation gaps.” Those answers confirm you’re finding diagnoses through coding changes, not clinical interventions.

The Retrospective-Driven RAF Inflation

You analyze where your RAF growth came from. 72% of incremental HCCs were identified through retrospective chart review. Only 28% came from prospective or concurrent identification at the point of care.

This distribution is a red flag. If most of your growth comes from looking backward through charts rather than forward through clinical encounters, it suggests documentation mining rather than population health management.

CMS scrutinizes organizations where retrospective programs drive RAF growth. They’re looking for patterns where diagnoses get added months after encounters without clear clinical basis.

The defensible approach: RAF growth should primarily come from prospective and concurrent identification. Retrospective should clean up documentation gaps, not drive significant growth. If retrospective is identifying major gaps, that reveals underlying documentation quality problems that need fixing, not celebrating.

The Diagnosis Without Utilization Pattern

Your CHF coding increased 35%. But hospital admissions for CHF remained flat. Emergency department visits for CHF didn’t increase. Cardiology referrals stayed constant. Diuretic prescriptions were unchanged.

CMS asks: If you’re identifying more CHF patients, why isn’t their healthcare utilization reflecting disease burden?

This pattern suggests you’re adding CHF diagnoses without corresponding clinical activity. Either the diagnoses aren’t real, or they’re not being managed clinically, or they’re so mild they don’t require treatment (in which case, should they be coded?).

The defensible pattern: diagnosis increases should correlate with appropriate utilization. More diabetic nephropathy coding should align with more nephrology visits, more GFR monitoring, more ACE inhibitor prescriptions. Otherwise, the coding looks disconnected from clinical reality.

The Coding Intensity Versus Quality Metrics Divergence

Your RAF scores increased 6.2%. Your HEDIS quality scores decreased 3%. Your hospital readmission rates increased. Your care gaps grew. Your medication adherence rates declined.

CMS notices this divergence. You’re coding higher acuity, but quality metrics suggest you’re not managing that acuity well. This pattern raises questions about whether the coded acuity is real.

Organizations with legitimately sicker populations typically show corresponding challenges in quality metrics. But they also show intensive management efforts: more care coordination, more case management, more outreach programs.

The defensible pattern: RAF growth accompanied by increased care management intensity, even if quality metrics temporarily decline due to higher baseline acuity. The key is demonstrating you’re actively managing the acuity you’re coding.

The Provider Outlier Concentration

You analyze which providers drove your RAF growth. Dr. Martinez’s patients account for 18% of total RAF increase despite representing only 4% of your membership. Three other providers account for another 40% of growth.

This concentration pattern triggers investigation. When a small number of providers drive disproportionate RAF growth, CMS asks whether those providers are documenting differently or coding more aggressively than peers.

The defensible explanation requires clinical justification: Dr. Martinez is a specialist who sees complex patients, or Dr. Martinez’s practice focuses on specific high-acuity conditions, or Dr. Martinez participates in intensive care management programs.

The indefensible explanation: Dr. Martinez received special documentation training, or Dr. Martinez’s practice was targeted by your retrospective vendor, or Dr. Martinez responds well to prospective alerts.

The HCC Category Imbalance

Your RAF growth concentrated in specific HCC categories. Diabetes with complications increased dramatically. CKD staging improved significantly. Vascular disease coding expanded.

Meanwhile, other predictable high-prevalence conditions remained flat. Depression coding didn’t increase. Obesity coding was stable. Substance use disorder coding unchanged.

This imbalance suggests selective focus on high-value HCCs rather than comprehensive documentation of all conditions. CMS recognizes this pattern as financially motivated coding rather than holistic patient assessment.

The defensible pattern: RAF growth should be relatively balanced across condition categories. If you’re improving documentation quality generally, you should see improvements across multiple conditions, not just the ones with highest financial value.

The Year-Over-Year Volatility

Your RAF scores grew 1.2% two years ago, 1.8% last year, and 6.2% this year. The acceleration is dramatic and unexplained by demographic or epidemiological trends.

Sudden acceleration without clear clinical drivers suggests program changes: new vendors, new technology, new compensation models, new documentation requirements.

CMS views sudden acceleration skeptically. Gradual, consistent growth is more defensible than sporadic spikes.

What Actually Works

Building defensible RAF scores requires aligning coding growth with clinical reality.

Document demographic and epidemiological changes that drive population acuity. Connect diagnosis increases to specific clinical screening or intervention programs. Ensure RAF growth comes primarily from prospective and concurrent identification, not retrospective mining. Demonstrate that diagnosis increases correlate with appropriate utilization patterns. Show care management intensity that matches coded acuity. Distribute RAF growth broadly across providers, not concentrated in outliers. Balance growth across condition categories, not just high-value HCCs. Maintain steady, consistent growth patterns rather than sudden spikes.

The organizations with defensible RAF scores can answer the question: “Why did your population get sicker?” with clinical evidence, not coding program changes. If you can’t answer that question convincingly, your RAF growth is a liability, not an achievement.